12th International Symposium On District Heating PDF

12thInternationalSymposiumonDistrictHeating
andCooling

September5thSeptember7th,2010
Tallinn,ESTONIA
ISBN: 978-9949-23-015-0
PREFACE
th
The 12 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,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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 most equipment currently used in district heating

substations is antiquated.
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.
Fig. 1. Evolution of wireless sensor networks. Although
A system with open standards facilitates the
the scalability of the sensor network has increased, many
introduction of new energy services such as individual industries still use vendor-specific cable solutions. (The
measurements and improved space heating control. figure was obtained from the literature [1])
In this study, we found that resource-limited battery-
powered devices possess a life expectancy of over If heat meters, control systems, and other non-district
10 years, using small batteries while participating in heating equipment could communicate, new services
IPv6 compatible communication. that have impact on both economy and the
environment could be developed.
INTRODUCTION The infrastructure required to achieve wireless device
Embedding low-power wireless devices in district communication may be attained with low-power
heating substations and surrounding equipment such wireless technology. Small sensor platforms with
as temperature sensors could provide useful services direct internet access through standardized wireless
to consumers and producers. Currently, many technology can provide a solid platform for new
different substation control systems on the market can services.
connect to the internet and have various wireless A lack of standardized communication protocols is
sensor reading systems. However, these systems commonly encountered when connecting electronic
tend to be specialized and are only compatible with devices from different vendors. In general, devices
equipment from the same manufacturer. Moreover, manufactured by different companies use different
internet-compatible control systems are often also communication protocols, which limits the functionality
relatively expensive, and provide bad scalability. of the substation.
In general, commercially available heat meters cannot District heating substations can be divided into
communicate through the current infrastructure; thus, sections based on metering, space heat control, and
specialized communication methods such as mbus, tap water control. For a visual overview of a common
pulse, and infrared readings must be employed. parallel connected district heating substation, see
Therefore, poor communication standards limit the Fig. 2, this is also the substation type used in the
current usage of heat meters and other equipment in study. Typically, information is not shared between
the substation. However, by sharing information with these sections; thus, each system can only be locally
other devices in the substation, the heat meter could optimized. To achieve complete substation
provide useful feedback and sensing information, optimization, information must be shared between
which can be used to improve the substation control sections. To this end, wireless sensor-platforms were
functionality. installed in temperature-sensors, heat-meters,
Fig. 1 provides an overview of the development of circulation pumps, and control valves, and new control
sensor networks over the last 20 years. Unfortunately, methods and services were tested. This empowers us
4
th th
to develop new control methods, and implementing temperature of the returning distribution medium
new services to heat suppliers, building owners and should be minimized.
end users.
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
Fig. 2. A systematic overview of a parallel coupled temperature also affects the ideal radiator supply
district heating substation divided into three sections: temperature and the radiator circuit flow. The
metering, heating and hot water system. relationship between outdoor temperature and primary
supply is often assumed to be linear (colder outdoor
SERVICES air leads to a warmer primary supply). However,
significant deviations from the ideal curve are
To control or reduce their energy bill, district heating common. More information on the effect of primary
customers require specific information to determine supply temperature and radiator control on the indoor
the appropriate action. Currently, the only information air temperature and T of the system can be found
available to the customer is the information provided in [2].
in the bill or on the heat-meter display.
Adaptive radiator control is another intelligent way of
If information on all devices was available online, controlling the radiator circuit and obtaining a high T.
customer could easily monitor their usage and interact More information on this method can be found in
with the substation. Examples of services that could previous studies by Lauenburg [3].
be provided by the substation are explained in the
following sections. Fault detection
Control valves in the district heating substation often
Improved substation control
possess inappropriate dimensions, resulting in
Combined heat and power plants are becoming more intermittent control, pressure shocks, and high return
common; thus, the importance of the distribution temperatures. Due to the high thermal time constant
system T is increasing. In a combined heat and of a building, the indoor temperature is not directly
power plant with a flue-gas condensation system, a affected. Therefore, an error in the control valve may
high T is even more important to obtain satisfactory go unnoticed for a considerable amount of time.
fuel efficiency.
Error identification can be achieved by evaluating high
To maintain high energy efficiency, the hot water frequent meter readings, which to some extent are
produced by the plant must be delivered to customers done today.
with a minimal heat loss. Once the hot water is
transported to the customer, a maximum amount of A fouling valve that is stuck or does not move in
energy per volume of water should be extracted and accordance with the control signal may also be difficult
used for heating purposes, such as hot tap water and to detect. A direct comparison of the valve control
space heating. To achieve a maximum T, energy signal with the heat meter, which measures the
transfer between the distribution medium to the point primary flow through the district heating substation,
of consumption should be maximized, while the can be used to identify a broken fouling valve [4].
5
th th
Individual measurements NETWORK TECHNOLOGY

Individual measurements are common in some A common method of visualizing a network
countries and are gaining interest in others. To obtain communication protocol is in the form of stack. A stack
measurements of each apartment, tap point, or consists of layers that are separated by function; thus,
radiator, new metering devices must be installed. The a communication stack contains different layers of
most straight forward method is to install flow meters tasks related to data transportation. The layers can be
at each tap point and/or radiator. In general, high divided and visualized in many ways. For example,
resolution flow meters are quite expensive; thus, the five-layer internet model has been used
installing one on every tap point/radiator can be cost- extensively in previous studies and is displayed in
prohibitive. Fig. 4 [7]. In this paper, only the layers that are
An alternative method has been evaluated by Yliniemi significant to the results of this research will be
[5]. In this method, temperature sensors were discussed. Thus, the network, link, and physical layers
installed at each tap point, and one central flow meter are considered in more detail.
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 Fig. 3. Performance of a district heating substation
independent auction system is used to decide which visualized on a map. The red square can represent the
supply/return temperature, energy usage, or heat flow in
buildings will be shut down or provided a limited
the connected building.
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
Fig. 4 A generic five-layer internet model and its
all substations could be plotted in a graph, table or
implementation in an IEEE 802.15.4 wireless network.
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 IP (Network Layer)
the area. Moreover, the map provides the utility The internet protocol (IP) is the most well-known and
company with an overview of the network and commonly used network protocol in the world. All
improves the detection of leaks and short circuits. traffic on the internet is currently routed through IP.
Moreover, the utility company can identify Today, there are two co-existing versions of IP,
deteriorating substations or individual installations that including IPv4, the older version of IP, and IPv6, the
perform poorly. latest version.
6
th th
IPv4 IPv6 prefix from the router to the local link layer
Currently, internet protocol version 4 is the most address, creating a complete IPv6 address. To ensure
widely used IP, and almost all computers connected that another device does not possess the same IP
to the internet use this version. address, the device broadcasts a neighbor solicitation
message to search for a duplicate address. If another
An IPv4 address is 32 bits long and is typically written device has the same IP number, the new device shuts
in 4 sections divided by dots (e.g., 192.168.100.123). down.
The theoretical number of IPv4 addresses is 232
(approximately 4.2 billion); however, a fraction of
6LOWPAN
addresses is reserved and cannot be used for online
purposes. The total number of usable IPv4 addresses 6LoWPAN is an adaptation layer that separates the
is approximately 3.7 billion. As the number of devices network and data link layer of the protocol stack. The
connected to the internet increases, IPv4 addresses purpose of the layer is to compress IPv6 headers and
are beginning to run out. minimize unnecessary data transmission while
maintaining IPv6 compatibility. According to the
Technology such as network address translation
literature, [8] the 6LoWPAN header uses less than
(NAT) and port address translation (PAT) have
10% of the total energy used during packet
postponed the depletion of IPv4 addresses; however,
transmission.
the number of available IPv4 addresses decreases
every day. IEEE 802.15.4 physical and data link layers are often
used in combination with 6LoWPAN; however, other
IPV6 standards can also be applied.
IPv6 was developed to compensate for the limited 802.15.4 (Link and Physical Layer)
number of IPv4 addresses. IPv6 uses a longer
The most common data link and physical layer used
address than IPv4 and has several convenient
with 6LoWPAN networks is IEEE 802.15.4; however,
features. IPv6 uses a 128 bit address, which means
6LoWPAN is also compatible with other layers.
that there are 2128 possible addresses. Thus, the
Moreover, IEEE 802.15.4 is also the basis for ZigBee,
number of address per square millimeter of the earths
Wireless HART, and MiWi. The IEEE 802.15.4
surface is 6.71017. Hopefully, the addresses
standard specifies operation at low frequency bands
obtained through the implementation of IPv6 will last
such as 868 MHz (EU), 915 MHz (US), and 950 MHz
for a long time.
(JP), and high frequency bands including 2.4 GHz
With the additional address space, it is possible to (World Wide) [9]. The main practical differences
give every small device its own unique IP number between low and high frequency bands are the
without implementing NAT. Thus, direct bandwidth and communication range. The 2.4 GHz
communication over the internet can be achieved band supports a higher bandwidth but the range is
without any special gateways. However, the new limited, especially in armored concrete buildings. The
address space increases the overhead of data low frequency bands have a moderate bandwidth and
packages, which negatively impacts small, low-power a considerably larger range. In a district heating
devices because more battery energy is wasted on substation, bandwidth usage can be minimized
header data in every wireless data transmission. because rapid changes are uncommon (compared to
However, a new adaptation layer (6LoWPAN) was many other control/measurement situations) and low
developed to limit the amount of lost energy. More frequency groups are preferred. However, only 2.4
information on 6LoWPAN can be found in the next GHz sensor platforms were available at the beginning
section and in [1]. of this study; thus, these platforms were used in most
of the tests.
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
7
th th
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 physical connection between the Ethernet network

The networking hardware used in this study included and the IEEE 802.15.4 wireless network is done at the
sensor nodes, a Linksys WRT54GL router, and an edge router (in Fig. 5). Here is where the 6LoWPAN
Ethernet to IEEE 802.15.4 edge router. A schematic header compression & decompression is performed to
depiction of the experimental setup is provided in Fig 5. the passing IPv6 packages. This can also be viewed in
To achieve a network setup that was compatible with Fig 6, where a stack view of the experimental setup is
IPv6, some reconfiguration was necessary. Currently, depicted. As a comparison to Fig. 6, Fig. 7 illustrates
every internet service provider in Sweden does not an example of how vendor specific products currently
supply native IPv6 support. Unfortunately, the ISP that connect to the internet, using proprietary protocols,
was available at the test site did not support native which makes them incompatible with devices from
IPv6. However, IPv6 internet can be accessed by other manufacturer. The scalability is hence limited to
constructing an IPv6 to IPv4 (6to4) tunnel using devices from the same manufacturer, which tends to be
customized firmware for the Linksys broadband router. a short tem solution.
In this study, such a tunnel was established in the
router to supply global IPv6 functionality to the LAN Sensor platforms
side. A router advertisement daemon (radvd), was also The sensor platform used in the study, is the Mulle v5.2
installed on the router; thus, IPv6 enabled devices were [10]. Other similar available sensor platforms are
configured through stateless autoconfiguration. As a among others Micaz [11], AVR Raven [12] and
result, full functional IPv6 internet access was provided Sensinode [13]. We choose to work with the Mulle
to all devices. platform since it has a very good performance to
energy ratio.
8
th th
To obtain IPv6/6LoWPAN functionality in the Mulles, a part of the 6LoWPAN standard. Additionally, TinyOS
the lightweight operating systems Contiki [14] and uses short addressing, while Contiki employs long
TinyOS [15] have been successfully ported to the Mulle addressing. The type of addressing and header
platform. Both operating systems were specifically compression used by the OS can be changed, but in
designed to be compatible with resource limited this particular test, default settings were used.
embedded systems such as Mulle. Moreover, Contiki
For payload sizes greater than 60/90 bytes, the IP
and TinyOS both support IPv6 and 6LoWPAN.
packet had to be divided into two separate 802.15.4
However, TinyOS was selected for this study because
frames because the maximum frame size of IEEE
stability issues due to edge-routing problems with
805.15.4 is 127 bytes. The separation of IP packets
Contiki.
increased energy usage and decreased the expected
lifetime of the sensor. Thus, software developers
Sensor platform energy usage
should consider the maximum frame size if absolute
Obtaining an acceptable life expectancy is one of the maximization of sensor lifetime targeted. However
biggest challenges to battery powered, wireless increased payload sizes can of course be
devices. In Sweden, heat meters are inspected every 5 compensated with a larger battery.
to 10 years, depending on the size of the meter. The
life expectancy of wireless devices should be As shown in Fig. 8, the fixed transmission interval was
equivalent to the inspection period to avoid frequent set to 15 minutes, and the effect of transmission
and expensive battery replacements. All sensor nodes interval on the expected lifetime of the sensor was
do however not need to be battery powered. In the analyzed. Additionally, sensor lifetime was evaluated at
case of available electric power in close proximity, e.g. various transmission frequencies and a fixed payload
for platforms mounted in pumps or valves there is no of 80 bytes, as shown in Fig. 9. In accordance to theory
explicit need for batteries since there are electricity the results indicated that a low transmission frequency
available. At other sensor platforms, battery power is has a positive effect on sensor lifetime. In the case of
the only feasible solution, for instance outdoor context aware sensors, which only transmit data when
temperature sensors. required e.g. when a measured temperature exceeds a
set threshold, sensor life expectancy will in most cases
To determine the amount of energy used by a wireless be increased. However, the impact of the
sensing device, the current at the sensor platform sleep/standby energy usage will make up a larger
associated with IPv6/6LoWPAN communication was percentage of the total energy usage, which hence will
measured. To measure the current used by the device, mean that the importance of keeping the sleep current
a 1 ohm high precision resistor was connected in series low will be even bigger.
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 Fig. 8. The effect of payload size on the expected lifetime
of a sensor platform at a transmission rate of
different sizes at consecutive time intervals to observe
4 transmissions per hour (1 to 100 bytes).
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
9
th th
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
Fig. 9. The effect of transmission frequency on the platforms during the spring/summer of 2010.
expected lifetime of a sensor platform at a payload
A lifetime of 10+ years can be achieved with 500 mAh
of 80 bytes.
battery and an average transmission interval of
The predictive life expectancy calculations did not take 15 minutes using IPv6 compatible communication;
into account the fact that batteries loose energy over thus, the life expectancy of battery powered sensors
time, even if they are not in use. Depending on battery did not have a negative effect on integration.
type, this can significantly reduce the expected lifetime
of a sensor. CONCLUSION
Integrating an IPv6/6LoWPAN wireless network in a
SENSOR INTEGRATION district heating substation can significantly increase the
To provide wireless accessibility to devices in the functionality and scalability of the substation and supply
district heating substation, some simple interface new services to both producers and consumers.
electronics were developed to integrate Mulle with Using an open, well documented, and tested protocol
device hardware. As shown in Fig. 10, a heat meter increases the possibility of interoperability between
was integrated with a Mulle in the bottom module products of different manufacturers. This study
location. revealed that available technology can be used to
When digital communication interfaces were available achieve IP-based wireless communication. However, a
(heat meter and circulation pump), the corresponding considerable amount of work on smart application
application protocols were kindly provided by the layers must be conducted before wireless sensor
vendors (Kamstrup and Grundfos). The control valve networks in district heating substations can be
(Siemens SQS-65) was not equipped with any digital deployed and used to its full potential.
communication interface; however, an analog 010 V
input used to control the position of the valve and a FUTURE WORK
010 V output used to read the position of the valve
were available. 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 DPWS1 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.
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.
1
Device Profile for Web Services
10
th th
REFERENCES [6] F. Wernstedt, P. Davidsson, and C. Johansson,

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

OF PARALLEL BURIED PIPES FOR DISTRICT HEATING
1 2
I. Weidlich , M. Achmus
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 Fig. 1. Typical trench condition for DH-pipes after
small-diameter pipes. With larger pipe diameters a FLOSS [2]
stress increase was identified between the pipes.
However, with regard to the average radial pressure The distance between the two pipes depends on the
only slight differences between single pipes and requirements of the laying technique and procedure.
parallel buried pipes were found. For small distances between the two pipes an
interaction between the two pipes is to be expected.
INTRODUCTION
PREVIOUS WORK
As a part of the underground infrastructure of modern
Previous theoretical investigations were based on the
settlements, district heating pipe networks are an
calculation method developed by Leonhardt, taking into
important medium of economic heat transportation. Hot
consideration the deformation behaviour of pipe and
water is pumped in a flow pipe from the supply station
soil and their influence on each other [3]. Leonhardt
to the consumer at a high temperature and under high
introduced the shear resistant beam on elastic
pressure, and the used water is pumped back to the
bedding theory, in which the backfill above the pipe is
supply station in a return pipe.
considered to be a shear resistant beam, which is able
For buried district heating pipes the earth pressure on to transfer shear loads, but no bending moments.
the pipe, respectively the radial contact pressure, is an Using this model it is possible to determine the shear
important value for the design, since it affects the forces activated by the deformation of the shear
friction forces which may be mobilized. The friction resistant beam caused by different stiffnesses of the
forces determine the axial deflections of the pipe and pipe and the surrounding soil, which leads to a
the distribution of normal stresses, which are induced redistribution of stresses in the soil with corresponding
by the temperature loading of the pipe. According to concentration factors .
the European Standard EN 13941, the normal stress
For practical application in Germany regulation
on the pipe coating is calculated for single pipe trench
ATV A 127 was published employing Leonhardts
conditions dependent on the overburden weight of the
theory for buried pipes [4]. This regulation can be
soil, the diameter, the pipe weight and an earth
applied analogously to all kinds of buried pipes. The
pressure coefficient [1]. However, in practice flow and
special application of regulation ATV A 127 for buried
return district heating supply pipes are buried side by
preinsulated district heating pipes was first investigated
side in the same trench. Fig. 1 shows a typical situation
by Beilke [5].
for buried district heating pipes according to Floss [2].
12
th th
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. 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,
Fig. 2. Shear resistant beam for parallel buried pipes 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
The concentration factors B. B for the area beside and flow and return pipes installed on the same
between the pipes were found with the following underground level. The presented study was therefore
assumptions: carried out without employing the Hornung and Kittel
calculation method.
The influenced area for the determination of
the concentration factor B beside the pipe
(B1) is defined by a line with an inclination of NUMERICAL INVESTIGATIONS
60 shown in Fig. 3. This angle coincides with Numerical calculations were carried out with the two
the theoretical slope inclination of a non- dimensional finite element program PLAXIS, version
cohesive soil with an internal angle of friction 8.6. Two standard situations with different outer pipe
of ' = 30. diameters D (DN65, D=140 mm; DN250, D= 400 mm)
of two parallel buried district heating pipes were
Between the pipes (B2) the full interspace is
investigated. The distance between the pipes was
taken as the area of influence.
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. 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 Fig. 5. Finite element mesh for the case DN65, H/D=3
pipes and the concentration factor B1 for parallel The installation process was simulated by a staged
buried pipes. However, a significant reduction of the construction process, considering a retained trench
vertical stresses (i.e. B2<B) was determined between and the backfilling procedure with several layers. The
the pipes. As an example, the ratio of the stress factors compaction process was accounted for by applying a
is shown in Fig. 4, dependent on the relative static distributed load of p=10 kN/m on each of the
overburden height H/D and the relative distance of the layers. Ground water was not considered in this
pipes A/D. investigation.
13
th th
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 0.3 Fig. 7. Horizontal effective stresses h (DN250 pipe,
Internal angle of friction [] 40 H/D=3)
Angle of dilatancy [] 10 In Fig. 8 the stress concentration is shown by the
Interface friction Rinter [1] 0.536 distribution of the radial contact pressure for the left-
hand pipe. In the springline area a maximum value of
Between pipe and soil, the Coulomb friction law with a r=21.44 kN/m for the radial pressure was obtained.
Compared with the calculated average radial pressure
of
according to Eq. (1).
r,avg,calc= 18.81 kN/m the deviation is about 12.2%.
tan i Rinter * tan ' (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. 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).
D 1 k
r , avg,13941 * H * (2)
2 2
In Table II the results of the numerical investigation are

Fig. 6. Contact pressure on the DN65 pipes, H/D=3 compared to the expected radial pressure from the DIN
EN 13941 regulation.
However, in the second numerical model of pipes with
Table II. Average contact pressure r,avg for H/D=3.0
an outer diameter of D=400 mm (DN250), a stress
concentration between the pipes was evident. The DN Single pipe according to Parallel buried pipe
distribution of horizontal effective stresses acting after DIN EN 13941 according to numerical
the installation process is shown in Fig. 7. The stresses results
are significantly larger between the pipes than beneath 65 6.15 kN/m 7.25 kN/m
them. 250 17.58 kN/m 18.81 kN/m
Regarding the average radial contact pressure the

difference between the expected values from the DIN
14
th th
EN 13941 regulation for single pipes and the calculated REFERENCES

values from the numerical simulations for parallel
buried pipes is rather small. The values for the parallel [1] DIN EN 13941, Berechnung und Verlegung von
buried pipes are slightly larger. werkmig gedmmten Verbundmantelrohren fr
Fernwrme, Deutsches Institut fr Normung e.V.,
Beuth Verlag Berlin, 2003.
CONCLUSION
[2] R Floss, Handbuch ZTVE-StB 94/1997,
The earth pressure on district heating pipes is an
Kommentar mit Kompendium Erd- und Felsbau, 3.
important design value and should be determined as
Auflage, Kirschbaum-Verlag, 2006.
exactly as possible. In the presented work the earth
pressure on parallel buried pipes was investigated. [3] ]G. Leonhardt, Belastung von starren
Rohrleitungen unter Dmmen, PhD Thesis,
The evaluation of the radial stresses on the pipe in
Institute of Soil Mechanics, Foundation
numerical calculations showed a stress concentration
Engineering and Waterpower Engineering,
between two pipes buried in the same trench for short
University of Hannover, 1973.
pipe distances and large diameters. However, former
theoretical investigations led to a reduction of radial [4] ATV A 127, Richtlinie fr die statische Berechnung
contact pressure between the two pipes, which was von Entwsserungskanlen und -leitungen,
observed in the numerical simulations for small Arbeitsblatt A 127 der Abwassertechnischen
diameters and small overburden heights. Vereinigung e.V., 2000.
Because typical trench conditions with two parallel [5] O. Beilke, Interaktionsverhalten des Bauwerks
buried pipes are not considered in current design FernwrmeleitungBaugrund, Institute of Soil
directives for district heating pipes the numerical results Mechanics, Foundation Engineering and
were compared with the values derived from the Waterpower Engineering, University of Hannover,
current design regulations. For the observed systems 1993.
only small deviations regarding the average normal
[6] V. Rizkallah, M. Achmus, Zur Gre der
pressure between single pipe and parallel buried pipes
Reibungskrfte an erdverlegten Fernwrmelei-
were found. Thus, as long as the exact distribution of
tungen, Forschungsvorhaben Wechselwirkungen
stresses along the pipe perimeter is not of particular
Fernwrmeleitung Bettungsmaterial, Institute of
relevance, current calculation directives are also
Soil Mechanics, Foundation Engineering and
suitable for parallel buried pipes. Only for conditions
Waterpower Engineering, University of Hannover,
with large pipe diameters and small distances between
1993.
the pipes and also relatively large overburden heights
is a significant deviation to be expected. [7] K. Hornung, D. Kittel, Statik erdberdeckter
Rohre, Bauverlag GmbH Wiesbaden und Berlin,
Furthermore, inhomogeneous backfill compaction,
ISBN 3-7625-2039-9, 1983.
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.
15
th th
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 Investments
In some countries the flat station concept is becoming a Reference for comparing the flat system concept with a
common way of realising heating and domestic hot conventional concept is based on modern way of
water (dhw) installation in blocks of flats. Anyhow, in making block pipe distribution systems [1]. In both
other countries it is at the very beginning. Experience cases it is a horizontal pipe layout in flats with a vertical
from those countries reveal a number of questions when pipe tunnel for distribution. Pipe distribution systems are
understanding and evaluating the flat station concept. shown in fig. 1. Main differences are to be seen in the
A number of parameters can be addressed to and be number of pipes installed. Since dhw is prepared
evaluated to disclose qualities and performance of the decentralized in flats, dhw pipe and dhw circulation pipe
flat station concept in relation to traditional concepts for are eliminated. Centrally located dhw station in the
heating and dhw installations. basement is replaced by decentralized flat stations.
Balancing valves for heating as well as for dhw
This paper aims at analysing main parameters distribution is saved for the flat station concept.
regarding quality (comfort) and performance of the flat Regarding metering then the dhw meter is eliminated,
station concept, covering block distribution system, flat since the primary supply to the flat station covers flat
station itself and flat installation. Parameters in focus heating and dhw as well. According to measurements of
are: riser system, instantaneous dhw principles, heat more than 2500 dwellings in Denmark, including
losses, comfort of dhw, investments and energy detached houses as well as multi storey buildings,
savings, metering and hygienic issues for dhw. individual metering, say individual billing, resulted in
savings of 1530%, [2]. Therefore, this analysis
INTRODUCTION assumes metering of all thermal energy deliveries to
Areas of district heating distribution systems, building flats.
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 Fig 1. Pipe distribution systems in blocks of flats. C:
Energy meters Modern reference principle. F: Flat station principle.
Energy Savings: A recent investment example comparing flat station
eat loss in primary distribution system concept (F) with traditional system (C) is included in
dhw circulation pump consumption fig. 2. Data are based on a Danish case from rhus
Comfort: area where a block, built in 4 levels and a basement
dhw temperature stability and variation level consisting of 24 flats, will generally be modernised.
dhw recovery time after idle period Investments compared are based on concepts
Hygienic issues presented in fig. 1. Main conclusion is that investment
considerations on Legionella related to the systems level is approx. break-even for the two systems, for this
physical layout.
16
th th
typical Danish case. For other countries implying other demand for a 1970 Danish block building (not including
components/costs levels, level could change. In energy for dhw).
general, the experience is that flat stations are on
Table 1. Energy losses for traditional system C, and flat
break-even cost level or slightly higher. This is valid for system F based on the rhus case.
new buildings as well as for renovation projects.
Concept Pipe T T E E Net E Energ. Energ.
length pipe amb. loss loss/y loss/y price costs
Investments comparison Flatstation concept
[m] [W/mK] [C] [C] [W] [kWh] [kWh] [ /kWh] [ /year]
rhus Case, block of 24 flats
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
Saved eneregy meter for Flat st. con.F Sum. flow 120 0.20 55 20 840 7358 3679 0.05 184
dhw Flat st. con.F Sum. return 120 0.20 30 20 240 2102 1051 0.05 53
Saved balancing vavles
Trad. con. C Unit heat loss 1 pcs. 300 W/unit 2628 1314 0.05 66
for dhw circulation Flat st. con.F Unit heat loss 24 pcs. 25 W/unit 3816 1908 0.05 95
Saved balancing valves
Item
for heating distribution dhw circ. C Summer 240 0.20 53 20 1584 13876 6938 0.05 347
1 dhw circ. elec. Sum. + win. - - - - 30 260 - 0.25 65
Saved dhw pipes, incl.
circulation Trad. c. C Total 10880 544
Saved dhw preparation Flat st. c. F Total 6638 332
centrally located Diff. C-F Total (ex. electrical consumption) 4242 212
Invested in Flat Station
0
5000
10000
-5000
-20000
-15000
-10000
Secondly, a situation is analysed where heat loss is not

utilised in the building distribution system at all. Winter
Euro
energy losses for the flat station is assumed to be
usable and no floor heating is active during summer.
Fig. 2. Investment balance for traditional system C and flat
system F. Block of 24 flats. Table 2. Energy losses for traditional system C, and flat
Energy savings system F based on the rhus case.
Main contribution to energy saving is originated from Concept Pipe T T E E Net E Energ. Energ.
installed hot distribution pipes. To begin with, it is

length pipe amb. loss loss/y loss/y price costs
assumed that half the yearly distribution energy loss is [m] [W/mK] [C] [C] [W] [kWh] [kWh] [/kWh] [/year]
net loss (summer time), meaning not contributing to 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
heating up the building. Wintertime temperatures are
Flat st. con. F Sum. flow 120 0.20 55 20 840 7358 3679 0.05 184
assumed to be identical for the two concepts, because Flat st. con. F Sum. return 120 0.20 30 20 240 2102 1051 0.05 53
for this period the heating system defines temperature Trad. con. C Winter flow 120 0.20 70 20 1200 10512 5256 0.05 263
levels. To quantify losses a room temperature of 20 C Trad. con. C Winter return 120 0.20 30 20 240 2102 1051 0.05 53
is assumed. Danish Technical Insulation Standard [3] 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
requires minimum allowable heat loss constants (W/m),
Trad. con. C Unit heat loss 1 pcs. 300 W/unit 2628 2628 0.05 131
depending on temperatures, annual operation time and Flat st. con. F Unit heat loss 24 pcs. 25 W/unit 3816 1908 0.05 95
pipe diameter. These constants turn out to be quite dhw circ. C Sum. + win. 240 0.20 53 20 1584 13876 13876 0.05 694
similar to all pipes in question. To simplify preconditions dhw circ. elec. Sum. + win. - - - - 30 260 - 0.25 65
a heat loss coefficient of 0.20 W/mK has been chosen Trad. c. C Total 22811 1141
Flat st. c. F Total 12946 647
for all hot pipes. Table 1 shows a comparison of pipe
temperatures, heat loss and electrical dhw circulation Diff. C-F Total (ex. electrical consumption) 9865 493
pump.
Flats in this first case are provided with floor heating in Comparing the two systems regarding heat loss, then
bathrooms; therefore, heating is active all year. Due to favour is again towards the flat station concept. For the
floor heating, temperatures for the traditional concept rhus case it means approximately 9900 kWh/year
are lower during summer season compared to the flat savings corresponding to 490 Euro/year (ex. pump.
station concept, since floor heating typically operates at costs). This means a saving of approx. 4 kWh/m2/year.
lower temperatures. For the flat station concept a dhw This represents a saving of approx. 4% of the yearly
temperature at 45 C is assumed, demanding a primary heat demand for a 1970 Danish block building.
temperature of 55 C. Additionally, as for the flat station concept there is no
need for dhw circulation pump, thus no need for the
Comparing the two systems regarding heat loss, then electric energy of 260 kwh/year. A part of this saving is
favour is towards the flat station concept. For the rhus anyhow spent for the flat station concept due to
case it means approximately 4200 kWh/year savings additional circulation of primary water. It is assumed that
corresponding to 210 Euro/year (ex. pump. costs). This this is approx. half the electric energy for dhw circulation
means a saving of approx. 2 kWh/m2/year. This pump of 130 kwh/year.
represents a saving of approx. 2% of the yearly heat
17
th th
When looking at annual energy consumption savings in dynamics are heavily influenced by idle bypass
percent, figures might appear rather low and of minor thermostat setting. Also pump dynamics are influencing,
impact. In this respect it has to be remembered that meaning how fast is the primary circulation pump
energy saving relates to a typical 1970 building. reacting on rapid changes of hydraulic conditions, say
Present building codes require energy savings in the opening of primary valve.
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 Fig. 4. Dynamic control performance (idle recovery) for
to mixing of temperature layers. A typical question thermostatic and pressure controlled heat exchanger for
regarding instantaneous prepared dhw is how stable are dhw production. Heat exchanger is cold during idle. [6]
temperatures when applying dynamics. Regarding
dynamic control performance an example is included in Fig. 4 shows a flat system with cold heat exchanger.
fig. 3: 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
Fig. 3. Dynamic control performance (step test) for reach a minimum demanded level of 45 C. After
thermostatic and pressure controlled heat exchanger for
dhw production [6] 5 meter of pex pipe of 22 mm additional delay is
approx. 7 sec. By this the total delay from tapping the
Fig. 3 shows that stability, temperature peaks at load start to reach 45 at the tap is approx. 17 sec. In this
change and total dhw temperature (T22) variation is example a very long idle branch pipe length is used,
limited to 34 C. Regarding oscillations at low tapping more realistic would be 02 m, resulting in a primary
flow it should be noted that T22 is measured at heat side delay of not more than a few seconds. Also
exchanger outlet. As example a 5 m 22 mm pex pipe diameter of secondary dhw pipe is rather big and
reduces peaks and amplitudes additionally, dependent represents a typical shared pipe dimension,
on frequencies, but typically 50%. This example is for representing one pipe for several taps.
very high primary supply conditions. Oscillations appear
at tap flow of 100 l/h or below. This level shall be seen Anyhow, this delay is only relevant for the first tapping,
in relation to the fact that a typical tapping flow for one since thermal capacities combined with efficient
tap is 200400 l/h. insulation is maintaining temperature, typically with time
constants of 12 hr.
Another relevant question is how fast dhw temperature
is on desired level if supply is in idle condition. Here
18
th th
Comfort level is increased by applying a higher bypass For simulations a branch pipe flow (Q1) of 800 l/h is
thermostat setting and/or a hot heat exchanger during assumed. This represents a situation where the
idle. Fig. 5 shows an example of flat station with hot thermostat is fully open until the desired set temperature
heat exchanger and thermostatic controlled heat is reached. Further a step vice flow change from zero to
exchanger [7]. Idle temperature is approx. 50 C Q1 or zero to Q2 is assumed. Tapping flow is assumed
corresponding to dhw tapping temperature. 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.
Delay until reaching 45C

L2=5m & 10m (internal 10mm) - Heat Exchanger hot & cold at idle
16
hot - dt at T2 - L2=5m hot - dt at T3 - L2=5m
14 hot - dt at T4 - L2=5m hot - dt at T4 - L2=10m
cold - dt at T3 - L2=5m cold - dt at T4 - L2=5m
12 cold - dt at T4 - L2=10m
10
dT [sec]
8
2
Fig. 5. Dynamic control performance (idle recovery) for
thermostatic controlled heat exchanger for dhw production. 0
Heat exchanger is warm during idle. [7] 0 1 2 3 4 5 6
L1 [m] (internal 20mm)
Fig. 5 shows a flat system with hot heat exchanger at
idle. Bypass temperature setting corresponds to a Fig. 7. Dynamic simulation for hot and cold heat exchanger
primary supply temperature (T11) of 58 C and primary during idle. Delay (dt) for dhw temp. of 45 C.
return temperature (T12) of 44 C. This setting is the
high end, meaning in high end regarding comfort. For Heat exchanger simulated is Danfoss XB06H-40 [6]. It
this system there are no primary delays, and dhw can be seen from figure 7, that influence on hot or cold
tapping temperature at the flat station is available after heat exchanger is in the range of 2 sec. delay. Branch
approx. 2 sec. Additional delay due to dhw piping pipe length (L1) has minor impact on time delay. This is
towards tap would be similar to previous example. due to the fact that temperature is maintained with a
temperature gradient along pipe during idle, reflecting
In many practical matters a compromise between the T1 to T2. Basically water in branch pipe is heated to a
two examples regarding idle temperature setting fulfils certain level already before tapping. Anyhow, due to
demands for good comfort with reasonable energy energy loss and return temperature, idle bypass
consumption. temperature is lower than dhw tapping temperature in
In the following a general trade off is included between this case.
branch pipe length, dhw pipe length, idle condition for Main influence on time delay is dhw pipe diameter and
heat exchanger and temperature delay on dhw, based length (L2). Connection in flats shall be of star
on dynamic simulations. Pipes are simplified by simple coupling principle where every tap has its own supply
delay models with no heat loss. Heat exchanger is pipe with a small inner diameter. Temperature in dhw
based on a lumped capacity model described in [5]. 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:
Fig. 6. Basic application for flat station, including boundary

conditions for dynamic simulations.
19
th th
Delay until reaching 40C rather Legionella bacteria concentration in the dhw.
L2=5m & 10m (internal 10mm) - Heat Exchanger hot & cold at idle Facts influencing on potential for Legionella
16
hot - dt at T4 - l"=5m hot - dt at T4 - L2=10m
concentration growth are dhw temperature, exchange
14 cold - dt at T4 - l2=5m cold - dt at T3 - L2=5m rate of hot water in distribution pipes, and volume of
12
cold - dt at T4 - L2=10m dhw water in the entire hot system. Also other factors
are influencing, e.g. systematic cleaning of shower
10
outlets, but this will be not addressed to here, since the
dT [sec]
8 effect is similar for concepts compared.

6
Comparing volumes of dhw in pipes for concepts, the
4 flats station solution has significantly lower volume
2 compared to the traditional system. Furthermore dhw
0
pipes should be star connected, meaning one small
0 1 2 3 4 5 6 (diameter) pipe from the flat station to each individual
L1 [m] (internal 20mm) hot tap. This eliminates problematic dead end or low
flow areas.
Fig. 7. Dynamic simulation results for hot and cold heat
Typically volume of heat exchanger is 0.25 to 0.50 litre.
exchanger during idle. Delay (dt) for dhw temp. of
Typical dhw pipe volume is 0.10 l/m, equal to 1.0 litre for
40 C.
10 m pipe. In total this is a volume of 1.5 to 2 litre
pr. flat. The comparable centrally placed dhw system
First of all it can be seen that time delay for reaching with dhw distribution will have a volume of 57 litre pr.
40 C at tap is only a bit shorter than reaching 45 C. flat. By installing a dhw storage tank this will increase
This is due to the fact that the T4 temperature profile significantly. The German DVGW regulations states that
has an almost step vice nature, i.e. if temperature goes heating dhw up to 60 C, due to e.g. Legionella, is not
up after the dhw pipe is flushed through, it goes almost required if volumes of heat exchanger or volume of dhw
like a step. pipes is less than 3 litres [8]. Based on those physical
concept differences Legionella bacteria risk is reduced
Different dhw controllers have different performance for the flat station concept.
regarding time delay. In case of pure proportional
control for dhw system, time delay is longer at part load. Future energy supply/demand perspective
This is because primary flow is proportional to One important challenge for DH is to convert to 4th
secondary flow, and the lower the flow the longer the generation DH systems. Intention is to realise efficient
waiting time. Looking at the example for Q1=800l/h, DH systems for urban areas where heat demands will
Q2=400 l/h, L1=4 m, L2=5 m then time delay (dt) at T4 decrease due to modernisation and new building energy
is 6.9 sec to reach 45 C dhw temperature. In case of saving codes. In this context one way to go is to reduce
proportional controller with parameters Q1=400 l/h, temperatures in DH networks [9]/[10]. This allows for
Q2=400 l/h, L1=4m, L2=5 m then dt=11.0 sec to reach cost effective geothermal sources as well as other
45 C. This has considerable effect on time delay as L1 renewable low temperature sources. For dhw, normal
gets longer. In case of a thermostatically controlled dhw temperature level is 45 to 60, where higher
system or a combination of a thermostatically and temperatures typically are based on considerations
proportionally controlled dhw system, time delay is towards Legionella. A way to reduce temperature levels
shorter because no matter how small tapping is, as long in DH networks is to set dhw temperature at 45 C. By
as the desired set point temperature is not reached, the this a primary temperature at sub station of 50 to 55 C
primary valve will be fully or almost fully open resulting will be sufficient. A precondition for this is to use heat
in high primary flow. Regarding delay to reach a dhw exchangers for dhw production, like the flat station
temperature of 40 C this is only related to dhw pipe concept.
dimension since 40 C is the bypass temperature if heat
exchanger is hot during idle. In case the heat exchanger
CONCLUSION
is cold during idle, then this introduces an additional
time delay as described above. In all cases, time delay The two pipe flat station concepts, consisting of
is dependent on dhw flow, resulting in delay in the dhw decentralised instantaneous dhw production, open the
pipe. possibility of reducing general DH net work temperature,
which for the future will be even more relevant due to
Hygienic considerations decreasing building heat demand and increased
Legionella is a well-known bacterial risk for dhw availability of renewable energy. For building owners,
systems. Normally it is not the question whether the investigated case shows that the flat station concept
Legionella is present in the dhw system or not, but is on brake-even investment level compared to
20
th th
traditional systems. The flat station concept has a net [5] Thorsen, J. E. Control Concepts for DH Compact
energy saving due to less installed hot pipes. Energy Stations Investigated by Simulations, The 9th
savings are in the range of 2 to 4 kWh/m^2/y for the International Symposium on District Heating and
investigated cases. Comfort level has been Cooling 2004.
investigated, revealing well acceptable dynamic control
performance. Dhw temperature recovery after an idle [6] http://www.danfoss.com/Products/Categories/List/H
period for the instantaneous preparation of dhw is, E/Temperature-Controllers/Temperature-
however, a trade-off between comfort and energy controllers/IHPT-and-XB-06/b1c8a73c-59f1-4fef-
saving. Related to Legionella, then risk can be reduced 8b52-f49c97b6019b.html
when installing flat stations as presented in this paper. [7] http://www.danfoss.com/Products/Categories/Group
/HE/District-Heating-Substations/Substations-
REFERENCES Direct-Heating/Flat-Stations/8f81605b-bab9-4644-
961b-51a3f0503f05.html
[1] Kristjansson, H. Comparing Distributions Systems
in Blocks of Flats, SDDE 2009, Slovenia [8] DVGW regulations, Germany, Arbeitsblatt W551,
April 2004
[2] Gullev, L., Poulsen, M. The Installation of Meters
Leads to Permanent Changes in Consumer [9] Olsen, P.K., Lambertsen, H., Hummelshj, R.,
Behaviour, the magazine News from DBDH, Bhm, B., Christiansen, C.H., Svendsen, S.,
#3/2006. Larsen, C.T., Worm, J. A new Low-Temperature
District Heating System for Low-Energy Buildings,
[3] DS 452, Code of practise for thermal insulation of The 11th International Symposium on District
technical service and supply systems in buildings, Heating and Cooling 2008.
2. Revision, Dansk Standard, 1999
[10] Paulsen, O., Jianhua, F., Furbo, S., Thorsen, J. E.
[4] Thorsen, J.E. Cost considerations on Storage Tank Consumer Unit for Low Energy District Heating Net
versus Heat exchanger for htw preparation, The Works. The 11th International Symposium on
10th International Symposium on District Heating District Heating and Cooling 2008.
and Cooling 2006.
21
th th
IMPROVED TEMPERATURE PERFORMANCE OF RADIATOR HEATING SYSTEM

CONNECTED TO DISTRICT HEATING BY USING ADD-ON-FAN BLOWERS
1 2
Per-Olof Johansson , Janusz Wollerstrand
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 level. In order to reduce the temperature demand in

existing buildings the idea of using small add-on-fan
District heating (DH), which is the most common heat
blowers placed under the radiator to increase the heat
source in multifamily houses and commercial buildings
output due to an increased share of forced convection
in Sweden, can be produced in several different type of
came up.
production units.
In order to gain thermal efficiency in a DH system it is Objective
important that DH supply and return temperatures are The field study presented in this paper investigates the
kept low. The temperature demand in the DH system is, possibility to reduce the space heating temperature
during the heating season, dependent on the program and estimates the impact on the DH supply
temperature level in the heating system of the DH and return temperature. Possible reduction of the DH-
connected buildings. Many production units benefit from flow rate is also calculated.
a lowered DH return temperature, while others are more
affected by a reduced supply temperature. In a CHP- This paper is focusing on buildings indirectly connected
station the heat to power ratio will increase when the to the DH network through a substation with heat
DH supply temperature is decreasing. In order to reduce exchangers (HEX).
the temperature demand, low temperature heating
systems are of interest, as well as systems resulting in a DESCRIPTION OF ADD-ON-FAN BLOWER
low DH return temperature.
The add-on-fan blower that is tested in this study
To increase the heat output in an existing radiator consists of several regular DC motor driven fans,
heating system, the radiators can be complemented originally used for cooling, mounted under a radiator,
with small electric fans resulting in an increased share see Fig. 1. In the study, two different kinds of radiators
of forced convection in the heating system. Field studies were tested, a panel and a column radiator.
have shown that the heat output, with constant supply
temperature and mass flow through the radiator, can Tss
increase with more than 50%.
ms
INTRODUCTION
Space heating
Outer wall
For many years, return temperatures in DH networks radiator

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 Tsr
electricity produced in a CHP station could increase. Add-on-fan
This would lead the way towards an increased share of
Increased air flow
electricity produced by non fossil fuels. In Sweden more
than 30 % of the DH is produced in CHP stations [4]. Floor
In many reports the gain from a reduced temperature
Fig. 1 The add-on-fan blower mounted on a panel radiator.
level in the DH network has been discussed and
quantified in economic terms, see e.g. [12], [13]. The add-on-fan blowers in this study are provided by a
Swedish company: A-energi AB (the product is called
The DH supply temperature level in the DH network is, flktelement in Swedish). The company describes the
during heating season, dependent of the temperature features of the add-on-fan blower as a possibility to
demand in the DH-connected buildings heating system. reduce the temperature program without replacing the
In modern buildings low temperature heating systems radiators, with the aim to reduce the electricity demand
are common, which may allow reduced DH temperature for buildings supplied with heat from heat pumps [5].
22
th th
THEORY Qrad C Tm3 (7)

In this section a theoretic analysis of the impact of
forced airflow on heat output from radiator with a length Convection
of (L) 1 m and height of (h) 0.59 m is described. The The convection that arises due to the temperature
radiator is in this study approximated by a flat vertical difference between the radiator surface and the
plate. The indoor temperature is assumed to be surrounding air is a function of the Nusselt number (Nu),
constant at 21C and equal to Tinf. see equation 8.
conv Nu / h (8)
Heat output
The heat output from the radiator to the room arises The heat output due to convection is divided into three
from convection and radiation. The heat transfer sections, natural, mixed and forced convection.
process from heating water to the room through a
radiator is summarized in equation 1 [7], [8]. For natural convection, the Nu number is dependent on
the Rayleigh number (Ra), which is a product of the
s c p (Tss Tsr ) (k A)
Qm (1) Prandtl number (Pr) and the Grashof number (Gr). For
air, Pr can be considered constant, Pr=0.71, wile
k is the heat transfer (convection) from the water to the
surrounding metal, conduction through the metal and h3
Gr g (9)
convection from the outer surface of the radiator to the 2
room according to equation 2. 1 / Tinf 1 / Ti
1 1 1 (2) where g is the gravity force, is kinematic viscosity and
metal
k watermetal metal conv rad is the coefficient of expansion.
The dominating parameters in this equation are the Several empirical relations describing Nu are available.
convection and radiation between the radiator and the In this study a relation described by Churchill has been
room (conv and rad), while the other terms, in this case, used [9], see equation 10 and 11.
can be neglected. This results in a new equation for
0.67 Ra 0.25 Ra 109 (10)
energy output, see equation 3. Nu 0.68
[1 (0.492 / Pr)9 / 16 ]4 / 9
Q Qrad Qconv conv Aconv rad Arad (3)
0.5 0.387 Ra 1 / 6 Ra 10 9 (11)
Nu 0.825
The temperature is the logarithmic mean [1 (0.492 / Pr) 9 / 16 ]8 / 27
temperature difference according to equation 4.
For forced convection the Nu number is calculated by
T Tsr (4) equations described by Holman [10], see equations 12
ss
T Ti and 13.
ln ss
Tsr Ti
Nu 0.664 Re 0.5 Pr1/3 Re 5 105 (12)
Radiation Nu Pr 1/3 (0.037 Re 0.8 871) 5 105 Re 10 7 (13)

According to Trschel [8] the heat output from radiation and the Reynolds number, Re, is described as:
can be estimated according to equation 5.
uL (14)
Qrad rad Arad Re

rad (5)
4 Tm3 Arad The product of Gr/Re2 describes the dominating type of
Arad
rad (1 rad ) convection. If Gr/Re2>10, natural convection is
Aradiator
dominating, if Gr/Re21, both natural and forced
Where the temperature, Tm, is the mean temperature of convection is of importance and if Gr/Re2<<1, forced
the radiator surface and the surfaces in the rooms, see convection is dominating. When a mix of forced and
equation 6. For a panel radiator the Arad/Aradiation=1 [8]. natural convection occurs, the Nusselt number is
calculated according to equation 15 [11].
(Tsf Tsr ) / 2 Ti
Tm Tradiator Troom,surface
2 (6) 3 3
Nu ( Nu forced Nu natural)1 / 3 (15)
Since the Arad and emissivity, rad, are constant for a
specific radiator, the relation can be simplified to
equation 7.
23
th th
Impact of air speed on space heating temperature T ss0= 60 C , Qrel= 100 %

60
Results from the theoretical analysis, using the T sr0= 45 C , Qrel= 100 %
equations above, are shown in Fig. 2 to Fig. 4. The heat 55 T ss0= 43.1 C , Qrel= 50 %
output for a radiator designed for the temperature 50

T sr0= 35.6 C , Qrel= 50 %
program 60/45 C is illustrated as a function of the air T ss0= 33.6 C , Qrel= 25 %
Temperature ( C)
T sr0= 29.9 C , Qrel= 25 %
speed in Fig. 2. The supply temperature and the mass 45
flow through the radiator are kept constant. Two cases 40

have been derived, one with heat output only from
35
convection, and one with heat output from both radiation
and convection. With =0.9, the share of heat output 30
from radiation will be 65% at DOT.
25
T sf ( C)
20
T sr,conv ( C) 0 5 10 15 20 25 30
air velocity (m/s)
% additional Q, conv
T sr,rad&conv ( C) Fig. 3 Possible Tss and Tsr to for three heat load situations
% additional Q, rad&conv
at different air speeds. Qrad=65% at DOT.
150 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
C and %
100
relative heat load.
50 NEW TEMPERATURE PROGRAM
65
60 T sf0 U= 0.0m/s
0
0 2 4 6 8 10 12 14 16 T sr0 U= 0.0m/s
55
air speed (m/s) U= 0.5m/s
30 50 U= 1.0m/s
Temperature ( C)
U= 2.0m/s
(W/m K)
20 45 U= 3.0m/s
2
10 40
35
0
0 2 4 6 8 10 12 14 16 30
air speed (m/s)
25
Fig. 2 Calculated heat output improvements at Tss=60 C, 20

Tsr0=45 C with increasing airspeed. ms is kept constant.
15
0 20 40 60 80 100
As seen, the additional heat output from the radiator is relative heatoutput (%)
increasing rapidly when the air flow is increased. With Fig. 4 New space heating temperature programs at
radiation taken into account, the increase is somewhat different air speeds. Red lines: Tss, Blue lines: Tsr.
lower since the mean temperature, Tm, is decreased, Qrad=65% at DOT.
see equation 7. In the calculations performed, the radiator is assumed to
In Fig. 2 the heat output is increasing. In Fig. 3 and have the same heat output from both sides of the
Fig. 4 the supply temperature to the radiator is reduced radiator. The air flow is assumed to be uniformly
instead to keep the heat output constant. New Tss and distributed through the length and height of the panel
Tsr can now be calculated under the assumption that the radiator. This is not the case in the real add-on-fan
total heat output and the mass flow (ms) through the blower applications, however, one can expect results
radiator are constant. The impact of the air flow is following the same pattern.
described for three different heat loads (Qrel=100%,
50%, 25%) with standard 60/45 C temperature EXPERIMENTAL STUDY
program as a reference. See Fig. 3.
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 (ms) 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
24
th th
radiators in the building is running at 60/45C at DOT * Calculated for new radiators of the same dimensions
(represented by narrow black lines in Fig. 8 trough manufactured by Lenhovda radiator factory [3]
Fig. 13). The radiators are located in traditional office
The radiator types tested were: environment in a building built in 1960.
Panel radiator, see Fig. 5 Data acquisition

Measured parameters in the test were: secondary
Column radiator, see Fig. 6.
supply and return temperature (Tss and Tsr), indoor
temperature (Ti) and outdoor temperature (Tout).
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
Tsr,0=39 Tsr,Fan=34
Fig. 5 Add-on-fan blower mounted on a panel radiator.

Tss (C)
Tsr (C)
Ti (C)
Tout (C)
Ufan (V)
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 Tpr is
Fig. 6 Add-on-fan blower mounted on a column radiator. decreasing, causing an additional heat output since ms
For each radiator the fans have been run at two is kept constant. See Fig. 8.
different rotation speeds. The net electric power
60
consumption (Pfan) has been measured. See Table 1 for Tss
Pfan and the design heat energy output at DOT. Note 55
Tsr,0
that the electric power to the add-on-fan blower is 50 Tsr,Fan
constant and not dependent on the relative heat output.

Temperature [C]
45
Original cooling in radiator
Table 1 Radiator and add-on-fan blower design. 40
Additional cooling with
Radiator Pfan Q @ DOT, 60/45C 35 add-on-fan-operation
type (el) (Heat) 30
Panel 2.7 W 360 W* 25

1.9 W
20
0 0.2 0.4 0.6 0.8 1
Column 3.0 W 430 W* Relative heatload
2.2 W
Fig. 8 Increased cooling of secondary system.
25
th th
The relative heat output from the radiator with and Panel radiator Pfan = 2.7 W
without fan operation is calculated from equation 1. 60

Tss,0
55
Qrel ,0 m s C p (Tss Tsr 0 ) QDOT (1)
Tsr,0

Qrel ,Fan m s C p (Tss Tsr ,Fan ) QDOT 50 Tss,Fan
Tsr,Fan
45
Since the temperature drop in the radiator is increasing
T [C]
with the fan in operation, the radiator now could be 40
s
considered oversized. Then, with the same type of
35
reasoning as in e.g. [2], the Tss program or ms needs to
be adjusted in order to avoid overheating of the building. 30
In this paper, the ms has been considered constant, 25

allowing us to compute the new relative space heating
load for a given Tss according to equation 2. Qrel,0 is 20
0 0.2 0.4 0.6 0.8 1 1.2 1.4
computed using the original space heating temperature Qrel
program.
Fig. 10 Modified temperature program for panel radiator,
(T Tsr ,Fan ) Pfan=2.7 W.
Qrel ,Fan ss Qrel ,0 (2)
(Tss Tsr 0 ) Panel radiator Pfan = 1.9 W
60
Knowing Qrel,Fan, a new temperature curve, which will Tss,0
result in correct heat output from the radiator with the 55 Tsr,0
fan in operation, can be calculated. The curve appears 50 Tss,Fan
to the right in the diagram, see Fig. 9. Tsr,Fan
45
T [C]
60 40
s
T
ss,0
55 T 35
sr,0
50 Tss,Fan
30
T
sr,Fan
45 25
s
40 20
T
0 0.2 0.4 0.6 0.8 1 1.2 1.4

35 Qrel
30 Fig. 11 Modified temperature program for panel radiator,

Pfan=1.9 W.
25 Column radiator Pfan = 3.0 W
60
20 Tss,0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Qrel 55 Tsr,0
50 Tss,Fan
Fig. 9 Modified secondary temperature program. Tsr,Fan
45
New space heating temperature program - results
T [C]
The procedure described above has been applied to all 40

s
collected data. Results are shown in Fig. 10 and Fig. 11 35

for the panel radiator, and Fig. 12 and Fig. 13 for the
column radiator. 30
As seen in the figures, the temperature program is now 25
significantly reduced for both the panel and the column 20

0 0.2 0.4 0.6 0.8 1 1.2 1.4
radiator. The new temperature program shows a similar Qrel
pattern for both types of radiators. For the panel
radiator, the measured return temperature is more Fig. 12 Modified temperature program for column radiator,
concentrated, especially at the higher fan speed. This Pfan=3.0 W.
could be explained by a more favorable air flow pattern
due to the physics of the radiator.
26
th th
Column radiator Pfan = 2.2 W DH primary supply temperature

60 110
Tss,0
55 Tsr,0
100
50 Tss,Fan
Tsr,Fan
45 90
Tps [C]
T [C]
40
s
80
35
30 70
25
60
20 1 0.8 0.6 0.4 0.2 0 0
0 0.2 0.4 0.6 0.8 1 rel heatload
Qrel
Fig. 13 Modified temperature program for column radiator, Fig. 14 DH primary supply temperature.
Pfan=2.2 W.
Results
INFLUENCE ON DH NETWORK
The first control strategy is in Fig. 15 Fig. 20 noted as
Knowing the reduced temperature level on the
Tps unchanged, and the second strategy is noted as
secondary side of the HEX, the impact on the DH
mp unchanged.
network can be estimated. The impact is calculated
based on two different strategies: In Fig. 15 and Fig. 16 the possible reduction of DH
supply temperature is shown.
1. Primary supply temperature (Tps) is kept at the
Panel radiator T saving
same level as before 15
ps
Pfan = 2.7 W mp unchanged

2. The primary flow (mp) through the HEX is kept P = 1.9 W m unchanged
constant
fan p
By applying the first strategy, both Tpr and the mass flow 10
saving [C]
in the DH network is reduced. The second strategy

results in a reduced Tps and Tpr without changing the
flow rate in the DH network.
ps
T
5
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 0
1 0.8 0.6 0.4 0.2 0
with 20 apartments. The substation is providing the rel heatload
building with heat and domestic hot water (DHW) and

Fig. 15 Resulting Tps reduction with panel radiator.
DHW circulation. The assumed DHW usage is Column radiator T saving
ps
125 l/apartment&day, space heating load at DOT is 15
Pfan = 3.0 W mp unchanged
3 kW/apartment. The heat loss from DHW circulation is
P = 2.2 W m unchanged
assumed to be 0.1 kW/apartment. For each space fan p
heating load a flow-weighted mean value for Tps and Tpr

is calculated for a time period of 24 h, including heat 10
saving [C]
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.
ps
T
0
1 0.8 0.6 0.4 0.2 0
rel heatload
Fig. 16 Resulting Tps reduction with column radiator.

In Fig. 17 and Fig. 18 the reduction of Tpr is shown. As
seen the reduction of Tpr is of the same magnitude
independently of which control strategy is used.
27
th th
However, by keeping the DH supply temperature Column radiator m saving (% )

p
constant (strategy 1) the flow rate in the DH network is

20
affected, see Fig. 19 and Fig. 20.

15
Panel radiator T saving
m saving [%]
pr
15
Pfan = 2.7 W Tps unchanged 10
P = 1.9 W T unchanged
fan ps
p
P = 2.7 W m unchanged
fan p
10 Pfan = 1.9 W mp unchanged 5

saving [C]
fan ps
fan ps
0
1 0.8 0.6 0.4 0.2 0
pr
rel heatload
T
5
Fig. 20 Resulting mp reduction with column radiator.
Annual gain in Tps and Tpr during heating season

0 In order to evaluate the annual impact on the primary
1 0.8 0.6 0.4 0.2 0
rel heatload temperature level, the outdoor temperature has to be
considered. In this case measured values for the
Fig. 17 Resulting Tpr reduction with panel radiator.
outdoor temperature in Malm have been used, see
Column radiator T saving Fig. 21.
pr
15
40
Pfan = 3.0 W Tps unchanged
P = 2.2 W T unchanged 35
fan ps
P = 3.0 W m unchanged 30
fan p
10 Pfan = 2.2 W mp unchanged 25

saving [C]
20
T out ( C)
15
pr
10
T
5
5
-5
0 -10
1 0.8 0.6 0.4 0.2 0 May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May
rel heatload
Fig. 21 Outdoor temperature in Mam

Fig. 18 Resulting Tpr reduction with column radiator. st th
Panel radiator m saving (% ) 1 May 2006 30 April 2007.
p
20 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
15
calculations we assume a maximum heat output (QDOT)
at -15 C and the balance temperature, when no space
m saving [%]
10
heating is needed, +17 C.
Table 2 Reduction in annual primary temperature level

p
during heating season

5
Tps - mp -
P = 2.7 W T unchanged unchanged unchanged
fan ps
P = 1.9 W T unchanged T ps T pr
0
fan ps
Tpr
1 0.8 0.6 0.4 0.2 0 2
rel heatload
radiator radiator
Panel Column
Pfan= 2.2 W -2.2 C -2.4 C

Fig. 19 Resulting mp reduction with panel radiator.
Pfan= 3.0 W -5.7 C -6.6 C
Pfan= 1.9 W -0.8 C -0.9 C
Pfan= 2.7 W -2.5 C -2.7 C
28
th th
Note that the annual flow-compensated mean ACKNOWLEDGEMENT

temperature in Table 2 is based on results from the field
This work is part of the Primary Energy Efficiency
study where measured values for low relative heat load project of Nordic Energy Research.
are missing, which makes the values in the table
somewhat underestimated.
NOMECLATURE
CONCLUSION AND DISCUSSION Abbreviations

CHP Combined heat and power station
By installing the add-on-fan blower application on
existing radiators the temperature level in the heating DH District heating
system can be substantially reduced. This will also have DHW Domestic hot water
impacts on the DH network and DH production units.
DOT Design outdoor temperature
The impact on the DH network can be applied based on
HEX Heat exchanger (DH substation)
two principles:
1) DH supply temperature kept at the same level as Variables
without the add-on-fan blowers. This will result in
reduced primary flow rate. Heat transfer Gr Grashof number (-)
coefficient (W/m2.K)
2) Reduced DH supply temperature while primary flow
coefficient of h Height (m)
rate is kept constant.
expansion (K-1)
The first strategy could be applied immediately, since Thickness (m) k Heat transfer
the primary supply temperature is kept as the same coefficient (W/m2.K)
level as before. This means that not all heating systems
Conductivity (W/m.K) L Length (m)
connected to the DH network need to be modified in
order to apply this method. The lowered secondary Emissisivity (-) m mass flow (kg/s)
temperature level results not only in reduced DH-return
Kinematic viscosity Nu Nusselt number (-)
temperature, but also in a reduction of the DH flow rate. (m2/s)
The reduced flow rate could be used to increase the
number of buildings connected to the DH network, or to Stephan-Boltzman P Electric power (W or
constant kW)
avoid bottlenecks in the DH network. The magnitude of
the reduction of the DH supply temperature is between Logarithmic mean Pr Prandtl number (-)
9 and 12 C at DOT and at the same time the flow rate temperature
difference (K)
is decreased with more than 10 %. On annual basis the
possible reduction of temperature level in the DH A Area (m2) Q Heat output (W or kW)
network is in the magnitude of several degrees Celsius. cp Heat capacity (J/kgK) Ra Rayleigh number (-)
In order to apply the second strategy the demand for a C Constant Re Reynolds number (-)
high temperature level in the DH network needs to be 2
g Gravity force (N/s ) T Temperature (C or K)
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 Subscripts
DH supply temperature can be reduced with about 0 Design condition p Primary (side)
10 C at DOT without affecting the DH flow rate. At the (without fan)
same time the DH return temperature will be reduced
Fan Add-on-fan blower in r Return
with as much as 10 C at DOT.
operation
The performance of the tested add-on-fan blowers i indoor rad radiation
corresponds to the pattern of theoretical calculations.
m Mean rel Relative
However, the results are not comparable since the air
flow in the pilot project has not been measured. out outdoor s Secondary (side) or
Supply
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.
29
th th
REFERENCES [7] J. A. Myhren, S. Holmberg, Design consideration

with ventilation-radiators: Comparisons to traditional
[1] Svensk Fjrrvrme, District heating substations
two-panel radiators, Energy and buildings 41, p.
design and installation, Technical regulations
92-100, 2009
F:101, The Swedish District Heating Association,
2008 [8] A. Trschell, Vrmesystem med luftvrmare och
radiatorer, En analys av funktion och prestanda,
[2] P. Ljunggren, P-O. Johansson, J. Wollerstrand,
Licentiate Thesis, Chalmers, Gteborg, 1999
Optimized space heating system operation with the
aim of lowering the primary return temperature, in [9] S. W. Churchill, Correlating equations for laminar
Proc. of the 11th International Symposium on and turbulent free convection from a vertical plate,
District Heating and Cooling, 2008, Reykjavik Int. J. Heat Mass Transfer, Vol. 18, p. 1323-1329,
1975
[3] http://www.lenhovdaradiatorfabrik.se/display_sub.a
sp?apid=20, 2010-04-16, Downloaded spread [10] J. P. Holman, Heat transfer, 9th edition, 2002
sheet for calculating heat output.
[11] Discussion with professor B. Sundn, April 2010
[4] Swedish Energy Agency, 2008, Energy Indicators
[12] P. Selinder, H. Walletun, Modell fr ndrade
2008, Theme: Renewable energy, 2008
frutsttningar i fjrrvrment, Rapport 2009:50,
[5] http://a-energi.jetshop.se/, 20010-04-20 Svensk Fjrrvrme, 2009
[6] EN 15316-4-5:2007, Heating systems in buildings. [13] S. Werner, FVB Sverige AB, Nytta med svensk
Method for calculation of system energy fjrrvrmeforskning, FoU orientering 2004:9,
requirements and system efficiencies, CEN, Svensk Fjrrvrme, 2004
Brussels, 2007
30
th th
PRIMARY ENERGY EFFICIENCY AND SYSTEMS ENGINEERING

1 1 2
M.Berner , R. Ulseth , J.Stang
1
Norwegian University of Science and Technology (NTNU)
2
SINTEF Energy Research
ABSTRACT
Transformation
Transmission
Distribution
Processing
Generation
Extraction
Transport
The revised Energy Performance of Building Directive
Storage
(EPBD) [1] emphasizes that the energy performance of
a building shall be calculated by use of Primary Energy
Factors (PEF). Calculation of CO2 emission will not be
mandatory so far. Thus EPBD will reduce the use of
Figure 1 A Typical energy chain
non-renewable energy, incite the use of energy from
combined heat and power generation (CHP) and
reduce the energy consumption in the building sector. Methodology
An energy chain might consist of several elements or
A simplified method that enables comparison of the
processes from extraction, through processes such as
PEF from different energy chains is required. However, drying, storage, transport, power/heat/cool generation,
calculation of all the parameters affecting the PEF and distribution to the end user. In order to ensure that
values like energy used for extraction, transportation, there is a correct PEF, all elements that influence the
power and heat generation etc. is time-consuming. The energy flow have to be accounted for.
method described in EN 15603 [2] is rather general l The energy balance or calculation of the energy
and provides PEF values for 13 energy carriers and efficiency of a process focuses primary on the energy
chains. This is based on average European values. Life input in the form of fuel and the output in kWh, and
Cycle Assessment methods include several of the lacks information on the energy used to build
relevant steps, but a complete LCA often imply infrastructures such as the power plant, distribution net,
collection of more than 6000 parameters. transportation and the extraction.
The systems engineering method used here have Life Cycle Assessment (LCA) might contribute to
demonstrated the feasibility of developing a generic provide such information in a generic method.
method that provides credible data for calculating However, the number of input parameters, often more
primary energy efficiency. It applies the generic method than 6000 in an ordinary LCA analysis demonstrates
on energy chains in the Nordic region which is relevant the need for an easily accessible method.
to CHP plants utilising bio based fuel. Systems engineering is a method that has been
developed gradually with increasing complexity of
INTRODUCTION projects and systems. Systems engineering is often
considered to have started at Bell Laboratories in the
Background
1940s, later applied in organizations such as NASA
The terms Primary Energy, Primary Energy Efficiency and formalized as a separate engineering field with the
and Primary Energy Factors (PEF) are introduced [3] - formation of INCOSE [9] in 1990. The benefits of
[8] in order to compare different energy sources and systems engineering is the possibility to treat complex
chains based on losses and a calculated environmental systems with several subsystems. Therefore, as a first
impact. step in the development of a method a systems
Primary energy is energy that has not been subject to engineering approach has been chosen. The main
any conversion or transformation process. The use of objective is to develop systems and methods that
primary energy factors takes into account the energy enable a sufficiently reliable calculation to be made of
that are used from the extraction of the energy carrier the primary energy factor (PEF) in general and for
and all of the losses until energy is delivered to the end different energy chains with required level of details.
use in the desired form such as heat, cooling or At present systems engineering approaches have not
electricity . been found to have been previously applied on the
The primary energy factor (PEF) expresses how much development of generic PEF methods for different
primary energy is needed to deliver 1 unit of power, energy chains.
heat or cooling to the end user. The term primary
energy efficiency (PEE) therefore is used to describe Objective
the total use of energy from extraction to the end user. The objective of this paper is to show how systems
engineering can be used as a tool to reveal important
31
th th
parameters when a model for calculation of the PEE of 3. Measures of effectiveness (MOE)
different energy chains is developed. The paper will The definition of MOE are: A small subset of the
show an overall approach and will not describe all the requirements that are so important that the system will
necessary iterations in detail. fail if they are not met and will be a huge success if
they are met [11].
SYSTEMS ENGINEERING
4. Development of information models
The system engineering process
The different information models describe the observed
A systematic approach such as systems engineering is system in relation to legislation, physical architecture
essential to be able to develop a generic model and a system interface model. Four separate models
describing a complex system with several subsystems. are developed
The intention with the systems engineering process is
to analyse and describe complex systems. Often the Requirement traceability model
method is used in the design process, to make sure
System architecture model
that the subsystems are connected properly, that the
process is optimized and that the different components Behaviour model
are described, implemented and integrated precisely.
System interface model
A common feature of all systems engineering
5. Trade-offs
processes is an indefinite number of iterations at all
different steps. The trade-off phase is essential in the development of
a method. Each of the steps is carried out in iterative
Systems engineering principles are often applied when loops gradually increasing detailing level. After
a new system or products are developed. The satisfactory trade-offs have been performed and
methodology alters slightly between development and consistent information models obtained, a theoretical
re-engineering. method is developed. Real data are collected and trade
Re-engineering methods are applied when an existing -off between the model and the gathered data are
system is described. The energy chains considered are performed.
already designed and built, and a re-engineering
6. Documentation
technique is selected in order to develop a method that
calculates the PEF for different kind of energy chains. The developed method will be then documented by
actual case studies before a final reporting.
CHOSEN METHODOLOGY
The system re-engineering process consists of the Iterate to find feasible No
4 feasible
following six different tasks according [2]. Some of solution
Create solution
them might seem unnecessary, but they all contribute
requirement
traceability model
to the decomposing of a system and development of a Feasible
solution
method. 4
1 2 3 Create system 5 6
Establish Asssess Define architect. model Trade- Document
1. Establish problem statement; problem available effective- offs current
statement inform- ness and
This comprises the definition of the problem approach, ation measures 4
system
design
which includes development of a problem statement Create behaviour
model
describing the problem/challenge, its importance and a
state of the art. To be able to establish the problem 4
statement; four questions must be answered: Create context
model
What is the problem?
Why is it important Figure 2 The system re-engineering process described
as a functional block diagram (FFBD), ref. [10]
What have others done?
What must be done? ESTABLISH PROBLEM STATEMENT
What is the problem?
2. Assess available information assessment
Provide available information including an overview of Use of Primary Energy Factor (PEF) will provide
possible stakeholders. information on the energy losses and consequently the
environmental impact of different kind of energy
sources, power production processes and energy
32
th th
transport systems. At preset there exists no easy and process lines (chains) primary in Norway and the
accessible calculation method. Nordic countries. Detailed data must be provided such
as efficiency and loss from the different systems and
Different countries have different energy chains and
mix of systems, or at least provide the necessary
energy supply systems. Analysis of even the most
parameters. Since the systems engineering approach
actual processes and process lines does not exist
is chosen, the problem approach must be defined, a
neither for Norway or Europe [12] [13]. In order to
theoretical method developed and data collect. This
compare and choose different energy chain there is a
includes performing of a trade-off between the
need for standardized methods. The lack of objective
theoretical model and available information. The
and reliable data of the different elements in the energy
method shall be tested by selected case studies and
chain might prevent an efficient use of energy, and
finally adjusted.
contributes to wrong choices and unnecessary CO2
emissions. [14] [15].
Main hypothesis
The method is principally described in EN 15603 [2] As a part of the systems engineering process, one or
and provides only single PEF values for 9 energy several (systems engineering) hypothesis is developed.
carriers and 4 energy chains, and is based on average The success of a system engineering process is
European values. Without an easy accessible method related to the fulfilment of the hypothesis. In this project
or methods is it not possible to compare PEF values the system engineering method must prove two main
and calculate the actual environmental impact of hypotheses;
different energy chains. Some studies [16] -[21] have
described parts of this topic, but they lack a holistic 1. It is possible to develop a generic method that
view of the energy chains from cradle to grave, often provides credible data for calculating primary
the chosen system boundaries are different, time scale energy use by use of PEF values.
varies, detailing level different and the, approach/
2. It is possible to apply the generic method on
method varies. Results from different studies therefore
energy systems in the Nordic region for CHP
are not comparable.
plants utilising bio based fuel.
Why is it important
Stakeholder analysis
PEF is a key indicator to be able to evaluate energy
A stakeholder is a party having a right, share or claim
use (for different purposes) especially with regards to
in the system [16]. The intention with the stakeholder
the goals of the EPBD [1]. PEF is an over all energy
analysis is to reveal the different kinds of stakeholders
efficiency indicator which makes it possible to compare
since they might have requirements influencing a
and collocate different energy sources and energy
possible method in a legal way. Stakeholders with
carriers by a single number. The same method can be
mutual interest are aggregated in groups; some of
used to calculate the CO2 emission.
them might not be in incompliance with each other.
What have others done?
Energy producer, distributors, energy companies;
Different CEN standards describe, and partly discuss, Business profitability is the main issue by
the theory. In the EC-mandated CEN standards related optimizing production from different energy carriers
to EPBD mainly one single reference are referred [14] according to cost-benefit
whilst the PEF values have been gradually changed
over time. An extended literature survey has showed Investors (energy and building); The electricity
discontinuity between some of the studies performed markets are opening gradually throughout Europe,
and lack of details in the calculations. e.g. Nord Pool Financial Marked and investments
in Power production and the introduction of so
Methods developed to provide PEF values for heating called Green Electricity Certificates might be a new
systems in buildings might be useful, but they will not or extended business area. Investors in the
totally comply with a whole energy chain approach. Life building marked might be interested in the actual
Cycle Assessment (LCA) might also contribute to a PEF values and primary energy use when
generic method, but the vast number of input choosing between different investment objects.
parameters, often more than 6000 in a traditionally LCA
demonstrates the need for a more easy accessible Building owners, end user; Correct calculation of
method. PEF values and primary energy use is supposed to
have significant importance for the choice of
What must be done? energy supply system, building services, insulation
In order to develop a method a systems engineering level, especially for new buildings and major
approach will be used. The most important task in this rehabilitation projects. Future operating cost might
context is the identification of relevant energy systems
33
th th
depend of PEF since taxes might be dependant of INFORMATION MODELS

the primary energy use and/or the CO2 emission.
The requirement traceability information model
Developers& Building and construction industry; In systems engineering shall the requirement
Technical equipment in the building and design traceability information model aim to show the break
strategies depends on the actual use of the down of requirements from source documents to final
specific energy carrier. The use of PEF values in allocation functions to stakeholders [2]
the primary energy calculations might change the
value of traditionally installed equipment due to This model is an important tool to keep track of the
overall energy costs and also create a demand for different requirements, source documents and
more energy flexible solutions. eventually what the system accomplishes and who or
what are in charge. Usually an Entity-Relationship-
Politicians, government, Regulators, Community Attribute method is used [23], where the entities
planning; Most European countries have affiliated (objects) represent the legislation, requirements, etc.
the Kyoto Protocol, and a possible method to whilst the relationship shows the association between
increase the use of renewable energy policy tools the system/process.
and subsidy schemes might be based on the use
Working Environm. Act EPBD Energy Landfill Directive 96/62/EC
of PEF values for the different solutions, besides
Planning and
Building Act Performance. Ambient Air
Quality
possible tax on systems with high primary energy Source Source Source Source
Documents
Source
Documents Documents
use. Building EN 15316-4- N 15603:2007 Waste regulations 1999/30/EC Limit
National regulators mandatory monitor and report

regulation TEK 4:2007) values
Source Source Source Source NOx
Requirements
emission levels and this influences national Incorporates Incorporates
legislation, local and urban planning Building Guide

REN
Pollution Control
Act
Source Requirements
Research groups, Universities; Different research Specifie Specifie Specifies
communities might be interested in development of NS3031 s s

Disharge permit Discharge permit
other PEF calculation methods or adjustments of function Requirements Requirements
methods and development of new solutions and Allocated Allocated to Specifies

to
Building Permit NOx emissions Internal control
systems system
stakeholder Boiler function
Allocated to
NOx emission
DESCRIPTION OF MEASURES OF EFFECTIVENESS
1.2.2.1
Stakeholder
As earlier stated the measures of effectiveness (MOE) Figure 3 Selected part of the requirement traceability
should be independent of any solutions and not model, case Norway
concerned with internal details [22]. It might also be
fruitful to develop MOE for the different kind of Most EC directives are enforced and implemented in
stakeholders since they often might have a different laws, directions, regulations and guidelines, both in the
opinion regarding MOE. EU and associated EEC countries, hence the order of
entities in Figure 3. Several directives influence the
In this context MOE are primarily described for the national laws and regulations. Since the directives
ongoing Nordic PhD-project Primary Energy Efficiency usually are enforced through national laws, the law
(PEE). A further detailing level, by including the includes requirements from more than one directive like
stakeholders, might provide valuable information, but the Norwegian Planning and Building Act [26], which
that is considered to lie outside the scope of this work. includes requirements from EPBD [27], Directive on the
The methods (tools) developed during the project promotion of the use of energy from renewable sources
should be suitable for different kind of energy chains. [30], Directive on the landfill of waste [29], The
Pollution Control Act [30] amongst others.
The results should be utilised by the different
kinds of stakeholders e.g. the building owner, the The requirements traceability model provides important
architect/designers of the building, the energy information about constraints regarding an energy
supplier and producer and finally politicians and chain. Some of the elements such as the the Working
governments. Environment Act [31] might seem irrelevant, but the
regulations set limits for the pollutant inside the working
The methods will enable the different stakeholders
area, introducing need for e.g. conveyor belts.
to choose between different energy systems and
furthermore be able to reduce primary energy use Each function consists of several entities for instance
and CO2 emissions from stationary energy the discharge permit from regulators like The
purposes. 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.
34
th th
The requirements affect most of the stakeholders, for An energy chain consists of several sub systems as
instance the Planning and Building Act will affect both described in Figure 4. A more detailed architecture
end users, construction industry and energy information model is also developed.
distributors.
Each of those sub elements can be spilt up into sub
elements as shown in Figure 5. The final or basis
Architecture information model
element can be described as Figure 6.
The architecture information model shows the physical
components of a system with subsystems.
In order to describe the possible physical systems a
Energy
Transformatio
n system
generic model is developed [11], detailed description of Consist of Consist of Component Consist of
some of the most relevant energy chains are carried Energy Energy Waste
out in the actual PhD thesis, Figure 1 shows a principal Feeding

system
Production
system
handling
system
descriptions of a single energy chain. An end user will Component

Consist of
Component
Consist of
Component
typically be supplied with energy from a various Consist of Combustion Energy

process Transformatio
number of energy chains, and each element might n system
Component Component
represent parallel processes. Consist of
Consist of
Purification Combustion Internal Electricity Heat
A CHP utilizing biomass might consist of the following system camber Control
System
production unit production unit
elements: Component Component Stakeholder Component Component

Consist of
Built of
Fertilizing, cultivation, logging, logging track, loop

Filter Heat Heat Storage
Transport system
of twigs, trimming, transport

system
Component Component Component
Chipping, packing, transport, local roads Figure 5 Segment/selection of part of the architecture
Intermediate storage, transport regional roads information model.
Transport central and regional roads

Building, operation demolition of power plant, Additional PEF
technology, efficiency, part-load, size, Lifetime,
waste treatment, gas cleaning supply of additives,
internal transport
PEFin
Infrastructure, Operation and Demolition PEF out
Transformation to central net, building, operation, buildings, maintenance
demolition of infrastructure, heat/power loss machinery etc.
Transmission to local net, building, operation,

demolition of infrastructure (pipes, high-tension Loss
lines heat/power loss (insulation, temperature Figure 6 Architecture information model, basis element
levels (supply, return, ground), twin/single pipes,
length)
Since the Primary Energy Efficiency (PEE) of an
Distribution to end user, building, operation, energy chain consist of all of the elements from
demolition of infrastructure (pipes, lines, extraction to delivery the PEF for a chain can be
substations) heat/power loss
calculated by
E Chain E Fuel E Extraction E Processing (1)
Construction Dismantling E Storage E Transport E Generation
Elec Construction
E Transformation E Transmission
prod
E Distribution
Other Cooling Col
(chem
prod d Construction
etc) Construction Construction
Where E is the primary energy input to the
Fuel Heat
Dismantling
Hea Dist.
Cold
Sub
Heat
Cool
system
in (1) t Heat
prod net station
Fuel Dismantling Dismantling Dismantlin

The Power Bonus Method
g
Waste
handlin
In [13] the power bonus method is applied to calculate
g the PEF value for a district heating system with CHP.
Electricit
Figure 4 Architecturey information model for a part of the E P ( Edel,i f P,del,i ) ( Eexp,i f P,exp,i ) (2)
energy chain from generation including distribution, based
on [25]
where
EP Primary energy input to the system
Edel,I Delivered energy, energy carrier i
35
th th
fP,del,i Primary energy factor, delivered energy carrier i Figure 4 and Figure 5. This is an iterative process and
Eexp, Exported energy, energy carrier i the detailing level is the first steps gradually increasing,
fP,expl,i Primary energy factor, exported energy carrier i until the analysis (trade-off) of the different factors
influencing the PEF value enables a removal of factors
Power exported from the CHP plant multiplied with the
with an impact of 1% or less.
PEF value for the replaced power shall be subtracted
from the delivered primary energy to the buildings
System interface information model
when calculating the PEF-value for the for the district
heating system [7]. The power bonus method is The system interface model also denoted the context
enforced in order to promote CHP, and the subtraction information model shows the systems interface with its
of power produced and delivered outside the system surroundings and the environment. The model provides
boundary significantly reduces the PEF value for the information on the core system and other
energy chain. This implies that the PEF value for a interconnecting systems; this means a description on
CHP will be dependent on the power to heat ratio. how things relate to each other.
The context is according to [10] the interrelated
Behavioural model conditions in which something exists or occurs.
The behavioural model is another information model
the what it does[10], but also described as the way in Energy Energy
prod. Raw
which an organism, organ, or substance acts, material
especially in response to a stimulus [23]. Energy Energy
source
A behavioural model consists of functions, inputs and carrier
Requirements Energy Energy
outputs and control operators. This implies that it is productio
supposed to provide information on what is happening, n
Requirements Energy Energy
in which order and what kind of iterations are transport
performed. system
Requirements End user
Requirements
Establish problem Figure 8 A simplified context information model.
statement
The system boundary is drawn with a dashed line, and
Assess available
the system assessed lies within. Since this is a
information
simplified model the relation towards investors, national
Define MOE regulators, constructors etc. are not shown. In this
system energy source/carrier is closely connected to
And Extraction, Energy Source consists of storage and
transport, Energy production corresponds with
Energy source Energy Energy transport Generation and Energy transport system to Trans-
production system
formation, transmission and distribution in Figure 1.
And The main issue has been to show the connection

between the energy chains from production to end use.
Trade-off Politicians and national regulators might have specific
requirements on each level. The building industry,
Develop generic
constructors may likewise have interest on several of
model
the levels, but a final listing is not possible to provide
Collect data within this paper.
case study
Another important question is the definition of the
Test, evaluate system boundary. Precise definitions of the system
method boundaries are essential in order to be able to compare
different studies. The system boundaries must
Simplify method
distinguish between what is included and what lies
Publish model outside of the task of the LCA, since the method must
also rely on data collected by other parties and the use
Figure 7 Simplified behaviour information model of the
of different constraints might influence the quality of the
model developing process.
method.
A more detailed partition of the energy chains have In order to provide information of the whole energy
been applied in the development of the method. The chain, all major elements have to be included i.e. the
energy chain is divided in subsystems as shown in extraction phase is an integrated part of the chain.
36
th th
Trade-offs often significant lower than the actual exchange rate

In the ongoing PhD-project of the first author a cut-off e.g. pipelines might have a twice times higher - more
rule is set to 1%. That implies that factors with less than 60 years. The use of yearly average efficiency and
than 1% impact on the final result can be removed. appurtenant power-to-heat ratio will often increase the
The trade-off considerations are still an ongoing PEF value for the whole system due to the impact of
process, and it will be presented and documented in a the power bonus method.
later paper. According to the main hypothesis this is not The reliability of the method will be influenced by
meant be developed as an optimization tool, since the possible lack of detailed data, but based of average
intention of the ongoing PhD-project is to develop a data a reliable comparison of different energy chains
method describing different energy chains. might be performed.
A complete trade-off could preferably [32] be performed More standardized values for some the different
by use of computerizes programme like CORE. parameters needs to be developed, like lifetime, heat
[http://www.vitechcorp.com/solutions/]. The complexity load curves and extraction of biomass. Some
of the different kind of systems shows the utility value adjustment will be necessary for instance for extraction
of more than manual tools, which has been applied. where the transport distances are an important
parameter. The resulting model can form a basis for
Document current system design
future optimization tools, since only elements with
The results of the iterative process are described in the major influence on the PEF-values are included.
figures mentioned above. Only selected parts of the
chosen design are illustrated in this document due to
ACKNOWLEDGEMENT
limitation in size.
This paper is developed as a part of the PhD-project
The final system design is carried out according to
Primary Energy Efficiency (PEE) and the work title for
Figure 6 for each element.
thee PhD-Theses is "System, methods and credible
data for calculating primary energy efficiency in general
CONCLUSION
and for energy systems in the Nordic region with special
By performing a system engineering process focus on energy systems applying CHP-technology with
describing different energy chains an outline of the bio based fuel in particular".
model have been developed. The method has proven
The project is financed by Nordic Energy Research
to be efficient in structuring the thoughts and will
with financial support from the industry and includes six
hopefully reduce mistakes in the future development of
PhD-studies carried out in the respective countries;
the model.
Estonia, Finland, Sweden, Iceland and Norway. The
The decomposition process in different subsystems is projects objective is to contribute to the effort of
valuable, and the generic model will be able to treat enhancing the primary energy efficiency (PEE) and
different kind of energy systems and chains. reducing CO2-emissions in the energy sector.
The systems engineering process have demonstrated
that; FURTHER INFORMATION
1. It is possible to develop a generic method that PhD.student Monica Berner, Norwegian University of
provides credible data for calculating PEF-values Science and Technology (NTNU).
and the primary energy efficiency.
Address: Monica.Berner@ntnu.no
2. It is e.g. possible to apply the generic method on
energy systems in the Nordic region with CHP
REFERENCES
plants utilising bio based fuel
The system engineering process provides a new [1] Proposal for a Directive of the European
approach to the design and development of a generic Parliament and of the Council on the Energy
model describing PEF-values for energy systems with Performance of Buildings Recast SEC (2008)
different kind of energy carriers. The method might be 2820, SEC (2008) 2821)
used for more than systems using CHP-technology [2] EN 15603 Energy performance of buildings
since the model development are generic and thereby Overall energy use and definition of energy ratings
utilizes different kind of energy carriers.
[3] EN 15603:2007 Energy performance of buildings
The method can provide detailed data (e.g. efficiency, overall energy use and definition of energy ratings
loss etc) from the different energy chains and mix of [4] EN 15316-1: 2007 Heating systems in buildings
chains. A major challenge is the data collection, some Method for calculation of system energy
of the parameters lack standardization. The life time of requirements and system efficiencies Part 1:
different equipment varies, the economical lifetime is General
37
th th
[5] EN 15316-2-1:2007 Heating systems in buildings Systems, ITEV-Report 1999:06, Dr.ing Thesis
method for calculation of system energy 1999:117, NTNU
requirements and system efficiencies part 2-1 [18] Sarigiannis D.A., Triacchini G., Meso-scale life-
space heating emission systems cycle impact assessment of novel technology
[6] EN 15316-2-3:2007 Heating systems in buildings policies: The case of renewable energy, Journal of
Method for calculation of system energy Hazardous Materials 78, 2000 p. 145-171
requirements and system efficiencies Part 2-3 [19] Alanne K., Salo A., Saari A., Gustafsson S., Multi-
Space heating distribution systems: criteria evaluation of residential energy supply
[7] EN 15316-4-4:2007 Heating systems in buildings systems, Energy and buildings 39, 2007 p 1218-
Method for calculation of system energy 1226.
requirements and system efficiencies Part4-4 [20] Eriksson O, Finnveden G, Ekvall T, Bjorklund A,
Heat generation systems, building-integrated Life cycle assessment of fuels for district heating: A
cogeneration systems comparison of waste incineration, biomass- and
[8] EN 15316-4-5:2007 Heating systems in buildings natural gas combustion, energy Policy 35, 2007
Method for calculation of system energy p.1346-1362.
requirements and system efficiencies Part 4-5 [21] Mnster M., Lund H., Use of waste for heat,
Space heating generation systems, the electricity and transport Challenges when
performance and quality of district heating and performing energy system analysis. Energy 34,
large volume systems 2009 p. 636-644
[9] INCOSE, International Council on Systems [22] Lenzen M., Life cycle energy and greenhouse gas
Engineering, A Consensus of the INCOSE Fellows, emissions of nuclear energy: A review, energy
www.incose.org Conversion &Management 49, 2008 p.2178-2199
[10] Dahl H J, Information modelling and systems re- [23] Sproles N, Coming to Grips with Measures of
engineering an efficient approach to assessing Effectiveness, John Wiley & Sons, Inc Syst Eng.
complex current Norwegian natural gas transport 3:50-58, 2000
operations, Proceedings of the Tenth Annual [24] Olivier, Merrian Webster 1981
International Symposium of the International [25] Berner M, Primary Energy Concept and Life Cycle
Council on Systems Engineering (INCOCE), July Assessment (LCA), Report no: 2009/001, June
2000 2010, The Norwegian University of Science and
[11] Olivier DW, Kelliher TP, Keegan JG, Engineering Technology
complex systems with models and objects, ISBN [26] Act of 14 June 1985 No. 77 the Planning and
048188-1, McGraw-Hill, 1997 Building Act, The Ministry of the Environment and
[12] Joelsson. A. Primary Energy efficiency and CO2 the Ministry of Local Government and Regional
mitigation in Residential buildings, Doctoral Thesis Development
58/2008, Mid Sweden University (Dissertation [27] Directive 2002/91/EC of the European Parliament
3.October 2008) and of the Council of 16 December 2002 on the
[13] Berner M., Ulseth R., The Primary Energy energy performance of buildings.
Concept, The 11th International Symposium on [28] Directive 2009/28/EC on the promotion of the use
District Heating and Cooling, August 31 to of energy from renewable sources and amending
September 2, 2008, Reykjavik, ICELAND and subsequently repealing Directives 2001/77/EC
[14] Frischknecht, R, Jungbluth et al, 2007, and 2003/30/EC
koinventare fr energiesysteme Grundlagen fr [29] Council Directive 1999/31/EC of 26 April 1999 on
den kologishen Vergleich von Energiensystemen the landfill of waste
und den Einbezug von Energiesystemen in [30] Act of 13 March 1981 No.6 Concerning Protection
kobilanzen fr die Schweiz , ETH, Zrich 1996 Against Pollution and Concerning Waste, [The
[15] CEN/ CLC BT JWG, Energy Management, 2005) Pollution Control Act]
[16] Nrsteb V., Application of systems engineering [31] Act of 17 June 2005 No. 62 relating to working
and information models to optimize operation of environment, working hours and employment
gas export systems, Systems Engineering archive, protection, etc. as subsequently amended, last by
Volume 11 , Issue 4 (November 2008), p: 329- Act of 23 February 2007 No. 10, (The Working
342, 2008, ISSN:1098-1241 Environment Act)
[17] Sther S, Thermal Heat and Power Production [32] Purves B, Information Models as a Prerequisite to
with models for local and Regional energy Software Tool Interoperability, Incose Insight, 1998
38
th th
ENHANCED DISTRICT HEATING AND COOLING SYSTEMS

REALISATION OF THE LOW-EX CONCEPT
1 1 2 1 2 1
Stefan Bargel , Clemens Pollerberg , Armin Knels , Li Huang , Dirk Mller and Christian Dtsch
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 domestic gas boiler used to provide space heating

wastes a huge amount of exergy, since the exergy
Since heating and cooling represent low-exergy energy
efficiency of such a system reaches only approximately
streams, high efficiencies can be obtained, if the
5%! This result is valid for arbitrary heating systems in
energy demand is covered by appropriate meaning
the supply target (room) itself. Therefore it is
also low-exergy (low-ex) input energy flows.
mandatory to use an integral evaluation approach to
In order to be able to employ great potentials of low- decide whether an energy system is efficient or not.
exergy heat from many different sources, it is important
With respect to district heating and cooling networks as
to develop technologies for the supply and the use of
energy supply systems two findings are important.
energy that allow network temperatures close to
First, it can be shown that the network subsystem itself
ambient temperature in return as well as in supply
as depicted in figure 2 reaches optimal exergetic
pipes. Two possible technologies are phase change
efficiency at quite low temperatures since the heat
slurries (PCS) and capillary tube mats (CTM).
losses dominate the pumping electricity effort.
PCS are discussed as heat transfer fluid, which has an Secondly the overall energy supply system efficiency
increased heat capacity compared to water. The use of can be greatly enhanced by utilising low-exergy input
PCS in energy supply networks instead of water leads energy flows such as industrial waste heat.
to an improved energy transport capacity, which results
In order to be able to employ great potentials of low-
in a reduction of the necessary temperature difference
temperature waste heat from many different sources, it
of the transfer fluid. To ensure the transfer of energy
is important to develop technologies for the supply and
from the supply network into the building while the
the use of energy that allow network temperatures
temperature difference between network and building is
close to ambient temperature in return as well as in
low, large heat transfer areas are required, which can
supply pipes.
be achieved by the use of CTM.
Today, district heating and cooling networks use water
This paper discusses opportunities for the realisation of
as heat transfer fluid. The heat is transported as
cold supply networks and low-ex systems and presents
sensible heat and the transport capacity of the
exemplary technologies for their realisation.
networks is determined by the heat capacity of water
and the temperature difference between forward and
INTRODUCTION backward flow. In cold supply networks as well as in
Temperature levels in district heating and cooling low temperature heating networks, high volumetric flow
networks have long been discussed. During the last rates are necessary to provide the required transport
years a tendency towards low temperature networks capacity due to the comparably small temperature
can be observed. From a scientific point of view difference between forward and backward flow. To
answers to the question for the optimal temperature overcome these restrictions, a new heat transfer fluid
levels can be given using exergy efficiencies as for with an increased heat capacity is under development
example discussed in [1]. The main advantage of this as an alternative to water, phase change slurries. PCS
evaluation parameter is the thermodynamically correct are mixtures of dispersed phase change material and a
distinction of thermal (low-exergy) and non-thermal continuous liquid phase, which can be used as heat
(high-exergy) energy flows. transfer fluid in district heating and cooling networks.
PCS possess an increased heat capacity due to
Since heating and cooling represent low-exergy flows, additional latent heat of fusion occurring during the
it is of uttermost importance to cover these demands by phase transition of the phase change material. The use
appropriate meaning also low-exergy input energy of such a dispersion in energy supply networks leads to
flows. For example a heating system based on a an improved energy transport capacity, which in turn
39
th th
results in a reduced temperature difference or mind that the annual average outdoor temperature for
volumetric flow rate of the transfer fluid needed to the heating period e.g. in Germany is about 3.5 C, it
transfer a given amount of heat. The application of becomes apparent that the exergy to energy ratio of
PCS for thermal energy transportation is investigated the target energy flows - passing the building envelope
and discussed for example in [2]. at 20 C is very small (approx. 7%). On the other
hand, exergy to energy ratios of conventional input
An improved transport capacity is one important point
energy flows are usually 100% as combustible fuels or
for the realisation of the low-ex concept; another
electricity is used.
important point is the use of the energy on the
consumer side. To ensure the transfer of the energy The low-ex concept acknowledges the fact that
from the supply network into the building while the demand flows are low-ex - meaning that they possess
temperature difference between network and building is small exergy to energy ratios. Hence the concept
low, large heat transfer areas are necessary. These demands to supply energy on appropriate exergy
heat transfer areas can be realised by using capillary levels, instead of wasting exergy by transforming high
tube mats integrated into the walls, the floors and the exergy flows into low exergy ones. In doing so, this
ceilings of buildings. approach opens up a totally new dimension of
enhancement potential since it deals with the quality
The E.ON Energy Research Center of the RWTH
aspect of the energy flows under consideration.
Aachen and Fraunhofer UMSICHT investigated the
Therefore, within the low-ex concept energy is no
possibilities to realise district heating and cooling
longer one-dimensional. In addition to decreasing the
networks as low-ex systems. These investigations
amount of energy demanded by the consumers
include system modelling and analysing as well as the
leading to insulation efforts a kind of exergetic
development and testing of technologies.
suitability has to be taken into account and the task at
hand becomes a two-dimensional problem (cf. fig. 1).
1. Exergy as evaluation parameter
Consequently, the concept aims at maximizing the
1.1. The low-ex concept exergy efficiency of an energy supply system, which
Exergy can be understood as the theoretical maximum allows to utilize potentials in both dimensions, quantity
of mechanical work that can be utilised by equilibrating AND quality.
an energy flow whilst considering its ambient The exergy efficiency can be defined as:
conditions.
Consequently this property distinguishes between
ex
exergy (demand ) (1)
types of energy that can theoretically be transformed
into each other without any losses - like mechanical exergy (supply)
work, electrical energy or combustible fuels - and
In applying this efficiency the demand flows and
thermal energy. The possibility to transform the latter
particularly the supply flows have to be defined
into any other type of energy is limited by the second
carefully (cf. chapter 1.2.).
law of thermodynamics and therefore inevitably
connected to losses. exerget ic suit abilit y
enhancem ent low -ex
This distinction is of importance if one analyses a
exerget ic qualit y
concept
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. insulat ion
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 energy dem and (quant it y)
temperature differences between supply flow and
target2. The real temperature differences occur due to
Figure 1. Energy as two-dimensional concept. Orange (light
heating and cooling techniques applied which are grey): conventional system, green (dark grey): optimal
mainly limited by finite heat transfer areas. Keeping in system
2
This statement is analogously true for cooling applications.
40
th th
net w ork building

conversion heat heat
losses losses losses
heat
prim ary f low heat
energy dist rict f low T = 20C
heat
heat ing
generat or
net w ork heat ing
surf ace
pum ping heat
elect ricit y f low
pow er plant
conversion losses
prim ary energy
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 Instead thermal input flows as industrial waste heat3 or
When evaluating a system it is important to specify the geothermal energy should be applied. On the other
evaluation boundaries (cf. fig. 2). It has to be pointed hand, if combustible fuels are used to meet thermal
out that an integrated system evaluation is mandatory demands, at least Combined Heat and
since otherwise results are ambiguous and misleading. Power generation (CHP) with a maximum electrical
This can be demonstrated by assuming e.g. evaluation degree of efficiency should be utilized. This allows
of the building subsystem only. If two systems are transforming part of the high-exergy fuel into high-
compared, one consisting of a target room equipped exergy electric current. Heat is produced as waste
with space heating and the other one with a target product of this conversion.
room equipped with conventional heating, one could Optimization potentials within the distribution
arrive at the conclusion, that the system utilizing space subsystem are basically indirect. At first glance, the
heating is more efficient. However, assuming both distribution system has no influence at all since the
systems are also equipped with an identical network acts as connection between heat generation
condensing gas boiler providing the heat, an evaluation and heat consumption. Consequently, no thermal flows
comprising the total system (consisting of heat exist that pass the overall system evaluation
generation and heat transfer to the target) would arrive boundaries. However, two aspects remain and need to
at a totally different conclusion. In this case, both be accounted for. One is heat losses occurring
systems possess the same exergy efficiency, which is throughout the network that have to be compensated
approximately 5% for the outlined case. This is by additional heat generation. The other is pumping to
because a potentially more efficient heating system is maintain the heat transfer medium circulation, which is
not put to use as the same input and supply flows met by an unalterable high-exergy input (electricity).
occur in both cases. The main problem is that concepts, which lead to
decreasing heat losses cause increasing pumping
1.3 Efficiency enhancement potentials efforts and vice versa. Nevertheless, heat losses are
The complete energy supply system can be divided the exergetically dominant influence, therefore the
into three subsystems generation, distribution and focus should be to confine these losses. Heat losses
building (representing the consumption). These depend on the driving temperature difference between
subsystems possess different potentials to enhance medium and surrounding ground and on surface area.
overall system efficiency. Minimization of the losses can most easily be achieved
by reducing the network temperatures since pipe
Currently heating demands are met by burning high- dimensions are affixed due to demands so that surface
exergy fuels, great enhancement potentials are areas are not a modifiable parameter. This approach is
available within the generation subsystem. Firstly, fuels even more rewarding since it allows employing low
should not be used to directly satisfy thermal demands
at all since this embodies pure exergy destruction.
3
Industrial waste heat in this sense is heat that can no more be
put to any use within the industrial production process.
41
th th
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 Figure3. Photograph of a paraffin/water dispersion
high-temperature cooling devices. Inlet and outlet
temperatures of the heating/cooling device The increase of the heat transport capacity of a supply
simultaneously define constraints for the distribution network using a PCS instead of water can be described
network subsystem, which in turn set constraints for the by a thermal capacity enhancement factor (TCEF),
generation. In the end supply temperatures close to the which is calculated according to equation (2).
target temperature form the basis for a low-ex ready
consumer. Without this step an exergetically optimal PCS wh f , PCM c p , PCM T 1 wc p , w T (2)
energy supply system would be greatly hindered. TCEF
w c p , w T
2. Applicable technologies for the realisation The TCEF is a function of the densities of the PCS PCS
2.1. Phase Change Slurries and water w, the mass concentration of the PCM w,
the specific heat capacity of PCM cp,PCM and water cp,w
The most used heat transfer fluid in district heating and
the heat of fusion of the PCM hf,PCM and the
cooling networks is water. In supply networks, the heat
temperature change T of the fluids. The TCEF is
is transferred as sensible heat with a temperature
difference between forward and backward flow. The calculated and plotted in the diagram figure 4 for
temperature differences T between the forward and
heat transfer capacity of a network is determined by the
temperature difference, the mass flow and the heat backward flow of 10 and 15 K as function of the mass
concentration w.
capacity of the heat transfer fluid. The temperature
difference and the temperature level of the network are
TCEF [-]
limited by technical restrictions and determine the
necessary mass flow of the heat transfer fluid. To 3.5
overcome these restrictions, fluids with higher heat

capacities than the heat capacity of water are under 3 delta T = 10 K
development. An alternative to water could be PCS. delta T = 15 K
PCS are mixtures of dispersed phase change material 2.5
and a continuous liquid phase, which possess an
increased heat capacity due to the additional latent 2
heat of fusion occurring during the phase transition of
the phase change material. The PCS remains 1.5
pumpable even when the phase change material is
frozen. Thus, the PCS can be used as heat transfer 1
fluid in supply networks. A promising PCS for heat or 0 0.2 0.4 0.6 0.8 1
w [-]
cold supply networks is paraffin/water dispersion.
Figure 3 is a photograph of a paraffin/water dispersion. Figure 4. TCEF PCS compared to water for temperature
Paraffin is the phase change material, which can be differences 10 and 15 K, diagram calculated with the
chosen according to the desired temperature of the properties of water and RT-42 of the company
Rubitherm [5]
phase transition, and water is the continuous phase of
the dispersion. In [4] paraffin/water dispersions are Using PCS with a mass concentration w of 0.4 would
investigated and their properties presented increase the heat transport capacity of the supply
network to 1.5 times of the value compared to water, if
42
th th
the temperature difference of the supply network is In view of the heat release in the room, the heat
15 K, and even 2 times, if the temperature difference is capacity Q can also be described by equation (4) and
only 10 K. Furthermore, the diagram shows that with is related to the heat transfer coefficient U, the heat
rising temperature difference the gradient of the TCEF exchange area A and the temperature difference
is lower, which means that the advantage of the PCS between the mean temperature of the heat release Tm
compared to water disappear at higher temperature as well as the room temperature Tr.
differences. At the point where the gradient of the
TCEF is 0, the water system and the PCS system have Q U A Tm Tr (4)
the same transport capacity. At that point, the mass
concentration of paraffin w has no influence on the
TCEF. The mean temperature of the heat release Tm is
calculated by equation (5).
The use of PCS in energy systems leads to an
improved energy transport capacity, which results in a Tin Tout
reduction of the necessary temperature difference or Tm (5)
T
volumetric flow rate of the transfer fluid needed to ln in
transfer a given amount of heat. Tout
Another technical issue of PCS systems is the
Based on the equations (3) to (5), it is possible to
increased pressure drop in the pipes due to the higher
calculate the NTU, which characterizes the heat
viscosity of the PCS. A calculation methods and
release in the room, according to equation (6), which is
measurement data can be found in [6, 7 and 8]. The
only a function of the inlet and outlet temperature Tin/out
viscosity of PCS is related to several influence
of the heat supply, the mean temperature Tm of the
quantities and can cause an incensement of the
heat release and the room temperature Tr.
pressure drop up 100%. PCS are non-newtonian fluids.
U A Tin Tout
NTU (6)
Tm Tr
2.2. Capillary Tube Mats
m c p
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 The NTU values have been calculated for a convective
should be small, so that also the heat exchange radiator system and a CTM system. The assumed
surface is small and the heating system must be temperatures for the calculation and the results are
operated on a high temperature level to ensure the given in table I.
heat transfer from the heating system into the room. An Table I. NTU for both heat release systems:
alternative to convective radiators are floor heating conventional radiator and CTM
systems. Floor heating systems consist of a capillary
parameter convective radiator CTM system
tube mat, which is installed in the upper layer of the
floor. Because of the bigger heat exchange surface Tin [C] 80 37
compared to the convective radiator, the temperature
Tout [C] 60 31
level of the heating system is lower. A new approach to
realise heating and cooling of buildings is via CTM, Tr [C] 20 20
which are integrated in the floors of the building, as well NTU [-] 0.4 0.43
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 The NTU value of the CTM system is 0.43 and as high
heat exchanger area, a low temperature difference as the NTU value of the convective radiator. This
between the heating system and room is possible. For means that both systems have the same heat release
the further discussion, the following simple model is capacity, although the inlet temperature Tin of the CTM
used to describe the heat release of the heating system system is lower and the temperature difference
in the building. The heating release system is between inlet Tin and outlet Tout of the CTM system is
evaluated by the number of transfer units (NTU). The smaller.
heat capacity provided by the heating network Q is
calculated by equation (3) with the inlet and outlet CONCLUSION
temperature Tin/out of the supply network, the mass flow
From the point of view of the low-ex concept the major
m and heat capacity cp of the heat transfer fluid.
task en route to an exergetically efficient energy supply
c p Tin Tout
Q m (3)
system is the replacement of the combustible fuel
boiler by utilization of low temperature thermal input
43
th th
flows such as industrial waste heat or geothermal ACKNOWLEDGEMENT

energy. To achieve this goal it is necessary to
This study was supported by the Project Management
decrease the medium temperatures within the
Juelich (PTJ) and the Federal Ministry of Economics
distribution networks first. A prerequisite is a low-ex
and Technology (BMWi) under 0327471A.
ready consumer that allows meeting the thermal
demands applying low temperatures. Comments of a highly constructive nature were
received from Daniel Wolf, Jorrit Wronski and Astrid
A possible realisation employs CTM in the heating
Pohlig.
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 REFERENCES
network, the utilization of PCS instead of pure water [1] C. Kemal et al., Evaluation of energy and exergy
enables the application of small temperature losses in district heating network, Applied Thermal
differences between forward and backward flow while Engineering, 24 (2004), pp. 1009-1017.
retaining the pipe dimensions. Since the backward flow
temperature mainly depends on the outlet temperature [2] H. Inaba, New challenge in advanced thermal
of the consumer system, small temperature differences energy transportation using functionally thermal
within the network automatically lead to low forward fluids, International Journal of Thermal Sciences,
flow temperatures. Consequently, the exploitation of 39 (2000), pp. 991-1003.
low temperature heat sources as input flows for the [3] M. Ala-Juusela et al., LowExergy Systems for
energy supply system is rendered possible. Heating and Cooling of Buildings, final report of the
Moreover, the decreasing temperatures in both forward IEA ECBCS Annex 37.
and backward flows of the network reduce the [4] L. Huang et al., Evaluation of paraffin/water
transportation heat losses. This leads in the end to a emulsion as a phase change slurry for cooling
reduction of energy input (quantitative aspect of the applications, Energy, 34 (2009), pp. 1145-1155.
low-ex concept) into the supply system.
[5] Rubitherm RT-42, datasheet 08/20/2009,
The only drawback suffered occurs in terms of an http://www.rubitherm.de, Rubitherm Technologies
increased pumping effort caused by a higher viscosity GmbH, Berlin (2010).
of the PCS in comparison with water. But, since heat
losses are the predominant factor over circulation [6] Yinping Zhang, et al., Experimental research on
pump energy, an overall benefit should be laminar flow performance of phase change
accomplishable. emulsion, Applied Thermal Engineering, 26 (2006),
pp. 1238-1245.
Summarizing it should be pointed out that applying
technologies such as CTM in the building heating or [7] A., B. Metzner et al., Flow of Non-Newtonian Fluids
cooling system and PCS as alternate heat transfer Correlation of the Laminar, Transition, and
medium for the distribution networks the low-ex Turbulent-flow Regions, American Institute of
concept can be realised, thus greatly enhancing the Chemical Engineers Journal, Vol. 1, No. 4 (1955),
efficiency of energy supply systems. pp. 434-440.
[8] R. Rautenbach, Kennzeichnung nicht-Newtonscher
Flssigkeiten durch zwei Stoffkonstanten, Chemie-
Ingenieur-Technik, 36 No. 3 (1964), pp. 277-282.
44
th th
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 Often, omitted from consideration is the quality of the

energy that is needed to provide comfort to the
The concept of energy efficiency, defined as useful
occupants of a building. While the heating
energy output as fraction of required energy input, has
requirements of a building can be determined (in GJ or
been used for years in technical systems assessments.
TJ), the nature or origin of this energy is not addressed
In addition to energy efficiency, there are benefits to
in energy efficiency calculations. The total amount of
using exergy efficiency to assess system performance.
Joules can be provided by oil, natural gas, electricity or
Whether systems will be installed or not is ultimately low temperature waste heat. While the first three
determined by their economic performance. This energy sources are considered high quality, and can be
performance is usually determined by comparing initial used to generate very high temperatures (over
investment cost and operational cost with revenues 1000 C), run equipment such as computers, radio and
throughout a systems lifetime in terms of payback time TV transmitters and receivers, waste heat is of low
or net present value. quality and has no other use. Comfort heating does not
require high temperatures and therefore using high
This paper describes a novel methodology that uses
quality fuel for low quality applications is considered
the concept of exergy and the thermoeconomic factor,
wasteful.
a ratio that compares investment-related cost and
exergy destruction cost, for the economic optimization Energy quality is often expressed as exergy. Exergy is
of a community energy system. It compares the cost of defined as the maximum useful work possible during a
exergy and the required capital and operational costs process that brings the system into equilibrium with a
including carbon taxes to accommodate this low quality heat reservoir. To illustrate the concept of exergy one
energy. In doing so it enables a quick way to properly can compare two different forms of the same amount of
assess the value of a systems ability to use low exergy energy: 100 kJ of energy is equivalent to:
energy inputs. The method is compared to a more
12 V/2.3 Ah stored in a car battery, or
traditional economic analysis.
1 kg of water at 43 C in a room with an ambient
INTRODUCTION temperature of 20 C.
In the last few years, we have become painfully aware Obviously, the energy contained in the battery is
of the effects of climate change. The burning of fossil considered more useful and therefore has the higher
fuels and the resulting emissions are thought to be a quality or exergy.
major contributor to the apparent increase of adverse
The ratio of Exergy (E) to Energy (Q) can be expressed
weather events. While people need energy for comfort,
as:
in some cases there may be a choice in the source and
nature of that energy. In addition to climate change,
E T
there is also a concern about the rapid depletion of the 1 ambient (1)
more valuable of fossil fuels, natural gas and oil. For Q Tsup ply
these reasons it makes much sense to re-evaluate the
sources of the energy we use and the effect of using where T is given in K.
them has on the environment. Equation 1 shows that when the supply temperature of
To lower energy requirements, energy efficiency has an energy source is high, the exergy converges to the
been practiced for many years. In terms of comfort energy value. Electricity and mechanical work are
heating in houses, most of the effort has gone into (nearly) perfectly convertible and the exergy content is
improving building insulation, better windows, building therefore equal to the energy content. Conversely,
orientation with respect to the sun, shading from solar when the supply temperature is closer to the
energy etc. In terms of energy conversion equipment, environmental temperature, the value of the exergy
improving the efficiency often meets natural limits, becomes (much) smaller than that of the energy.
such as those expressed by Carnots Law. 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
45
th th
of coins and bills, but neglecting their value. In an exergoeconomic factor is also found in other sources,
ethical society the value, worth and quality of different such as Temir & Bilge [5].
energy supplies should, as a minimum, be matched to
It is beyond the scope of this paper to provide a
the requirements of the different energy applications.
comprehensive literature overview of thermoeconomic
Methods to design low exergy buildings are available publications or even of the methods used in these
today. For instance, Schmidt [2] developed a method publications. The aim of this paper is to apply one of
and pre-design tool for low exergy buildings in which he these methods, using the above-mentioned
compared different heating systems, such as boilers, exergoeconomic factor to optimize building heating
condensing boilers, electric heating, GSHP and low systems connected to a district heating system. To the
temperature under-floor heating. However, this method best of the authors knowledge, so far this method has
does not directly address the effect of system heat only been applied to optimize individual components.
transfer surface area on the overall economics.
This work ties in with research into advanced low-
Also, there is an additional benefit realizing that a temperature district energy systems currently carried
building that can accommodate low exergy streams is out at the CanmetENERGY laboratories of Natural
ready for future hook-up to other, perhaps renewable Resources Canada in Ottawa, Canada.
energy sources: GSHP, solar, waste heat from
industry, energy from thermal storage to name a few. The system considered consists of buildings with their
This is a distinct advantage when the move to a heating system (radiators and cross-flow heat
sustainable society gains momentum, and the concept exchangers are considered), the energy centre with
of low-temperature heating should be incorporated in boilers and pumps and the pipeline to move the energy
building codes. in the form of hot water to the community. The
development of the methodology was the main object
This paper considers the cost of using the low quality of the study, not the optimization itself.
part of the energy source and the (increased) capital
cost and operating cost that are required to While the development of the optimization was related
accommodate low quality energy. A methodology has to economics, in other words, the least costly option, it
been developed to determine the optimal cost of should be noted that the concept of exergy opens up
operation, based on the capital cost, operational cost the notion of morals and ethics. For new
and the cost of the exergy. developments, the costs of resource depletion and
environmental destruction should be considered as
This type of analysis is considered part of the field of well. Just because a certain system is economic, it is
thermoeconomics, more in particular exergoeconomics. not necessarily the best moral or ethical choice. Just
Wikipedia defines thermoeconomics in a very because a certain system does not cause local
theoretical way as a school of economics that applies problems, that does not mean that (environmental or
laws of thermodynamics to economy. Valero et al. [3] other) problems caused by this system elsewhere can
operationalize this definition by describing two aims of be ignored.
thermoeconomics, (1) optimization to minimize cost of
a system or component, and (2) cost allocation of Traditional Optimizations
individual outputs of a plant producing a number of System optimization is often done by optimizing
outputs. systems separately, and not by considering the overall
Valero and coworkers [3] date this research field back efficiency of integrated systems. Often, an integrated
as far as 1932, when Keenan apportioned cost of heat approach leads to optimal solutions, as in electricity
and work taking into account irreversibility and generation using a back pressure steam turbine.
thermodynamic efficiency instead of enthalpy only. Accepting a lower efficiency of the turbine may lead to
However, they go on to say that Gaggioli, and Tribus the residual energy in the condenser being useful in
and Evans in the early 1960s started off real other applications, whereas in the separately optimized
development in thermoeconomics. Ever since, these version this thermal energy would be useless. In the
fields have received tremendous attention. Valero and latter case, the turbine back pressure is kept as low as
coworkers identify that an important problem in this possible, to extract the maximum electrical power. This
body of research is the variety of methodologies used makes the condensate of too low a temperature to be
with accompanying nomenclature. Between them and useful in other applications. Optimizing integrated
Tsatsaronis [4] they already name a fair amount of systems as a whole avoids this problem.
methods. In doing so, Tsatsaronis introduces the
exergoeconomic factor f, as a fraction that compares Exergoeconomic Optimization
two sources contributing to cost increases, investment- In an exergoeconomic optimization, the concept of
related cost and exergy destruction cost. This exergy is used to determine the best and most
economic solution to an energy conversion process or
46
th th
system. While the total quantity of transferred energy the design temperature for Ottawa. While this is an
remains the same, the exergy that delivers this energy over-simplification of reality, it neither hinders the
may vary. Analyzing the required exergy with respect to development of the methodology nor introduces
the energy transfer equipment will result in an optimum serious errors of consequence.
economic solution allowing for integration of the system
with other systems. Energy Transmission
In this exergoeconomic optimization, the system is The pipe diameters were estimated using the
treated as an integrated whole together with other RETScreen software tool. This means for diameters
systems. While the (comfort) energy supplied remains under 400 mm the pressure drop is kept below 200 Pa
the same in any given scenario, the exergy required for per meter of pipe and for larger diameters flow velocity
this scenario varies and the cost implications of this is maximized at 3 m/s [6]. As RETScreen has a limit of
variation are included in the analysis. Therefore, in this 13 sections for district heating systems, the 1000
analysis the consumer of energy does not pay for the homes were assumed to be located along twelve 80-
energy, but for the exergy, the real value of the energy home streets. The final 40 homes were located in a
supplied. separate street.
The energy transfer fluid is water. The pipes are
SYSTEM DESCRIPTION preinsulated steel or cross-linked polyethylene (PEX)
pipes. The first iteration of the methodology accounted
The system considered here to develop the
for heat losses from the pipes. Since it was found that
methodology is a district heating system supplying hot
this heat loss was a negligible fraction of the
water for space heating to a 1000 home community in
transmitted energy, it was omitted in subsequent
the Ottawa area in Canada. It was modelled using the
versions. For a thorough analysis, it is recommended to
RETScreen clean energy project analysis software tool
include heat losses, especially if the piping system is
[6] and in-house spreadsheet based models. The hot
extensive and the supply temperatures reach high
water is transported from an energy centre located
levels.
centrally in the community to the 1000 detached
homes. Inside the buildings, radiators or cross-flow A pressure drop analysis was used to determine the
heat exchangers (water-to-air fan coils) are employed required pump energy. The electric motor driving the
to provide space heating. pump was estimated to have 90% efficiency while the
pump was assigned an efficiency of 85%.
Energy Supply
End-Use
The building temperature set point is kept constant at
20C. The hot water supply temperature is determined The energy supplied to the pipeline was used to keep
by the outdoor temperature. If the outdoor temperature the building temperatures at set point. Therefore,
is above 5 C, the supply temperature is 70 C. When regardless of the (size of) building heating system
the outdoor temperature drops below -15 C, the used, the same energy was used to keep the buildings
supply temperature equals 90 C. Between 5 C and warm. However, the exergy used was dependent of the
-15 C, the supply increases linearly from 70 C to system in place and of its size. The larger the size, the
90 C. This is a common supply temperature profile lower the required water temperature and hence less
used in many European district heating systems. It exergy was required to achieve the same end result.
prevents excessive flows in the pipes at high loads and Two different technologies were used to model the
permits smaller heat transfer surfaces in the buildings transfer of energy into the building space: cross-flow
due to the higher temperature difference between heat exchangers, and radiators. Simulations were done
water and building air. When the heat transfer surface for both technologies separately, and the technologies
area was varied to reach an optimum solution the were never mixed. This was done to simplify the
supply temperature was adjusted by a constant value analysis. In reality, mixed systems will occur and
over the entire load range. The water return should be analysed as such. While this will increase
temperature was set at 30 C in all design calculations, the level of modelling complexity, it is not difficult to do.
but varied throughout the year according to the off-
design characteristics of the heating equipment used.
DESCRIPTION OF MODELING
The load of the buildings is related to the outdoor
temperature. The annual heat consumption was set at Climate
100 GJ per house, a typical value for detached homes The local climate has a significant effect on the design
in this area. The instantaneous load throughout the of a building heating system. A maritime climate may
year is simply calculated as a linear relationship have many degree days but not show the variability in
between zero and the maximum capacity, when the demand that a building in a continental climate with an
outdoor temperature varies between 20 C and -28 C, equal amount of degree days experiences. Even if both
47
th th
buildings use the same amount of energy per year, the temperature. The pumping energy required was
demand load in the building with the continental climate included in the modeling. Since electrical energy is
may be far greater. Therefore, the climate plays an equivalent to exergy, the pumping energy calculated
important role in the design of a heating system and from the pressure drop calculations (including
should, therefore, be considered in this analysis. efficiencies), was numerically counted as exergy.
Normally, during periods of no-load, the pumps keep
Supply Temperature
operating to keep a supply of design temperature water
To include the effect of the variability of the energy close to the load. This is done with a thermostatically
demand with time-of-day and the seasons, the operated by-pass valve. Since this valve represents a
statistical average hourly temperatures for the city of constant effect which does not affect the optimization, it
Ottawa were used. These temperature values are real was not modelled for simplicity.
values, with realistic variability (high and low
temperatures), using time-periods from different years Design of Cross-Flow Heat Exchanger
to provide for a correct average. In total, 8760 hourly
The design of the cross-flow (or fan-coil) heat
values of temperature were used in the spreadsheet,
exchanger was based on the assumption that the
as shown in Figure 1.
overall heat transfer coefficient U was 25 W/(m2K).
40
The F-factor was set at 0.94. To meet the design load,
heat exchangers with a combined area of 17,152 m2
30
were required.
20
Design of Radiator Heating System

Temperature (C)
10
0 The design of the radiators was done in a very simple

0 1000 2000 3000 4000 5000 6000 7000 8000 9000
-10 manner. It is acknowledged that better methods exist,
-20
but the development of the methodology did not suffer
because of this simplification. For any optimization,
-30
actual modelling of the equipment should take place.
-40
To determine the heat transfer from the panels, the

Hour (-)
Fig.1. Average hourly temperatures in Ottawa general radiation equation
Solar Radiation Q Tpanel

4
Troom
4
(2)
To simplify the spreadsheet calculations, the effects of
solar radiation, plug loads and occupancy gains were
neglected. When performing the optimization, these was used with the average panel temperature. The
effects remain constant and so have little effect on the surface emissivity was estimated at 0.9. Convection
final outcome. When using this method for design, from the surfaces was not separately considered.
these contributors to the exergy balance should be To meet the design load, 42,271 m2 of radiative
considered. surface was required to meet the design load.
Determination of Instantaneous Load Determination of Exergy Use

The maximum thermal load of the community for space As indicated in Section 1, the ratio of Energy to Exergy
heating was determined using the average hourly can be expressed as:
temperatures and the annual heat consumption per
house of 100 GJ. It turned out to be 10,640 kW. When E T
the outdoor temperature reaches -28 C, the Ottawa 1 ambient (1)
design temperature, the community requires the
Q Tsup ply
maximum thermal load. At the ambient temperature of
20 C, the load is nil. The modelling is set up so that Where E is exergy, Q is energy and T is the
between these ambient temperatures, the load varies temperature given in K.
linearly. For instance, at -4 C, the load equals 5.32
MW. For this study, knowing the energy supplied, Equation
(1) was used to calculate the supplied exergy for each
Pumping Power and Exergy hour interval. For each of the heating systems used
and each of their variations is size, the amount of
To meet the load, the water had to be pumped from the
energy supplied to the heated space remained the
supply source to the load. The amount of water
same. However, due to the supply temperature
pumped varied with the load and the supply
48
th th
requirements, and the different flow requirements, the per ton CO2eq. Carbon intensity factors of 0.050 ton
exergy used by each system was unique. CO2eq/GJ were used for natural gas and 0.054 ton
CO2eq/GJ for electricity (taken from RETScreen [6] as
Capital and Exergy Cost Assessments representative for Canada). This results in a $6.5/GJ
All cost numbers reported in this paper are in 2009 energy charge for heat, a $38.9/GJ exergy charge for
Canadian Dollars. An in-house costing tool was used to heat and an $18.6/GJ energy (or exergy) charge for
estimate cost for the district heating energy centre, electricity.
containing the pumps and boilers, and the buried
distribution piping. For the two heating systems Thermoeconomic Factor
considered, the costs were assigned as shown in The exergoeconomic or thermoeconomic factor f
Table I. For water-to-air fan coils (cross-flow heat compares two sources contributing to cost, investment-
exchangers) an installed cost of $250/m2 was related cost and exergy destruction cost. It is defined
considered representative, and for radiators $200/m2 here as the ratio of Capital Cost Rate (CCR, which
was selected as a typical value. includes O&M cost, but excludes heat and electricity
cost) and the sum of Exergy Destruction Cost Rate
Table I. Cost for heating technologies
(EDCR) and CCR. The CCR equals the cost per unit
Cross-flow heat
2
$250/m of heat transfer surface time for the installation, depreciation, maintenance, etc,
exchanger while EDCR is the cost of exergy.
2
Radiative system $200/m of exposed panel
CCR
f (4)
EDCR CCR
Future cash flows were discounted at a rate of 8% and
system lifetime was set at 40 years. Annual operating
Since CCR and EDCR have the dimensions of $/time,
and maintenance (O&M) cost other than cost for heat
f is dimensionless.
and electricity were set at a fixed fraction of 1% of total
investment cost. A high value for f indicates that the capital and
maintenance costs are dominant. Also, a high f value
To compare traditional optimization with exergo-
indicates good use of the exergy in the fuel. On the
economic optimization three types of analyses were
other hand, a low value for f indicates an inefficient
performed. The classical analysis applies the
use of fuel resources. For each heating system
traditional optimization where heat is valued based on
variation, the average annual thermoeconomic factor
energy content, at a rate of $5/GJ, which is considered
was calculated.
representative for heat from natural gas combustion.
Electricity cost has been set at $17/GJ (just over
$60/MWh). MODELLING RESULTS
In the exergoeconomic analysis heat and electricity are Base case design
priced based on the exergy content. The exergy charge Table II shows the main information for the base case
was determined at $30/GJ for thermal energy, based designs for both the radiator and cross-flow heat
on the above mentioned $5/GJ for heat, assuming a 1 exchanger systems. As expected, the distribution pipe
to 6 ratio of exergy to energy content (applies to a diameters, required pump capacity, annual space heat
temperature around 80 C). The electrical energy to consumption and annual heat cost are the same for
exergy ratio was taken as one, resulting in an exergy both systems.
charge of $17/GJ for electricity. At first glace it may
As the water return temperatures throughout the year
seem erroneous to charge more for exergy from the
are generally lower for the radiator system, the required
thermal source than that for the electricity for the pump,
water flows and consequently the annual electricity
but it must be remembered that the (thermal) exergy is
consumption are lower for the radiator system. As both
a fraction of the thermal energy.
systems have a design supply temperature of 90 C
The third type of analysis is a classical analysis (and thus also the same off-design supply
corrected for the difference in value of low- and high- temperatures throughout the year), the annual exergy
temperature heat, by assuming energy under 60C is consumption is the same for both. The lower return
available free of charge (as waste heat from a nearby temperatures for the radiator system also show in the
process). For energy over 60 C the charge is still higher fraction of energy provided under 60 C. In
$5/GJ. terms of cost, the radiators are clearly more expensive
resulting in higher annual investment and O&M cost,
To assess the influence of carbon taxes, two sets of
which is not offset by the somewhat lower electricity
results are presented. One assumes no carbon taxes
cost. Overall the more capital intensive radiator system
are in place and the other assumes a carbon tax of $30
has a higher f-factor than the cross-flow heat
49
th th
exchanger system. In a comparison between the two Alternative designs radiator system
systems, the cross-flow heat exchanger system works For both the radiator and the cross-flow heat
out cheaper using all three types of analysis due to the exchanger system alternative designs with increased
large difference in investment cost. and decreased surface areas were costed. The district
heat supply temperatures were modified accordingly,
Table II. Main information base case designs without and as noted before, the required water flows and thus
CO2 tax. distribution pipe diameters and pumping power
requirements were modified too. The effects of these
Cross-flow variations on cost were taken into account.
heat
Radiator exchanger
The results of all the modelling runs are shown in the
figures below in the form of the relationship between
Surface area (m )
2
42,271 17,152 the annual cost (the sum of capital investment, O&M
cost and energy or exergy costs) and the f-factor. An
Distribution pipe increasing f-factor means increasing surface areas
diameters (mm) DN80/DN65 DN80/DN65 (and thus increasing capital and operating and
maintenance cost) and decreasing heat supply
Required pump capacity
temperatures (and thus decreasing exergy cost).
(kW) 32.2 32.2
Figure 2 shows results for the radiator based heating
Annual electricity system with no carbon taxes in place. The slight jump
consumption (GJ) 115.2 147.6 in annual cost around an f-factor of 0.81-0.82 is caused
by an increase in district heating piping diameter from
Annual exergy DN65 to DN80 for the 80-house streets and from DN50
consumption (GJ) 16,810 16,810 to DN65 for the 40-house. All lower f-factors shown
have piping diameters of DN65 and DN50 and all
Annual space heat higher f-factors shown have DN80 and DN65
consumption (GJ) 100,000 100,000
respectively.
Fraction of energy < 60 C 68.4% 64.4% $2,900,000
$2,700,000
Installed cost heaters $8,454,298 $4,288,066 $2,500,000

Annual cost ($)
$2,300,000
Investment cost district $2,100,000
heating system $11,556,386 $11,556,386 $1,900,000
$1,700,000
Annual O&M cost $200,107 $158,445
$1,500,000
0.68 0.70 0.72 0.74 0.76 0.78 0.80 0.82 0.84 0.86 0.88
Annual charge investment f-factor (-)
and O&M cost $1,878,206 $1,487,163 Classical analysis Exergo-economic analysis Heat under 60C free analysis
Annual heat (energy) cost $500,000 $500,000 Fig. 2. Relation between f-factor and annual cost radiator
system, no carbon tax.
Annual heat (exergy) cost $504,193 $504,193 The classical analysis shows a continuous increase in
annual cost with increasing f-factor.1 This makes sense
Annual electricity cost $1,958 $2,509 because cost is not based on exergy but on energy.
Therefore, an increasing surface area means
Total annual cost classical increasing capital cost, but constant energy cost, so the
analysis $2,380,164 $1,989,672 lower exergy requirement does not offset the increase
in capital cost. The classical analysis would tell us to
Total annual cost
optimize the system with minimum capital expenses. In
exergoeconomic analysis $2,384,357 $1,993,865
reality there would be a limit as ever increasing
temperatures will mean that we are dealing with more
Total annual cost heat
under 60 C free analysis $2,038,164 $1,667,672 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
f-factor 0.814 0.782
1
Although exergy is not explicitly costed in the classical analysis,
we can still calculate an f-factor.
50
th th
will increase with increasing supply temperatures, Alternative designs cross-flow heat exchanger
which is not modelled here. system
The exergoeconomic analysis does take into account Figure 4 shows the results for the system with cross-
the increasing exergy requirements for systems with flow heat exchangers. Note again that the jump in
lower surface areas. Consequently, there is a minimum annual cost at an f-factor around 0.78 is due to the
annual cost (for an f-factor in the range 0.710.73). increase in district heating pipe diameter. As for the
Going to lower f-factors, exergy cost significantly radiator system, the classical analysis shows a steep
increase, which results in increasing annual overall slope with increasing f-factors as capital cost are
cost. Going to higher f-factors, the annual cost increase dominant and lower exergy requirements do not
again because the decreasing exergy cost are more translate into cost savings. Again the classical analysis
than offset by the increase in capital and O&M cost. would lead us to minimize the surface area (with the
same limitations as applied to the radiator).
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 $2,200,000
price for exergy in the calculations. As a result, the line $2,100,000
does not slope up as strongly with increasing f-factor $2,000,000
as the line pertaining to the classical analysis. Going to
Annual cost ($)

$1,900,000
higher f-factors, eventually all heat will be delivered $1,800,000
under 60 C, and all heat provided will be free. Going to $1,700,000
lower f-factors, eventually all heat will be supplied at $1,600,000
temperatures over 60 C and the green line will $1,500,000
coincide with the blue classical analysis line. The heat $1,400,000
0.65 0.67 0.69 0.71 0.73 0.75 0.77 0.79 0.81 0.83
under 60 C free analysis does not show an optimum f-factor (-)
and would suggest minimizing the f-factor. Like the Classical analysis Exergo-economic analysis Heat under 60C free analysis
classical analysis the economic analysis suggests that
capital cost are dominant. Fig. 4. Relation between f-factor and annual cost cross-
flow heat exchanger system, no carbon tax.
Figure 3 shows the effect of introducing a carbon tax of
$30/tonCO2eq on the exergoeconomic analysis. It is The exergoeconomic analysis shows a downward
clear that the carbon tax leads to higher annual cost sloping line. This is caused by the reduced capital cost
and lower f-factors for the same systems, both caused and O&M cost compared to the radiator system and
by the increased exergy cost. Both lines show an thus increased importance of exergy cost as fraction of
optimum for an f-factor in the range 0.710.73, but for the total cost. An increase in cost due to surface area is
the case without carbon tax the corresponding surface more than offset by a decrease in exergy cost.
area is lower than for the case with carbon tax. This Contrary to the radiator system, though, the
makes sense as increasing heat and exergy cost mean exergoeconomic analysis does not show a clear
a shift to a system with higher surface areas and lower optimum, although it clearly levels off at higher
heat and exergy requirements. As figure 3 shows, the f-factors. It is interesting to note here that the classical
capital and O&M cost as a fraction of total cost (and analysis and the exergoeconomic analysis lead to
consequently also the exergy cost as a fraction of total contradictory recommendations as to optimization.
cost) remain in the same range.
The analysis with free heat under 60 C shows a line
gradually sloping up, though far less pronounced than
$2,750,000
$2,700,000
the classical analysis line. Like the classical analysis
$2,650,000 line it would indicate that lower surface areas would
$2,600,000
optimize this system.
Annual cost ($)
$2,550,000
$2,500,000
$2,450,000
Figure 5 shows the effect of a carbon tax on the
$2,400,000 exergoeconomic analysis. As for the radiator system
$2,350,000
$2,300,000
the carbon tax means higher annual cost and lower
$2,250,000 f-factors for the same system due to increased exergy
$2,200,000
cost. As there is not a clear optimum in either line, we
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 (-) can not conclude that the optimum f-factor is the same
No carbon tax Carbon tax $30/tCO2 for both. However, it is clear that both level off in the
higher f-factors range.
Fig. 3. Relation between f-factor and annual cost radiator
system, with and without carbon tax.
51
th th
$2,500,000 combining heating technologies and possibly including

$2,400,000
other technologies such as under-floor heating provide
further opportunities to optimize system cost.
$2,300,000
Annual cost ($)
$2,200,000
$2,100,000
In addition, the application of the methodology
$2,000,000
developed in this study should be applied to a heat
$1,900,000
$1,800,000
pump, where the variations in COP with supply
$1,700,000 temperature would be included. This would result in the
$1,600,000 ability to match the heating equipment to the heat
pump, resulting in an optimum operation.
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 (-)
No carbon tax Carbon tax $30/tCO2
Fig.5. Relation between f-factor and annual cost cross- ACKNOWLEDGEMENT

flow heat exchanger system, with and without carbon tax.
During this work the authors have had very fruitful
From the foregoing, it is clear that useful comparisons conversations with many colleagues: Mikhail Sorin,
can be made using this methodology. The results from Evgueniy Entchev, Libing Yang, Ibrahim Dincer, Hajo
exergoeconomic analyses can significantly deviate Ribberink and Kirby Wittich. These discussions helped
from those obtained with a classical analysis. Which of focus the work and stimulated further thinking in this
the two is the more relevant one will depend on the interesting area of science. This is to thank all those
situation. For non-integrated systems, the classical who spent their valuable time listening and providing
analysis may be the one to follow, but for integrated valuable comments.
energy systems, which are expected to become more
and more important, the temperature level of heat REFERENCES
becomes important, and the exergoeconomic analysis
[1] G. Wall, Exergy and Morals, in Second law
seems more appropriate. Using the f-factor will help in
analysis of energy systems: towards the 21st
finding optimum solutions, especially for exergo-
century, E. Sciubba, M.J. Moran Eds, Circus,
economic analyses.
Roma (1995), ISBN 88-86662-0-9, pp. 21-29.
Variations in external factors, such as fuel costs or
[2] D. Schmidt, Design of Low Exergy Buildings
Government / utility incentives could change the shape
Method and a Pre-Design Tool, in International
of the curves to make the minimum more pronounced.
Journal of Low Exergy and Sustainable Buildings,
Vol. 3 (2003), pp. 120-126.
CONCLUSION AND SUGGESTIONS FOR FURTHER
WORK [3] A. Valero, L. Serra & J. Uche, Fundamentals of
Exergy Cost Accounting and Thermoeconomics.
From the results of testing the methodology of Part I: Theory, in Journal of Energy Resources
exergoeconomic optimization using the f-factor, it is Technology, Vol. 128 (2006), pp. 1-8.
clear that it is a useful tool to determine the effects of
different heating technologies and heat transfer surface [4] G. Tsatsaronis, Application of Thermoeconomics
sizes of these technologies on the annual overall to the Design and Synthesis of Energy Plants, in
operational costs. This is especially true if the heating Exergy, Energy System Analysis, and
system is integrated with other energy systems. It is Optimization, [ed. Christos A. Frongopoulos], in
also true if the temperature level of the heat is Encyclopedia of Life Support Systems (EOLSS),
important for another reason. The methodology can be developed under auspices of the Unesco, Eolss
used to make informed choices regarding technologies Publishers, Oxford, UK (2007).
to be used for heating homes or buildings and [5] G. Temir, D. Bilge, Thermoeconomic analysis of a
regarding the size of these technologies. trigeneration system in Applied Thermal
To continue this development work, it is recommended Engineering, Vol. 24 (2004), pp. 2689-2699.
that more practical considerations will be incorporated [6] Clean Energy Project Analysis RETScreen
into the models and analyses. Increasing temperatures Engineering & Cases Textbook, 3rd edition,
do not just cost more in terms of exergy but also in RETScreen International, Natural Resources
more expensive materials, and steam based district Canada, Varennes (2005).
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
52
th th
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 installations. Specific to the secondary network are the

post-insulated steel distribution pipes and connections
This paper describes the innovative integral approach
to customer installations hanging in narrow crawl
improving district heating network efficiency, SlimNet.
spaces under blocks of buildings.
SlimNet consists of five phases which lead to annual
energy savings of about 227.000 GJ and almost 37.000 In the distribution process no heat exchangers are
ton CO2 savings for the city of Purmerend in 2015. used except from the production of hot tapping water in
the houses.
INTRODUCTION Hydraulics are controlled by decentralized pressurizing
valves, differential pressure valves and pumps
Company situation compensating for hydraulic deficiencies.
In 2007 the new company Stadsverwarming
The supply temperature from production is directly
Purmerend B.V. (SVP) took over the responsibilities of
related to the ambient temperature (i.g. 95 C at
the district heating network from the municipality in
Ta=-10 C and 75 C when Ta=15 C). The maximum
Purmerend, the Netherlands. With 25.000 customers
supply pressure to the primary network is 6,8 bars and
the grid is the fourth largest grid of the Netherlands.
to the secondary network 4,5 bars.
District heating Purmerend started in 1980. The
network expanded organically following the city
expansions. While daily operations were outsourced to NETWORK CONDITION
external and changing partners, the final responsibility Part of the business analysis was an extensive
stayed with the municipality. technical research program covering all technical
A comprehensive business analysis performed by the aspects of the grid and finally entire district heating
new management in 2008 showed severe problems. In chain. The main conclusions were:
the present state the company would remain 1. The network characteristic had become
structurally loss giving, (future) heat delivery was not uncontrollable: Network builds out has occurred
ensured, and sustainability and customer satisfaction without a master plan. Effectively SVP had no
were below benchmark standards. Fall 2009 a new control on the characteristics of customer
business plan was presented that sets course for a installations. Furthermore, hydraulic problems in
future proof company, based on sustainable, cost- the grid had been masked with decentralized
effective and 80% renewable heat. On the technical pumps and control systems.
side this is achieved by two major project programs, a.
improving network efficiency, SlimNet, and b. 2. Heat production capacity was critical, reaching a
incorporation of sustainable energy sources, the critical limit under the conditions of the winter of
Energy transition. The company mission is to become 2008. There was certainly no spare capacity to
the most sustainable district heating company of the facilitate the planned expansion of the grid and
Netherlands. thus the heat demand as shown in Fig 1.
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
Fig. 1 Required heat production
secondary network to the 25.000 customer
53
th th
3. In 2008 the network showed a heat loss factor of Primary network

33,6% (with a Dutch benchmark of 25%). requiring Most substations in the network are provided with a
32.683 m3 of water replenishment in the same SCADA6 system. This data in combination with a newly
year. developed network model made it possible to
4. Parts of the network showed excessive heat loss calculating annual heat loss at 100.706 GJ.
and repairs, mainly due to high ground water table, According to [3] about 14% of this heat loss is caused
exposing the pipes in crawl spaces directly to by cross-linked polyethylene (PEX) piping material
water for most of the year. Repairs with standard used in the early 90s.
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)
Fig. 2 IR scan of a PEX pipe constructed in 1990
C. Developing analyzing tools
D. Developing and defining measures Considering that those PEX pipes are applied in only
3,5% of the primary network, these may be referred to
E. Quantifying KPI results from SlimNet
as hotspots.
In the following those phases will be discussed.
Secondary network
KNOWING WHERE THE HEAT FLOWS With four public housing companies, SVP conducted
For SVP the heat losses are defined as: research on failures in the district heating related
systems in Purmerend [3]. It became clear that during
Qloss Q produced Qsold (1) the period 2006-2008 74% of the unplanned repairs
were caused by the high ground water level in the
The heat losses in the network, Qloss, were 427.158 GJ crawl spaces where post-insulated steel pipes with
(33,6%) in 2008. Causes for those losses5 are: Armaflex insulation are installed. In total research
1. Losses in buffering tanks identified areas of 4000 houses, where heat loss was
extreme, i.e. hotspots.
2. Losses in primary network
This research confirmed the conclusion of an earlier
3. Losses in secondary network
research [4] that the thermal conductivity k for the wet
4. Undefined losses insulation in the crawl spaces will be close to 0,1 W/mK
None of the above can be determined exactly within and 0,2 W/mK instead of the 0,02 or 0,03 W/mK for the
the boundary conditions of the network but the current pre-insulated pipes. The total of heat losses in
following describes the results of the research the secondary network are estimated at 304.041 GJ.
performed on this matter and the localization of
hotspots, parts of the grid with excessive losses. Conclusion addressing heat losses
Table 1 gives the overall results of the heat loss
Buffering tanks analysis.
In [1] an estimated calculation was made for the heat Table 1: Overall results of heat loss analysis
losses due to the buffering tanks, 5.562 GJ annually.
Main network part Loss(GJ) % of total
There are four buffering tanks with a 4.000 m3 capacity
in the network which are used for peak shaving. A Buffering tanks 14.032 3,3 %
check upon this calculation [2], based upon an IR-scan Primary network 100.706 23,6 %
of one of the buffering tanks resulted in an estimate of Secondary network 304.041 71,2 %
14.032 GJ annually which is considered to be a Undefined losses 8.170 1,9 %
maximum value. Total 427.158 100%
5 6
Losses from heat plants are not taken into account. Supervisory Control And Data Acquisition
54
th th
It was concluded that replacing the PEX pipes in the DEVELOPING ANALYZING TOOLS
primary network and post-insulated pipes in the crawl
Research had located the hotspots of unplanned
spaces of houses in the areas identified as hotspots
repairs and heat loss in an area of 4000 houses. These
was the most effective strategy for heat loss reduction.
hotspots were responsible for 50% of the unplanned
repairs. In order to define and implement a suitable and
DEFINING KEY PERFORMANCE INDICATORS cost-effective replacement strategy a set of tools was
The main goal of SlimNet is improving network developed.
efficiency as part of the new business plan that sets
course for a future proof company which provides Upgraded network diagram
sustainable, cost-effective and 80% renewable heat. Analyzing networks requires reliable and
The Key Performance Indicators (KPIs) can be divided comprehensive network diagrams. All required
in four main criterions: information such as dimensions, age, depth etc. should
be available in the diagram. Many network diagrams
1. Economics
are drawn using CAD-software. Analyzing from those
2. Sustainability
drawings is costly. It therefore was chosen to revise the
3. Reliability
diagram completely and apply the possibility to add
4. Customer Satisfaction
element attributes to the drawing connected to an
Economics integral database system. The upgraded network
Every GJ of heat lost in the network cannot be sold and diagram had a catalytic effect on two other models, the
has therefore a negative effect on the balance sheet. network model and the grid valuation model.
Consequently the heat loss in the DH-network is an
obvious and important KPI. Network model
In 2009 SVP replaced the outdated and inadequate
Another parameter that has a negative effect on
network with a validated dynamic model (TERMIS),
profitability is the amount of water that is replenished.
developed by 7-Technologies with COWI as system
integrator. With the upgraded network diagram SVP
Sustainability
had the first and validated model of the primary
The avoided CO2-emissions are and should be an network within five months.
important driver for DH grids. According to subsequent
directives in the Netherlands for assessing energy In combination with a new CRM system, operational
performance of buildings NEN 7120, the avoided since 2010, SVP will soon be able to tap into the
CO2-emissions has to be determined on the required information on customer behavior and consumption.
primary energy sources and by referring to common This will allow SVP to dynamically calculate the current
state-of-the-art technologies. The HR-107 type (107% state of flow, pressure and temperature throughout the
LHV efficiency) is the required and accepted common network at a configurable cycle time. Additionally, every
state-of-the-art reference technology. real-time model calculation cycle will include a forecast
simulation for a given period. This allows SVP to be
Reliability abreast of demands, enabling optimization of
operations and planning of the future.
The condition of the network in terms of reliability
presents itself in the amount of times that mechanics
Valuation model
have to deal with unplanned repairs. It was apparent
that SVP was facing an increasing trend curve. The The upgraded network diagram supplied database
actual deprecation of the replaced piping provided a information on lengths, dimensions, age and type. With
another criterion for assessing system degradation. the following equations added to the database it was
possible to develop a valuation model, that could help
Customer satisfaction to prioritize and direct renovation efforts.
Reducing off time, during replacement was an
important element of the SlimNet approach. network
KPI summary
X R
x 1
x Lx (2)
1. Heat loss
2. Water replenishment
3. Avoided CO2 emissions network
D Ax
4. Unplanned repairs Y
D
R x Lx (3)
5. Network degradation x 1
6. Off-time during replacement
55
th th
network
D Ax 1 Strategic metering
Z Y x 1

D
Rx Lx

(4) It was concluded that actual data on heat loss on
smaller scale (houses and clusters of houses) would
X = value network in new state () facilitate decision making on future renovation projects
and grid management. To get hold of this information
Y = current network value () SVP installed heat meters with radio transmission
Z = required annual maintenance costs () modules on strategic positions in the network. Together
with the metering data from heat meters in customer
x = pipe installations this firstly gives accurate data on the heat
Rx = construction costs per meter pipe dimension x () loss in the corresponding part of the grid. This setup
will also provide us with empirical data on the long term
Lx = length of pipe x (m) results of network improvement measures.
D = lifetime expectancy (year) In order to make the data comparable, two areas where
Ax = age of pipe x (year) chosen. One with the new SlimNet approach (Using
polybutene pipes and new construction techniques)
The network degradation is defined as factor : and one with conventional material and construction
techniques. First comparative results will be available
Y
(5) by the end of 2010.
X
From consultation with amongst others COWI, it was Leak detection
concluded that networks with a < 0.5 are in a critical Most producers of pre-insulated pipe systems offer the
stage. possibility of leak detection wiring. Using a master plan
with proper zero and recurrent measurements this
For the entire grid the was above the threshold. would be a reliable method of leak detection.
Discriminating the for separate grid sections helped Unfortunately this is not applicable to the situation in
to identify the hotspots and monitoring will help to Purmerend.
determine the effect of SlimNet.
With one of its partners SVP developed a method using
Sustainability assessment model tracer gas to detect leakages. The detection devices
proofed to be very sensitive and with this method
To assess the current sustainability results of the
almost 2.500 houses have been inspected this year
network, SVP developed a sustainability assessment
and last year. Leakages were detected in 3% of those
model in accordance with Dutch law and guidelines,
cases, mostly in an early stage, that otherwise would
resulting in Fig. 3 [5]. This model can also predict the
only have been detected through visual sighting of
effects of optimization in the chain from production,
damp.
distribution and delivery to customer installations
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
Fig. 3 CO2 reduction DH-network Purmerend in past be gained by digging in the gardens, meant huge
inconvenience for customers.
It appeared that the ratio of CHP operation to the total
of heat produced and the heat loss factor have the Parallel to this SVP had challenged pipe manufactures
biggest impact on the sustainability results. to come up with innovative material construction
methods, suitable for the Dutch situation (groundwater
56
th th
and retrofit in narrow crawl spaces). The only viable SlimNet part I: Renovation and smart redesign
solution came from Flexalen of Thermaflex, using Applying Flexalen means an improvement of k from 0,1
flexible polybutene (PB) carrier pipes. The producer of of the wet post-insulated steel pipes to theoretically
the PB material offers a 50 years plus life guarantee [6] 0,031 W/mK (manufacturer information, at 50 C).
for the pressures and temperature profiles of the SVP
network. Key to the SlimNet approach was smart redesign.
Calculations in TERMIS showed that many parts of the
A pilot with Flexalen was conducted in September DH-grid in Purmerend are generally oversized, and that
2009. The pilot used prefabricated joints of Flexalen, the common circular grid can easily be changed into a
called Flexalinks, which were under research and star shaped grid, whilst reducing pipe lengths. Using
development at that time. The pilot did meet all the TERMIS redesign focused on reducing radial
requirements. Costs were reduced by 30% compared dimensions and pipe lengths by deleting obsolete
to the steel flex pilots, 16 houses were overhauled pipes.
within a week and access could be gained by the crawl
space hatches. The results for the part of the grid that is replaced this
year, Fig. 4 and Fig. 5, gave, Table 2 [8]:
On the basis of this pilot decision has been made to
retrofit 4000 houses within four years. Works has Table 2: Results from redesign 2010 area
currently started at the first 309 houses, at a speed of Heat demand Heat loss
30 houses a week.
Current situation 100,0 % 100,0 %
The second part of the challenge (b): improving
New dimensions 93,0 % 76,3 %
network efficiency, is furthered by SlimNet through
optimizing pipe dimensions and lengths (smart grid Finger system 91,0 % 69,5 %
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]:
(Tin To u t)
Qlo ss _ p ip e 2 k L (6)
r
ln o u t
rin Fig. 4 Existing network part to be renewed
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.
Fig. 5 Redesigned and renewed network
57
th th
The actual effect of SlimNet on heat losses will be houses that have a 9050 C characteristic during
closely monitored in the grid, through the strategic design conditions (-10 C). In most areas before that
metering project. time SVP found return temperatures that are
structurally higher than the required 50 C. Hence the
SlimNet part II: Smart chain management flows in those areas are also much higher than
The last heat loss parameter, fluid temperature (Eq. 6), necessary.
can only be changed by modification of the complete
The high return temperatures and corresponding high
chain.
flows are caused by absence of pressurizing valves in
To start at the production side, the current supply the customer installations and defective control valves
temperature is dependent on the ambient temperature, in the hot tapping water installations. By the end of
95 C at Ta=-10 C and 75 C at Ta=15 C, Fig. 6. 2010 SVP starts a campaign to encourage house
Lowering this curve, while still meeting the owners to improve or renew their installations, also for
requirements of customer installations, would reduce their own benefit. This campaign will make use of local
the average network temperature hence the heat approved installers of customer installations. Research
losses. It was calculated through the network model indicated that in certain areas the peak flow can be
that the alternative temperature curve in Fig. 6 solely reduced with 60% [9].
would reduce the heat losses with 4%. Further
research will focus on matching the most effective QUANTIFYING KPI RESULTS FROM SLIMNET
temperature curve with production characteristics.
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
Fig. 6 Existing and alternative temperature curve 5. Implementing cascaded heating installations
6. Encourage house owners to improve or renew
This research will also look upon the possibilities of their installations in accordance with SVP
implementing demand-driven heat production. This is guidelines.
achieved by using a real time network model 7. Eliminating arrears of maintenance an
connected to the substations and production SCADA. implementing a structural preventative
The model uses the weather forecast with customer maintenance program.
information to adjust the temperatures and pressures
just to meet the requirements of customer installations. Heat losses will reduce from 33,6% in 2008 to 22,1% in
It is expected that this will reduce the average fluid 2015. While heat consumption prognoses stays the
temperature even more. same, the corresponding required heat production falls,
Fig. 8. This results in a energy saving of 227.000 GJ
Further research is done to implement cascading that year. In Fig. 9 the results of the sustainability
heating services, i.e. using the latent heat in the return assessment model are shown regarding CO2 savings.
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
58
th th
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-
Fig. 7 Required heat production with SlimNet 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)
Fig. 8 CO2 savings with SlimNet
[7] A. D. A. Kaminski and B. M. K. Jensen,
Replacing the post-insulated steel and PEX pipes Introduction to thermal and fluid engineering,
together with a maintenance program including leak John Wiley& Sons, Hoboken, USA (2005), pp 103
detection will have a positive effect on the water [8] T.A. stergaard, New dimensions for O16,
replenishment. The leak detection actions have already COWI A/S, Aarhus, Denmark (2010)
resulted in a 30.285 m replenishment in 2009, which is
a 7% reduction compared to 2008. [9] A. B. Zitoony and B. E. Roukema, Rapport
inregelstatus onderstations Stadsverwarming
It is expected that al measures will result in a 50% Purmerend 10.001.V2, Roukema B.V., Groningen,
reduction in 2015. Unplanned repairs will also reduce the Netherlands (2010)
50% and consequently is expected to improve
significantly.
59
th th
A DIRECT HEAT EXCHANGER UNIT USED FOR DOMESTIC HOT WATER SUPPLY IN
A SINGLE-FAMILY HOUSE SUPPLIED BY LOW ENERGY DISTRICT HEATING
1 2 3 4
Marek Brand , Jan Eric Thorsen , Svend Svendsen and Christian Holm Christiansen
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 will form areas with lower heat demand than nowadays.
Currently used DH networks will not be able supply
The increasing number of new and renovated buildings
these areas in economical way, because the ratio
with reduced heating requirements will soon make
between network heat losses and heat consumption in
traditional District Heating (DH) systems uneconomic.
buildings would be unacceptable and thus cost of heat
To keep DH competitive in the future, the heat loss in
for end users will increase and DH systems will loose
DH networks needs to be reduced. One option is to
concurrency with other solutions, e.g. heat pumps.
reduce the supply temperature of DH as much as
Recently, research in DH is focused to find the way how
possible. This requires a review of the behaviour of the
to use DH in areas with low energy buildings and how to
whole domestic hot water (DHW) supply system with
increase ratio of heat produced by renewable sources of
focus on the user comfort and overall costs. This paper
energy as solar heat plants or heat pumps driven by
describes some practical approaches to the
electricity from renewable sources.
implementation of this Low Energy District Heating
(LEDH) concept. It reports on the testing of the dynamic One of interesting application of renewable energy in
behaviour of an Instantaneous Heat Exchanger Unit DH is use of decentralised heat sources as e.g. solar
(IHEU) designed for DHW heating and space heating in collectors installed on roofs of individual buildings,
detached family houses supplied by LEDH ensuring an supplying heat to DH network, but it still needs more
entry-to-substation temperature of 51 C. We measured time and work to develop new substations and new
the time it takes for the IHEU to produce DHW with a concept of DH networks to be able to handle these new
temperature of 42 C and 47 C when the tap is features. The solution for future development of DH is to
opened. Measurements were made for control reduce heat losses of DH networks by means of pipes
strategies using internal and external by-pass and no with better insulation properties e.g. twin pipes, use
by-pass. Our results show the importance of keeping better concepts of network design (circular network
the branch pipe warm if comfort requirements are to be configuration, possibility of using circulation line for main
fulfilled, but this involves higher user costs for heating. pipes) and to reduce the supply temperature of district
To increase user comfort without increasing costs, we heating water to lowest level as possible.
propose the whole-year operation of floor heating in
The District Heating Systems designed due to this
bathrooms, partly supplied by by-pass flow.
philosophy are called Low Energy District Heating
Systems (LEDH). The main focus in LEDH system is to
INTRODUCTION reduce heat losses from network as much as possible,
District Heating (DH) is a well known concept of exploit more sources of renewable energy for heat
providing buildings with heat for space heating (SH) and supply and still maintain or improve level of comfort for
Domestic Hot Water (DHW) heating in economical and users, because without high level of comfort this
environmentally friendly way. Nowadays, building concept cant be successful. LEDH concept was
regulations have been introduced worldwide and are reported e.g. in project Development and
pushing to reduce energy consumption in buildings, Demonstration of Low Energy District Heating for Low
because 40% of all energy consumption takes place in Energy Buildings [2], where theoretical case study
buildings. The energy policy of European Union is documented, that LEDH concept is a good solution for
recently focused on energy savings, reducing future and even in sparse housing areas is fully
production of CO2 and increasing the ratio of renewable competitive to heat pumps. This article is focused on
energy [1]. DH is one of the most suitable solutions to application of LEDH for DHW heating. Considerations
achieve these goals for building sector and it gives high related to use of LEDH for space heating will be
priority for further development of DH. But recenlty used reported in future in another article.
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
60
th th
LOW TEMPERATURE DISTRICT HEATING investigation of Legionella in DHW system using IHEU,
CONCEPT producing DHW with temperature below 50 C and
reduced volume of the system below 3L.
Reduced risk of Legionella by use of system with
minimal volume of DHW For single family houses with appropriate close location
Since LEDH is mainly developed for low energy of tapping points, volume of DHW in IHEU and pipes will
buildings already designed with low temperature space be lower than 3 L and thus temperature of 50 C on
heating, the lowest acceptable forward temperature of primary side will not cause Legionella problems. For
LEDH system is defined by requirement for DHW supply multi-storey buildings, district heating substations for
temperature. The hygienic requirement for heating of each flat is a state of the art solution [8]. In this case,
DHW is due to recent standards 50 C for single-family each flat has own completely separated DHW system
houses and 55 C for multi-storey buildings [3] where (with volume of water below 3 L) and thus has
DHW circulation is used. In case of using circulation, increased users comfort and no huge DHW systems
temperature of recirculated water should never fall with circulation, where Legionella is forming and
below 50 C. These requirements are based on need to spreading [9]. The other advantage of using flat station
avoid Legionella growth in DHW pipes and storage in multi-storey buildings is individual metering of each
tanks. It is widely believed, that Legionella grow in flat and complete control over space heating and DHW
temperature range between 46 C 20 C, in systems preparation, which is positively affecting energy savings.
with high volume of water. Mentioned temperature With properly designed DHW building installations,
levels are made in order to assure comfort and hygienic supply temperature of LEDH will be defined by
requirements in furthest tap away from a heat source. It requirements for users comfort. These requirements are
is important to say, that there is high level of discussed in following text.
discrepancy among different results and national
standards focused on Legionella. Users comfort in DHW supplied by LEDH
Another important question, when concerning DHW
Due to German Standard W551 [4], temperature of
systems is level of user comfort. From comfort point of
DHW can be below 50 C and not cause Legionella
view, requirements for temperature and waiting time for
promotion, if total volume of DHW system connected to
DHW can be specified. Due to Danish Standard DS439
one heat source is lower than 3 L. From literature
Code of Practice for domestic water supply
studied, it can be concluded that requirements to
installations, [10] temperature of DHW should be 45 C
produce DHW with temperature higher than 50 C are
in kitchen and 40 C in other taps, provided with
defined for an old fashion DHW building installations,
nominal flowrate and desired temperature reached
which can be characterized as systems with vertical
within reasonable long time, without significant
riser, branched pipes with bigger diameter (increasing
temperature fluctuations. It is a question, if requirement
water volume of the system), using DHW circulation.
of 45 C degrees for kitchen tap is not too high, but
For new and renovated buildings, DHW installations are
argument of problems with fat dissolving from dishes
designed in much better manner, with individual
can be objected and should be investigated. Based on
connection of DHW pipes between each tap and source
mentioned standard, desired temperature of DHW
of DHW and with maximally reduced pipe diameter,
flowing from fixture is 45 C. But in order to define
defined by requirements for noise propagation and
desired forward temperature of LEDH system, we
pressure drop.
should be aware of temperature drop in DH network, in
Due literature, danger of Legionella growth in DHW users substation and in DHW installations in building.
system is influenced by temperature of DHW, nutrients The temperature drop in DH network is not in focus of
in DHW, laminar or turbulent flow in the DHW pipes and this paper, so our goal is to find needed temperature
water stagnation [5]. Several on site measurements level at the entrance of substation to produce 45 C
were performed in buildings using DH for DHW heating. from tap in building. Desired temperature will be found
From results of Martinelli [6] and Mathys [7] can be by experimental measurement of LEDH substation later
concluded, that Instantaneous Heat Exchanger Unit in article.
(IHEU) tend to have much less problems with Legionella
Beside temperature requirements, users comfort is
than traditional units with DHW storage tank. Both
influenced by time needed for DHW to reach a fixture
studies concluded, that these findings are caused by the
after tapping was started. This waiting time is in
fact that in IHEU, DHW is produced with temperature
following text called tap delay. Due to DS439,
60 C, while in case of storage units only with
suggested value for tap delay is 10 sec and it is defined
temperature 50 C. But is necessary to mention, that in
in order to avoid wasting of water and to protect users
case of traditional DHW storage tanks, overall volume of
against too long waiting times for DHW. In large multi-
DHW in a system is much higher than in case of IHEU
storey buildings with centralised preparation of DHW,
system. Due to our knowledge, there is not reported
short tap delay and measures avoiding Legionella
61
th th
growth are assured by circulation line of DHW, but not Tap delay on primary side
properly designed or maintained DHW circulation is A transport delay on primary side consists of delay in
quite often responsible for increased risk of Legionella branch pipe and delay in DH substation. While tap delay
[11]. Another disadvantage of DHW circulation is big in DHW installations in building is for DHW system
heat losses, sometimes even bigger than net energy without circulation uniquely determined, tap delay on
needed for DHW heating [8]. The 10 sec waiting time is primary side varying as control strategies for substation
not rule and for some people it is a long time, for some control varies. From energy consumption point of view,
people short, but this value is used to evaluate tested the best solution is a control strategy without by-pass
concepts if they are fulfilling requirements for high level (see Fig. 1). In this case, DH water staying in the branch
of users comfort or not. An overall tap delay can be pipes is cooled down to temperature of ambient ground
studied from different angles. From dynamic point of (if tapping wasnt performed for long time) and DH water
view, tap delay consists of transportation time needed in substation to room temperature. In general, waiting
for new volume of water travel to tap and dynamic time for DHW is influenced by controller used in
thermal behaviour of passed components, i.e. pipes and substation. Basic principles of controllers are
substation. From point of view related to location, it proportional flow controller and thermostatic controller.
consists of three parts, tap delay in branch pipe (pipe Each controller has own advantages and
from DH pipe in street to users substation), in DH disadvantages, thus best solution is to combine both
substation and in DHW system in building. A tap delay controllers [12]. In case of proportional flow controller,
in branch pipe and substation are related to DH network ratio between primary and secondary flow is fixed to
and substations control system strategy, while tap provide DHW with desired temperature and it means in
delay in DHW pipes in buildings without DHW case of using LEDH primary and secondary flow will be
circulation are defined only by thermal capacity of pipes, very similar. If proportional flow controller is used for
volume of water in individual pipes, nominal flow and to setup without by-pass, user will face long waiting time
some extend also by their insulation. for DHW. Waiting time for this case can be seen from
Table 2. For branch pipe with inner diameter 15 mm (as
Tap delay in DHW system in building is designed in Lystrup for IHEU), even transport delay to
For DHW systems with individual feeding pipes and reach substation for nominal flow for basin, kitchen sink
overall volume of pipes lower than 3 L, DHW circulation and shower will be 31.6, 17.7 and 12.6 sec,
is not needed, because waiting time for DHW with respectively. This solution is from comfort point of view
desired temperature is not critical. In Table 1, transport and water savings completely unacceptable. If we
delays for individual fixtures in typical house built in pilot decrease inner diameter of branch pipe to 10 mm,
LEDH project in Larch Garden - Lystrup, Denmark [11] transport delay is decreased roughly to one half of value
are presented. It should be mentioned, that data are for pipe with inner diameter 15 mm, but it is still long
only transport delay, without dynamic behaviour of time. In case of combined proportional flow controller
cooled pipe. From Table 1 can be seen, that reasonably and thermostatic controller, from beginning of tapping
designed close locations of fixtures, not so far away thermostatic part assures opening of valve on
from substation, lead to maximal transport delay around approximately full capacity until desired temperature of
6 sec, for basin. The total volume of DHW system DHW is reached.
consists of 0.99 L in pipes and 1.1 L in HEX (type
Table 2 Transport delay for nominal flows for individual
XB37H-40). It means, that it is possible to install longer fixtures due to DS439, in branch pipe, 10 m long, for typical
pipes or more fixtures and still fulfil requirement of DHW house in Lystrup, data simulate using proportional flow
system with volume lower than 3 L. The velocity of controller without by-pass
flowing water is below 2 m/s and thus problems with inner
noise propagation during tapping are avoided. nom. volume
pipe velocity transp.
fixture .flow in pipes
d (m/s) delay (s)
Table 1 Transport delay for nominal flows for individual (L/min)
(mm)
(L)
fixtures due to DS439, in DHW system in typical house in basin 3.4 15 1.77 0.3 31.6
Lystrup, for pipes with inner diameter 10 mm kitchen 6 15 1.77 0.6 17.7
shower 8.4 10 0.79 1.8 5.6
length volume
nominal transp. shower 8.4 15 1.77 0.8 12.6
to in velocity
fixture flow
fixture pipes (m/s)
delay bath 12.6 15 1.77 1.2 8.4
(L/min) (s)
(m) (L)
shower 8.4 2.2 0.17 1.8 1.2 Full opening from beginning of tapping leads to much
basin 3.4 4.1 0.32 0.7 5.8 higher flow rate on primary side than on secondary and
kitchen 6 6.3 0.49 1.3 4.9 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
62
th th
primary side defined by DH provider by means of flow case of external by-pass, DH water enters substation,
restrictor or by available differential pressure in DH but not enters heat exchanger and is sent back to DH
network. To reduce tap delay on primary side, control return pipe and thus branch pipe is kept on desired
concepts with by-pass, avoiding cooling of DH water in temperature. Desired temperature is controlled by
branch pipes and substations, and thus reducing thermostatic valve situated in by-pass loop. Increased
substantially waiting time for DHW are available. There level of comfort expressed by reduced tap delay can be
are two concepts of by-pass in relation to the heat adjusted independently on temperature of DHW on
exchanger: external and internal by-pass (see Fig. 1). In 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 Fig. 2 Combined by-pass concept, with possibility of use
heat losses. If substation is installed in room with need by-pass flow in space heating loop
of space heating, heat losses are considered only In order to run by-pass without drawback of insufficient
outside of heating season. cooling of DH water and wasted heat also outside of
Contrary to external by-pass solution, where it is not so heating season, it is proposed to use by-pass flow for
important if space heating loop is installed in series or floor heating, installed in bathroom and operate it all
in parallel to DHW heat exchanger, in case of internal year. From preliminary calculations it looks, that flow
by-pass it is in importance. If space heating loop is needed to keep bathroom floor surface temperature on
connected in parallel to DHW heat exchanger in 24C will be enough as by-pass flow. Considering the
traditional way, by-pass water just pass through DHW use of renewable sources of heat, the problem of
heat exchanger and is sent back to DH network with insufficiently cooled DH water is related to reduced
still high return temperature, without any other use. If efficiency of these sources and whole year using of
space heating loop is connected in series to DHW heat floor heating for comfort in bathroom is reasonable.
exchanger or in parallel but with possibility to sent by-
pass water flown through internal by-pass to space Supply supply recirculation
heating loop (see Fig. 2), this solution provides high As an alternative solution for customers who dont
level of comfort for users as well as proper use of heat want to use whole year bathroom floor heating,
needed for by-pass operation. 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
63
th th
heating water is supplied by pipe 1 to substation, lower heat loss. On site measurements were started in
circulated through HEX or external by-pass (see Lystrup to evaluate performance of both types of DH
Fig. 3) and then sent back to district heating network substations, but no detailed measurements requiring
(DHN) supply by pipe no.3. This concept is in early short time steps are performed to evaluate level of
stage of investigation but it looks promising. The main users comfort. The measurements more focused to
question will be related to flow of DH water in branch users comfort are planed to be performed this year in
pipe in order not cool it down too much before will be Danish Technological Institute and Technical
sent back to DH supply pipe in the street. 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

Fig. 3 Supply supply recirculation with external by-pass
As a first part of measurements planed to be
This solution is expected to be favourable mainly for performed at DTU, the time needed for IHEU to
circular shapes of DH networks, but it should be produce DHW with temperature of 42 C and 47 C
mentioned, that re-heating stations will be probably was measured, after tapping of DHW was started. The
needed in point of DH network, where temperature of tap delay was investigated for two control strategies,
DH water decrease bellow defined value. one using internal and second using external by-pass.
The measurements were performed for different initial
Full scale demonstration of LEDH
conditions before tapping was started to simulate in
Full scale demonstration of LEDH is recently running in realistic way users behaviour. Finally, the period
Larch Garden in Lystrup, Denmark [11], where 40 low between two by-pass flow operations was measured.
energy houses class 1 and 2 are connected to LEDH
system, with designed forward temperature from heat Experimental setup and instruments
plant 52 C. For primary side of substation, forward Tested DH substation was prototype of Instantaneous
temperature of 50 C and return temperature of 25 C Heat Exchanger Unit (IHEU) developed specially for
are designed. The DH network is built from highly LEDH pilot project in Larch Garden Lystrup,
insulated single pipes (for main pipes) and main pipes Denmark. The IHEU is a type of district heating
with smaller diameter, distribution and branch pipes substation consists of a heat exchanger (HEX) without
are built from twin pipes. Two types of district heating storage tank. DHW is heated instantaneously in HEX
substations providing houses with DHW and space only when tapping is performed and then supplied
heating are tested by customers in real conditions. The directly to DHW taps by individual feeding pipes, while
first concept is 29 Instantaneous Heat Exchanger Units space heating is using direct connection without heat
(IHEU), second is 11 District Heating Water Units exchanger, i.e. concept typical for Denmark.
(DHWU). IHEU is classical concept of substation with Substation is same concept as regular IHEU for
instantaneous heat exchanger, only with enlarged traditional DH. The difference is in increased number
number of plates. IHEU units have external by-pas, of plates in heat exchanger assuring better heat
with set point temperature of 35 C for customers transfer. Water volume of primary and secondary side
situated not at the end of street and 40 C for is 1.1 L each and the heat exchanger is not insulated.
customers situated at the end of the street. DHWU is The experiments were focused only on dynamic
new concept of DH substation, reported e.g. by behaviour of substation related to DHW heating and
Paulsen [13]. DHWU consist of buffer tank for district thus space heating loop wasnt connected and space
heating water and when DHW is needed, DHW is heating valves in substation closed. Desired
heated in instantaneous heat exchanger as in previous temperatures of DHW were chosen in accordance with
case. Advantage of concept with buffer tank is peak- requirements in DS439 for temperature of DHW for
shaved demand of DH water during charging and use kitchen sink and other fixtures. Required temperatures
of branch pipes with lower diameter, connected with mentioned in DS 439 are 45 C and 40 C. In order to
64
th th
cover additional temperature drop in building DHW temperature of 51 C started to flow in the substation
installations, 2 C were added. This addition is based and flew through external by-pass, until closing
on experience from previous measurements. During temperature was reached and by-pass flow stopped.
the experiments, temperatures of four different flows Then we wait until by-pass was opened again. Time
passing through the DH substation were measured. between two by-pass openings as well as volume and
On primary side it was temperature of DH water temperature of DH water passed through by-pass was
supplied to substation (T11) and temperature of DH written down and after by-pass was closed again, we
water returning back to DH network (T12) and on waited a little bit shorter time than was needed to open
secondary side it was temperature of cold potable by-pass flow again and we start tapping on secondary
water entering substation (T21) and temperature of side with flow rate 8.4 L/min. In this way, most
heated DHW (T22). All temperatures were measured unfavourable condition for substation with by-pass, i.e.
by thermocouples type T, installed directly in pipes, in highest recovery time, was measured. After tapping of
flowing water, so they do not have any practical time DHW was finished, we wait 5 minutes and we
delay for the measurements. The time constant to performed one more tapping to simulate short time
reach 90% of step change was less than 1 second. step between two subsequent tapping of DHW.
The distance of thermocouples from substation flanges
2. For measurement of internal bypass concept, IHPT
was 5 cm and thermocouples were previously
controller was used. In case of IHPT, by-pass set point
calibrated. We also measured surface temperature of
temperature cant be adjusted independently and is
HEX in upper (HEX-UP) and bottom part (HEX-
defined by desired temperature of DHW, i.e. 47 C for
DOWN) and temperature of air in the testing room.
our measurements. IHPT controller was developed for
Temperatures were measured and collected by
traditional DH networks operating with forward
multifunction acquisition unit every second. For
temperatures around 70 C. For traditional DH, by-
authentic simulation of DH network, DH water with
pass opens when temperature in HEX falls 57 C
constant temperature of 51 C was necessary. It was
below set point of DHW, but in case of LEDH with
solved by connecting of IHEU to source of DHW in
forward temperature 51 C, by-pass opens 1 C below
laboratory of DTU, where DHW is supplied by DH
DHW set point temperature, i.e. 46 C in our case.
system. DHW system of DTU is big enough, to assure
stable temperature 51 C without any fluctuations. In The testing procedure was similar to measurements
order to prevent cooling down of pipes supplying DHW with external by-pass. After supply valve on primary
to laboratory in periods when there was not flow side of substation was opened, DH water with
through substation (stopped by by-pass controller), temperature of 51 C started to flow in the substation
small guard flow, just before entrance to substations and temper HEX, until by-pass closing temperature
was kept to maintain DHW always on 51 C and was reached. Then we wait until by-pass was opened
drained to sink. again and we performed tapping of DHW just before
next by-pass opening was expected. In following steps
Experimental procedure was procedure same as in case of external by-pass.
As a first step, both controllers were adjusted to Moreover, we also performed measurements of time
provide 47 C on DHW side with supply temperature of delay in IHEU for control concept without by-pass.
DH water 51 C. Then we measured time delay in the
substation, i.e. time needed for substation to produce
RESULTS
DHW with temperature 42 C and 47 C on secondary
side outlet from the moment when DHW tap is opened. Time delay for IHEU with PTC2+P controller and
external by-pass adjusted to 35 C to start supply
The measurements were performed for different initial
DHW water with temperature 42 C and 47 C after
conditions and secondary flowrate was always
long idling period just before opening of external by-
8.4 L/min, which is nominal flow for shower.
pass was expected, can be seen from Fig. 4 and is 11
1. For measurements of concept with external by-pass, and 22 seconds, respectively. This measurement
substation was controlled by PTC2+P controller with represents condition with the longest time delay for
by-pass set point temperature adjusted to 35 C. This PTC2+P controller. Temperature of room, where IHEU
setup is exactly the same as is installed in Lystrup pilot was installed was 22.2 C. For this case, temperatures
project. The testing procedure was made in following of produced DHW in first 10 sec after tapping was
steps. Substation was left idle for long time in the started are listed in Table 3.
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
65
th th
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 In this case, tap delay in substation to produce DHW
long idling just after by-pass flow was stopped, time with temperature 42 C and 47 C was shorter, 7 and
delay decreased to 8,5 and 16,5 seconds. In this 14 seconds.
measurement, temperature of substation and thus
For room temperature around 22 C, external by-pass
water standing in the HEX was little higher than
was opened roughly every 30 minutes. The by-pass
ambient air temperature. It is expected that time delay
was in average opened 2.5 minute and volume of DH
will be slightly longer, if substation will have real
water needed to close the by-pass was in average 3 L,
ambient temperature but still shorter than in case 2.
i.e. when substation is idle, by-pass uses 6 L of DH
We also performed measurement of tap delay five
water per hour.
minutes after previous DHW tapping was finished.
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
(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 The average flow of internal by-pass was 24 L/hour
internal by-pass adjusted by requirement of DHW to and average return temperature to DH network was
47 C was 6 and 14 seconds to reach 42 C and 47 C 45 C. When internal by-pass is once opened, the time
on outlet for situation when tapping was performed just delay in substation decrease substantially to 1.5 and
before by-pass was expected to open. The internal by- 7 seconds to produce DHW with temperature 42 C
pass opens 3 minutes after previous tapping is finished and 47 C. The condition with expected longest time
and when is once opened never closes, only when delay was solution without by pass. In this case time
another tapping is performed, but again only on delay to produce DHW with temperature 42 C and
3 minutes. 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
T11 42 45 47 T12 T12AVG THEX-UP THEX-DOWN

case number and description
(C) (sec) (sec) (sec) (C) (C) (C) (C)
NO BY
50.1 12 18 25 16.2 19.5 20.4 21
PASS 1 after long idling, no by-pass (BYP)
2 after long idling, just before BYP was
49.6 11 16 22 30.1 19.3 21.5 21.4
EXTERNAL expected to open again
BY-PASS 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
5 just before BYP was expected to open (3
50.5 6 10 14 19.5 19.1 22.6 38
INTERNAL min after prev. tapp. finished))
BY-PASS 6 anytime, when BYP was already in
49.3 1.5 3.5 7 47.3 18.4 44 45.5
operation
66
th th
DISCUSSION only difference is that in case 5 (internal by-pass), tap

delay is again reduced because tapping was
Focused on level of users comfort and proper cooling
performed 3 minutes after previous (to prevent
of DH water during idling, time delay of LEDH
influence of by-pass) and thus HEX was warmer.
substation to supply DHW with temperature 42 C and
47 C was measured. Three different control strategies If the requirement is to fulfil 10 sec tap delay for less
related to tap delay were investigated. Obtained favourable fixture, i.e. in our case basin (see Table 1),
results represent case of IHEU used in single-family DHW should leave DH substation with temperature
house in period when space heating is not in 42 C in 4 sec after tapping was started, because it will
operation. Explored concepts can be evaluated from take 6 second to reach the tap. This requirement was
two different points of view, due to highest advantages reached only by concept with internal by-pass and only
for customer and for DHN. when by-pass was already opened. On the other hand
from Table 3can be seen, that even for concept with
The solution without by-pass is from energy savings
external by-pass and tapping after long idling and just
point of view very interesting because doesnt need
before expected bypass opening, DHW at a
any DH water for idling, but from users comfort point of
temperature 26 C leaving substation in 3 sec. DHW
view is very poor because of reduced comfort and
with this temperature is not sufficient for taking a
problems with wasting of water during waiting for DHW
comfortable shower for which temperature 371 C is
with desired temperature. Solution without by-pass can
preferred, but for washing hands this temperature
be probably used for substations equipped with
should be enough. The values in Table 3 are for flow
combined thermostatic and proportional flow controller,
rate used for shower, but it can be used to explain that
for customers with short branch pipes or for customers
it is time to rethink the suggested value of tap delay
with low requirements for level of users comfort. If
from 10 sec to another value and consider also
solution without by-pass will be used for substation
nominal flows and use of tapped water. The different
controlled only with proportional flow controller, even
standards for the different use of DHW based on new
transport delay in 10 m long branch pipe for nominal
solutions in DHW supply systems and results from test
flow for basin will be 32 sec. For period when space
panels are needed, because it may have some
heating is operated, branch pipe will be kept warm
influence on design of optimized DHW systems.
from flow needed for space heating and time delay for
Nevertheless, for customers requiring DHW in very
solution without by-pass will be very similar to solution
short time e.g. continuously or discontinuously (only
with external by-pass. Anyway, in non-circularly
during rush hours) operated trace heating elements
shaped DH networks, by-pass should be installed at
can assure almost no tap delay by keeping DHW
least at the end of a street, so it is better to find
staying in pipes on desired temperature.
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 CONCLUSION
operation of floor heating in bathrooms to increase Based on literature study it can be concluded that
comfort for customers and at the same time solve hygienic requirement of DHW with 50 C on outlet of
problem with by-pass flow which otherwise increasing DHW heater is not needed for systems with a total
return temperature to DH network. volume of the DHW lower than 3 L.
From user comfort point of view, better solution than From results of our measurements and evaluation of
solution without by-pass, but consuming more energy, IHEU supplied by LEDH, only substation with external
is substation equipped with external by-pass. By by-pass with set point 46 C is able to produce 47 C
comparison of results of concepts without by-pass DHW in time bellow 10 sec. The easiest step how to
(case 1) and solution with external by-pass, for case decrease waiting time also for other concepts is to
when tapping is performed after long period of idling insulate HEX. This measure will reduce time delay for
just before by-pass opens again (case 2), we can see DHW tapping and also will decrease heat losses from
that time delays are almost the same (see Table 4). DH substation. The lower waiting times for DHW can
Difference is only that for external by-pass are pipes in be also achieved by further optimisation of HEX in way
DH substation kept on higher temperature and it made of decreased number of plates reducing volume of
slightly faster reaction. In the case 3, time delay is water in HEX and thus transport delay, and by
even more reduced since pipes in substation were increased thermal efficiency of HEX (followed on the
warmer by just finished by-pass flow. For control other hand by higher pressure loss). These
concept with external by-pass and tapping repeated modifications can lead for higher temperature of DH
5 minutes after previous one, time delay is again water returning to DH network, but during all our
reduced, since HEX is still hot from previous tapping. experiments, average return temperature was below
The time delay for case 4 and 5 are almost the same, 20 C, what is 5 C less than is designed for LEDH.
67
th th
Traditional control concepts of DH substations are [4] DVGW, W551 - Trinkwassererwrmungs- und
always trade-off between users comfort and reduced Trinkwasserleitungsanlagen ,1993, Bonn, (in
cooling of DH water during idling and thus customer German)
should have to some extent possibility to choose which
[5] Z. Liu, Effect of flow regimes on the presence of
solution prefers. In case of traditional concepts,
Legionella within the biofilm of a model plumbing
decision is between longer waiting time for DHW and
system, 2006, Journal of Applied Microbiology,
energy savings or vice versa, if by-pass in substation
Vol. 101, pp 437-442
is used. In non-circularly shaped networks, by-pass
should be used anyway at least at the end of a street [6] F. Martinelli, A Comparison of Legionella
line. The one of possible solutions how use by-pass pneumophila Occurrence in Hot Water Tanks and
flow in better way can be proposed innovative concept Instantaneous Devices in Domestic, Nosocomial,
of whole year operated floor heating in bathrooms or and Community Environments, 2000, Current
supply-supply recirculation. Both solutions will increase Microbiology, Vol. 41, pp. 374-376
level of user comfort and at the same time also energy
[7] W. Mathys, J. Stanke, et. al., Occurrence of
efficiency of DH system.
Legionella in hot water systems of single-family
LEDH is a promising solution for providing buildings residences in suburbs of two German cities with
with DHW and space heating regarding fulfilling special reference to solar and district heating,
requirements of modern society with reduced CO2 2008, Int. J. Hyg. Environ. Health, Vol. 211, pp.
emissions and energy consumption. More detailed 179-185
investigations by testing of different parameters and
[8] H. Kristjansson, Distribution Systems in
numerical simulations are needed in order to optimize
Apartment Buildings, Published at the 11th
LEDH concept.
International Symposium on District Heating and
Future work Cooling, August 31 to September 2, 2008,
It will be very interesting to compare time delay of Reykjavik, ICELAND
substation for traditional DH with time delay for DHW [9] T. Persson, District Heating for Residential Areas
produced by LEDH substation. It is expected that with Single-Family Housing, paper IV, 2005,
timed delay for LEDH will be higher because dynamic Doctoral Thesis, Lund Institute of Technology,
response is slowed down by lower temperature Lund
difference between DH water and desired temperature
of DHW, but on the other hand, lower temperature [10] Dansk Standard, DS 439 Code of Practice for
difference is in some extend compensated by bigger domestic water supply installations, 2009
HEX. It is also suggested to rethink 10 sec tap delay [11] P.K. Olsen, Low-Temperature District Heating
suggestion for different taping flows and purposes of System for Low-Energy Buildings, 2009,
DHW use. http://www.fbbb.dk/Files/Filer/Peter_Kaarup_Olsen
_-_COWI_29-10_2009.pdf
REFERENCES [12] H., Boysen, J.E. Thorsen, Control Concepts for
[1] S. Froning, Low energy communities with district DH Compact Stations, Published in Euroheat and
heating and cooling, PLEA 2008 25th Power IIII 2004
Conference on Passive and Low Energy [13] O. Paulsen, Consumer Unit for Low Energy
Architecture, Dublin, District Heating Net, Published at the 11th
[2] Hovedrapport, Udvikling og Demonstration af International Symposium on District Heating and
Lavenergifjernvarme til Lavenergibyggeri 2009, Cooling, August 31 to September 2, 2008,
(in Danish) Reykjavik, ICELAND
[3] EUROHEAT & POWER, Guidelines for District

Heating Substations, 2008, downloaded from
www.euroheat.org in October 2009, pp 8
68
th th
CHALLENGES ON LOW HEAT DENSITY DISTRICT HEATING NETWORK DESIGN

1 1
M. Rm and K. Sipil
1
VTT Technical Research Centre of Finland PB 1000, FI-02044 VTT, Finland
ABSTRACT heating. The expansion of mature and large scale

systems take place in areas with lower heat
While district heating is an energy efficient solution to
consumption. This transition to more demanding
provide heating to areas with high heat consumption,
operational environment both technically and financially
mature systems extending out to more demanding
represents challenges to district heating network design.
operational environment face challenges maintaining
This is also true in small scale systems of limited
competitiveness over alternative heating systems. As
consumption separated from a larger system.
the heat density falls below a certain level, district
A careless network design in these circumstances can
heating is no longer economically feasible. Studying the
lead to deterioration of the advantages of district
possibilities of extending this threshold by means of
heating; efficiency and reliability. An annual heat loss of
district heating system design and pointing out the
5% in district heating distribution is considered a good
operational challenges while approaching it are the
result, but the case in question the heat losses can
main topic of this paper.
easily reach 10% or even tens of percents if the
The problem is investigated in a representative case of characteristics of low heat density areas are not taken
a low heat density area bordering a more extensive into account in design.
district heating network. A node-and-branch type
network simulation model is used study the operation of LOW HEAT DENSITY AREA
the network and a simulation period of one year is used
to get a realistic view of the system in a normal A detached house area consisting of 56 identical 150
operational cycle. m2 houses with energy consumptions in compliance of
todays building standards is studied. Dedicated heat
Not taking into account the characteristics of a low heat exchangers between the network and the consumer
density area in network design can result in inefficient exist for both heating and domestic hot water. Total
distribution system. Operational problems, especially energy consumption for the houses is 18.75 MWh/year
maintaining the temperature level in summertime, must of which domestic hot water has a share of 20 percent.
be solved. Only concentrating on minimizing the heat
losses will not result in best possible design. The district heating network studied is presented in
Figure 1. The detached house connections are marked
The temperature level issue can be solved with a by- as green dots and the connection to the main district
pass valve, auxiliary heating or accumulators, but in heating network as a red rectangle. The connections
overall more efficient system requires steps to be taken have 1, 2 or 6 detached houses as consumers,
in the houses. Floor heating and a heat pump coupled indicated by the size of the dot.
with an accumulator enables the use of low temperature
design where the heat losses can be cut significantly.
INTRODUCTION
50 m
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 Figure 1. District heating network studied.
heating systems. From the consumer point of view,
The total trench length in the area is 2 390 m of which
district heating is considered as a reliable and carefree
the service pipes (DN 15-25) account for 1 300 m. The
source of heating energy and is also often an
pipe size distribution is illustrated in Figure 2. The dark
economically sound choice.
blue coloured bar (DN 65) represents the pipe
Areas with high heat consumption i.e. economically the connecting the area to the main district heating network.
most attractive areas will be connected first to district
69
th th
As the pipe diameters are quite small, twin pipes with The consumptions for both heating and domestic hot
insulation class IV are used in the area as water use were given as hourly time series as well as
recommended by Energy Industry [1], [2] in Finland. the radiator supply and return temperatures on the
The pressure drop design principle used here is roughly secondary side.
~1.5 bar/km.
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
700
assumed to be constant. When heat demand, both
600
supply and return temperatures on secondary side and
Pipe lenght (m)
500
supply temperature on primary side are given as input,
400
the primary return temperature and district heating mass
300 flow can be calculated.
200
100
Table 1. Design point for heat exchangers.
0
15 20 25 32 40 50 65 Description Value
Pipe size (DN)
Primary side temperatures 115/45 C
Figure 2. Pipe size distribution
Radiator heating 70/40 C
The linear heat density is 0.44 MWh/m which makes the Domestic hot water 55/10 C
area a low heat density area by definition [3]. Design heating load 8 830 W
The heat demand around the year is presented in Design DHW load 2 060 W
Figure 3. The peak demand for the area is 507 kW. As
expected, in the summertime the load consists almost The design loads for domestic hot water are low
solely of domestic hot water consumption. compared to a real life design load of a heat exchanger
in normal detached house in Finland, 50 kW is a
600
common choice. This is due to the simulation model
Total heat demand (kW)
500 taking hourly data originally calculated for a multifamily

400 house as input so the domestic hot water demand is
also flatter than it really is. However, from the network
300
design point of view hourly data is considered accurate
200 enough.
100 Other input data used were the undisturbed ground
0 temperature of 5 C, assumed to be constant, and the
0 50 100 150 200 250 300 350 supply temperature from the main district heating
Days network as a function of outdoor temperature. The
Figure 3. Heat demand of the simulated area. outdoor temperature time series used described a
typical year in Southern Finland. The supply
temperature reaches its maximum value of 115 C in an
SIMULATION MODEL
outdoor temperature of -26 C and its lowest value of
A node-and-branch type simulation model [4] was used 75 C in 5 C. Between these two points, the relation is
to study the case in hand. The model calculates linear.
temperatures and pressures for the nodes and flows
and heat losses for the pipes, i.e. the branches. From SIMULATION RESULTS
these results pumping power can also be calculated,
although a constant efficiency of 0.5 is used for the The most interesting results concern the heat losses
pump. The pressures are calculated separately from and the temperature variations within the network. The
temperatures. The temperature calculation is dynamic pumping needed (less than 1 MWh) in a network of this
while the flow and pressure calculation is not. A size is quite low and thus negligible.
minimum 0.6 bar pressure difference over a consumer In the initial simulation runs it was noted that the system
is assumed. was struggling to maintain high enough temperature
When defining the network, each pipe is given a start level in the summertime when the load consist solely of
and an end node, a pipe type (twin, single), an domestic hot water demand. This problem was met by
insulation standard (class I to IV) and length. defining a flow through valve at the consumer, opening
70
th th
when the supply temperature on primary side dropped relative heat losses in the heating season are
too low (< 65 C). The valve allowed a constant mass acceptable, they reached 47 % in the summertime. The
flow (0.015 kg/s) to go past the heat exchanger on the high heat losses are partly because of the by-pass valve
primary side. This solution helped the situation letting hot water past the heat exchangers. The by-pass
significantly although not without ill effects as can be valve is also responsible for small cooling, i.e. the
seen from the heat losses presented below. difference between supply and return temperatures,
within the system in summertime (Figure 6).
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 1.0
lower cooling; all of which are undesirable outcomes. 0.9
Relative heat losses (-)

One possibility to solve the problem is just to accept the
0.8
0.7
flaw and to use additional electrical heating element to 0.6
raise the temperature of domestic hot water to the 0.5
required level. As the temperature boost needed is for 0.4
most of the time quite small and is only needed in 0.3

0.2
summertime, the increase in electricity consumption is 0.1
reasonable. 0.0
I II III IV V VI VII VIII IX X XI XII
Because of the high capital costs of district heating, the Month
pipes should basically be sized as tight as possible Figure 5. Monthly relative heat losses.
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 90
with a simplified example by showing the temperature

80
70
on supply side service pipes if there is no flow for three
Cooling (C)
60
different pipe sizes. The temperature drop of 15 C, for
50
example, takes 5 times longer with a pipe size DN 50 40
than with a small DN 15 pipe. The calculations assume 30
a constant return side temperature of 30 C and a 20
ground temperature of 5 C. 10
0
DN 15 DN 25 DN 50 0 50 100 150 200 250 300 350
70 Days
60
Figure 6. Difference between supply and return
Temperature (C)
50
temperatures at the border of the area.
40
30
20 The most obvious way to cut heat losses in already

10 reasonable insulated network is to lower the supply
0
temperature. In the simulated system, this would cause
0 2 4 6 8 problems because aforementioned issues concerning
Time (h) domestic hot water demand in summertime, and during
the heating season because of the traditional radiator
Figure 4. Temperature drop in three pipe sizes when no heating design temperatures of 70/40 C. However, if
flow is introduced.
more significant changes would be possible, a floor
heating system and a heat pump coupled with an
The use of smaller pipes reduces the heat losses in
accumulator handling the higher temperature level
W/m and this is accentuated if the temperature level
required domestic hot water would enhance the
drops as described above. As a result, looking solely on
efficiency of the distribution system at a price of a very
heat losses when designing a low heat density area
modest increase in electricity consumption and higher
network on common design principles can lead to
investment costs for the consumer because of the
reliability issues as the system cannot supply the heat
accumulator, heat pump and floor heating. If the
required by the consumers.
domestic hot water demand takes 3.75 MWh/year,
The relative heat losses (that is, heat losses per needed 20 percent of the total consumption of 18.75 MWh/year,
production) for the simulated case are 13.8 % in a year. the electricity consumption would be a very reasonable
The monthly values can be seen in Figure 5. While the
71
th th
1.25 MWh with an average COP of 3. With this setup, REFERENCES

supply temperature would need to be just 40 C.
[1] Lappeenranta University of Technology,
Kaukolmpjohtojen optimaalisen eristyspaksuu-
CONCLUSIONS
den tarkastelu / Investigation of the optimal
The use of traditional district heating network design insulation thickness on district heating pipes,
principles can lead to an inefficient area heating system Energy Industry, 2009, 36 p.
in areas with low heat density. Special attention must be
[2] Preinsulated district heating pipes,
paid on operation of the system to ensure reliability, one
Recommendation L1/2010, Energy Industry, 2010,
of the advantages of district heating.
44 p.
When aiming for an efficient system, one goal is to
[3] Zinko, H., Bhm, B., Kristjansson, H., Ottoson, U.,
minimize the heat losses. However, concentrating solely
Rm, M., Sipil, K., District heating distribution in
on this can make another problem, maintaining high
areas with low heat demand density, IEA DHC
enough temperature level for domestic hot water in
Annex VIII, 2008, 117 p.
summertime, even worse. The problem can be solved
using a by-pass valve, but this causes unwanted [4] Ikheimo, J., Sderman, J., Petterson, F., Ahtila, P.,
effects; worse cooling and an increase in heat losses Keppo, I., Nuorkivi, A., Sipil, K. 2005. DO2DES
and pumping power. Other solutions are auxiliary Design of Optimal Distributed Energy Systems,
heating (electrical heating or a heat pump) or the use of Design of district heating network. bo Akademi.
an accumulator and with it, aiming for a steady domestic Report 2005-1.
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.
72
th th
DESIGN OF LOW TEMPERATURE DISTRICT HEATING NETWORK WITH SUPPLY

WATER RECIRCULATION
1 1 1
Hongwei Li , Alessandro Dalla Rosa , Svend Svendsen
1
Civil Engineering Department, Technical University of Denmark
ABSTRACT technologies become barriers to further increase the

market share [2]. In order to sustain the economic
The focus on continuing improving building energy
competiveness and realize the long term sustainable
efficiency and reducing building energy consumption
development, the concept of design and operation of
brings the key impetus for the development of the new
DH system needs to be re-examined under the new
generation district heating (DH) system. In the new
energy regulation and development trends. This is the
generation DH network, the supply and return
main impetus for the development of the new
temperature are designed low in order to significantly
generation DH system. Based on previous studies, in a
reduce the network heat loss. Meanwhile, the low
properly designed in-house substation system, the
network operational temperature can make a better
network supply temperature at 55oC and return
utilization of renewable energy and further improve the
temperature at 20oC can meet the consumer space
CHP plant efficiency.
heating and domestic hot water demand [3].
Though the designed return temperature is low, it may
The low return temperature has the advantages to
increase considerably when the heating load becomes
reduce the network heat loss, increase CHP plant
low and the by-pass system starts to function. The aim
power generation capability, and utilize direct flue gas
of this paper is to investigate the influence of by-pass
condensation for waste heat recovery. However, the
water on the network return temperature and introduce
return temperature can become much higher than the
the concept of supply water recirculation into the
designed value when the heating load becomes low
network design so that the traditional by-pass system
and the by-pass system at the critical user starts to
can be avoided. Instead of mixing the by-pass water
function. In this paper, the influence of by-pass water
with return water, the by-pass water is directed to a
on network return temperature was examined for a
separated circulation line and returns back to the plant
reference residential area. The concept of supply water
directly. Different pipe design concepts were tested and
recirculation was introduced to avoid the mixing of by-
the annual thermal performances for a selected
pass water and the return water. Three network design
residential area were evaluated with the commercial
methods were tested. The annual thermal performance
program TERMIS. The simulation program calculates
was evaluated with the commercial district heating
the heat loss in the twin pipe as that in the single pipe.
network hydraulic and thermal simulation software
The influence of this simplification on the supply/return
TERMIS [4]. The simulation program calculates the
water temperature prediction was analyzed by solving
heat loss in the twin pipe as that in the single pipe. The
the coupled differential energy equations.
influence of this simplification on the supply/return
water temperature prediction was analyzed by solving
INTRODUCTION the coupled differential energy equations.
In European Union, one of the major energy
development targets is to reduce the building energy SUPPLY WATER RECIRCILUATION
consumption and increase the supply of renewable
The solution to overcome the excessive temperature
energy. The introduction of European Energy
drop along the supply pipe due to reduced flow rate is
Performance of Building Directive (EPBD) poses
to install by-pass system at the critical user in the
stringent requirement for the member countries to
network. Figure 1 shows the principle of supply water
effectively reduce their building energy consumption.
by-pass. Extra flow is called based on the temperature
According to the national energy policy, the building
measurement at the critical user until the minimum
energy consumption in Denmark will drop to 25% of
supply temperature requirement is met. This extra flow
current level by the year 2060, while the renewable
is then by-passed and sends back to the return pipe.
energy share will increase from 20% to 100% at the
As the by-pass flow rate may be considerable and its
meantime [1].
temperature is high, the mixing with return water will
District heating (DH) benefits from economic of scale significantly increase the return water temperature
with mass production of heat from central heating which causes both increased heat loss in the return
plants. The significant reduction of building energy pipeline and decreased power generation capability in
consumption and wide exploitation of waste heat and the CHP plant.
renewable energy, however, makes the current DH
73
th th
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. 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
Fig. 1 Schematic for hot water by-pass system
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
Fig. 2 Schematic for by-pass water recirculation avoid the legionella problem, the design temperature
for DHWS is higher than HE, at 65 C /30 C for supply
and return respectively.
74
th th
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
Fig. 4 In-house heat exchanger (HE) in DH system user along the network critical route which is shown in
green color.
Three network design scenarios were investigated for
Table 1 Simultaneous Factors 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,
Fig. 5 Domestic hot water storage (DHWS) in DH system while at the virtual point adjacent to the end user in
case 2.
75
th th
Network Heat Loss Calculation coefficients corresponding to the temperature

The reference network was designed with twin pipes by difference between the flow and the ground.
placing the supply and return pipe in the same casing. The temperature variation along the pipeline was
Two types of twin pipes were considered in the calculated as internal flow with isothermal boundary
simulation: AluFlex multilayer flexible pipe and straight condition. The downstream temperature in the pipe is
steel pipe. The pipes were selected with continuous expressed as [4]:
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 [3]
for the 3rd recirculation line. This 3rd pipeline can be
assumed being placed in the same trench along the Td, Tu and Ta represent the downstream fluid
twin pipes. The thermal interaction between the twin temperature, upstream fluid temperature, and ambient
and the single pipe is assumed negligible. temperature respectively. M and K are parameters
The heat loss in the twin pipe was calculated according include the overall heat transfer coefficient. As the
to the reference [7,9] overall heat transfer coefficients have to be calculated
beforehand, the influence of flow temperature variation
on Us and Ur along the pipeline is neglected. It is a
reasonable assumption when the thermal by-pass
temperature is set close to the plant temperature,
[1]
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 oC) is higher than ground temperature,
[2]
the net heat transfer in the return pipe may absorb heat
from surrounding which makes Ur negative. However,
negative Ur 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
Fig. 6 District heating network smaller pipe was suggested by the program, the Alx16
single pipe was selected as the minimum diameter pipe
The supply and return pipe are assumed identical and available on the market. It was assumed that the
placed horizontally in the same depth from the ground. recirculation pipe can be used as water supply in winter
The linear thermal transmittance Uij reduces to in case 3. Therefore, the main pipes in the reference
U11=U22=U1 and U12=U21=U2. In addition, the thermal line were re-designed with considering that a portion of
conductivity of insulation foam was assumed constant. supply water goes through the recirculation line. It can
U1 and U2 were then calculated with the analytical be seen that the supply pipe has smaller diameter than
solution developed from the multi-pole method [10]. return pipe in some sections in the twin pipe line.
The simulation program cannot handle two heat
transfer coefficients in the same pipe, Us and Ur were
derived to represent the overall heat transfer
76
th th
Table 2 Selected pipe types and length in Case 13 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
Figure 7 shows the pressure profile along the critical
feasible, according to the simulation results. However,
route. The network is designed for a 10 bar system.
it may be used as an alternative solution to supply
The minimum network static pressure is 2 bar and the
water in the 3rd pipeline under extreme whether
minimum differential pressure at consumer is 50 kPa.
condition, which otherwise has to raise the plant supply
The plant static supply pressure is 853 kPa in case 1 at
temperature to meet the increased heating demand.
design condition. In case 3, the designed plant supply
Furthermore, results in table 4 were limited to fixed
pressure head rise to 917 kPa, which is due to the
recirculation pipe diameters. The double pipe supply
increased flow rate indicated in Table 4. The pressure
concept may be economical feasible by free selection
drop along the reference line during summer is quite
both reference pipe and recirculation pipe diameter
low due to the reduced flow rate. However, extra
with the objective to minimize the annual network
pressure head has to be applied to overcome the
operational cost or exergy consumption. This study is
pressure loss along the recirculation line in Case 2.
out of the scope of current paper due to the limitation of
The required static supply pressure is 800 kPa during
the simulation program.
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. Fig. 7 Pressure profile on the critical route in Case 13
77
th th
Table 3 Simulation results in Case 1 Table 6 Simulation results in Case 1
Table 4 Simulation results in Case 2 (First 5 rows) and Table 7 Simulation results in Case 2 (First 5 rows) and
Case 3 (Last 3 rows with italic) Case 3 (Last 3 rows with italic)
Domestic Hot Water Storage Tank

Table 5 shows the pipe types and corresponding length Further Discussion on Heat Transfer
in the DHWS installation. Alx 14 was selected as
As shown in Eq. 13, the simulation program simplifies
branch pipe due to the smaller design heating load.
the calculation of the heat loss in the twin pipe as that
Similar to the HE, the by-pass flow rate exceed the
in the single pipe. The influence of the adjacent pipe
actual flow rate through the consumer in summer
was accounted through converting the linear thermal
season. The plant mixed return water temperature in
transmittance Uij to the overall heat transfer coefficients
case 1 is 46 oC. The introduction of the recirculation
Us and Ur , with pre-assumed constant network
line can keep the plant return temperature in reference
supply/return temperatures. To assess the influence of
line as low as 30 oC, while increases the return
this simplification on the temperature predication, the
temperature in the recirculation pipe to 54 oC at the
thermal interaction between the supply and return
plant. Extra heat loss has to be tolerated due to the
pipes was calculated by solving the coupled pipe heat
recirculation pipe in both case 2 and case 3.
transfer differential equations. The governing equations
Table 5 Selected pipe types and length in Case 13 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]
78
th th
The governing equations then change to: Table 8 Pipe temperature predication comparison (supply
o
outlet temperature is controlled at 50 C)
[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 by-pass water temperature in this study was set in

a conservative way. In many practices, the by-pass
The final solutions are given as: water can be set 10 C lower than the supply water
temperature. Even lower by-pass temperature is
[11] proposed for the low temperature district heating
network [3]. Table 9 shows the simulation results based
[12] on a10 C temperature drop along the supply pipe. It
shows the prediction errors increase in both supply and
Where : return pipes. The heat transfer was predicted in a
reverse trend in the return pipe at 4 C. Considerable
, [13] prediction error was found in the return pipe at high
ground temperature.
It is worth to be noted that the increase of supply
, [14]
temperature drop has more influence on the return pipe
temperature prediction than that of supply pipe. The
Tws- DN32, which is the longest main pipe in HE of reason can be explained from the expression of Us and
case 1, is selected for the assessment with U1=0.141 Ur in Eq. 12. As the magnitude of Ts-Tg is higher than
and U2=0.0523. The pipe length is assumed 500 m. Tr-Tg, the same amount of return water temperature
Ground temperature ranges from 0 to 15 oC. The inlet variation will have more influence on Ur than Us,
of supply and return temperatures are known as 55 oC therefore causes a larger prediction error in the return
and 22 oC respectively. The outlet temperature of pipe than in the supply pipe.
supply pipe is controlled as 50 oC and 45 oC, Table 9 Pipe temperature predication comparison (supply
o
respectively. outlet temperature is controlled at 45 C)
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 oC, 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 oC, the
net heat transfer effect in the return pipe is to absorb CONCLUSION
heat to the surrounding.
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
79
th th
introduced to avoid the mixing of by-pass water to the s = Supply

return water. Double pipe water supply concept was u = Upstream
tested to use the recirculation pipe supply water during
d = Downstream
winter season. Two different house installation modes
were considered in the analysis. Abbreviation
DH = District heating
The by-pass water significantly increases the return
water temperature in the traditional design. The mixed HE = Heat exchanger
return temperature can reach 35.5 oC for HE and DHWS = Domestic hot water storage tank
45.6 oC for DHWS. With applying the by-pass water
recirculation, this return temperature can be maintained REFERENCE
at 22 oC, while the re-circulated by-pass water can be
kept as high as 44 oC and 53.5 oC for HE and DHWS at [1] H. Lund, B. Moller, B. V. Mathiesen, A. Dyrelund,
the plant, respectively. It was found that the double The role of district heating in future renewable
pipe supply leads to the highest network heat loss. energy systems, Energy, 35, pp. 1381-1390,
However, the conclusion that whether the concept of 2010.
double pipe supply is inferior to other network design [2] Charlotte Reidhav, Sven Werner, Profitability of
methods can only be drawn after further network sparse district heating, Appliced Energy, 85, pp.
thermal-economic optimization. 867-877.
The simulation program simplifies the twin pipe heat [3] Udvikling og Demonstration af Lavenergifjern-
transfer prediction as a single pipe, and neglects the varme til Lavenergibyggeri, EFP 2007.
return pipe heat loss when the return pipe absorbs heat
from the surroundings. The temperature prediction [4] TERMIS Help Manual, Version 2.093,
errors due to the single pipe assumption were analyzed 7-Technologies A/S.
through solving the coupled supply/return pipe [5] Dansk Standard DS 439, 2000. Norm for
differential energy equations. The prediction errors vandinstallationer, Code of Practice for domestic
increase with increase the allowable temperature drop water supply installations, 3. udgave, www.ds.dk.
in the network. Considerable error was found for the
return pipe at high ground temperature. [6] Otto Paulsen, Jianhua Fan, Simon Furbo, Jan Eric
Thorsen, Consumer Unit for Low Energy District
Heating Net, The 11th International Symposium
NOMENCLATURE on District Heating and Cooling, 2008, Iceland.
cp = specific heat capacity [ J/kg.K] [7] P. K. Olsen, et.al, A new low-temperature district
q = Heat transfer rate [kW / m] heating system for low energy buildings, the 11th
s = Laplace transform variable International Symposium on District Heating and
Cooling, Iceland, 2008.
T = Temperature [ K]
U = Overall heat transfer coefficient [ kW /m.K] [8] Logstor. http://www.logstor.com/
Uij = Linear thermal transmittance [kW/m.K] [9] Benny Bohm, Halldor Kristjansson, Single, twin
and triple buried heating pipes: on potential
= mass flow rate [ kg/s]
savings in heat losses and costs, International
Greek Letter Journal of Energy Research, 29, pp. 1301-1312,
= Dimensionless temperature 2005.
Subscripts [10] P.Walleten, Steady-state heat loss from insulated
g = Undistributed ground pipes, Thesis, Lund Institute of Technology,
Sweden, 1991.
r = Return
80
th th
STEADY STATE HEAT LOSSES IN PRE-INSULATED PIPES

FOR LOW-ENERGY DISTRICT HEATING
1 1 1
A. Dalla Rosa , H. Li , S. Svendsen
1
Technical University of Denmark
ABSTRACT insulation series to reduce heat losses and thus saving

operational costs; however, this option would increase
The synergy between highly energy efficient buildings
investment and installation costs. The design principles
and low-energy district heating (DH) systems is a
for DH networks could instead be changed towards the
promising concept for the optimal integration of energy
use of media pipes with small nominal diameters, with
saving policies and energy supply systems based on
a higher permissible specific pressure drop. All-year
renewable energy (RE). Distribution heat losses
around lower supply temperature and return
represent a key factor in the design of low-energy DH
temperature constitute an effective option to reduce
systems. Various design concepts are considered in
heat losses [3]. These principles have a big potential
this paper: flexible pre-insulated twin pipes with
for heat supply to low-energy buildings, as explained in
symmetrical or asymmetrical insulation, double pipes,
[4] and they are investigated in this paper.
triple pipes. These technologies are potentially energy-
efficient and cost-effective solutions for DH networks in The total length of branch pipes can be significant in
low-heat density areas. We start with a review of proportion to the total length of the network, above all
theories and methods for steady-state heat loss in areas with a low-energy demand density. Moreover
calculation. Next, the article shows how detailed the temperatures in the critical service lines affect the
calculations with 2D-modeling of pipes can be carried temperature level in the whole network, so that the heat
out by means of computer software based on the finite losses and the temperature decay in building
element method (FEM). The model was validated by connection pipes are decisive for the overall
comparison with analytical results and data from the performance of the system. In this paper particular
literature. We took into account the influence of the focus was given to branch pipes.
temperature-dependent conductivity coefficient of
polyurethane (PUR) insulation foam, which enabled to State-of the art of district heating pipes
achieve a high degree of detail. We also illustrated the At present time DH distribution and service lines are
influence of the soil temperature throughout the year. based either on the single pipe system, where the
Finally, the article describes proposals for the optimal supply/return water flows in media pipes with their own
design of pipes for low-energy applications and insulation, or on the twin pipe system, where both pipes
presents methods for decreasing heat losses. are placed in the same insulated casing, or in a mixture
of them. All plastic pipe systems are characterized by
INTRODUCTION having the water medium pipe made of plastic (cross-
linked polyethylene (PEX) or polybutylene (PB)). They
The energy policy on energy conservation poses
are covered by insulation, usually polyurethane foam,
stringent requirements in the building energy sector, so
but in some cases of PEX foam or mineral wool; the
that the entire DH industry must re-think the way district
outer cover is formed by a plastic jacket. Durability of
energy is produced and distributed to end-users [1, 2].
plastic pipes is not a real issue, since it has been
This is a requirement to be cost-effective in low heat
proved that the expected life of PB pipes and PEX
density areas. Low-energy DH networks applied to low-
pipes is, respectively, more than 40 years and approx.
energy buildings represent a key technology to match
100 years [5]. As consequence of even lower average
the benefit of an environmentally friendly energy supply
operational temperature, longer lifetime can be
sector and the advantages of energy savings policy at
predicted according to Annex A in [6]. Studies have
the end-users side. Future buildings with a high
indicated that cross-linked polyethylene (PEX) pipes
performance envelope will lead to reduced space
have a cost advantage over steel pipes at pipe
heating load and therefore to a lower required
dimensions less than DN60, due to their greater
distribution temperature for heating. The introduction of
flexibility since the joints do not require welding [7].
low-energy DH networks is an appropriate and natural
Alternative design concepts must be considered in
solution to enhance energy and exergy efficiencies.
branch pipes from street lines to consumers
Distribution heat losses represent a key-point for substations: a pair of single pipes, twin pipes or triple
designing low-energy DH systems, due to the critical pipes. Traditionally most DH branch connections have
role they have in the economy of the system. The been built with two single steel pipes: one supply pipe
industry could meet the requirements of higher and one return pipe. Twin pipes can be made of steel,
81
th th
copper or PEX, with the supply and return pipe in the demand, although a non perfect cooling of DH water
same casing. The heat losses from twin pipes are occurs when tapping of DHW starts. The concept
lower than from single pipes, considering same based on twin pipes and a substation with
dimensions and temperatures. instantaneous production of DHW in a heat exchanger
is an optimal solution, if certain conditions are
Furthermore commercially available twin pipes, with
respected. The first requirement is that the control
dimensions up to DN200 for traditional steel media pipe
method gives priority to DHW preparation over space
or up to DN50 for PEX media-pipes are usually less
heating; the second condition is that the space heating
expensive to install than single pipes [7]. This
load during summer, to keep a high level of comfort in
technology has been introduced in Nordic countries
bathrooms for example, has to guarantee a sufficient
(and it is used in daily operation in many DH networks.
cooling of the return water. As a result media pipes with
Triple pipes might be considered in the near future, due
inner diameters as small as 10 mm can be applied in
to flexibility in the way the system can operate and
the primary loop and the water return temperature can
lower heat losses in case of optimal configuration. The
be kept sufficiently low, even in summer conditions.
choice of house connections depends mainly on the
length of the branch pipe, on supply and return The triple pipe system is applicable in three different
temperatures, building heating load and type of operational modes. The first one (mode I) occurs in
substation. The latter is decisive with regard to energy case of DHW demand, when pipe 1 and pipe 3 both act
performance and thermal comfort. The types of as water supply pipes; the second operational mode
substations are typically divided into three concepts: (mode II) is activated when an idle water flow is
unit with domestic hot water (DHW) storage tank, supplied by pipe 1 and pipe 3 acts as re-circulation line
where the tank is the secondary-loop and consumer to the supply distribution line, while the return line (pipe
unit with DH water tank, where the tank is placed in the 2) is not active: this is often the case when there is no
primary loop. In this paper branch pipe solutions are demand for space heating, but a small amount of water
considered for the concept of a consumer unit with heat circulates in the DHW heat exchanger, keeping the
exchanger and no storage tank. Two possible loop warm to satisfy the instantaneous preparation of
configurations of user connection to the distribution line DHW in the required time. This system avoids an
are shown in Figure 1. 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
Figure 1: Sketch of a user connection with heat 1, 2 active; pipe 3 not active.
exchangers: twin pipe connection with/ without booster
pump (12) and triple pipe connection (1-2-3).
1: supply METHODS
2: return
3: supply/re-circulation Theory of steady state heat loss in buried pipes
In order to calculate steady-state heat losses in DH
A simple and cost-effective configuration is composed buried pipes there are analytical methods [8] and
of the control system and two heat exchangers for, explicit solutions for the most common cases [9]. A
respectively, space heating (SH) and domestic hot complete review of the available literature about
water (DHW). The main disadvantage of such type of steady-state heat losses in district heating pipes has
substation unit is that only rather short lengths of been carried out in [10]. Here the methods are
service pipes can usually be applied; otherwise it would presented with reference to the present status of the
not be possible to assure the required DHW technology in the district heating sector. Furthermore
temperature at tapping points in the required time, due key-points and critical aspects are discussed; finally,
to the unsatisfactory transportation time. A modified improvements in the methodology of how to calculate
unit is therefore proposed and it is equipped with a steady-state heat losses are proposed, with particular
booster pump which assures quicker response to DHW focus on low-temperature and medium-temperature
82
th th
applications. Low-temperature district heating systems thermal coefficient, which is function of the temperature
are defined as networks where fluids at a temperature in this case. U-values are dependent both on
below 50 C are used, while a medium-temperature temperature and time. If the time-dependency due to
district heating system is defined as using fluids at the ageing of the foam can be restrained by introducing
temperatures not higher than 70 C [11, 12]. effective diffusion barriers, that is not true for the
Steady-state heat losses from pre-insulated buried intrinsic dependency on temperature. It is practice to
pipes are generally treated by use of the following evaluate the steady state heat loss applying a thermal
equation [10], which is valid for each pipe-i: conductivity value that corresponds to a hypothesized
mean temperature of the insulation. Nevertheless we
(1)
need models based, for example, on the finite element
where Uij is the heat transfer coefficient between pipe-i method (FEM) when complex geometries or a high
and pipe-j, Tj is the temperature of the water in pipe-j degree of detail are requested.
and T0 is the temperature of the ground. In case of two
buried pipes, which is the most common application in Temperature dependant thermal conductivity of
the DH sector, the heat losses can be calculated as PUR insulation foam
follows, respectively for the supply pipe and the return In this paragraph the authors want to explain and
pipe, where T1 is the supply temperature and T2 is the demonstrate the importance of taking into account the
return temperature. 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%
Supply pipe: (2) 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
Return pipe: (3) 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-
Equations (2) and (3) show how the heat transfer from dependency of the thermal conductivity of the
each pipe can be seen as linear superimposition of two insulation foam. The calculations in this paper use the
heat fluxes, the first one describing the heat transfer following expression, if not differently stated. It derives
between the pipe and the ground, the second one from experimental data [15]:
representing the heat transfer between the supply pipe (T) = 0.0196734 + 8.0747308.10-5.T [W/(mK)] (1)
and the return pipe. The equations can also be
re-arranged in the following way:
Supply pipe: (4)
Return pipe: (5)
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.
Equations (4) and (5) show how the heat transfer from
each pipe can be calculated by use of only one linear
83
th th
Temperature field in the soil around the pipe Combined heat and moisture transfer is disregarded.
In this paper we address the question of how to create The material properties are homogeneous and phase
a simple yet detailed FEM model for steady state heat changes, i.e. freezing and thawing were not
loss calculations. The overall heat transfer resistance considered. Table 1 lists the material properties, used
between the DH water and the environment is mainly as input values also for the following models; a sketch
composed of the thermal resistance of the insulation of the slab-model, where the boundary conditions are
and the thermal resistance of the soil; compared to described, can be seen in Figure 3.
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 Figure 3: Sketch of the model. Dimensions are in [mm].
temperature influences heat losses from DH pipes. The
soil layer around the heating pipes slightly warms up FEM model
around the pipes. The evaluation of the temperature A rectangle representing a semi-infinite soil domain
field in the soil is a prerequisite to create a realistic (width: 1020 m, height: 2040 m) is the most used
model for calculations of heat losses. Finite Element geometry to model the ground in heat loss calculations
Method (FEM) simulations were carried out and [18, 19]. In this paper a finite, circular soil domain was
temperature conditions in the soil around a typical DH applied, instead. Its diameter is 0.5 m and it is equal to
service pipe, suitable for low-temperature applications the distance between the surface and the centre of the
were evaluated over a 10-year period. casing pipe. Calculations show that the introduced
Table 1: Thermal properties of materials. simplification hardly affects the accuracy of the results.
The mesh model and an example of the temperature
[W/(mK)] [kg/m] Cp [J/(kgK)] field in a small size twin pipe are shown in Figure 4.
soil 1.6 soil 1600 Cp_soil 2000
PE 0.43 PE 940 Cp_PE 1800
PU
PUR 0.023 60 Cp_PUR 1500
R
PE
PEX 0.38 938 Cp_PEX 550
X
Ste Cp_Stee
Steel 76 8930 480
el l
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, Figure 4: Mesh model of a pre-insulated twin pipe
respectively for the supply pipe and the return pipe. embedded in the ground (top and left). Temperature field
in Aluflex twin pipe 16-16/110 (bottom-right); temperature
The heat transfer coefficient at the ground surface was
supply/return/ground: 55/25/8 C.
assumed to be 14.6 W/(m2K), including convection and
radiation [16]; we set the outdoor air temperature In [3], where FEM simulations were performed, it is
during the year according to the harmonic function valid stated that for media pipes size from DN 50 to DN 400,
for the Danish climate [17]: the deviation of the lineal thermal coefficient between
the piggy-back laying (arranging the supply pipe below
M the return pipe) and the traditional system (horizontal
Tair 8.0 8.5 sin 2 (6)
12 laying) is less than 1%. The same conclusion can be
stated for twin pipes; this is confirmed by calculations
84
th th
with the multipole method in [20] for two examples of FEM model: geometry of the ground and of the
twin pipe (DN 20 and DN 80) and by [10]. For twin pipes
pipes of even smaller size, such as in branch We considered the geometric model of the pre-
connections, the heat losses occurring in case of insulated Aluflex twin pipe type 16-16/110; the
vertical layout are only slightly more favorable than the temperatures of supply/return/ground are 55/25/8 C.
losses occurring on horizontally arranged pipes; this We calculated the heat losses for vertical or horizontal
result is shown with an example in the results section. placement of the media pipes inside the casing, which
was embedded in a rectangular or a circular model of
RESULTS AND DISCUSSION the ground. The same calculations were repeated for
other twin pipe size, up to DN 32 and other medium
In this section we discuss the influence of the soil
pipe materials, i.e. steel and copper. The results
temperature on heat losses; next, we present the
confirm that the vertical placement of the media pipes
validation of the FEM models; finally we apply the
inside the insulation barely affect the heat transfer,
method to show the potential for energy saving in the
being the difference between the two configuration less
case of asymmetrical insulation of twin pipes, in the
than 2% for the considered cases.
case of double pipes and triple pipes.
Table 2: Heat loss for various placements of the media
Temperature field in the soil pipes and various model of the ground.
Temperature conditions in the soil around a typical twin Heat loss Heat loss Heat loss
pipe, type Aluflex 1616/110, were evaluated over a Ground Media
pipes supply return total
10-year period. Figure 5 shows the all-year model layout [W/m] [W/m]
[W/m]
temperature profiles of the outdoor air and of the
ground at depth equal to 0.5 m, at three horizontal A Vert. 3.79 -0.17 3.62
distances from the centre of the casing, during the first A Horiz. 3.80 -0.18 3.62
year of operation. No notable differences in the yearly
B Vert. 3.84 -0.18 3.66
profile were noticed in longer periods of time.
We found that in state-of-the-art well insulated twin A: Semi-infinite, rectangular (width x depth: 40 m x 20 m)
pipes (series 2 or 3) a certain amount of soil is slightly B: Finite, circular (diameter: 0.5 m)
heated up by the warm twin pipe; nevertheless the
Steady-state heat loss in commercial pipes
level of such heating can be neglected because its
effect is not noticeable in comparison to the fact that The model was validated by comparing the results from
the uncertainties about the thermal properties of the FEM simulation to the analytical calculation for pre-
soil usually have a bigger impact. Considering yearly insulated pipes embedded in the ground [14].
average temperatures, the magnitude of the soil Calculations were carried out for four different sizes of
heating is about 1 C for distances of around 0.2-0.3 m Aluflex twin pipes (size 1414, 1616, 2020, 2626)
from the centre of the casing, and less than 0.5 C by and for chosen sets of supply (50, 55, 60 C), return
0.5 m. The temperature raise is considered in (20, 25, 30 C) and ground (8 C) temperatures. The
comparison to the undisturbed temperature of the selected pipes are suitable to be used as branch pipes
ground at a distance of 10 m. 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
Figure 5: All-year temperature profiles of the outdoor air material; the lambda-value of the insulation is then
and of the ground at depth equal to 0.5 m and 3 horizontal calculated as a function of such temperature. An
distances from the centre of the casing. average temperature of the ground is similarly
calculated. The calculation is repeated until the mean
temperature difference for the insulation material, pipe
85
th th
shell and surrounding soil is less than 0.005 C for two (Aluflex: DN 26, steel: DN 50) the best design is to
consecutive calculations. Thestandard and put the supply pipe in the centre of the casing, assuring
advanced model are available online [14]. In the the best possible insulation for the supply pipe. This
FEM advanced model we directly implemented strategy guarantees also the lowest temperature drop
equation (1) in the insulation domain, instead. The in the supply side, which is a critical figure in
results indicate that in case of low-temperature low-temperature applications.
operation, lower total heat losses are calculated if the
For bigger sizes (Aluflex: DN 26, steel: DN 50) the
temperature-dependency of the insulation lambda-
best design is achieved by moving up the media pipe
value is taken into account. Moreover the heat transfer
layout and at the same time by keeping the same
between the pipes in twin or triple pipes can be
distance between the media pipes as in the
properly evaluated.
symmetrical case.
7.0
6.0
Double pipes
Total Heat Loss [W/m]
5.0 A double pipe consists of a pair of media pipes of

4.0
dissimilar size, co-insulated in the same casing. It is a
further development of the twin pipe concept. A sketch
3.0
of a possible application of the double pipe concept is
2.0
shown in Figure 7. Though these measures, network
1.0 heat loss reduction is possible, in case of operation
0.0 during low heating load periods.
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
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
Figure 7: Sketch of the possible application of the double
drop); next, the heat loss from the return pipe can be
pipe concept in a simple district heating network.
close to zero, maintaining isothermal conditions in the
return line. If commercial available casing sizes are The space heating demand in summer is diminished,
kept, we suggest two design strategies, depending on except for the energy requirement in bath room
the size of the pipes. For small pipe sizes heating. According to the energy balance, the reduced
heating load requires less
Table 3: Comparison between asymmetrical and symmetrical insulation in twin pipes. network flow rate as far as the
The centre of the casing is the origin of the Cartesian system. designed building temperature
drop is sustained. However, the
Coordinates Heat loss asymm.- reduction of network flow rate
(x; y) [mm] [W/m] symm. [%]
will increase the supply water
Size
(DN)
Mat. Sup. Ret. Sup. Ret. Tot. Sup. Tot. temperature drop along the
pipeline due to heat loss. As a
14 (0; 0) (0; 27) 3.24 0.01 3.25 -7.6 2.0
consequence, the supply
16 (0; 0) (0; 28) 3.56 -0.01 3.55 -5.1 1.1 temperature at the end user
20 Alx. (0; 0) (0; 30) 4.16 -0.04 4.12 -4.2 -0.3 may lower down below the
minimum requirement. This
26 (0; 0) (0; 36) 4.67 0.00 4.67 -5.1 1.9
problem is relevant to tow-
32 (0; -16) (0; 28) 5.54 0.00 5.54 -5.8 -2.5 energy DH systems with an
already low supply temperature.
50 (0; -25) (0; 55) 5.69 -0.03 5.66 -7.7 -2.4
Steel This design is based on the fact
65 (0; -36) (0; 60) 6.70 -0.02 6.68 -7.8 -3.2 that the supply line acts also as
re-circulation line during low
86
th th
heating load periods; hence by-pass at the critical Triple branch pipes
consumers are not necessary and the exergy loss due The development of an optimized triple pipe solution for
to the mixing of warm water into the return line is low-energy applications is reported to show the
avoided. Furthermore the water flow in the return line potentiality of utilizing detailed models for steady-state
has the same direction as in the supply line (clockwise heat loss calculation. In this survey focus was given on
in the example), so that the smallest size for the return the choice of media pipes diameters as small as
pipes are expected in correspondence to the biggest possible. The triple pipe geometry is based on
size for the supply size, and vice versa. This results in modifications of the 14-14/110 (outer diameters in [mm]
lower local pressure differences between supply and of respectively supply pipe, return pipe, casing) twin
return lines and savings in operational costs, thanks to pipe design which has been reported in [18]. Four
lower heat losses. This is shown in Table 4 and Table geometrical variations have been considered (see
5, by means of two examples: the first one refers to a Figure 8) and the Cartesian coordinates describing the
small to medium-size distribution network, the second placement of media pipes inside the casing are listed in
one to a bigger one, being capable to supply four times Table 6.
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 Total
Sup. Ret. Tot.
(DN) (system) [%]
40-40 -6.24 0.04 -6.20 Twin:
80-80 -7.66 0.07 -7.59 -13.79
6.1
40-80 -5.55 0.05 -5.58 Double:
80-40 -7.41 0.05 -7.36 -12.94
Figure 8: four different geometries for a triple service pipe
type Aluflex 14-14/110.
Table 5: Comparison between a distribution network
based on twin pipe (DN100-100 and DN200-200) with a
Table 6: placement of media pipes inside the casing for
distribution network based on double pipe (DN100-200
four triple pipe geometries, type Aluflex 14-14-20/110.
and DN200-100). Supply/return/ground: 55/25/8 C.
Coordinates (x, y) [mm]
Heat loss [W/m]
Pipe 3
Size Total Variation Pipe 1 Pipe 2
Sup. Ret. Tot. [%] (Sup. or
(DN) (system) (Sup.) (Ret.)
re-circ.)
100-100 -7.83 -0.55 -8.39
Twin: A (14;-14) (0;20.5) (-14;-14)
200-200 -8.92 0.24 -8.68 -17.06 B (10;-14) (0;20.5) (-21;-7)
11.8 C (3;-14) (0;20.5) (-21;-7)
100-200 -6.4 0.08 -6.36 Double: D (0; 0) (0;25) (0;-28)
200-100 -8.07 -0.03 -8.69 -15.05
The results of FEM simulations are listed in Table 7 for
We considered an optimal placement of the media the four geometries (A, B, C, D) and the three
pipes in case of double pipes, thus asymmetrical operational modes (I, II, II), previously described. Since
insulation is applied. The total amount of insulation is mode II occurs in case of no demand of space heating
used both in the twin pipe-based distribution network and then outside of the heating season, simulations
and in the double pipe-based one, so that the were additionally performed with a more realistic
investment costs are equal in both cases. Results show temperature of the ground during that period
that the heat loss can be reduced by 6% by means of (14 C),considering Danish weather. This gives also an
double pipes instead of twin pipes for the low to insight in the effect of ground temperature throughout
medium-size distribution network. Even higher energy the year.
savings (around 12%) are possible in the case of the
large-size distribution network.
87
th th
Table 7: Steady state heat losses of triple pipes type mode II is the most likely outside heating season, the
Aluflex 14/14/110 for 4 geometries and 3 operational conclusion is that geometry D is preferable.
modes. Temperature supply/recirculation/return/ground:
55/55/25/8 C.
CONCLUSIONS
Heat loss [W/m]
Geo
Mode m. Pipe Pipe Pipe The soil temperature at 0.5 m below the surface varies
Tot.
1 2 3 between 2 C in January-February and 14 C in
A 2.67 -0.08 2.67 5.30 JulyAugust, for Danish conditions. This knowledge
I can be used to better predict the winter peak load and
B 2.91 -0.29 2.75 5.38
(DHW the temperature drop in the distribution line during
tapping) C 2.52 -0.22 2.74 5.06 summer.
D 2.46 0.05 2.74 5.24 The slab-model for steady state heat loss calculations
A 2.67 / 2.67 5.34 can be replaced, in case of small size
II distribution/service pipes, by a model where the effect
(supply-to- B 2.69 / 2.85 5.55
of the soil is represented by a circular soil layer around
supply re- C 2.48 / 2.70 5.18
circulation) the district heating pipe.
D 2.49 / 2.75 5.25
The results confirm that the vertical placement of twin
A 3.46 0.48 / 3.95 media pipes inside the insulation barely affects the heat
III B 3.39 0.43 / 3.83 transfer, in comparison to the horizontal placement; the
(space difference between the two configurations is less than
heating) C 3.41 0.35 / 3.76 2% for the considered cases.
D 3.53 -0.01 / 3.53
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
Table 8: Steady state heat losses of triple pipes type
Aluflex 14/14/110 for 4 geometries and operational mode accuracy of the heat transfer calculation among pipes
II. Temperature supply/recirculation/ return/ ground: embedded in the same insulation.
55/55/25/14 C.
We applied the model to propose optimized design of
Heat loss [W/m]
Geom.
twin pipes with asymmetrical insulation, double pipes
Pipe Pipe Pipe
Tot. and triple pipes. We proved that the asymmetrical
1 2 3
insulation of twin pipes leads to lower heat loss from
A 2.35 / 2.35 4.70
II the supply pipe (from -4% to -8%), leading to a lower
(supply-to- B 2.37 / 2.51 4.88 temperature drop; next the heat loss from the return
supply re- C 2.39 / 2.63 5.02 pipe can be close to zero.
circulation)
D 2.20 / 2.42 4.62 It is possible to cut the heat losses by 612% if an
optimal design of double pipes is used instead of
We conclude that an absolute best design for the traditional twin pipes, without increasing the investment
service triple pipe does not exist, but it depends on the costs.
operational mode that is chosen as critical. In fact the The development of an optimized triple pipe solution
results reported in Table 7 and Table 8 show that was also reported. It is suitable for low-energy
geometry C gives the lowest total heat loss for applications with substations equipped with heat
operational modes I and II, while geometry D has the exchanger for instantaneous production of DHW.
best thermal performance for operational mode III and
for operational mode II, if a temperature of the soil of
REFERENCES
14 C is considered. It has to be underlined that,
considering the operational mode III, geometry D [1] S. Froning, Low energy communities with district
shows no heating of return water; this is a situation heating and cooling, 25th Conference on Passive
always desirable, although it has a slightly higher heat and Low Energy Architecture, Dublin (2008).
loss from the supply pipe than the other geometries. It [2] S. F. Nilsson et al., Sparse district heating in
is proved that usually operational mode I occurs for Sweden, Applied Energy 85 (2008), pp. 555564.
less than 1 h/day [20]. Moreover the temperature drop [3] F. Schmitt, H.W. Hoffman, T. Gohler, Strategies to
in the supply pipe to the DHW heat exchanger is critical manage heat losses technique and economy,
in low-temperature applications, so that it is strongly IEA-DHC ANNEX VII, (2005).
recommended to minimize the heat loss from this [4] P.K. Olsen, B. Bhm, S. Svendsen et al., A
media pipe. Considering all this and the fact that mode new-low-temperature district heating system for
III is the most likely during the heating season and
88
th th
low-energy buildings, 11th international symposium [13] J. Korsman, S. de Boer and I. Smits, Cost benefits
on district heating and cooling, Reykjavik (2008). and long term behavior of a new all plastic piping
[5] M. Klompsch, H. Zinko, Plastic pipe systems for system, DHC ANNEX VIII (2008).
DH, handbook for safe and economic application, [14] www.logstor.com (March 2010).
IEA-DHC ANNEX V (1999). [15] Udvikling og demonstration af lavenergifjernvarme
[6] DS-EN 253:2009, District heating pipes Pre- til lavenergibyggeri (development and
insulated bonded pipe systems for directly buried demonstration of low energy district heating for low
hot water networks - Pipe assembly of steel service energy buildings), 2007.
pipe, polyurethane thermal insulation and outer [16] B. Kvisgaard, S. Hadvig, Varmetab fra
casing of polyethylene. fjernvarmeledninger (Heat loss from pipelines in
[7] H. Zinko, GRUDIS-tekniken fr vrmegles district heating systems), Copenhagen (1980).
fjrrvrme (The GRUDIS technology for low heat [17] DS418:2002, Calculation of heat loss from
density district heating), Swedish District Heating buildings.
Association, Stockholm (2004). [18] H. Kristjansson, F. Bruus, B. Bhm et al.,
[8] J. Claesson, J. Bennet, Multipole method to Fjernvarmeforsyning af lavenergiomrder (District
compute the conductive heat flows to and between heating supply of low heat density areas), 2004.
pipes in a cylinder. Department of Building [19] T. Persson, J. Wollerstrand, Calculation of heat
Technology and Mathematical Physics, Lund flow from buried pipes using a time dependent
(1987). finite element model, 45th International
[9] P. Wallenten, Steady-state heat loss from insulated Conference of Scandinavian Simulation Society,
pipes, Lund (1991). Copenhagen (2004).
[10] B. Bhm, On transient heat losses from buried [20] B. Bhm, H. Kristjansson, Single, twin and triple
district heating pipes, International Journal of buried heating pipes. On potential savings in heat
Energy Research, 2000, Vol. 24, pp. 1311-1334. losses and costs, International Journal of Energy
[11] Terminology of HVAC, ASHRAE, Atlanta (1991). Research (2005), Vol. 29, pp.1301-1312.
[12] I.B. Kilkis, Technical issues in low to medium- [21] EN 13941:2003, Design and installation of pre-
temperature district heating, International Journal insulated bonded pipe systems for district heating.
of Global Energy Issues, 2002, Vol. 17,
pp. 113-129.
89
th th
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 standardised method used for determining the

steady-state thermal conductivity of district heating
The standardized methods to measure the thermal
pipes, the guarded hot pipe method, is only applicable
conductivity of straight district heating pipes are not
on straight single pipes. The method is based on [1],
applicable on flexible district heating pipes. This paper
described in [2] and [3]. A heater pipe is placed inside
presents a transient method determining the
the service pipe and the heat transferred through the
temperature dependent thermal conductivity of flexible
insulation is measured. The measurements must be
twin pipes.
conducted with a constant distance between the heater
A transient method to determine the temperature- pipe and the service pipe along the test specimen
dependent thermal conductivity of flexible single district which can not be achieved with flexible pipes. An
heating pipes is presented in this paper. A flexible pipe alternative method was presented in [4] and applied in
coil is immersed into cold water. Hot water is [5] where the thermal conductivity of flexible district
distributed in the coil. The temperature decline of the heating single and twin pipes can be determined. The
coil water is measured and calculated. Minimizing the temperature decline of hot water pumped in a flexible
difference between the calculated and measured pipe coil is measured. A long pipe coil is needed to
temperatures gives (T) of the flexible polyurethane have a sufficient temperature decline along the pipe.
foam. The method gives small errors. The Danish method is based on steady-state
measurements at different temperatures to get the
INTRODUCTION temperature dependence of the decline. The Finite
Element Method is used to determine the thermal
District heating is supplied to the customers in one pipe conductivity (T) of twin pipes.
and returned to the heat generation plant in another
pipe. The two pipes may be placed in separate casings A transient method to determine the temperature-
(single pipes) or in one casing (twin pipes), see Fig 1. dependent thermal conductivity of flexible single district
The temperature difference between the district heating heating pipes was presented in [6]. A pipe coil is
supply (~80-110C) and return temperature (~40-50C) immersed into cold water and the temperature decline
gives an internal heat flow from the supply pipe to the of hot water inside the coil is measured. The measured
return pipe in a twin pipe. The total distribution heat temperatures are compared to numerically calculated
loss from a twin pipe is lower than that of comparable values to characterize (T). A Kirchoff transform is
single pipes due to this internal heat flow. When used to simplify the calculations. Finally, the mean
distributing district heat to areas with single-family square difference of the measured and calculated
houses to heat sparse areas, the issue of distribution temperatures are minimized which gives (T). In this
heat losses is of special importance. The relative paper, a similar experimental set-up is used for a
distribution heat losses are considerably higher in flexible twin pipe. The numerical and mathematical
sparse areas than in more heat dense areas due to low model developed in [6] cannot be used for twin pipes
heat densities. Flexible district heating twin pipes are due to the complicated geometry of twin pipes. In [7]
widely used when single-family houses are connected and [8] a method was presented where heat losses
to district heating systems due to their light weights, from district heating twin pipes were calculated with
flexibility and long lengths. In the efforts of minimizing conformal coordinates describing the twin pipe
distribution heat losses, the possibility of determining geometry. In this paper, the conformal coordinate
the insulation capacity of flexible twin pipes is an model is used to calculate the temperature decline in a
important issue. 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.
Fig. 1 Cross-section of single (left) and twin (right) district

heating pipes
90
th th
EXPERIMENTAL MEASUREMENTS The twin pipe studied in this paper has two copper
service pipes, semi-flexible polyurethane foam and a
The experimental set-up is similar to that used when
slightly corrugated LDPE casing. The pipe is of
determining the thermal conductivity of single district
dimension DN 20 with the pipe dimensions described in
heating pipes in [6]. A flexible twin pipe of about 18
Table 1. The pipe producer declares a thermal
meters coiled with a diameter of 1.8 meters is
conductivity of the semi-flexible polyurethane foam of
immersed into a pool with circulating water. In this
50 = 0.0255 Wm-1K-1of a newly produced pipe of
experiment, the pool water was about 17 C. Previous
this kind. The pipe has no diffusion barrier. The density
tests show that air is unsuitable as surrounding media
of the polyurethane foam was = 60 kg/m3.
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
Table 1. Dimensions of the twin pipe studied
steady-state is established in the insulation, at time
t = 0, the circulation is stopped. Then, the temperature Casing outer diameter (mm) 91
decline of the stagnant loop water is measured at one Casing thickness (mm) 2.2
position in the coil. The thermocouples are placed at
three positions of each service pipe, see Fig. 2. One is Service pipe outer diameter (mm) 22
placed on top of the service pipe, one on the side and Service pipe thickness (mm) 1.0
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 initial coil temperature was T0 = 81.3C. The water
the insulation was probably insufficient and it appears temperature at the service pipe (Tw,meas(t)) decreases
as if pool water permeated after about 5.5 hours and during the 16 hours of measurements. The pool
disturbed the measurements. temperature was initially T1 = 17.4 C and increased
slightly to T1 =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 Tw,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.
Fig. 2 Experimental set-up and positions of thermocouples

at the service pipes
91
th th
90
80
Chosen interval
70
60
Temperature (C)
50
40
Tw, meas
30
20
T1, meas
10
0
0 2 4 6 8 10 12 14 16
Time (h)
Fig. 3 Measured coil Tw, meas and pool temperature T1, meas.
18,0
17,9
Chosen interval
17,8
Temperature (C)
17,7
17,6
17,5
17,4
0 2 4 6 8 10 12 14 16
Time (h)
Fig. 4 Measured pool temperature T1, meas
92
th th
MODELLING USING CONFORMAL COORDINATES water and the right-hand boundary against the poll
water. The heat flux in the vertical v-direction is zero on
It is rather complicated to calculate the temperature
the horizontal boundaries due to symmetry.
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:
Fig. 6 Initial temperature distribution in the cross-

section of a pipe quarter in the u, v-plane.
T T T
c ( (T ) ) ( (T ) )
t x x y y (1)
In the numerical solution, the region is divided into a
rectangular mesh. The area factor is now the area of
The (x,y)-coordinates ( z x i y) are transformed to
each of the cells shown in Fig.5. They are shown in
suitable conformal coordinates (w u i v) with the Fig. 7. The largest cell is the one in the lower left corner
aid of line sources and so called multipoles. 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 Tw = 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
Fig. 5 A quarter of a twin pipe in x-y-plane geometry water in the twin pipes are calculated
The density and the heat capacity c of the
The heat equation in the conformal coordinates is: polyurethane foam are assumed constant in the
temperature interval studied. The boundary
T T T temperatures at the casing are given by the pool
c A(u, v) ( (T ) ) ( (T ) ) (2)
t u u v v temperature.
The thermal conductivity (T) of the polyurethane foam
is determined by the thermal conductivity at 50C 50
Here, A(u,v) is the area factor in the conformal
(W/mK) and a coefficient to account for a linear
transformation.
temperature dependence.
The considered region shown in Fig. 5 is transformed
to a rectangular region in the u, v-plane, see Fig. 6. In (T ) 50 1 ' (T T50 ) (3)
the figure, the left-hand boundary lies against the coil
93
th th
EVALUATION OF MEASUREMENTS The difference between the calculated and measured

temperatures for the optimal parameter values of 50,
The temperature-dependent thermal conductivity of the
and c are showed in Figure 9. The saw tooth
polyurethane foam is obtained by calculating the
disturbance from the measurements is seen. The
temperature decline of the coil water. Certain values of
difference giving the best fit lies in the interval -0.20 to
the thermal conductivity of the polyurethane are
0.25 (C). The error is small.
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-1K-1) of the polyurethane
foam is input to the calculations. Literature references
for the heat capacity of polyurethane foam varies, 1300
Jkg-1K-1 at 50C in [9], 1400 Jkg-1K-1 in [10], 1400-
1500 Jkg-1K-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 Tw (C) and measured coil temperatures
Tw,meas (C).
Fig. 9 The difference between the measured and
D max Tw, calc (t ) Tw, meas (t ) for t1 t t2 (4) calculated temperatures for the optimal parameter
values
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 The optimal thermal conductivities 50 and give the
due to the disturbances in the measurements probably final result for (T), as shown in Fig.10.
caused by penetration of pool water at the position of
three thermocouples. The precision decreases also 0,031
with time due to the decreased difference between the 0,03

coil temperature and pool temperature. The problems
(T) (W/mK)
0,029
at the start can be an effect of not having steady-state 0,028
conditions before starting the temperature decline. 0,027
The maximal values of D for the interval studied are 0,026
compared for different combinations of the parameters. 0,025
A single optimum point is obtained with the lowest 20 30 40 50 60 70 80
value of D as illustrated in Fig. 8. Temperature (C)
' (W/mK) Fig. 10 Estimated thermal conductivity of the studied

semi-flexible polyurethane foam
50 (W/mK)
D = 0.55 o 50 = 0.029
' = 0.0072 o D = 0.60 DISCUSSION AND CONCLUSIONS
In this paper, the twin pipe method is evaluated only
D= 0.43
o o
D= 0.28
o
from the experiences of one experiment. However, this
c (J/kgK)
c = 900 D = 0.23 c = 1500
experiment is part of a greater context with experiences
from previous experiments on single pipes.
D = 0.43
o 50 = 0.0285 W/mK
50 = 0.43 The experimental procedure can be improved. In this
o D = 0.69 = 0.0054 W/mK
' = 0.0036 c = 1200 J/kgK experiment, the reattachment of the polyurethane foam
at the positions of the thermocouples was probably
insufficient. Penetration of pool water lead to a certain
Fig. 8 The maximum difference D for different change of slope at about t = 5.5 h. The problem with
combinations of 50, and c in the chosen interval.The penetration could have been solved by measuring the
optimal parameter values giving the lowest D are showed coil temperature inside the copper pipe instead.
in the box.
94
th th
The sawtooth variation seen in the measurements in [6] [5] District Heating Association (2008), Heat plan
is seen in this measurement as well and should be Denmark, Ramboll Danmark A/S and Aalborg
further investigated. University, (2008), In Danish, available at Dansk
A large difference between the pool and coil Fjernvarmes F&U-Konto, www.danskfjernvarme.dk
temperature is desirable to minimize the relative errors. [6] C. Reidhav and J. Claesson, A transient method to
It is also important to assure that steady-state determine temperature-dependent thermal
conditions are established before starting the conductivity of polyurethane foam in district heating
temperature decline. pipes, Building Physics 2008 - 8th Nordic
Symposium, Copenhagen, Denmark, (2008)
The final result, the obtained thermal conductivity:
[7] C. Persson and J. Claesson, Prediction of heat
(T ) 0.0235 10 105 T (C), 50 0.0285 (5) losses from district heating twin pipes, The 11th
is in reasonable agreement with the declared 50 =
Cooling, August 31 to September 2, Reykjavik,
0.0255 Wm-1K-1 for a newly manufactured pipe. This
Iceland, (2008)
pipe piece had been in store for some time and had no
diffusion barrier. The temperature-dependent part of [8] C. Persson and J. Claesson, Numerical solution of
the thermal conductivity is in well agreement with [12]. diffusion problems using conformal coordinates.
Application to district heating pipes, Report
Department of Civil and Environmental
REFERENCES
Engineering, Chalmers University of Technology,
[1] U. Jarfelt, Test apparatus of pipe insulation. Gteborg, Sweden (2008)
Doctoral thesis. Chalmers University of [9] S. Peng, P. Jackson, V. Sendijarevic, K.C. Frisch,
Technology, Gteborg (1994) G.A Prentice, A. Fuchs, Process Monitoring and
[2] European standard EN 253:2009, District heating Kinetics of Rigid Poly(urethane-isocyanurate)
pipes - Preinsulated bonded systems for directly Foams, Journal of Applied Polymer Science,
buried hot water networks Pipe assembly of steel (2000) Vol 77, 374-380
service pipe, polyurethane thermal insulation and [10] R. Zevenhoven, Treatment and disposal of
outer casing of polyethylene, Brussels, Belgium. polyurethane wastes: options for recovery and
(2009) recycling, Helsinki University of Technology,
[3] European committee for standardization. European Report TKK-ENY-19, Espoo, Finland, June (2004).
standard EN ISO 8497:1996, Thermal insulation- [11] BING, Federation of European Rigid Polyurethane
Determination of steady-state thermal transmission Foam Associations, Thermal insulation materials
properties of thermal insulation for circular pipes, made of rigid polyurethane foam (PUR/PIR),Report
Brussels, Belgium. (1996) No1 October (2006)
[4] Danish District Heating Association. Development [12] U. Jarfelt and O. Ramns, Thermal conductivity of
of an experimental set-up for measuring the heat polyurethane foam best performance,
conduction properties of flexible pipes, Project nr. th
10 International Symposium on District Heating
2006-05, rhus, Danmark. (2006), In Danish, and Cooling, Sept 3-5, Hanover, Germany, (2006)
available at Dansk Fjernvarmes F&U-Konto,
.
www.danskfjernvarme.dk
95
th th
DISTRICT HEATING PIPES 200 MM BELOW SURFACE

IN A STREET WITH HEAVY TRAFFIC
1 2
Anders Fransson and Sven-Erik Sllberg
1
Gteborg Energi AB, Sweden
2
Building Technology and Mechanics, SP Technical Research Institute of Sweden
ABSTRACT sales (i.e. less district heating, d.h. to be sold). To

connect new district heating customers in the future,
This article reports the results from a field experiment
with the competition of other heating suppliers, it is not
initiated by Gteborg Energi AB with an extreme
enough to use just smaller pipes because of the
shallow burial of district heating pipes 162/76,1 (DN 65)
smaller demands. Building the grid and maintaining the
casaflex under a street with heavy traffic designed for
grid needs to become more cost efficient.
an average of 20004000 passes of vehicles a day and
line. The pipes were laid only 200 mm below the The purpose of this article is to inspire and if possible
surface. The backfill was of 040 mm particle size. help whoever is interested in making district heating
and cooling in the world more cost effective using the
Several consecutive measurements were done to study
ideas or test results from this article.
the effects from instant and long term loads from the
traffic. The tests were done on a test pipe prepared
1.2 Cost-cutting due to shallow burial in roads
with displacement gauges and on operating pipes.
When reducing costs, it is important to maintain the
The aim is that the results will inspire and give input for qualities that are required. The road owner needs the
making district heating and cooling more cost effective. road to be functional and has its standards. The district
The tests showed that both the instant and long turn heating supplier is responsible for its pipes and
deformation of the pipes are small at the actual laying deliveries of heat and has its standards. Finally there
depth and also that the acceleration in the ground as are workers (contractors and maintaining staff) who
heavy vehicles passes does not seem to be alarming. need acceptable working conditions.
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 Fig. 1 Standard shaft section
legislation from 2006 allowed new buildings in In a standard shaft section the drainage may be taken
Gothenburg to use a maximum of 110 kWh/m2 away in roads. A properly built road has a hard top and
externally supplied energy for heating (,cooling) and is drained as it is. You do not need to drain it any
producing domestic hot water. Today the municipality more.
of Gothenburg wants new buildings to use 60 kWh/m2
at most. It is also possible to make the shaft more narrow and
maintain acceptable working conditions if either long
These changes are not unique. New houses are using pipes with no joints are used or if the joints are welded
less and less energy per square meter. There are on top of the shaft.
already households that are not using but producing
energy. The former energy suppliers in Europe are Less coverage is also an alternative. Earlier studies
finding themselves not as suppliers but distributors, [1][3] shows that the pipes are solid enough to be
buying and selling energy. Climate changes are global placed with very little coverage (180 mm) and in rough
and have already measurable effects on district heating materials. It is also shown that there is less settling in
96
th th
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. 4 Marking of the location of the gauges.
The test shaft
Fig. 2 Left; Supporting leg with no cover, Right; Rock Three displacement gauges were placed in a test pipe
curb. containing air (see section 2.1) beside two operating
pipes (see Fig. 35). Two accelerometers were placed
If the existing fraction is used as backfill transports can
in separate boxes near the test pipe. The gauges were
be reduced which lower the costs and the
monitored through wires at the bicycle path beside the
environmental influence.
street.
Normally the district heating pipes have no problems
The test was done with two single district heating pipes
coping with the traffic load. The extra pressure and the
162/76,1 (DN 65) casaflex buried in a fraction with
movement in the soil are making extra loads that are
grain size 040 mm. (Normal standard is a fraction of
quite negligible compared to the thermal load, the inner
sand 0.216 mm.) The distance from the top of the
pressure load and the load from the outer pressure
pipes to the top of the fraction was 60 mm. (Normal
from the soil.
standard is 460 mm.) The distance from the top of the
To get an idea of if the graphite gaskets used in the pipes to the top of the asphalt was 200 mm. (Normal
pipe joints (for casaflex) can stand the traffic load standard is 600 mm.)
200 mm below the surface in a street with heavy traffic
As extra protection, the operating pipes were wrapped
the traffic load was empirically measured in the test
in a grid of polyethylene, PE.
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. 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
Fig. 3 Drawing over the test area. the thermal loads and the working conditions (i.e. the
97
th th
length 140 m of a single pipe means that there is no 100 100
need to work with the pipe in the shaft).

Indicator 1
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 Indicator 2
Indicator 3
countries. The pipe is supposed to be used as ordinary
pipes with sand as backfill. Fig. 8 Locations of displacement gauges inside test pipe.
The media pipe is made of corrugated stainless steel During the test period that lasted for one year
and surrounded by CFC-free polyisocyanurate foam. indications from the displacement gauges were
The foam is wrapped in a multi layer barrier foil at the measured twelve times. During the test period the
outside covered with a corrugated low density temperature varied between summer temperatures to
polyethylene casing. Inside the insulation along the winter temperatures.
pipe there are three surveillance wires. The casaflex
pipe can be delivered in very long lengths. The pipes 2.2 Instant deformation of the pipe and
used in this test were 140 m. accelerations from traffic load
Two accelerometers were placed 200 mm respectively
To connect different casaflex pipes a system with
600 mm below the asphalt surface (see Fig. 5), close to
flanges, bolts and gaskets are used. The gaskets are
the test pipe, to measure the vibrations in the road
made of graphite.
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.
Fig. 6 Left; Casaflex pipe, Right; Casaflex pipe with a
joint. It was arranged so that a heavy lorry passed over the
test area several times at different speed (20 and 40
2.1 Deformation of the pipe over time km/h) while the vibrations in the road structure were
The test pipe was 1.66 m long and prepared with three registered with the two accelerometers. The weight of
displacement gauges inside to measure the radial the lorry was 26 400 kg.
deformation in three directions. The displacement To investigate the instant deformations in the test pipe
gauges were installed at the half length of the pipe with when heavy vehicles pass over the pipe the indications
a distance of 100 mm in between. One displacement from the displacement gauges (see section 2.1) were
indicator measured on the upper side of the pipe measured at the same time as the vibrations in the
casing, the second on the underside of the pipe casing road structure were registered.
and the third at the side of the pipe casing. The
displacement gauges were fixed to the media pipe to 2.3 Radial and axial stiffness of pipe
measure the changes in the pipe casing.
In laboratory the physics of the test pipe were tested
Before the test pipe was installed reference concerning radial stiffness and axial stiffness. The tests
measurements were done at the laboratory to create were done on a 165 mm long test specimen from the
zero values for the displacement gauges. same pipe as the test pipe. The arrow in Fig. 9 shows
the direction of the applied load during the test.
Fig. 7 1.66 m long district heating test pipe of type

Casaflex 162/76.1 (DN 65).
Fig. 9 Arrangement for test of radial stiffness.
98
th th
The axial stiffness in the pipe was tested in three ways. 2.5 Leak test of the pipe casing
Fig. 10 describes the three arrangements for applying To discover moisture or even water in the insulation,
the load in the tests, (a) the applied load acts on the there are different indicators on the market.
whole cross section, (b) the applied load acts on only The typical indicator system used in Gothenburg is the
the steel pipe and (c) the applied load acts on the outer so called Nordic System. The Nordic System is a
steel net including the pipe casing. The arrow in the system which is using two naked cupper wires inside
figures indicates the direction of the load. the insulation along the pipe at 10 am and 2 pm.
The casaflex pipe uses the Hagenuk System. That
(a) (b) (c)
system uses three wires
a) Ni Cr,
b) Cu, insulated and
c) Cu, not insulated.
165 In this test different pipes and different systems were
connected. The Ni Cr wire in the Hagenuk System was
left disconnected.
90 140 The resistance was measured with an ordinary ohm

meter, BM 400.
Fig. 10 Three types of arrangement for applying the load
for test of axial stiffness. The pipes were also three times tested with a, State
meter, Time Domain Reflectometer (TDR) from
2.4 Pipe prolonging while pressurizing Stateview.
A casaflex pipe does not expand because of the
thermal load. It is self compensating. But because of 2.6 Test of degree of compaction of the street
the geometry of the media pipe it expands when it gets To get the permission from the road owner to do this
pressurized. On the other hand the multi layer barrier test in the street there were certain standards to follow
foil in the pipe holds the expansion back. Because the [5] and [6].
pipe is flexible, it will still be able to expand, but only
Before the asphalt could be put on the shaft there were
until the multi layer barrier foil stops the expansion.
to be some tests of the degree of compaction of the
To see how much the pipe expands because of the street with certain limits. It is a German test that is also
pressurization, a distance indicator, Hilti PD4, was fixed used in Sweden [7]. Basically the soil gets compressed
on the pipe before it was installed and pressurized with a known load over a known area and one
while it was still on the ground. The distance was measures the Youngs modulus Ev for the soil two
measured three times against an iron angle which also times. The demands were that;
was fixed on the pipe. After the pressurization the
a) Ev2 / Ev1 < 2,8
distance was measured again three times. b) Ev2 > 50 MPa
c) At least 4 out of 5 tests should be correct.
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:
Fig. 12 Test of degree of compaction of the street
Fp = PA (1)
Where Fp is the prolonging force [N], P is the internal 2.7 Visual control of the surface of the street
over pressure [Pa] and A is the maximum inner area of As an extra precautionary measure, the street was
the pipe [m2]. 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
99
th th
find potential frost action damages. The inspections In Fig. 16 and 17 the diagrams show the vibration
were documented with photos. velocity (m/s) in the ground when a heavy lorry pass
over the test area at a speed of 40 km/h. The vibration
3. RESULTS velocity is calculated from the acceleration signal by
integration.
3.1 Test results from deformation of the pipe over
time The diagrams in Fig. 18 and 19 show the maximum
amplitude of the acceleration in the ground as a
The measured pipe deformations during the test period
function of the speed of the lorry when it passes over
turned out to be very small. The diagram in Fig. 13
the test area in 20 km/h and 40 km/h, respectively the
describes the measured changes in the casing since
maximum vibration velocity as a function of the speed
installation and average air temperatures during the
of the lorry.
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 200 mm below surface, vehicle speed 40 km/h
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
Time (s)
squeezed out approximately 0.1 mm. The under side
(red curve) was squeezed out approximately 0.1 mm Fig. 14 Vertical acceleration 200 mm below the road
during the period between the first and second surface when a lorry passes at 40 km/h.
measurement results. During the rest of the test period
the casing have been pressed back in 0.1 mm.
600 mm below surface, vehicle speed 40 km/h
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.
0,9 20
Average temperatures (C)
0,8
0
0,7
Time (s)
0,6 -20
Under Side
0,5 Fig. 15 Vertical acceleration 600 mm below the road
0,4
-40 surface when a lorry passes at 40 km/h.
mm
0,3 -60
Upper Side
0,2
Side -80
0,1 200 mm below surface, vehicle speed 40 km/h
0 -100
0 100 200 300 400
Velocity (mm/s)
Days since installation
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 Time (s)
process at 200 mm, the same depth as the test pipe,
and 600 mm below the road surface as acceleration Fig. 16 Vibration velocity 200 mm below the road surface
(m/s2) in the ground when a heavy lorry pass over the when lorry passes at 40 km/h.
test area at a speed of 40 km/h.
100
th th
600 mm below surface, vehicle speed 40 km/h

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.
Velocity (mm/s)
0,12
0,10 Side (mm)

0,08 Up (mm)
0,06 Down(mm)
0,04
Time (s)
mm
0,02
Fig. 17 Vibration velocity 600 mm below the road surface 0,00
when lorry passes at 40 km/h. -0,02
-0,04
-0,06
0,00 1,00 2,00 3,00 4,00
time (s)
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
Fig. 18 Maximum amplitude of acceleration as a function forces should be in the test when a heavy lorry passes
of speed.
over the test area (see Fig. 20) it can be established
that the instant forces from passing vehicles is small.
1600
1400
1200
1000
800
Load, N
600
400
Test 1
200
Test 2
0
0 0,5 1 1,5 2
Deformation, mm
Fig. 19 Maximum amplitude of vibration velocity as a
Fig. 21 Diagram radial stiffness of a casaflex pipe.
function of speed
Feed speed 3 mm/min

The acceleration of the ground increases with the 20
speed of the traffic. And the effect is more sensitive the 18

Load case (a)
closer you are to the surface (see Fig. 18). 16
14
The vibration velocity also increase with the speed of

Axial load, kN
12 Load case (c)

the traffic. The effect is not as sensitive as for the 10
acceleration when it comes to the coverage (see

8
Load case (b)
6
Fig. 19). 4
2
In Fig. 20 the diagram describes the instant 0
deformations in the test pipe when the heavy lorry 0 0,5 1 1,5 2
passes over the test area at a speed of 40 km/h. It can Axiell compression %
be seen from the red and the violet curves that the pipe
Fig. 22 Axial stiffness of a casaflex pipe
casing is squeezed together approximately 0.17 mm
from top to bottom at the instant when the lorry passes
101
th th
3.4 Test result of pipe prolonging while The diagram (Fig. 23) below contains two different TDR
pressurizing measurements. It is one graph per wire and test. If
The test results are as follows: there are no changes in the impedance there are no
changes in the profile in the graph. And there are no
Table 1. Test result of pipe prolonging while pressurizing changes in the profiles.
Tests
1 2 3 Average
[mm] [mm] [mm] [mm]
Distance
before 4 360 4 360 4 360 4 360
pressurizing
Distance after 4 365 4 365 4 365 4 365
pressurizing
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%. Fig.23 TDR graph for the supply pipe.
his could be compared to the more common steel pipe
3.6 The results from the test of degree of
for district heating. If that pipe would be loaded with a
compaction of the street
thermal load of 100 C it would prolong itself 0,12%.
The different tests were plotted in diagrams and gave
The force with which the casaflex pipe is expanding the different Youngs modulus Ev1 and Ev2 for different
because of the inner pressure would for 11400 kPa be places. The places were documented in a photo. The
7,3 kN according to the supplier. That would mean that results can be read in the table below.
the diameter would be 81,5mm. In real life the diameter
was measured to be 83,9 mm. The corresponding force Table 2. - Results from test of degree of compaction
for the diameter 83,9 mm would be 7,7 kN. Spot Ev1 Ev2 Ev2/
2
If the pressure would have been 1 600 kPa and the (MN/m )
2 (MN/m ) Ev1
diameter would have been 81,5 mm then the
corresponding force would have been 8,3 kN.
1 38,75 91,44 2,36
This could again be compared to the steel pipe with the
thermal load of 100 C. This pipe would prolong itself 2 18,23 25,25 1,39
with the force of 164,9 kN. 3 61,29 126,4 2,06
So the casaflex pipe expands with a force that is 4 51,53 100,65 1,95
approximately 100*8,3/164,9 = 5,0% of the force from a
5 29,30 70,08 2,39
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 For every test the division Ev2/Ev1 are approved. Spot 2
overcome is negligible. Ev2 is to low but the other four spots are approved so
overall the test is ok.
3.5 Test results from the leak test of the pipe
casing 3.7 Results from the visual control of the surface of
the street
Different TDR graphs have been made in May 2009, in
June 2009 and in April 2010. The strength of the road
Through metering the resistance and making TDR There was no change in the surface of the street due to
graphs it is proven that: the shallow laying of the district heating pipe what so
ever the first eleven months. In spring after an
a) It can be done to connect the two different unordinary cold winter one could see a small crack
systems (The Nordic System and the Hagenuk
System). (approximately 12 cm) in the street along the pipes
extension. As this article is getting written it is not
b) There are no leaks in the test area, neither in investigated why the crack has appeared nor of the
the supply pipe nor in the return pipe.
importance of it. The street has much worse injuries
from frost action damages outside the test area.
102
th th
Heat loss The acceleration and vibration velocity are also

negligible under the traffic load from e.g. a heavy lorry.
The snow did not melt over the pipes in the test area. If
Probably there will be no problems using graphite
one would study other older district heating pipes they
gaskets also with only 200 mm of coverage.
would reveal themselves by melting the snow over
them. This effect never happened in the test area. 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.
Fig. 24 Left; Test area in January, Right; Test area in The demanded levels for the degree of compaction of
February. 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
Fig. 25 Left Test area in March, Right; D. H. chamber a couple of key persons. There had been no test of
revealing itself in February. this kind without their support and permitting. The
However the heat loss is of course bigger compared persons are:
with normal standard because of that the pipes are
Mr. Bo Andersson Planing Manager at
placed closer to the air.
Trafikkontoret Gteborgs Stad,
Frost action damages
Mr. Lars Ljunggren Manager at Gteborg
In theory one could imagine that the street on both Energi AB and
sides of the district heating pipe would erect during the
Mr Gran Johnsson Technical Manager at
winter if there were soil that could frost heave. This
Powerpipe Systems AB.
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. 6. REFERENCES
There was no notable difference in the height of the [1] Molin J., Bergstrm G. and Nilsson S. (1997).
street over the district heating pipes compared to the Kulvertfrlggning med befintliga massor, Swedish
street beside the test area during the winter. District Heating Association FOU 1997:17, (in
Swedish)
4. CONCLUSION [2] Bergstrm G., Nilsson S. and Sllberg S-E. (2001),
The article probably describes the first operating d.h. Tthet hos skarvar vid terfyllning med befintliga
pipes placed in backfill of 0-40 only 200 mm below the massor, Swedish District Heating Association FOU
surface in a street with heavy traffic. As expected, the 2001:58, (in Swedish)
pipes are working nicely. The loads that have been [3] Nilsson S, Sllberg S-E, Bergstrm G, (2006)
measured are acceptable or even low for the d.h. pipe. Grund frlggning av fjrrvrmeledningar, Swedish
As it seems also the street is satisfactory working even District Heating Association FOU, FOU Vrmegles,
though there are d. h. pipes close to the surface. 2006:25, (in Swedish)
The pipe deformations are negligible with respect to the [4] Gudmundson T. F-Processdesign AB, (2002),
pipes function both over time and under an instant Casaflex-rr i Malm 2001,. Swedish District
traffic load. Heating Association, FVF 021241, (in Swedish)
103
th th
[5] Trafikverket, (2005) Allmn teknisk beskrivning fr 9AJyMvYwMDSycXA6MQFxNDPwtTo2Anc_2CbE

vgkonstruktion ATB Vg, dFABCTfUM!/?WCM_GLOBAL_CONTEXT=/wps/
http://www.vv.se/Startsida-foretag/vagar/Tekniska- wcm/connect/goteborg.se/goteborg_se/Foretagare/
dokument/ATB-Allmanna-tekniska- Upphandling_staden%20som%20kund/Specifik%2
beskrivningar/Vagteknik/Aldre-versioner/ATB-Vag- 0upphandlingsinformation/art_N400_FOR_Up_SU
2005/, visited 2010-04-27, (in swedish) _Trafikkontoret, visited 2010-04-27, (in swedish)
[6] Gteborg Stad Trafikkontoret, Bestmmelser fr [7] Trafikverket, (1993) Publikation 1993:19
arbeten inom gatu- och spromrden i Gteborg, Bestmning av brighetsegenskaper med statisk
http://www.goteborg.se/wps/portal/!ut/p/c0/04_SB8 plattbelastning Metodbeskrivning 606:1993, (in
K8xLLM9MSSzPy8xBz9CP0os3gjU- swedish)
104
th th
STUDY ON THE HEAT LOSS REDUCTION METHOD FROM THE SECONDARY

PIPELINES IN THE APARTMENT COMPLEX
1 2
Byung-Sik Park , Yong-Eun Kim
1 1 1 1, 3
Sung-Hwan Park , Yong-Hoon Im , Hyouck-Ju Kim , Dae-Hun Chung Mo Chung
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 individual heating, central heating and district heating.

At the moment there is a lot of potential for district
This study aims to suggest better methods for reducing
heating and cooling. Korea has seen about a 10%
heat losses from the pipelines installed as secondary
supply of DHC among total residential houses which is
heating pipes in the apartment complex in which hot
very low compared to that of European countries which
water is being supplied for space heating and hot water
supplies over 50% DHC.
by a district energy supply company. Right now the
district heat supplier is responsible only for the primary If we are to increase green growth with low carbon, it is
district heating pipelines just before the substations in crucial to supply DHC, which has higher energy
the apartment complex. That is why the heat loss efficiency than any other method, in dense regions of
reduction becomes more important in the secondary population. The recent Korean government has shown
pipelines after the substation in the Korean apartment effort in making a point of energy efficiency throughout
complex. main energy consuming sectors including building
area. However, the supply policy of DHC is now being
Several methods to reduce the heat loss from the
crippled due to various reasons. Makers or consumers
secondary pipelines were set up and compared by a
of individual heating devices do not have positive
simulation technique. One of the methods is to
attitudes toward DHC. Therefore it is important to draw
combine the hot water heating pipes and space heating
attention to the multitude of benefits and merits of
pipes. Another method is to install a small heat
DHC.
exchanger in each house to supply hot water from the
single space heating pipeline. In this case we can There is certainly some heat loss from the pipelines
easily change the means of heat supply and the right installed under the ground to supply the district energy
choice of end users can be ensured for the means of from the power plant to the consumers. To reduce the
heat supply. heat loss from these primary pipelines many
innovations and advancements have been made for a
In this study the preferable method to reduce the heat
long period since the district energy was supplied in the
loss in the secondary pipelines has been suggested.
northern European countries. The heat loss generally
The simulation result has shown about 30% heat loss
differs according to the network type of pipelines. The
reduction compared to the existing scheme for the
more compact that the network is, the less heat loss
simple change of methods and much more reduction
occurs. But the type of network cannot be made
for the optimization of pipe diameter and insulation
arbitrarily by the designer. The designer can simply
thickness or surface enhancement by low emissivity.
optimize the network in view of geological and
environmental conditions, such as population density,
INTRODUCTION not the type of network. Although heat loss exists with
Korea is characterized as having four distinct seasons. the primary pipelines, it can be controlled and
Apartment complexes became a typical type of maintained effectively by the district heat supplier. On
residence in urban areas after the recent rapid the other hand, the heat loss from the secondary
industrialization of last 30 years. At the moment over pipelines cannot be controlled properly by the building
half of the population chooses to live in apartments owners who are responsible for.
rather than in individual houses and the trend will The apartment complex is a unique housing system in
increasingly continue in the future. There are three Korea. It contains many high rise buildings of over 10,
typical heating methods for apartment complexes - often over 20, stories high. In many cases it has over
105
th th
one thousand homes. However, the basic structure is to be measured and evaluated on the heat loss from
almost the same as that of western apartment buildings the secondary pipelines. Many thermocouples and
except for the pipeline network between buildings and flow-meters were installed in the region of the pipelines
substation. In the past there was one substation in one to collect information on heat demand pattern,
apartment complex. The substation has a minimum of temperatures and heat loss from certain regions to be
two heat exchangers which are in general shell and evaluated.
tube type or plate type. Nowadays the number of
201 207 208
substations grows bigger and bigger. That means that
the designer plans to install the heat exchanger 202 205 206
separately and respectively according to the buildings
203 204
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
HWR
HWR
HWR
HWR
HWR
HWR
HWR
HWR
HWR
HWR
HWR
HWR
HWS
HWS
HWS
distinguish between positive heat gain and heat loss Supplementary
Water
from the pipelines installed within the buildings. If the

pipelines are installed in the center of the building, the Calorimeter
heat loss from the heating pipes or hot water supply

City water
M DHWR
DHWS
pipes can be regarded as positive heat to the

Fig. 2 Secondary pipeline network from the substation
consumer. But if the pipelines are installed near the
building surface, the heat from the pipes can be
regarded as loss. ANAYSIS OF HEAT CONSUMPTION PATTERN
Measurement and analysis of the heat loss from
1) Space heating water flowrate
several apartment complexes in Korea has been tried.
The heat loss data from the several sites has been Space heating amount is being measured daily. Toom
stored and accumulated throughout the year. A temperature does not differ much between the homes
simulation method has been set up and the accuracy of in the apartment complex. Thus the temperature
the simulation has been investigated. Some difference (T) between inlet and outlet of the pipeline
alternatives to reduce the heat loss have been of individual homes remains fairly constant except in
prepared from the existing scheme. The simulation the summer season. Therefore the heating water flow
method and results have been presented in this paper. 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
TYPE OF APARTMENT COMPLEX
example of this is shown in Fig. 3.
The apartment complex was built and opened in
November 2007. It has 8 buildings which are
comprised of 518 homes. Each home has 112 m2 of
heating area. Fig. 1 shows the location and overall
shape of the apartment complex which was chosen
Fig. 3 Temperature difference between supply and return

North
KOREA
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
South
KOERA
JAPAN
CHINA
homes for the year 2009 was extracted from the
comparison of the total measured amount and the
individual house measurement. Fig. 4 represents the
Fig. 1 The location and the shape of the apartment
annual heating water flow rate. Some differences exist
complex
during the cold winter season.
106
th th
12th floor
11th floor
10th floor
9th floor
T high
8th floor
T out 7th floor
6th floor
5th floor
4th floor T low

3rd floor
Primary pipe line Secondary pipe line 2nd floor
1st floor
DHWS
T base T sb
DHWR
Fig. 4 Heating water flow rate
Fig. 6 Category of temperature charicterization.
2) Hot water flow rate
Hot water consumption is measured by ton from the Tout: Temperature outside the building
general water flow meter. Hot water supply line is
Tbase: Temperature of the underground space
designed to have a supplementary recirculation line in
frequently open to the outside surrounding.
order to supply instant hot water. By adding the water
which was used by individual homes to the heat Tsb: Temperature of the underground space closed
exchanger, the flow rate can be constantly maintained. to the outside surrounding.
Tlow: Temperature of low-rise region in the building
Thigh: Temperature of high-rise region in the building
Fig. 5 Hot water flow rate
3) Various temperatures outside pipes

The temperatures outside the pipeline were
categorized into four cases taking into consideration Fig. 7 Varous temperature of the surroundings of
the atmosphere outside the pipeline. The first one is pipelines.
the underground space which is fairly open to the
outside of the building. The second one is the
SIMULATION METHOD
underground space which is not so open to the outside
of the building. The other two are the spaces of low and For the heat loss simulation commercial tool,
high regions of the building which is not open to the Flowmaster of 1 D system analysis has been used.
outside of the building. These temperatures, measured Flowmaster is a program which can analyze the
according to the categories, were applied in the thermo-hydrodynamic characteristics of pipe systems if
simulation in view of the pipelines outside the following items are given such as the physical
characteristics. 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
107
th th
installed underground between buildings and VARIOUS SCHEMES FOR SIMULATION

substation as well as in the buildings themselves in the
apartment complex. The simulation was performed 1) Present Scheme
daily in view of calculation time. 12th floor
11th floor
HWS Heating water supply

10th floor
(1)
9th floor
HWR Heating water return
8th floor
HWR Hot water supply
HWR
7th floor
HWR
HWS
HWR
HWR Hot water return
(2)
6th floor
5th floor
4th floor
(3) 3rd floor
Primary pipe line Secondary pipe line 2nd floor
1st floor
HWS
DHWS
HWR
DHWR HWR
HWR
Fig. 8 Pipeline configuration of present scheme

fluid temperature, C
ambient temperature, C The present scheme is composed of 4 pipelines, two of

which are for heating water supply and return and the
convection heat transfer coefficient, other two of which are for hot water supply and return.
radiation heat transfer coefficient, Heating water is supplied to each home and returned
from each home and resultantly the flowrate of supply
internal heat transfer coefficient, and return are equal. On the other hand hot water has
a certain amount of recirculation in order to keep
insulation thermal conductivity,
supply water hot. Hence the same amount of
pipe thermal conductivity, consumed hot water in each individual home should be
supplied to the hot water heat exchanger to guarantee
external pipe diameter, instant hot water supply.
external insulation diameter,
2) Alternative A
internal pipe diameter,
Alternative A is a scheme which removes the hot water
pipelines and combines with the heating water
For accurate simulation, individual flow rate was used pipelines. Thus there are only two pipelines of supply
respectively and differently based on total measured and return from substation to each building. These
flow rate and read amount of individual flow meter of supply and return pipelines have two functions of
518 homes throughout the year. By doing this, the flow heating and hot water supply and return. By reducing
rate in the individual pipelines can be determined from 4 to 2 pipelines the heat transfer surface area can
according to the usage amount of heating water and be decreased. However, this scheme has
hot water. This similarly leads to actual flow rate in the disadvantages in summer when the heating water
pipelines. From this complicated process, the number supply has been closed. If the pipelines should be used
of individual flow rate of hot water and heating water for the supply of hot water, the resultant water speed in
comes to 378,140. Macro which was combined with the pipelines would be very small.
Excel and Flowmaster was used for a 365 day 12th floor
analysis. This process requires 32 hours for 8 buildings 11th floor
for only one case. HWS Heating water supply 10th floor
HWR Heating water return 9th floor
HWR Hot water supply 8th floor

HWR
HWS
HW
R
7th floor
6th floor
5th floor
4th floor
3rd floor
2nd floor
Primary pipe line Secondary pipe line
1st floor
DHWS
P-1
HWS P-9
P-4
P-6
E-2
E-1 P-3
DHWR
E-3
HWR
P-2
Fig. 9 Pipeline configuration of Alternative A

108
th th
For hot water, this scheme is effective in heat 5) Use of PEX PIPES
exchange because the heat exchanger acts for the PEX is being used in western countries as district
consumed hot water only and can avoid extra heating pipes for low temperature service from
recirculation pipelines. But it has the drawback of renewable resource application. For this reason PEX
supplying cold water or non heated water when can be used as secondary pipelines which normally are
intermittently using hot water. under service of low temperature.
3) Alternative B
VALIDATION OF SIMULATION RESULT
12th floor
E-10
To validate the simulation result, the measurement
value was compared with the prediction result by the
11th floor
HWS Heating water supply 10th floor

E-11
E-15
present simulation method. The measured value of
HWR Heating water return
9th floor
8th floor
E-14
December 2009 was used for heat loss reference
value. As seen in Table.1 the simulation result fairly
E-12
HWR
7th floor
HWS
E-13
agreed with the measured data. And the simulation

6th floor
E-7
5th floor
4th floor
E-8 method can be used without much modification. For
3rd floor
E-9
more accurate prediction it needs slightly more
supplementation in the numerical modelling of heat
E-6
2nd floor
Primary pipe line Secondary pipe line E-5
DHWS
P-1
HWS
1st floor
E-4
transfer phenomena of outer pipe surface and
environment.
E-1
DHWR HWR
P-4
Fig. 10 Pipeline configuration of Alternative B Table1 Comparison of measured and predicted

values[unit: MWh]
Alternative B is the same as Alternative A in the point Measurement heat Hot water
of unifying the heating water and hot water pipelines. Heat loss rate loss 20.2 11.2
But it is different in the point of the individual hot water
heat exchanger being installed in each home among 9.44% supply 264.7 67.8
the heating water pipelines. This scheme is said to Simulation heat Hot water
resemble the pipelines installed in the apartments of loss 17.0 14.29
Heat loss rate
European countries.
9.90% supply 252.6 63.53
SIMULATION RESULT
4) Alternative C
Fig. 12 shows typical heat supply and heat loss for the
24 hours of 11.11.2009. The simulation result of each
12th floor
11th floor
HWS Heating water supply 10th floor

scheme is shown from Fig. 13 to Fig. 17. Each graph
HWR Heating water return 9th floor shows similar patterns and the heat loss comparison of
HWR Hot water supply 8th floor
each scheme is summarized in Table 3.
HWR
HWS
HW
HW
R
7th floor
6th floor
5th floor
4th floor
3rd floor
2nd floor
Primary pipe line Secondary pipe line
P-9 1st floor
DHWS HWS
P-7
P-14
E-5 P-8
DHWR HWR
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. Fig. 12 Hourly heat supply and heat loss of 11.11.2009
109
th th
Fig. 13 Annual heat supply and heat loss (Present

scheme)
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%

Fig. 14 Annual heat supply and heat loss (Alternative A)
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
Fig. 15 Annual heat supply and heat loss (Alternative B)
Table 3 shows a comparison among parameters which
affect heat loss. There is frequent excessive design for
Alternative B should supply heating water in the the nominal diameter of pipelines which are installed in
summer season when it is not required for the supply of both the underground and the buildings. In this
heating water in order to supply hot water to the comparison the increases of insulation thickness by
individual home. From the comparison of Fig. 14 and 10 mm and 20 mm were considered. Also the decrease
15, Alternative B is more efficient than Alternative A in of nominal diameter by 1 level was considered. It was
the cold region. 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.
Fig. 16 Annual heat supply and heat loss (Alternative C)
110
th th
Table 3 Heat loss comparison due to various More enhancements in heat loss can be extracted from
parameters heat loss reduction by the selection of optimum pipe
diameter, good insulation material, increasing
Heat loss comparison insulation thickness and changing surface emissivity of
insulation material.
Present pipe of insulation thickness 40 mm 100%
Downsizing nominal diameter by 1 level 88.7% ACKNOWLEDGEMENT

The financial support from KDHC made this work
Insulation thickness 50 mm 86.1% possible. This paper is based on the results of an
ongoing research project which will be completed at the
Insulation thickness 60 mm 75.5%
end of 2010.
Closed-Cell Elastomeric thermal insulation 89.3%
REFERENCES
Pipe diameter downsizing + insulation
76.7%
thickness 50 mm [1] W. F. STOEKER, DESIGN OF THERMAL
Pipe diameter downsizing + insulation SYSTEMS 3rd edition, Mc Graw Hill, pp. 53-160
67.6%
thickness 60 mm
[2] Incropera, HEAT TRANSFER 5th,WILEY
[3] Byung-sik Park et al, Study on the Reduction
CONCLUSION method of Heat Loss from the Secondary Pipelines
installed in the Apartment Complex, 2008
Present scheme for the secondary pipelines is
evaluated to have 14% annual heat loss based on the [4] Byung-sik Park et al, Study on the Reduction
total heat supply to the apartment complex. This is a method of Heat Loss from the Secondary Pipelines
very large amount when we consider that the primary installed in the Apartment Complex, 2009
district heating pipeline has only about 4 to 5% annual
[5] Flowmaster, Flowmaster heat transfer manual
heat loss in dense population urban areas.
Heat loss by Alternative A can be reduced about 30% [6] Manfred Kl psch, Plastic pipe system for DH,
compared to that of the present scheme which has Handbook for safe and economic application, IEA
been widely adopted in Korea until now. However, R&D Programme on District Heating & Cooling
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.
111
th th
HEAT LOSS OF FLEXIBLE PLASTIC PIPE SYSTEMS,

ANALYSIS AND OPTIMIZATION
1 2
EJ.H.M. van der Ven , R.J. van Arendonk
1
Thermaflex International Holding B.V
2
Liandon B.V.
ABSTRACT for a small-sized pipe to about eight hours for the

largest sized diameter.
A newly developed, in-house, test rig for measuring
heat loss of pipe systems allows the user to analyse The novelty of the test rig is described in the paper
various systems in a short timeframe. This allows quick Verification of heat loss measurements conducted on
insights into heat loss variables and mathematical (semi) flexible pipe systems [3].
analyses. The effect of alternative compositions of
The novelty for product improvement is that due to the
insulation and other layers can be evaluated within a
reduced time required for a test run the effect of
short time span.
alternative systems can be mathematically analysed
This already led to improvements of the production and evaluated in a short time. In this way the analyses
process and of the insulating foam. of alternative production methods has a short
feedback. Optimization of the product can be effected
INTRODUCTION in a short time.
Liandon developed a heat loss testing rig for In the near future the test rig will be used for quality
Thermaflex to test their produced flexible pipe systems. control of the production process. This test will partly
Within a short time the pipe system, undergoing a heat replace other currently applied standard tests, such as
loss test, tends towards the controlled temperature in density and cell size measurements.
the sections of the sample, the added power reaches
equilibrium and the test results can be collected. METHOD DESCRIPTION
Due to the short time required for testing, the results of
In the Flexalen 600 pre-insulated pipe a PB medium
alternative production methods are easily available.
pipe is encapsulated in insulating foam, which is
Due to the short response time the test is a great help
protected against wear and tear in a corrugated hard
in the search for product and production improvements.
cover pipe. The pipe product has a solid bonding
The objective of this paper is to present the results of between the insulation and cover and no bonding
the research to the overall heat loss performance of a between the insulation and the medium pipe.
flexible plastic pipe product, Flexalen 600.
According to EN 15632, the European Standard for
The objective of the research is: pre-insulated flexible pipe systems, this pipe system is
classified part 3: Non bonded system with plastic
1 Find correlations between heat loss and other
service pipes. The Flexalen 600 plastic pipe system
parameters of the pipe system such as outer
differs in some areas significantly from most other
diameter, inner diameter, foam surface and foam
systems in this class:
structure. These correlations are determined by the
mathematical analysis of practical heat loss 1 Physical bonding between foam and outer casing,
measurements.
2 One layer of foam, filling the complete space
2 Find possibilities for the improvement of the pipe between service pipe and cover,
parameters by analysing the heat loss correlations.
3 Next to other connection methods the service pipes
can also be connected by welding.
NOVELTY AND MAIN CONTRIBUTION
Annex D of part 1: Classification, general requirements
The actual heat loss of pre-insulated pipe products is and test methods give rules for calculation of the heat
determined under controlled, similar conditions for an flow to ambient (heat loss) from measured values,
entire diameter range. This range comprises various making the heat flow of various parameters
outer diameters, various inner diameters and various comparable.
compositions in materials and pipe systems. The time
required for one single test run varies from half an hour The heat loss calculations of annex D are based on the
thesis of Wallentn as published in Steady-state heat
112
th th
loss from insulated pipes; Lund Institute of Technology, systems. Test spools can be extracted directly from
Sweden [2]. production. In this way tests can be executed with fresh
product.
The described heat loss calculation is valid for the pre-
insulated pipes. Branches and connections are Also testing of cured piping and piping that is aged and
excluded. 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
Figure 1, Thermaflex heat loss equipment on (semi) flexible pipe systems (van der Ven et al).
In the new developed test rig (figure 1) a test spool

(figure 2) is put in a slim fitting sleeve. MANUFACTURING PROCESS
The test spool is heated internally in three sections. The Flexalen 600 product has been developed by
The middle section of the spool is the test section. Thermaflex, located in The Netherlands. The
Heating in this section is controlled to obtain the development started in 2002 and resulted in a first
required test parameters. The two ends are heated to small-scale commercial production in 2005. During the
compensate for the heat loss from the ends of the production of Flexalen 600 four different production
middle test section. In this way an endless pipe is techniques are combined, partly simultaneous and
imitated. partly sequential:
The outer side of the sleeve is water-cooled to obtain 1 Production of PB service pipes optionally covered
heat transport from the test spool. with an EVOH oxygen barrier layer.
During the start of the test, heat is lost into the heating 2 Production of LDPE insulation foam to fill the area
of the pipe system and into the surrounding cooling between medium pipe and outer casing.
water. When heat losses have reached equilibrium, the 3 Production of outer casing of HDPE.
steady state heat transfer can be measured.
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
Figure 2, Longitudinal section guarded end heating probe
Netherlands. Unique for the process is the ability to
The time span required for testing in the test rig is produce continuous lengths. For practical reasons the
rather short. The time to reach equilibrium lies in the lengths produced depend on the outer casing of the
order of hours, depending on the diameter and product and the size of the reel. The maximum length
insulation thickness. Comparable tests often require produced can reach up to 2000 meters.
time spans in the order of days. The complete Flexalen 600 pipe system includes pre-
Containing various diameters, which are based on the insulated pipes, couplings, sleeves, pre-insulated T-
standard production outer diameters, the test rig connections, etc. The production range is described by
enables heat loss tests for various diameters of piping Engel and Baars. [5]
113
th th
Production of PB service pipes extruder. This foaming agent is a hydrocarbon that

All service pipes are made of Poly-butene. This is a causes the expansion of the LDPE.
plastic with a special combination of properties. Poly- The quality of the foam, moreover the insulation
butene has excellent heat and creep resistance, properties of the foam, depends on parameters such as
flexibility and strength at a long lifespan and is fully density, cell size and chemical composition. All
recyclable. In accordance with the temperature parameters are measured and adjusted within limited
duration profile mentioned in the BRL5609/EN15632, tolerances to meet the specifications. The best
PB is suitable up to a maximum temperature of 95 C. efficiency is further improved when the space between
All PB-pipes are weldable by socket fusion, electro the service pipe and the foam is filled better.
fusion and butt welding, which allows for an all plastic As there is no bonding between the service pipe and
distribution system without metal parts that are prone to foam, there is no risk of damaging the foam and
corrosion. The service pipes are produced via state-of- properties by expansion of the pipes due to thermal
the art extruders. fluctuations in the applications.
The production line consists roughly of an extruder,
calibration tools for adjusting pipe size, cooling baths, Production of corrugated outer casing (HDPE) and
and marking and cutting equipment. assembly of complete product
The outer casing is applied by an extrusion process
Pipe dimensions are checked inline every second
and thermally welded to the foam after the pipe has
during and after production with ultrasonic
been inserted. The outer casing is also corrugated to
measurements. PB-pipes can be produced up to an
optimize the flexibility of the finished product. Now the
outer diameter of 225 mm.
product is ready for coiling.
If desired an outer oxygen barrier layer may be applied
After production the final product must cure for 5 days.
via co-extrusion up to an outer PB pipe diameter of 90
During this curing period the degassing of the foaming
mm.
agent starts, while the insulation foam is still stabilizing
After production the PB-pipes are stored for a minimum
period of 5 days and cured to create the correct Production testing and controlling
crystalline polymer structure. After curing the pipe is During the production, parameters are checked and
used for the Flexalen 600 production. controlled, such as:
Every batch produced is verified by the in-house QC 1 Chemical composition of the foam.
department according to Dutch directives
2 Settings of all extruders involved (foam extruder and
- BRL-K5609, for PB pipes with oxygen barrier or extruder for corrugated outer casing).
- BRL-K17401 for PB pipes without oxygen barrier.
Product and manufacturing processes are checked 6-8 3 Density and cell size of the foam.
times a year and certified by independent agencies 4 Dimensions of the foam (outer diameter and inner
such as Bureau Veritas, KIWA and CSTB and le diameter).
Centre Scientifique et Technique du Btiment. The
quality of the QC department is validated by these 5 Line-speed of all involved products (Foam / PB-pipe
checks and by internal and external audits. / End-product).
6 Thickness of the corrugated outer casing and the
Production of LDPE insulation foam connection of the corrugated outer casing to the
Thermaflex has now approximately 35 years of foam.
experience in the production of LDPE foam via
7 Since a few months: in-line production control of
extrusion techniques.
Heat Loss of the pre-insulated pipe system.
Most of the raw materials that are used are tailor-made
The most prominent factors to influence heat loss are
mixtures according to Thermaflex specifications.
parameters one to four. The foam lambda is directly
Through these specifications and proper production
affected by these factors as represented in equation 1.
quality control the companys philosophy related to
core business is also realized for raw materials.
total convection conduction (1)
During the heating, melting and mixing of the raw
materials the foaming agent is injected into the radiation blowingagent
114
th th
RESULTS AND CORRELATIONS DISCUSSION OF PARAMETERS

It is again the companys philosophy that paves the Table 1 summarises the results of measurements and
way for innovations. One of these innovations is the calculations of tests on Flexalen 600 pipes directly from
development of an in-house device for testing pipe production. Various diameters are tested and
systems, directly from production, cured and degassed. calculated according to EN 15632 for a surface
This proved to be a suitable device for production temperature of 10 C and a common medium
control, but also for gaining more insights into product temperature of 70 C (Instead of the maximum medium
parameters. temperature of 95 C).
The objective of testing is to improve the knowledge of The table indicates the relationships between product,
the produced pipe systems in order to optimize: cross sectional area of the foam, foam density, cell size
of the foam, remaining foaming agent, calculated
Production methods:
thermal conductivity and the calculated heat loss of a
Machine data can be adjusted based on test results.
buried piping system.
Test results may lead to new production methods with
new equipment. The products 50A25, 63A32, 75A40 and 90A50 are
newly developed. These products are not necessarily
Chemical and physical composition of layer material:
District Heating products. However, they are produced
Knowledge of various composition materials may lead
using the same process and have their application in
to an improvement of insulating values.
the connection between the district heating network
Cell structure and gap between insulating foam and and the building or house. It is also applicable in case
medium pipe: of low temperature differences, cooling or in-house
Cell size influences values, test result based heating or cooling.
improvements are possible. The gap is a bad insulator.
The quest for a minimal gap started with testing. Table 1: Test results of fresh, uncured piping systems
Improvement and minimization of this gap was an
achieved challenge in the testing period. Foam
50,calc Heat Loss*
Product section density cell size agent
Thermaflex is a lean and mean organisation that mm kg/m (mm) % mW/m.K W/m
responds quickly to new insights. Therefore new 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
insights were applied even before the complete range
75 A 40 3.044 38,0 0,40 46 44 17,8
of production testing was performed. 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
For the company, improvements of product and
125 A 63 9.155 39,0 0,88 70 56 22,0
production have the highest priority. Although the 160 A 75 15.688 40,3 1,20 81 54 21,0
production range is wide, the insight into specific and 160 A 90 13.745 35,0 1,30 85 61 25,2
general parameters increased considerably. 200 A 110 21.913 45,0 1,60 81 68 27,4
*) calculated heat loss of buried system at temperature difference of 60 K
The research provides the prominent variables to
improve insulation performance. Practical heat loss In Graph 1 foam density and cell size are related to the
determination, in combination with analytical studies, cross sectional surface.
results in a clear understanding of heat loss behaviour In Table 1 the foam density varies from about 35 kg/m
in single and twin flexible pipe systems during their to about 50 kg/m. Graph 1 shows hardly any
entire lifetime. relationship with the surface of the cross section.
As a result of the tests the manufacturing process is Table 1 shows that cell size varies from 0.47 to 1.60
improved in two steps. mm. Graph 1 shows that cell size is directly related to
The emphasis of the first step was to diminish the cell the cross sectional surface, however less than 1 to 1.
size of the foam. This succeeded in a decrease of cell This relationship is influenced by physical production
size by some 20%. parameters.
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.
115
th th
60 1,8 Graph 3 shows the relationship between outer pipe

system diameter size and the percentage of foaming
50 1,5 agent directly from production. With increasing
diameter the foaming agent increases, possibly to an
Average cell size [mm]
Density [kg/m]
40 1,2
asymptotic value.
30 0,9 An additional interesting factor is the ongoing process
in the foam during and after production. As described
20 0,6
before the final step in the production is a 5-day curing
10 0,3
stage.
During the curing stage the foam expands and part of
0 0,0
the foaming agent releases from the foam. As the foam
- 5.000 10.000 15.000 20.000 25.000
Cross section of foam [mm]
is locked by a hard outer shell, expansion is directed
inwards. By this the gap between the foam and the PB
Graph 1: Density and cell size in relation to foam cross
medium pipe, typical for our production method, is
section
decreased.
Table 2 shows the effect of curing and degassing on
40,0 80
both the contents of foaming agent and the calculated
Heat loss of buried piping [W/m]
heat loss.
35,0 70
Conductivity [mW/m.K]
Even when forced, degassing takes time. The number

30,0 60 of degassed samples manufactured in the same way
as the fresh samples is therefore limited. Table 2 is
25,0 50 short due to a lack of adequate and comparable
samples.
20,0 40
Table 2: The effect of time on curing and degassing
15,0 30 Fresh 6 days curing Degassed
50 100 150 200 Heat Heat Heat
Product Agent Agent Agent
Outer diameter pipesystem [mm] Loss Loss Loss
% W/m % W/m % % W/m %
Graph 2: 50,calculated and calculated heat loss of buried 63 A 32 52 15 23 17,5 15 0 17,1 12
pipe in relation to outer pipe size 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
Graph 2 shows the influence of outer pipe size to

calculated conductivity 50 and heat loss of buried pipe 20
systems.
It also shows that part of the increase of the heat loss 19
with the diameter is caused by increase of conductivity.
18
90
17
Percentage foaming agent
80
16
70
15
60 0 10 20 30 40 50 60
Percentage of foaming agent
50
Graph 4: Relationship between heat loss and foaming
agent
40
50 100 150 200 Graph 4 shows that there is a tendency of decreasing
Outer diameter pipesystem [mm] heat loss with increasing foaming agent. This tendency
has seems weak. The spread is large over the entire
Graph 3: Foaming agent content in relation to outer
graph.
diameter size
116
th th
RESULT BASED FUTURE DEVELOPMENTS 31.0
The tests have led to improvements to the flexible 29.0

Heat loss

plastic pipe products produced by Thermaflex. In the 27.0 Heat loss Degassed
near future we expect improvements in: 25.0
Production methods: 23.0

Starting June 2010 a change in the machine 21.0
configuration will be implemented as an extra step. The
19.0
new configuration improves temperature control in the
extruder, which leads to better cell structure. 17.0
15.0
Chemical and physical composition of layer material: 50 70 90 110 130 150 170 190
Up to a certain degree the anti-radiation agent Outer diameter pipe system [mm]
improves isolating values. With trial and error the anti-
radiation agent content is increased. Up till now the Graph 5: Degassed heat loss values of a buried system at
maximum content has been limited by production a temperature difference of 60K
methods. Research is required to investigate maximum
desired value for insulating effects. Examples of product improvement
Cell structure and gap between foam and medium pipe: Further research has led to product improvements.
In the tests we see variations in cell structure and gap Based on these improvement proposals Thermaflex
width. Future research will aim at acquiring more has been able to produce new pipe system samples.
detailed knowledge of these phenomena. As represented in graph 5 the new samples have a
heat loss decrease up to 16 percent compared to the
Up till now the heat loss performance on single pipes previous results.
has been measured and analysed. This has resulted in
an understanding of the heat loss principles in district 40.0
heating systems. Twin pipe systems will soon be
tested, analysed and evaluated. 35.0 Heat loss Heat loss New
In this paper only the heat loss of the Flexalen 600 pre-
insulated pipe product has been handled. Information 30.0
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 25.0
economical connection solutions (Engel).
20.0
CONCLUSION
The results of testing are reliable. Knowledge of the 15.0
product and production has led to promising 50 70 90 110 130 150 170 190
improvements of both. Outer diameter pipe system [mm]
Further research will certainly lead to further Graph 6: New heat loss values of a buried system at a
developments. temperature difference of 60K
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.
117
th th
ACKNOWLEDGEMENT REFERENCES
We would like to acknowledge P. van Rijswijk for his [1] EN 15632 District heating pipes, Pre-insulated
dedication to all the heat loss measurements flexible pipe systems, Requirements and test
performed during this research. methods
[2] P. Wallentn; Lund Institute of Technology,
FURTHER INFORMATION
Sweden; 1991, Steady-state heat loss from
Questions concerning the paper may be addressed to: insulated pipes
Thermaflex International Holding B.V, [3] E. van der Ven, F. Duursma, H. Korsman, I. Smits;
Veerweg 1 Paper on DHC, Tallinn; 2010, Verification of heat
5145NS Waalwijk loss measurements conducted on (semi) flexible
The Netherlands pipe systems
www.thermaflex.com
[4] H. Korsman; Paper on DHC, Tallinn; 2010, Heat
Liandon B.V. loss optimization of flexible plastic piping systems,
Dijkgraaf 4 life time heat loss performance
6920AB Duiven
[5] C. Engel and G. Baars, New economical
The Netherlands
connection solution for flexible piping systems,
www.liandon.com
12th ISDHC 2010.
118
th th
COMPARISON OF COMPETITIVE (SEMI) FLEXIBLE PIPING SYSTEMS BY MEANS

OF HEAT LOSS MEASUREMENT
1 1 1 2
I.M. Smits , J. Korsman , J.T. van Wijnkoop and E.J.H.M. van der Ven
1
Liandon B.V.
2
Thermaflex International Holding B.V.
ABSTRACT Notice that this study does not compare entire district
heating systems. For system comparisons see Heat
Different types of pre-insulated pipes are tested on
loss analysis and optimization of a flexible piping
their heat loss values. Three flexible pipes and a rigid
system by J. Korsman et al. [2].
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 NOVELTY AND MAIN CONTRIBUTION
heat loss values are measured according to EN 15632 Where most studies only focus on one product this
and published as declared values. The declared study compares different types of flexible pre-insulated
values are calculated according to EN 15632 Annex pipes on their practical heat loss values and gives an
D1-D3. explanation of the practical heat loss values. It also
The flexible pre-insulated systems, with PE and PE-X compares flexible pipes with rigid pre-insulated pipes
foams, show a variance of up to 5 W/m in the heat loss on an equal basis.
values. These absolute differences in the system are
caused by the outer casing dimensions of the pre- BRIEF METHOD DESCRIPTION
insulated pipes. Recalculation to the same outer casing
First a brief description of different types of flexible pre-
diameters shows a slight advantage for the PE system
insulated pipes and a rigid pre-insulated pipe is given.
in service pipes of 32 and 63 millimetres
This chapter highlights the differences and similarities.
The flexible piping system with the PUR insulation The different types of foam for plastic pre-insulated
foam on the other hand performs better compared to pipes are described in a separate paragraph.
equally dimensioned flexible PE and PE-X insulation
Secondly, the method of testing is briefly addressed.
foams.
Thirdly, the different types of flexible pre-insulated
Flexible pre-insulated pipes have a higher heat loss
pipes are tested on their absolute heat loss just after
compared to rigid pre-insulated pipes. Recalculation to
production.
the same transport capacity [kg/s] and the same outer
casing diameter also shows that rigid pre-insulated Since heat loss of pre-insulated pipes can increase
pipes perform better. However the fact that smaller over time due to degassing of the insulation foam, a
diameters show a smaller heat loss difference between gas analysis is performed on all test samples.
rigid and flexible pre-insulated pipes is interesting.
In the second paragraph the absolute heat loss values
of the different types of flexible pre-insulated pipes are
INTRODUCTION & OBJECTIVE compared on the basis of service pipe dimensions. The
The objective of this research is to compare different third paragraph defines a comparison on the basis of
types of competitive (flexible and rigid) pre-insulated insulation surface and service pipe dimension.
pipes on their differences in heat loss values. The In the fourth paragraph the comparison of flexible
comparison is based on an overall heat loss pre-insulated pipes versus a rigid pre-insulated pipe is
measurement under similar conditions. Overall heat described. The comparison in the third and fourth
loss is determined for different samples of pre-insulated paragraph is based on declared values. The defined
pipes, by using newly developed heat loss testing conditions are: (1) thermal conductivity of soil: 1.0
equipment as described in Verification of heat loss W/m.K, (2) thermal transmittance factor of earth to
measurements by J.T. van Wijnkoop et al. [1]. The ambient air: 0.0685 m2.k/W and (3) soil covering:
heat loss data of these flexible pipes will be compared 0.8 m.
with practical measurement on a rigid pre-insulated
pipe. 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
119
th th
pre-insulated pipes. The third paragraph founds the barrier (EVOH) is placed in Polyurethane (PUR)
second comparison by adding a heat loss value based insulation foam with a corrugated outer casing of high
on insulation surface. density poly-ethylene (HDPE). This product shows a
tight connection between the service pipe, the foam
Finally, both flexible and rigid pre-insulated pipes are
and the outer casing. This product therefore cannot be
compared, resulting in conclusions concerning flexibility
re-used once it is formed.
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;
Fig. 2. Section view of PEX/PUR pipe
Two different types of Cross linked Polyethylene
(PEX) service pipe with Cross linked PB service pipe with PE insulation
Polyethylene (PEX) insulation;
The third type of pre-insulated pipe is a flexible PB/PE
One type of Cross linked Polyethylene (PEX) pipe. Figure 3 shows the cross section view of the
service pipe with Polyurethane (PUR) insulation; PB/PE/PE pre-insulated pipe. A Polybutene (PB)
One type of Polybutene (PB) service pipe with service pipe with anti-oxygen barrier (EVOH) is placed
Polyethylene (PE) insulation. in a low-density poly-ethylene (LDPE) insulation foam
with a corrugated outer casing of high density Poly-
The rigid pre-insulated pipe system is;
ethylene (HDPE).
One type of Steel (ST) service pipe with
Polyurethane (PUR) insulation. The PB service pipe makes it possible to use electro
fusion welding with a PB coupling. This makes a strong
Firstly all types of pre-insulated pipes are functionally bond. Corrosion is not an issue, because PB is inert
explained. Secondly the different kinds of foam with water.
production methods are described. All types of pre-
insulated pipes described are commonly available There is no connection between the PB and foam plus
products used for district heating purposes in Europe. outer casing. Therefore it is possible to re-use both
elements in own production. The complete product can
PEX service pipe with PEX insulation be re-used.
A short description of the PEX/PEX systems is given. Information regarding the use of this product is given in
Figure 1 shows the cross section view of the New economical connection solutions for flexible
PEX/PEX/PE pre-insulated pipe. A cross linked Poly- piping systems (Engel) [7].
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. 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
Fig.1. Section view of the PEX/PEX pipe view of the ST/PUR/PE pre-insulated pipe. A steel (St)
service pipe is placed in Polyurethane (PUR) insulation
PEX service pipe with PUR insulation
foam with a smooth outer casing of high density Poly-
This paragraph describes the flexible PEX/PUR pre- ethylene (HDPE). Because of the steel service pipe,
insulated pipe. Figure 2 shows the cross section view the complete system has to be mechanically welded.
of the PEX/PUR/PE pre-insulated pipe. A cross linked Also, because of the combination of steel and water,
poly-ethylene (PE-Xa) service pipe with anti-oxygen there is a potential risk of corrosion.
120
th th
This system is called a rigid system with a tight extruder. Once the material flows out of the extruder,
connection between the service pipe, the foam and the pressure drop causes the expansion of the
outer casing. Once it is formed, this product cannot be hydrocarbon. The aggregation state of the molecule
re-used in own process. 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].
Fig. 4. Section view of Steel/PUR pipe
Foam production processes Cross linked Poly-ethylene foam (PE-Xa)
The insulation foams described in this paper are made Although cross linked PE foam (x-PE) is also made of
of Poly-urethane (PUR) foam, Poly-ethylene (PE) foam PE, there is a big difference compared to PE foam: the
or cross linked Poly-ethylene foam (PE-Xa). These type of Blowing agent.
foams have different properties. Some of these The foaming process to make x-PE foam is called the
properties influence the heat loss properties of the chemical foaming process. In this case a chemical is
complete product. mixed into the PE matrix. The blowing agent can for
instance be Azodicarbonamide. While heating the
Polyurethane (PUR) foam matrix, the chemical starts decomposing and gases are
PUR foam is a thermo-set foam. It is made out of two released. These gases are Carbon dioxide and
chemicals, a Poly-alcohol and an Iso-cyanate. These Nitrogen. The thermal conductivity of these gases is
materials react and the Polyurethane is formed. This more or less equal to the thermal conductivity of air. So
reaction is irreversible, so the material can never return the aging effect of this product in relation to the heat
into its original chemicals. The blowing agent for this loss is less.
kind of foam can be Carbon Dioxide, Nitrogen or
To make this foaming process possible, it is necessary
Hydrocarbon molecules, for instance Cyclopentane or
to connect the Poly-ethylene chains with each other.
Butane.
This is called the cross link. The complete process to
If Hydrocarbon gases are used, these gases strongly make x-PE foam is called the chemical foaming
influence the heat loss performance of the pre- process with cross link. To make the comparison with
insulated system. These gases have different thermal PE foam complete: this process is called the physical
conductivities compared with air. After production of the foaming process without cross link.
foam an exchange with air starts. A product that is
The blowing agent is not the only additive that
freshly made contains a high percentage of
influences the thermal conductivity of the foam and
Hydrocarbon gases. At this point in time the product
therefore the heat loss properties of the pre-insulated
will have the lowest heat loss possible. If the same
system. Also other additives can influence the thermal
product is for instance three years old it contains more
conductivity of PE foam.
air and less Hydrocarbon gases due to gas diffusion.
And so the product will have a higher heat loss Nucleating agents will influence the cell structure of the
compared to the fresh product. The process of foam. As a basic rule: the finer the foam the lower the
degassing is described in research papers Long term thermal conductivity. With this additive the convection
heat loss of plastic Polybutylene piping systems by S. part of the insulation material will be influenced.
de Boer et al. [3]. Another additive that influences the thermal
conductivity is an anti-radiation additive. By using this
Poly-ethylene (PE) foam
special kind of additive it is possible to create a
PE foam is a thermoplastic foam. Once it is formed, it reflection of radiation energy.
can go back to its original state by heating it above its
melting point. Because of this property, it is possible to
HEAT LOSS TEST METHOD
re-use these kinds of foams.
This chapter briefly describes the test rig and test
The foaming process to make PE foam is called the
method used to determine the absolute heat losses of
physical foaming process. A Hydrocarbon molecule is
the different types of pre-insulated pipes.
mixed into the PE matrix under high pressure in an
121
th th
Test rig Measured samples

The test rig is used to determine the absolute heat The table below contains all products and dimensions
losses. The test rig has been designed in compliance of outer casing and service pipe that are treated in this
with EN 15632 and the tests are carried out according paper.
to ISO 8497 and EN 15632.
Table 1: Measured products
The physical part of the Thermaflex heat loss
ST/PUR PEX/PUR PB/PE PEX/PEX PEX/PEX
equipment consists of three sections. The first is the II I
water cooled compartment in which all the tests are Dc/Ds Dc/Ds Dc/Ds
Dc/Ds Dc/Ds
performed. This compartment is kept at a constant
23 C during each measurement. -- -- 62A32 -- --
Heating probes are used as a heat source. These 90DN25 -- 90A32 90A32 90A32
heating probes are custom made by preparing a two 125DN50 -- 125A63 160A63 175A63
meter Thermaflex piping segment of all available
162DN80 162A110 200A110 200A110 200A110
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 Results of testing
temperature and power values. All results have been extracted from measurements
carried out by the Thermaflex testing rig. The new
Method of testing European standard EN 15632 has been used.
Different heating probes are used for the testing. The This standard describes in Annex D a method to
probe with the appropriate diameter is inserted in a test present the results of testing in end-use condition. This
sample and inserted in the cooled test section. means: the product is buried in soil. According Annex
The heat loss measurement is done by measuring the D.3 the following general values are used for the
energy required to keep the probe at a constant calculation:
temperature, by measuring the current at constant Soil covering
voltage in the heating coils and calculating the power
o 0.8 m
consumption. Since the middle/testing coil is exactly
one meter in length the required energy represents the Thermal transmittance factor of earth-air
exact heat loss through one meter of piping and o 0.685 m2.K/W
insulation in W/m. For this paper the heat loss is
Thermal conductivity of the soil
determined for multiple probe temperatures.
o 1.0 W/(m.K)
Information The heat loss is calculated using the following
For more information concerning the test rig and formulas:
method of testing see the paper Verification of heat
loss measurements (J.T. van Wijnkoop et Al. [1]) Tflow Tsurrounding W
(1) Q
Rsoil Rflow
m
Blowing agent analysis
All products that are involved in this paper have been d4
(2) Z H m
analyzed on quantity of blowing agent and type of 2
blowing agent. The following results were found:
(3) Zc Z R0 soil m
The samples of PEX/PEX I and PEX/PEX II
did not show any amounts of hydrocarbon 1 4 Zc m K
blowing agents; (4) Rsoil ln
2 soil
d4
W
The PB/PE samples contained a quantity of
hydrocarbon blowing agent over 50 percent;
All results are presented in W/m, measured and
The samples of ST/PUR and PEX/PUR
calculated at a temperature difference of 60 Kelvin.
products contained a mixture of hydrocarbon
This temperature difference is derived from inner
gases. These gases were analyzed. Both
service pipe temperature minus surrounding ambient
product types contained approximately 95% of
blowing agent.
122
th th
temperature (70 degrees Celsius minus 10 degrees The 32, 63 and 110 millimetre service pipes
Celsius). First the absolute heat loss is displayed, followed by
Table 2: Results according to EN 15632 at a temperature the insulation area analysis.
difference of 60 Kelvin.
Absolute heat loss
Product Type Heat Loss System
In this paragraph all absolute heat loss values are
Buried (W/m.K) compared for the 32, 63 and 110 millimetre service
system
pipes.
W/m
In Graph 1 the results are displayed for temperature
ST/PUR 90DN25 11.6 0.042 difference of 60 Kelvin.
ST/PUR 160DN80 16.0 0.033 PB/PE PEX/PEX I PEX/PEX II PEX/PUR
35.0
PEX/PUR 162A110 22.3 0.049
30.0
PB/PE 63A32 15.2 0.038
Heat Loss [W/m]

25.0
PB/PE 90A32 12.8 0.044 20.0
PB/PE 125A63 22.0 0.056 15.0
10.0
PB/PE 200A110 27.4 0.068
5.0
PEX/PE II 90A32 16.6 0.057
0.0
PEX/PE II 160A63 17.6 0.055 90/32 160/63 200/110
Diameter service pipe [mm]
PEX/PE II 200A110 31.1 0.073
Graph 1 Absolute Heat Loss 32, 63 and 110 mm service
pipe (dT = 60 K)
PEX/PE I 140A32 12.5 0.057
The products based on PE or PE-x foam show higher
PEX/PE I 175A63 17.6 0.059 heat losses for the 110 mm service pipe than the
system based on PUR foam. The difference is
PEX/PE I 200A110 28.8 0.051 approximately 20 percent.
The different test samples show a wide variance in the
COMPARISON OF FLEXIBLE PLASTIC diameter of the outer casing.
PRE-INSULATED PIPES
Therefore, only the results for the 32 millimetre service
This chapter compares the flexible pre-insulated pipes. pipe are comparable for PB/PE 90A32 and PEX/PEX II
The comparison is based on three diameters 90A32. For the 110 mm service pipe, a comparison can
representing the entire diameter range for plastic pre- be made between the PEX/PUR 200A110, PB/PE
insulated pipes. 200A110, PEX/PEX II 200A110 and PEX/PEX I
The comparison is expanded by evaluating the heat 200A110.
loss in correlation to the outer casing diameter (resp. Another difference in this comparison is the use of a
the foam area). PB pipe or a PE-x pipe. PB and PE-x have different
In Table 1 the flexible plastic pre-insulated pipes are thermal conductivities (0.19 W/m.K versus 0.40
defined. These are the products PEX/PEX, PEX/PUR W/m.K). However, this effect is already corrected by
and PB/PE. using the Wallentn equation [5], as shown in (1).
For more information concerning the PB/PE pre- Insulation area

insulated pipes see Heat loss of flexible plastic pipe
To compare the different kinds of flexible pre-insulated
systems, analysis and optimization by van der Ven et
pipes on their performance, all outer diameters are
al. [4].
altered towards 90, 160 and 200 millimetres. The
123
th th
corresponding heat loss is calculated using the thesis is expanded by evaluating the heat loss in
of Wallentn [5], as in (5). correlation to the foam area.

2 Tprobe Tcasing

i Rigid Pre-insulated pipes
1 d2 1 d3 1 d4 (5) The different systems and their corresponding
ln ln ln
s dimensions are represented in Table 1.
i d2 c d3
d1
The rigid pipe product that has been tested
Where:
according to EN 15632 was the ST/PUR product.
Tprobe/Tcasing = Probe / Casing temperature First the absolute heat loss is displayed, followed by a
d1 to d4 = inner/outer diameters of service recalculation towards transport capacity and finally the
pipe and casing insulation area analysis.
s, i, c = heat coefficient of service pipe, Absolute heat loss

insulation and casing In this paragraph all absolute heat loss values are
Graph 2 represents the comparison on the basis of compared. The rigid DN25 pipe service pipe is
the same outer casing. compared with a flexible PB/PE-x service pipe with an
outer diameter (OD) of 32 mm. DN50 is compared with
PB/PE PEX/PEX I PEX/PEX II PEX/PUR OD 63 mm and DN80 is compared with OD 110 mm.
35.0 In Graph 3 the results are displayed for temperature
differences of 60 Kelvin.
30.0
PB/PE PEX/PEX I PEX/PEX II PEX/PUR ST/PUR

Heat Loss [W/m]
25.0
20.0 35.0
15.0 30.0
Heat Loss [W/m]
10.0 25.0
5.0 20.0
0.0 15.0
90/32 160/63 200/110
10.0
Diameter casing/service pipe [mm]
5.0
Graph 2 Relative Heat Loss 32, 63 and 110 mm service
pipe, all with an equal outer casing (dT=60K). 0.0
Result analysis DN25-PB32 DN50-PB63 DN80-PB110
The flexible pre-insulated systems, with PE and PE-x Diameter service pipe [mm]
foams, show a variance in heat loss values. The
Graph 3 Absolute Heat Loss DN25/PB32, DN50/PB63 and
absolute differences in the system are caused by the DN80/PB110 mm service pipe (dT = 60 K)).
dimensions of the pre-insulated pipes and the quantity
and type of blowing agent that has been used. Also the The different test samples show a wide variance in the
recalculation to the same outer casing diameters diameter of the outer casing.
shows an advantage for the PE foamed system in PB
The heat loss for ST/PUR 160DN80 is much lower
service pipes of 32, 63 and 110 millimetres.
compared to the heat loss of the 200A100 flexible
piping products.
COMPARISON OF FLEXIBLE PLASTIC PRE-
INSULATED PIPES VERSUS A RIGID PIPING Even the difference with the PUR based PEX/PUR
SYSTEM system is high (28 percent). For the PE and PE-x foam
based products the difference is even higher
In this chapter the flexible pre-insulated pipes are (42 percent)
compared with a rigid piping system. The
The heat loss for ST/PUR 90DN25 is more or less
comparison is based on diameter. The comparison
comparable with the heat loss for PB/PE type 90A32
(9 percent). So it seems that for smaller sizes the
124
th th
difference in absolute heat loss is lower, compared to

the absolute heat loss difference for larger sizes.
40.0
Transport capacity
35.0
When comparing rigid steel service pipes with flexible
30.0
plastic service pipes there is a difference in transport
Heat Loss [W/m]

capacity for comparable diameters. 25.0
This paragraph calculates the amount of heat loss 20.0

when transporting water of 70 degrees Celsius through 15.0
one meter of steel DN25, DN50 and DN80
10.0
(k-factor = 0.07 mm, velocity = 1.0 m/s). Subsequently
the same amount of heat loss is used as a reference 5.0
for calculating the amount of water that can be
0.0
transported through a plastic pipe 32 and 63 90/DN25-PB32 160/DN50-PB63 200/DN80-
(k-factor = 0.007 mm, velocity = 1 m/s). For these PB110
calculations the thesis of Colebrook and White [6] is Diameter service pipe [mm]
used. The results of this calculation are displayed in
Table 3. Graph 4 Relative Heat Loss, all with an equal outer
casing and transport capacity.
Table 3: Calculation results transport capacity
Result analysis
DN2 PB32 DN5 PB63 DN8 PB110
5 0 0 Flexible pre-insulated pipes have a higher absolute
heat loss compared to rigid pre-insulated pipes.
Velocity
1.0 1.072 1.0 0.81 1.0 0.76 Recalculation to the same transport capacity [kg/s] and
[m/s]
the same outer casing diameter shows that rigid pre-
Head
insulated pipes perform better.
Loss 425 188 112
[Pa/m] The reason for this difference is the relative small inner
Flow diameter of the plastic service pipes. The low k-factor
0.64 0.57 2.33 1.68 5.35 4.86
[kg/s] can not compensate for the smaller diameter. Table 4
shows the steel versus plastic service pipe diameter
Next the absolute heat loss is recalculated to an equal dimensions.
flow per diameter. The basis is 0.57 [kg/s] for the Table 4: Service pipe diameter dimensions
DN25/PB32, 1.68 [kg/s] for the DN50/PB63 and
4.86 [kg/s] for the DN80/PB110. DN25 PB32 DN50 PB63 DN80 PB110
ID 28.5 26.0 54.5 51.4 82.5 90.0

Insulation area [mm]
To compare the flexible pre-insulated pipes and the
However the fact that smaller diameters show a smaller
rigid piping system on their performance, all outer
heat loss difference between rigid and flexible pre-
diameters are altered towards 90 and 160 millimetres
insulated pipes is interesting.
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 CONCLUSIONS
has the same outer casing and is not recalculated. This chapter briefly addresses each chapter and
outlines its conclusions.
Overall comparison
The comparison on basis of the same outer casing Test samples
and transport capacity is shown in Graph 4. 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.
125
th th
The rigid system in this paper; The flexible piping system with the PUR insulation
ST/PUR. foam on the other hand performs better compared to
flexible PE and PE-X insulation foams with equal
Method of testing
dimensions.
All heat loss tests are performed on a test rig that has
been designed in compliance with EN 15632. The Comparison of flexible piping system versus the
tests are carried out according to ISO 8497 and rigid pre-insulated pipes
EN 15632. Flexible pre-insulated pipes have a higher absolute
heat loss compared to rigid pre-insulated pipes.
Blowing agent analysis Recalculation to the same transport capacity [kg/s] and
All measured products are checked on type of gas and the same outer casing diameter shows that rigid pre-
gas content. The ST/PUR and PEX/PUR products insulated pipes perform better.
contain approximately 95 percent of blowing agent.
However the fact that smaller diameters show a smaller
The PB/PE product range has a quantity over 50% of heat loss difference between rigid and flexible pre-
blowing agent. insulated pipes is interesting.
In the products of PEX/PEX II and PEX/PEX I no To be comparable in heat loss some dimensions of the
Hydrocarbon gases were detected. flexible piping systems range need to be optimized.
However, other advantages of flexible pipe systems, for
Comparison of flexible pre-insulated pipes instance the potential decrease of service meters
A fair comparison is difficult because of differences in because of a curved layout-design, can partly
outer casing and other dimensions. These conclusions compensate the higher heat loss compared to the rigid
are therefore only valid for the products that have been system (see Heat loss analysis and optimization of a
tested for this paper. flexible piping system by J. Korsman et al. [2]).
In a buried condition the PB/PE pre-insulated pipe

ADDENDUM
shows for equally dimensioned pipes 90A32 and
200A110 the lowest absolute heat loss values for all Significant product improvement of the PB/PE/PE pipe
pre-insulated pipes based on PE or PE-x foam. system has led to a decrease in heat loss [4]. Graph 5
As mentioned before, the absolute differences in the is updated with these improvements resulting in the
comparison displayed in Graph 6. The new samples
system are caused by the dimensions of the pre-
insulated pipes. Recalculation of the same outer casing are displayed under the name of PB/PE II.
diameter shows also an advantage for the
PB/PE PB/PE II PEX/PEX I PEX/PEX II
PB/PE system in service pipes of 32, 63 and 110 mm.
PEX/PUR ST/PUR
See Graph 5.
40.0
35.0
40.0
Heat Loss [W/m]
30.0
35.0 25.0
30.0
Heat Loss [W/m]
20.0
25.0 15.0
20.0 10.0
15.0 5.0
10.0 0.0
5.0 90/DN25-PB32 160/DN50-PB63 200/DN80-
PB110
0.0
90/DN25-PB32 160/DN50-PB63 200/DN80-
Diameter service pipe [mm]
PB110
Diameter service pipe [mm] Graph 6 Relative Heat Loss 32, 63 and 110 mm service
pipe, all with an equal outer casing and transport capacity
Graph 5 Relative Heat Loss 32, 63 and 110 mm service (dT = 60 K)
pipe, all with an equal outer casing and transport capacity
(dT = 60 K)
126
th th
FURTHER INFORMATION [2] J. Korsman, I.M. Smits and E.J.H.M. van der Ven,
Questions concerning the paper can be addressed to: Heat loss analysis and optimization of a flexible
piping system, in Proc. of the 12th International
Thermaflex International Holding B.V. Symposium on District Heating and Cooling,
Veerweg 1 Tallinn, Estonia (2010).
5145NS Waalwijk
[3] S. de Boer, J. Korsman and I.M. Smits, Long term
The Netherlands
heat loss of plastic Polybutylene piping systems,
www.thermaflex.com
in Proc. of the 11th International Symposium on
Liandon B.V. District Heating and Cooling, Tallinn, Reykjavik
Dijkgraaf 4 (2008).
6920AB Duiven
[4] E. J .H. M. van der Ven and R.J. van Arendonk,
The Netherlands
Heat loss of flexible plastic pipe systems, analysis
www.liandon.com
and optimization, in Proc. of the 12th International
Symposium on District Heating and Cooling,
ACKNOWLEDGEMENT
Tallinn, Estonia (2010).
We would like to thank all involved employees of
[5] P. Wallentn, steady-state heat loss from
Thermaflex Isolatie B.V. who made this research
insulated pipes, Lund Institute of Technology,
possible (especially H. Leunessen and M. van Doorn).
Sweden, 1991
Special thanks go to P. Blom and P. van Rijswijk for the
[6] C. F. Colebrook, "Turbulent flow in pipes, with
dedication they showed in carrying out all the heat loss
particular reference to the transition region
measurements during this research.
between smooth and rough pipe laws", February
REFERENCES 1939
[1] J. T. van Wijnkoop and E.J.H.M. van der Ven, [7] C. Engel and G. Baars, New economical
Verification of heat loss measurement, in Proc. of connection solution for flexible piping systems, in
the 12th International Symposium on District Proc. of the 12th International Symposium on
Heating and Cooling, Tallinn, Estonia (2010). District Heating and Cooling, Tallinn, Estonia
(2010).
127
th th
EFFECTIVE WIDTH THE RELATIVE DEMAND

FOR DISTRICT HEATING PIPE LENGTHS IN CITY AREAS
1 1
Urban Persson , Sven Werner
1
School of Business and Engineering
Halmstad University, PO Box 823, SE-30118 Halmstad, Sweden
ABSTRACT effective width becomes the width of an analogous

rectangle with the trench length as the length and
One key concept when assessing network investment
where the rectangle area is equal to the given land
cost levels for district heating systems is the linear heat
area.
density. In contrast to a traditional way of expressing
this quantity entirely on the basis of empirical data, a The concept was introduced by Werner [3] and has
recently developed analytical approach has made it been further elaborated recently in model estimations
possible to estimate linear heat densities on the basis of distribution capital cost reactions to decreased heat
of demographic data categories. A vital complementing demands in four north European countries [2].
quantity in this analytical approach is the concept of Essential for calculations of anticipated investment cost
effective width. levels for future district heating systems, the effective
width constitutes an important model parameter
Effective width describes the relationship between a
indicating levels of network extensions in given land
given land area and the length of the district heating
areas.
pipe network within this area. When modelling
distribution capital cost levels by use of land area Since the concept of effective width itself is rather new,
values for plot ratio calculations, there is a potential with no previous analytical or statistical use, data on
bias of overestimating distribution capital cost levels in effective widths are in principal non attainable within
low dense park city areas (e < 0.3). national statistical sources. Effective width might be
regarded as an innovative model quantity with no
Since these areas often include land area sections
previous representation in the field of district heating
without any housing, avoiding overestimations of
research.
network investment costs demand some kind of
corrective mechanism. By use of calculated effective
width values, a compensating effect at low plot ratio AIM
levels is achieved, and, hence, renders lower The aim of this paper is to describe the concept of
anticipated distribution capital cost levels in low dense effective width and outline the basic properties of this
park city areas. quantity. On the basis of, although sparse, empirical
observations, preliminary statements concerning the
INTRODUCTION properties of effective width are made. The aim is
further to enlighten the theoretical environment in which
One key concept when estimating investment cost
effective width contributes when applying demographic
levels for district heating systems is the linear heat
quantities for estimations of district heating network
density, i.e. the quota of annually sold heat in a district
investment costs.
heating scheme and the trench length of the piping
system in this scheme (Qs/L) [1]. In contrast to a
traditional way of expressing this quantity entirely on LIMITATIONS
the basis of empirical data, a recently developed Due to a limited amount of empirical data, in principal
analytical approach has made it possible to estimate less than 100 observations, the specific result values
linear heat density on the basis of demographic data and relationships accounted for in this paper must be
categories [2]. A vital complementing quantity in this considered as preliminary. Although thorough in theory,
analytical approach is the concept of effective width. the concept of effective width needs to be supported
further by extended empirical data gathering. In order
BACKGROUND to be able to produce solid and reliable estimations of
effective width values in different kinds of city areas,
Effective width is a stand alone concept within district
such information is considered vital for future use of the
heating theory, describing the relationship between a
concept.
given land area, AL, and the length of the district
heating pipe network, L, within this area. Hence, the
128
th th
EFFECTIVE WIDTH p = P/AL [number/m2] (7)

Effective width is a measure indicating the district = AB/P [m2/capita] (8)
heating network extension level within a given land
area. The quantity effective width, which is symbolised P = Total population [number]
by use of the letter w, with the unit metres, expresses
the ratio between land area and the total trench length AL= Total land area [m2]
of the distribution network within a district heating
system [3] AB= Total building space area [m2]
w = AL / L [m] (1) The concept of effective width hereby plays a key role
in the reformulation of the traditional expression for
Being in this way the result of explicit area and grid
linear heat density, and hence, constitutes a central
properties, effective width can be used to describe
quantity in model estimations of the feasibility and
typical district heating properties in different population
viability of future district heating network. If linear heat
density areas and hence, give information on
density can be said to indicate the level of district heat
prerequisite conditions for future district heat
distribution system utilisation, the effective width
establishments.
indicates the distribution system coverage of the land
area at hand.
THE CONCEPT
In order to introduce the concept of effective width, it is THE PROBLEM
necessary to first understand some basic principals
From a district heating distribution point of view it is
regarding the linear heat density. The concept of linear
relevant to distinguish between two kinds of land area
heat density, being the division of total annually sold
low plot ratio situations. The land areas can, principally,
heat in a district heating system and the total length of
consist of either a wide dispersion of households
the district heating piping network, indicates the level of
spread out over the whole area (A), or households can
district heat distribution system utilisation. Furthermore,
be closely limited to only a fraction of the land area (B),
linear heat density is a denominator parameter when
see figure 1.
calculating district heating network capital costs.
Qs
LinearHeatDensity [GJ/m] (2)
L
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 Figure 1. Low plot ratio land areas, scenario A with wide
planning quantity plot ratio (e), which is suggested in dispersion of buildings and scenario B with high
[2], the concept of linear heat density can be concentration of buildings.
alternatively expressed as; In the first case (A), a district heating distribution grid
would have to cover all of the land area at hand in
Qs
qew [GJ/m] (3) order to deliver heat (at very low linear heat density),
L while in the latter case (B), the grid could be narrowed
down to the limited area fraction. If, when conducting
The three new parameters, specific heat demand (q),
district heating feasibility model analysis, plot ratios are
plot ratio (e) and effective width (w), are defined as:
extracted by means of (5), it would be relevant and
q = Q/AB [GJ/m2a] (4) recommended to somehow adjust the land area
magnitude in order not to include non-targeted area
e=p [1] (5) fractions. An adjustment to reach this purpose can be
achieved in several different ways, of which Effective
w = AL/L [m] (6)
Width compensation suggested in this paper is one
where option.
129
th th
DATA AND VALUES plot ratio values above 1. This would indicate that the
relationship between high dense inner city land areas
In the spring of 2009, the authors, both being lecturers
and the length of the required piping grid in such areas
at Halmstad University in Sweden, initiated a pre-study
is constant.
to be carried out by two Bsc-students at their
department [4]. The study was two-fold in regard of Still, if plotted explicitly, the function does not converge
gathered data. Partly it delivered previously assembled at any effective width value, no matter how far the plot
and crucial data on plot ratios, land areas and trench ratio value is extended, but the rate of divergence
lengths in 39 detached house districts heating schemes decreases with higher plot ratio values. Since plot
in Sweden [5], allowing estimations of effective widths ratios values above 3 are considered extremely rare,
in these districts, see Figure 2, and partly own collected effective width values within high dense inner city areas
data. (plot ratio values above 0.5) can be anticipated to be
found in the interval of 50 < w < 60 meters.
Effective
width (w)[m]
Effective
250 width (w) [m]
400
200 Detached 350 Detached
houses and MF
150 300
houses
250
100 Power
-0,3731 (Detached 200 Power
y = 27,802x
houses) 150 (Detached
50
y = 61,838x-0,1495 and MF
100 houses)
0 50
0,0 0,1 0,2 0,3 0,4
Plot ratio (e) 0
0 0,5 1 1,5
Figure 2. Effective width as a function of plot ratio in 39 Plot ratio (e)
district heating schemes in detached house districts in
Sweden. Source: [5] Figure 4. Effective width as a function of plot ratio,
combination of 39 district heating schemes in detached
The own collected data of the study refers to data from house districts and 34 in multi family housing districts in
34 district heating schemes in multi-family housing Sweden. Datapoints merged from figure 2 and 3.
districts in the Swedish cities of Halmstad and
Gothenburg, see Figure 3. Effective
Width (w) [m]
140
Effective 130
width (w) [m] 120
400
110
350 Multi-
100
family
300 90
houses
80
250
70
200 Power
60
(Multi-
150 family 50
y = 56,622x-0,41 40
100 houses)
0 0,5 1 1,5 2 2,5 3
50
Plot Ratio (e)
0
0 0,5 1 1,5
Plot ratio (e) Figure 5. Effective width as a function of plot ratio by use
of eq. (9).
Figure 3. Effective width as a function of plot ratio in 34
district heating schemes in multi family housing districts in For plot ratio values below 0.5, on the other hand
Sweden. Source [4]
(outer city area and park areas), the relationship is by
On the basis of these results, and when combined in no means constant, but diverges rapidly with increased
one common graph, see Figure 4, a power function effective width values as a consequence. At a plot ratio
were established and presented in [2]. Note that (e) value of 0.04 the effective width reaches a value of 100
refers to plot ratio values, not to the natural logarithm meters, and the curve reveals that the increase of
base (e); effective width values at even lower plot ratio values
below 0.04 renders values above 100 meters and
w 61.8 e 0.15 [m] (9) beyond.
As can be seen in Figure 4, the graph suggests a The graph characteristics of Figure 5 has significance
convergence at effective width values at 60 meters for for estimations of district heat distribution capital cost
130
th th
levels in park areas, since these areas often also REFERENCES

include land area fractions without any housing, i.e. not
to be targeted by district heating networks. When using [1] Frederiksen S. and Werner S, Fjrrvrme teori,
crude statistical land area values for plot ratio teknik och funktion (District Heating theory,
calculations, there is a potential bias of overestimating technology and function). Studentlitteratur, Lund
distribution capital cost levels in these suburban areas, 1993.
since actual habitations plausibly only occupy parts of [2] Persson U. and Werner S, The Future
the land area at hand. In these occasions, effective Competitiveness of District Heating, to be
width values arrived at by use of eq. (9). have a published.
compensating effect by rapidly increasing its value at
low plot ratio levels, and, hence, rendering lower [3] Werner S, Fjrrvrme till smhus vrmefrluster
anticipated distribution capital cost levels. och distributionskostnader (Sparse district heating
heat losses and distributions costs). Report
1997:11, The Swedish District Heating Association.
CONCLUSION
Stockholm 1997.
The main conclusion from this analysis is that the
[4] Netterberg H and Isaksson I, District Heating in
concept of effective width offers a new simple shortcut
Slough. BSc thesis from Halmstad University,
for quick estimations of capital investments for heat
Halmstad 2009.
distribution in virgin urban areas.
[5] [Andersson S et al, Nulge Vrmegles Fjrrvrme
This conclusion is especially valid if the effective width
(The current situation for sparse district heating),
has almost a constant value over a plot ratio of 0.5 as
The Swedish District Heating Association, research
preliminary stated from Figure 4. Further data collection report FoU 2002:74. Stockholm 2002.
will show how true this new finding will be.
131
th th
INTEGRATING RENEWABLE ENERGY INTO LARGE-SCALE DISTRICT HEATING

SYSTEMS
Peter Begerow, Dr. Stefan Holler
MVV Energie AG, Mannheim, Germany
ABSTRACT If the used technology just produces heat, a financial

aid from the BAFA (Federal Office of Economics and
Renewable energy for heating is mostly used in small
Export Control, Germany) or KfW (bank under control
systems for single-family houses. The existing district
of the Federal Republic, Germany) is possible. There
heating networks are generally run by large heating
are different regulations which have to be fulfilled by
plants or combined heat and power plants fired with
the project in order to be eligible for those subsidies [8].
fossil fuels.
Main criteria are the size and type of the investor and
To combine these two systems, a feasibility study was planner, the type of technology and the size of the heat
completed with a focus on the district heating grid in plant and of the storage tank.
Mannheim, Germany, and with a focus on solar thermal
Previous studies [2], [12], [16], [18] showed that solar
heat. Other renewable energy heat sources,
thermal energy is mostly used in single-family houses
geothermal heat and heat from biomass, are included
and smaller heating grids combined with seasonal heat
for a comparison.
storage systems. Those systems are still in
The study focuses on the heat price as a key figure to development and need financial support to be realized.
analyse the economic feasibility. The technical Most of these heating grids run with a lower flow
feasibility has been evaluated by using a simulation temperature and use either fossil fuels or heat from
model of a secondary district heating grid, which is biomass for an auxiliary heat generation. The largest
operated on a low flow temperature level of 70 C and solar thermal district heat system in Germany is located
which is connected to a central solar thermal energy in Crailsheim. It covers an area of approx. 7300 m of
plant. The paper describes which technical and solar collectors with two buffer tanks with a combined
economic framework conditions are necessary for volume of about 500 m as thermal storage. In addition
implementing renewable energy into large-scale district a seasonal geothermal storage has been built which
heating systems. The calculations show that in will cover 50 % of the heat demand for about 2000
comparison with other renewable heat sources solar residents. This research project has heat production
heat has the highest heat costs ranging from costs without any financial support of about 19 ct/kWh.
7,7 ct/kWh to 14,5 ct/kWh depending on the plant size, This sum will be reduced depending on the possible
the solar fraction and the use of a storage system. The subsidies. [12] Reported technical difficulties were
major technical problems for integrating solar heat into mostly in the thermal storage technology. There were
a heat grid are the pressure difference between the little problems with the collectors as common flat plate
flow pipe and return pipe and the low temperature the collectors were used which are commercially available
flat plate solar collectors are working with. and used in large numbers in smaller systems.
The project in Crailsheim has shown the technical
INTRODUCTION feasibility of a system with a seasonal thermal storage,
Based on the protocol of Kyoto and European but it also shows that considerable costs are involved.
regulations a high reduction of CO2 emissions in Furthermore, if a single-family house will install a
Germany is necessary. To achieve those goals, an seasonal storage to get a solar fraction above 50 %, it
expansion of renewable energy in the heat market is needs more than 10 m of hot water storage
required. In Germany the major aim to reach is a share (depending on the building type and planned solar
of 50 % renewable energy in the heat market by 2050. fraction). However, in common buildings there isnt
Furthermore, 50 % of the renewable heat is supposed enough room for that size of storage [17]. Those two
to be contributed by a heating grid. [13] aspects show that the use of a heating grid could
To achieve these goals, different governmental as well significantly reduce the costs of the solar thermal
as local support mechanism and financial subsidies are systems and could save space otherwise necessary for
available. If the heat production is combined with an a storage tank.
electricity production, the major financial support is The paper will give an overview of how the expansion
based on the EEG (German law for renewable energy). of renewable energy in the heat market will be possible
132
th th
by integrating renewable heat into a district heating is not part of the simulated area and on the reduction
network. A detailed simulation for a solar thermal through population, which has not a direct effect on one
integration was done by using RETscreen [14] as special housing area. The reduction through influence
simulation software. of temperature has a share below 5 % within 15 years
and is therefore not included within the simulation.
MATERIAL AND METHODS
Existing Systems
Evaluation of heat demand Between newly built and existing heat networks there
For planning a new heat production facility, the heat exist some main differences which have to be
demand of the connected consumers is necessary. If considered. If the network is designed especially for the
those are existing households, the heat demand from renewable energy source, it can be technically
the past can be used for calculations. For newly built specialized (e.g. forced low return temperature for
houses the heat demand should be exactly calculated building owners; special isolation of the used pipes).
with the standards named in DIN V 4108-6. Older heating grids on the other hand are normally
constructed for the heat production with fossil fuels and
If this is not possible, the yearly heat demand can be are normally designed for higher temperatures.
assumed by the given figures: Furthermore, in some heating grids a high temperature
Table 1: heat demand [15] is necessary either for thermal cooling systems (e.g.
absorption chillers) or for the heat transfer stations
Building size Heat demand heating demand within the houses which are built for high temperatures
(room heating) (hot tap water) (low flow temperatures need optimized heat transfer
[housing
units] [kWh/ma] [kWh/ma] stations [12]. In the following, the main aspects for the
integration of different sustainable heat generation
1-2 72,3 20
technologies are described.
More than 3 55,3 20
Heat grid for renewable energy
For the integration of renewable energy into heat grids,
These figures can be realized in buildings constructed
different possibilities for the connection exist.
between 2011 and 2020. [15]
Especially for the solar thermal energy production it is
Another factor for the planning of a heating grid is the assumed, that more than one heat plant will be
outlook into the future, because the payback period of connected.
a renewable heat production facility is very long.
The three options are:
The following graph shows the expected change in
heat demand for Germany focusing on different factors 1. Taking water from the return pipe, heat it and
of influence: 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.
Fig 1: development of heat demand [11]
To evaluate the necessary pump work a first estimation
Within the following simulation this development is not can be done with equation (1). It gives the pump work
further regarded. The major reduction within whole W depending on the necessary heat flow Q, the
Germany is based on renovation of old buildings, which
133
th th
pressure difference p and the temperature difference Used Software

T between flow and return. Included in this equation RETscreen is a program to make first feasibility studies
is, the pump efficiency as well as the density and of all kind of green energy projects. In terms for solar
the thermal capacity of water cp . 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
(1) collectors. The needed amount of heat can either be
calculated other ways or assumed by the software
For the integration of renewable energy into heat grids depending on the amount and size of buildings.
some aspects have to be regarded. Combining those input factors with others, the
For solar thermal energy the flow and return simulation tool gives a recommendation of the used
temperature of the network is a major problem. This number of solar collectors and the size of a thermal
should be lower than in the existing district heating grid storage system. If all input factors are included the
and run with a temperature of about 60 C / 40 C. [6] program calculates the yearly heat production and the
For using deep geothermal heat it depends on the used solar fraction. Beyond that, the program can be used to
technology. If it is combined with the electricity include a second heating system for the remaining
production, the waste heat after the power plant is needs to get the final payback period and the total
normally below 80 C. emissions. For the simulation of this paper the version
4 (November, 2009) of the named software was used.
Another major obstacle is the variation of the heat
production and the demand if using solar thermal heat.
Financial Calculation
During the summer months, the solar radiation is at its
peak, but the heat demand has its peak during the The calculation of the heat costs is based on the net
winter months. To cover a heat grid with a high solar present value method. For the internal rate of return the
fraction, a long term thermal storage system is given value was used, all other costs included and the
necessary. heat costs varied to get a net present value of zero.
This method gives the current heat price and a further
Description of selected site increase during the next years is included. This makes
it possible to compare the actual heat price to the given
For modeling the integration of solar thermal energy values of other systems. For the economical calculation
into a district heating network, a yet to be built housing in the conclusion of this paper, a competitive heat price
estate was selected. This housing area is planned with from now on was realized.
a district heating grid running at a flow temperature of
The named financial support which is included in the
about 70 C. This area is connected with a heat
calculations are subsidies on the capital cost. They
exchanger to the central heating grid of the city, which
depend, like mentioned in the introduction, on different
is run with flow temperatures between 90 C and
aspects. A research project like the one in Crailsheim,
130 C.
can get a higher support than commercial ones run by
large companies. [1]
Table 2: heat demand selected site
Building size Number of Total heating Solar thermal heat production

[housing units]
buildings demand [MWh/a] For the heat supply of the given housing area, different
(room heating + scenarios based on solar thermal energy were
hot tap water) developed. Using the RETscreen software tool the
technical parameters of the flat plate solar thermal
1-2 135 1561
collectors, weather data from a climate database and
More than 3 111 585 the given heat demand of the area was combined for
each scenario.
Sum 246 2146
The scenarios differ in the necessary amount of
collectors needed to achieve a solar fraction of the total
Table 2 gives an overview of the planned houses and heating demand of 50 % [scenario 1], a 100 % solar
their heat demand. The whole heating grid will have an heat production of the used hot tap water (which stays
length of about 1,3 km and the total heat demand will constant throughout the whole year) [scenario 2] and a
be 2146 MWh/a.
134
th th
50 % solar fraction of the total heating demand without For a comparison of the different scenarios the heat
a thermal storage system [scenario 3]. cost per kWh were calculated.
For the thermal storage a hot water system is The calculation of the emissions is based on the
assumed, because those are state of the art and can operation of the system and not on its total life cycle.
be used in most applications. Other systems have For solar thermal heat the CO2 emissions only arise
more specific requirements to the geological situation from the used electricity for the necessary pumps.
of the area. A geothermal heat storage for example Included in the calculation is only the pump energy for
does not work in an area with a flow of the ground the solar thermal collectors and, if necessary, to
water. For the simulation of scenario 2 a smaller increase the pressure for the integration into the
thermal storage compared to scenario 1 was assumed, heating grid flow pipe. The CO2 emissions for the
because there is no necessity for a seasonal heat German electricity grid are given with 506 g/kWh.
storage system.
For the calculation without a thermal storage system
For the simulation model a commercial solar collector [scenario 3] it was assumed that the produced solar
was taken (s. Table 3). It is a flat plate collector with a heat can directly be distributed throughout a district
anti-reflection glass and a gross area of about 2,6 m. heating network. This would make it possible to save
Its efficiency is 84,4 % (calculated according to EN the investments of a seasonal heat storage system and
12975). The simulation uses specific given parameters. also reduce the losses within the thermal storage
Those are shown in Table 3. system.
Table 3: used input parameters for solar simulation For those calculations the same heat amount was used
Annual heating energy Scenario 1: 1076 MWh than in scenario 2. But in this case it is not possible to
(calculated with given Scenario 2: 493 MWh cover 50 % of the heat demand of the total grid. Just a
method) Scenario 3: 1071 MWh small amount, for example the losses of the grid and
Scenario 4: 11,6 MWh the base load, can be produced with solar thermal
Scenario 5: 11,6 MWh technologies without a thermal storage.
Flow temperature 67 C
Another option would be to integrate small systems into
Return temperature 45 C the district heating grid. In this case the operator of the
Slope of collector 55 grid would not run the facility by itself. The heat
producer could use a solar thermal collector for its own
Azimuth of building -45 (southeast) heat demand but without a thermal storage system.
Instead of using an in-house thermal storage (what is
Type of collector WagnerSolar L20 AR
getting very large if a seasonal heat storage system is
Storage capacity Scenario 1: 1000 l/m used) the heating grid could be used. For the single
Scenario 2: 100 l/m house technology an internal rate of return of 5 % was
Scenario 3: 1 l/m used for the economic calculation (average percentage
Scenario 4: 100 l/m of building credit [4]). Furthermore the financial support
Scenario 5: 10 l/m
is a little different because of different regulations for
Heat exchanger 80 %
efficiency large and small systems. In the following those two
Miscellaneous losses 5 % if storage is used calculations are named scenario 4 for the heat
(smaller grid) production of a single-family house with a thermal
8 % if integrated into storage and scenario 5 for the calculation without a
large grid thermal storage.
Pump efficiency 40 %
(for grid integration) Summary of different scenarios:
Time period 20 year
Scenario 1
Internal rate of return Scenario 1: 8,5 %
Solar fraction of 50%
Scenario 2: 8,5 %
Scenario 3: 8,5 % Seasonal thermal storage included
Scenario 4: 5,0 %
Scenario 5: 5,0 % Scenario 2
Increase of heat price per 2% 100% heat production of hot tap water
year
Financial support 30 % Buffer heat storage included, but no seasonal
thermal storage
135
th th
Scenario 3 Heat production with biomass

Same amount of heat produced than in scenario 1 The heat production from biomass is technically very
No storage; connected to large district heating grid similar to the fossil fuel powered heating plants.
Therefore the integration into existing district heating
Scenario 4 grids is the easiest way compared to the other
Solar fraction of 100% for a single family house renewable energy sources.
Seasonal thermal storage included The exact technology depends on the used fuels and
therefore the economic calculation is mainly based on
Scenario 5 the price development of the biomass.
Same amount of heat produced than in scenario 4 The emissions of such a system are by way of
No storage; connected to large district heating grid calculation zero, because the emitted CO2 was firstly
bound by the biomass during its growing period. If the
Geothermal heat production biomass is planted in an area which was deforested for
The geothermal heat can be used in various ways for that, the emissions are not zero any more. The former
room heating. Using the shallow geothermal heat is forest was a CO2 sink which does not exist anymore
only possible in combination with a heat pump. and should be included in the calculation. Furthermore
Therefore, for a large integration into heating grids the the transport and processing of the biomass should be
deep geothermal energy is the favoured one. included. [19]
Furthermore in the upper valley of the river Rhein
(Oberrheingraben) the geothermal heat can be used for Fossil fuels for comparison
a combined heat and power production because of its In our days the district heating grid in Mannheim is fed
high temperature. In Germany this gives the possibility with heat from a fossil fuel fired CHP plant. The heat
to get a payment for the electricity based on the EEG prices from that system are much lower than the
which grows for 3 ct/kWh if the heat is used as well. renewable heat. Looking into the future it mainly
For a comparison to the solar thermal heat a depends on the price development of CO2 emissions
geothermal power plant in Landau, Germany is used as and the coal price. [8]
a reference. The emissions of such a system are very high, even if
This project began in 2004 and at the end of 2007 the the used heat is more or less waste heat. To reduce
power plant started its first electricity production. The those, a CCS technology can be implemented in the
first heat output was planned for 2009. future.
The power plant uses the ORC process (Organic

RESULTS
Rankine Cycle) to generate electricity. A drill hole with
a depth of 3000 m connects to thermal water with a The results of the simulation are shown in Tab. 4 and
temperature with up to 160 C which is cooled down Fig. 2 and 3.
during electricity production to 70 C. The whole yearly
In conclusion the heat price is lower if the collector area
energy output of the power plant is planned to be
increases (economy-of-scale). Furthermore the use of
22.000 MWh electricity and 9.200 MWh heat. One of
a district heating grid instead of a thermal storage
the major benefits of the geothermal heat production is
lowers the heat cost extremely.
the base load which is always available. On the other
hand this gives the problem that the heat is also For scenario 1 it is necessary to install a gross area of
available in the summer time and needs to be cooled 3080 m solar thermal collectors. 1076 MWh heat can
down in other ways. be produced in combination with a 2820 m hot water
storage. The heating costs calculated with the given
The calculated emissions of the power plant are
framework conditions are 11,2 ct/kWh. To operate the
0 g CO2/kWh because the electricity production has no
collector area, pumps are needed which consume
emissions and for the pumps the own electricity can be
electricity. The emissions of that electricity are, based
used. [6]
on the produced heat, 7,9 g CO2/kWh.
Currently the power plant runs with a limited output due
In scenario 2, 1916 m solar thermal collectors need to
to small earthquakes in the area of the drilling hole and
be installed. Combined with a hot water buffer storage
does not deliver heat until now. Additional geological
with a volume of 175 m, 494 MWh of heat can be
studies are done right now and a heat output should
produced. The financial calculation over 20 years lead
start after they are finished.
136
th th
to heating costs of 14,5 ct/kWh. The emissions of such

a system are 8,9 g CO2/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
Fig 2: CO2 emissions of different system [1], [5]
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 Fig 3: heat costs of different system [1], [20]
financial and ecological calculation where made. Those
results are shown in Figure 1 and 3. For the renewable
CONCLUSION
technologies the increase of 2% of the heat costs can
easily be included in the calculation. For the fossil (and To make the solar thermal heat production economical
the biomass) use, the heat price is highly dependent on compared to the other systems, different aspects have
the fuel price development. Therefore a price range is to be changed. In order to show the potential of cost
given on those systems. The reason for the range for reduction for solar thermal heat generation a sensitivity
fossil CHP heat CO2 emissions is that different analyses has been carried out. The following
references are used. parameters have been varied in order to reach a heat
prices of around 3,5 ct/kWh in the beginning year. This
Table 4: output parameters of simulation
is the actual heat price for private customers in
scenario Collector gross Produced Storage Mannheim.
area [m] heat size
Future heat price development
[MWh/a]
[m]
Change of investment
1 3080 1076 2820 Amount of financial support
2 1916 493,6 175 Internal Rate of Return
3 2542 1071 -
4 44 11,6 4
5 29 11,6 -
137
th th
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 8 % p.a. 8 % p.a.

development
Fig 4: influence on the heat costs of different factors 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
Figure 4 shows the influence of the different factors, if
the depth could be exploited or cheap biomass sources
the others stay the same. But it also shows, that by are available. Further research in the solar collector
changing just one aspect, a reduction of the heat cost
technology is needed to lower the capital costs and
to 3,5 ct/kWh is only possible with lowering the capital
equally within the thermal storage technology, as it
costs by 50% of the scenario without a thermal storage might get interesting in the future to include those even
[scenario 3].
in district heating grids with fossil fuels as heat source
Therefore a combination of different factors was done. to cover peaks in the demand and transfer a surplus
The capital costs are also influenced by the financial heat production from the summer into the winter
support und were calculated separately. season.
In order to reduce the heat costs down to about

3,5 ct/kWh in scenario 1, a reduction of the capital cost NOMENCLATURE
by 40% combined with financial support of 50% is W [J] work of pump
necessary, if the heat price will rise with 8% per year.
Q [J] heat flow
This is higher than shown in figure 4 and is based on a
high price assumption as reported in [10]. p [Pa] pressure difference (flow / return)
If a lower IRR is assumed (5 %), the capital costs have T [K] temperature difference (flow / return)
to go down to 70 % and a financial support of 40% of
cp [J/(kg*K)] heat capacity
the investment is necessary.
[kg/m] density
For scenario 3, a rise of the heat price and the lower
IRR (5 %) just need a reduction of capital costs of 10 % pump efficiency
to achieve heat costs of 3,4 ct/kWh. In this case the
assumed financial support of 30% stays the same. This REFERENCES
shows that an economical use of solar thermal energy
within district heating could be achieved. The financial [1] Begerow, P.; Integration von erneuerbaren
support of 30% can be possible based on a KfW Energien in Fernwrmenetze Eine technische
program, the capital costs of 90% of the base und wirtschaftliche Analyse aus Sicht eines
investment is possible within a feasibility study and the Fernwrmeversorgers, Diplomarbeit an der
5% IRR is an average figure for building loans. Universitt Flensburg, MVV Energie AG,
Mannheim; 2010.
The high requirements to make the solar heat profitable
show, that this technology is not advisable if other [2] Bodmann, M.; Mangold, D.; Nubicker, J.; Raab,
renewable energy sources are available. Furthermore S.; Schenke, A.; Schmidt, T.: Solar untersttze
for the technical integration a low temperature heating Nahwrmeversorgung und Langzeit-
grid is necessary. Wrmespeicher; Forschungsbericht zum BMWA
Vorhaben; Universitt Stuttgart; 2005
138
th th
[3] Bucar, G.; Schweyer, K.; Fink, C.; Riva, R.; [12] Mangold, D.; Riegger, M.; Schmidt, T.: Solar
Neuhuser, M.; Meissner, E.; Streicher, W.; Nahwrmeversorgung und Langzeit-
Halmdienst, C.; Dezentrale erneuerbare Energie Wrmespeicher; Forschungsbericht zum BMU
fr bestehende Fernwrmenetze; Vorhaben; Solites; Stuttgart; 2007. Page 14, 20
Bundesministeriums fr Verkehr, Innovation und
[13] Nitsch, J.; Wenzel, B.: Langfristszenarien und
Technologie; Wien, 2005. Page 15-16
Strategien fr den Ausbau erneuerbarer Energien
[4] Deutsche Bank; Zinslandschaft; in Deutschland; Leitszenario 2009; BMU; Berlin;
https://www.deutsche-bank- 2009. Page 53-57
bauspar.de/de/media/Zinslandschaft.pdf; 2010.
[14] RETscreen Version 4; Natural Resources Canada;
[5] Fielenbach, H.; Ohl, G.; Schwarzburger, H.: http://www.retscreen.net; 2009
Effiziente Wohnwrme und hoher Komfort; GBG
[15] Smolka, M.: kologisch-technische Auswirkungen
Mannheimer Wohnungsbaugesellschaft mbH;
dezentraler Energieversorgungsszenarien mit
2009.
Blockheizkraftwerken in elektrischen Verteilungs-
[6] Frey, M.; Milles, U.: Geothermische netzen; Verlagshaus Mainz GmbH; Aachen; 2009.
Stromerzeugung in Landau; BINE Projektinfo Page 18
14/07; Karlsruhe; 2007.
[16] Solarge: Marstal district heating Plant; Project
[7] Heidemann, W.: Solare Nahwrme und saisonale Summary; http://solarge.org/index.php?id=1235&
Speicherung; FVS LZE Themen; Berlin; 2005. no_cache=1; 14.03.2010
Page 36
[17] Sonnenhaus-Institut e.V.; http://sonnenhaus-
[8] Kaltschmitt, M.; Streicher, W.; Wiese, A.: institut.de/wohnhaeuser.html; 2010.
erneuerbare Energien; Springer Verlag; Berlin;
[18] Ulbjerb, F.: Large-Scale Solar Heating; Hot|Cool;
2006. Page 29
3/2008; DBDH; Frederiksberg; 2008
[9] KfW: Programm erneuerbare Energien;
[19] Watter, H.; Nachhaltige Energiesysteme;
http://www.kfw-mittelstandsbank.de/DE_Home/
Vieweg+Teubner; Wiesbaden, 2009. Page 168
Service/Kreditantrag_und_Formulare/Merkblaetter/
KfW-Programm_Erneuerbare_Energien_ 270_ [20] Vo, A.: Das Wachstumspotential der Nah- und
271_272_281_282.jsp; 2010. Fernwrme - wirtschaftliche und gesetzliche
Voraussetzungen fr den Ausbau; aus: Forschung
[10] Klpsch, M.; Besier, R.; Wagner, A.: Reicht fr
und Entwicklung Heft 10; AGFW; Frankfurt, 2005.
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.
139
th th
SOLAR DISTRICT HEATING (SDH): TECHNOLOGIES USED IN LARGE SCALE SDH

PLANTS IN GRAZ OPERATIONAL EXPERIENCES AND FURTHER
DEVELOPMENTS
1 1 1
M. Schubert , C. Holter and R. Soell
1
S.O.L.I.D. Solarinstallationen und Design GmbH, Puchstr. 85, A-8020 Graz,
m.schubert@solid.at
ABSTRACT First solar thermal plants for district heating were built
in the 1970s in Sweden. Since then, various plants
S.O.L.I.D. installed three large scale solar plants for
have been built mainly in Austria, Denmark, Germany
feeding into the citys district heating in Graz in recent
and Sweden.
years. These three solar plants have an annual heat
production of 15,8 PJ, the citys grid delivers 2800 PJ Most of these solar plants feed into rather small heating
per year. Therefore the integration of solar thermal in a grids or sub-grids with an annual heat delivery below
technical and economical feasible way has to meet the 50 GWhth(180 TJ). In Denmark, this market was
requirements of Graz existing district heating grid, growing rapidly in recent years and is now bigger than
which is one of the largest in Austria. the market for small-scale solar systems for single-
family houses.
The first plant, at stadium Graz-Liebenau with
1.420 m, has been now for seven years in reliable In Graz, Austria, solar thermal plants feed into a large
operations, with very good power output data. scale heating grid with an annual heat delivery of
830 GWhth(2,99 PJ) and a maximum power of
AEVG Graz, the largest plant in Graz at 4.960 m,
382 MW th. Technical parameters and operation
feeds into the gas power station (maximum power of
strategies in large scale heating grids are different to
250 MW) and from there the heat is distributed through
those in small scale grids and solar thermal technology
the district heating grid.
has to adopt to these circumstances.
The latest plant, at Wasserwerk Andritz with currently
Three solar thermal plants in Graz are presented and
3.860 m, has a buffer storage of 60 m and the
the way they are integrated into the citys heating grid.
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. SDH PLANT DESIGNS IN GRAZ
This paper presents operational experiences about 1. Feeding directly into the district heating grid
three different ways for feeding solar thermal energy plant at stadium Graz-Liebenau
into a large citys district heating grid. Recent
This plant is located on the roof of an ice-skating hall
developments like buffer management for combined
next to the citys football stadium (Fig. 1).
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 Fig. 1: Aerial view of solar plant Stadion Liebenau
begins to develop.
140
th th
The return medium of the heating grid is heated up and During first operation years, detailed monitoring was
transferred to the flow (Fig. 2) [2]. The adaption of solar done on the plants performance. Dependant on
thermal technology for the temperature and pressure climate condition, the annual yield of the plant was
levels of the district heating grid were challenging. This between 521 MWh/a and 569 MWh/a. This
project was realized with standard large scale corresponds to a specific yield of 370404 kWh/a per
collectors (1420 m collector area) of the Austrian square meter collector area. Also the return
manufacturer kotech and temperature levels in the temperature of the heating grid is of great importance
district heating flow of above 70 C have to be reached for the performance of the solar plant.
dependant on the ambient temperature.
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
141
th th
3. SDH for combined room heating and district Even lower temperature levels in shoulder seasons and
heating with buffer and heat pump plant in winter can be raised by a heat pump. The installation
Wasserwerk Andritz is planned for the end of 2010. COPs above 4 are
As solar thermal systems cant always generate the expected, i.e. when heat from the collectors of 26 C is
high temperatures as required for the district heating heated up to 55 C for room heating.
grid, other applications were found for temperature
levels below 75 C (Fig. 4). ACKNOWLEDGEMENT
Solar heat at low temperature level is stored into a This work is supported by the EU in the project
60 m buffer tank and later used for room heating of an SDHtake-off (IEE - Intelligent Energy Europe).
office building (low temperature floor heating). The
buffer is also fed by district heating and thus decreases REFERENCES
the required connected load of the office building.
[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.
Fig. 4: solar thermal plant Wasserwerk Andritz
142
th th
BIOENERGY COMBINES IN DISTRICT HEATING SYSTEMS:

PROSPECTS FOR A FUTURE GROWTH INDUSTRY?
1 2 2
E. Axelsson , A. Sandoff , C. Overland
1
Profu, Gothenburg, Sweden.
2
Department of Business Administration, University of Gothenburg, Sweden.
ABSTRACT One industrial branch that shows promising prospects

in this respect is bioenergy production, i.e. production
District heating offers opportunities for integration of
of various kinds of biofuel, biogas and solid biofuel.
bioenergy production (e.g. of biofuel). The aim of this
Integration of bioenergy production to district heating
paper is to assess the environmental benefit and the
production eventuates in a bioenergy combine were the
economic value of such integration, in order to evaluate
residual heat from the bioenergy production can be
the prospect for bioenergy combines in district heating
utilised for district heating. Moreover, the integration
systems. Since the detailed characteristics of the
can, in many cases, offer additional positive synergies,
district heating system are crucial for the feasibility for
e.g. regarding the use of steam and combustible
integration of bioenergy production, the assessment is
by-products.
based on four real district heating systems. The
environmental evaluation shows that the decrease in The fact that worldwide bioenergy production as well as
green house gas emissions from a combine are in the number of bioenergy products offered is increasing
proportion to the increase in output of CO2 neutral is a result of changing demand, which in turn offers
energy products. However, the CO2 reduction per used new business opportunities. However, one of the great
quantity of biomass is higher in conventional combined issues with large-scale production of bioenergy
heat and power production as long as marginal products is the growing concern over the negative
electricity is related to high CO2 emissions. Also the externalities (social and environmental aspects as well
economic evaluation show ambiguous results: two as resource efficiency). Since energy production and
cases had negative net present value even for low consumption shows strong path dependence [1], there
discount rates, while the two other cases showed to be is an urgent need to develop and establish production
more economically robust. In addition to this, a more technologies that help minimize the negative
detailed analysis of the industrial conditions for the externalities. Utilizing the taiga and deciduous forest
integration shows a need for achieving a fit regarding resources in the Northern hemisphere for this purposes
several operational, strategic and economic is, arguably, a promising alternative. The majority of
circumstances for this type of business ventures. Two these natural resources exist in harvested forests,
important conclusions that can be drawn from this is typically found in regions with, or suitable for, district
that: 1) not all district heating systems are suitable for heating.
bioenergy combines 2) there are many barriers for a
This paper investigates the prospects of using district
wide spread adoption of bioenergy combines.
heating production as a base for bioenergy production
and its potential to become a wide spread technology.
INTRODUCTION For this purpose, we use data from four existing district
District heating is a technology that receives increasing heating companies to which a bioenergy production
interest as it has great potentials in several ways. One unit is fitted. By acknowledging the complexity of this
unique characteristic of the district heating technology integrative business venture, it is possible to get
is the use of low temperature energy flows for large credible assessments of the magnitude in energy
scale energy distribution. In contrast to other energy efficiency, environmental gains and economic profits.
transformation technologies (e.g. condensing power or Equally important is the possibility to detect potential
distributed gas heating), district heating can interact limitations for bioenergy combines to become a
with energy flows that otherwise do not have any complement to district heating. Finally, conclusions are
alternative use (e.g. industrial residual heat). Although made to acquire clues to important restrictions to a
this is one of the competitive advantages of the wide spread adoption.
technology and a fundamental platform for its business
model, this can further enhance the scoop of the RESEACH DESIGN
business: by backward integration it is possible to
We argue that prospects for becoming a future growth
increase profitability in other industrial processes with
industry are dependent on the environmental benefits,
waste heat as a by-product.
economic attractiveness and fit with existing business
context. Hence, these three aspects of joint production
143
th th
are analysed. The environmental benefits are analyzed Below follows a description of the environmental and
with a system perspective on greenhouse gases (GHG) economic evaluation procedure. It is important to stress
emissions, taking into account both on and off site that the input data for these assessments only include
consequences of introduction of an energy combine; the change resulting from the integration of the
see Environmental evaluation below. Moreover, the bioenergy production. One implication of this approach
resource efficiency in the form of CO2 reduction per is that the environmental benefit of the heat produced
used quantity of biomass is evaluated for each (for district heating) is not included, since one base
combine. condition is that the heat deliveries are the same with
and without bioenergy production. Another implication
The economic benefits of the joint production set up
is that production units in the district heating system
are analyzed through both a short and long-term
that are not affected (e.g. base load and peak load
commercial lens. By using discounted cash flow
production units) are not included. This system
techniques as a base for this analysis, it is possible to
boundary is also pervading for the Description of the
account for both the yearly consequences as well as
cases to follow.
long term economic value; see Economic evaluation
below.
Description of the cases
Fit with existing business context is analysed with The four district heating systems with reference and
respect to input/output markets, production and system combine cases, respectively, are presented in brief
configuration and general business conditions below. The four objects for the evaluation are also
dominant in the host industry. The analysis focus on summarized in Table I. A more comprehensive
restrictions for short term fit; see Business context description can be found in ref. [3].
evaluation.
Table I. Overview of the reference and combine cases in
Since the detailed characteristic of the district heating the four district heating systems. Economic and energy
system is paramount to the feasibility for integration of data are given for both the reference and combine case,
bioenergy production, we base our investigation on four separated with a slash (ref./combine).
real district heating systems in Sweden with different CONFIGURATION
compositions. The chosen systems are all of equal size
(500-600 TWh of yearly heat deliveries) established in 1 2 3 4
towns with 40 000 to 80 000 inhabitants. These Heat deliv.

500 530 560 620
systems are in turn equipped with a bioenergy (TWh/y)
production unit that best suits ruling company strategy Ref. inv. Bio CHP None Bio CHP Bio CHP
as well as operational characteristics and maximizes
Combine Enzymatic Acid Gasi-
energy efficiency. In order to capture the additional Pyrolysis
technology hydrolysis hydrolysis fication
values of these investments, evaluation of each
Products Bio oil Ethanol1 Ethanol FTdiesel2
combine configuration is made in relation to a
reference case consisting of the existing system ECONOMIC DATA, reference/combine
(complemented with investments to maintain a
1 2 3 4
comparable level of production quality). The reference
and combine cases are further described in the Inv. (M) 74/60 0/144 116/310 146/473
Description of the cases below. O&M (M/y) 2.3/2.8 0/8.8 3.6/15.8 6.1/11.1
Much effort was put into indentifying efficient technical ENERGY CONSUMTION, (GWh/year), ref./combine
solutions that best take advantage of the site-specific 1 2 3 4
conditions in each system. This work included
Biomass 397/244 730/1537 470/1271 362/2970
everything from choice of equipment, appropriate size
of the integrated production unit and production Others 74/1353 - - -
strategies over the year regarding output of heat,
ENERGY PRODUCTION (GWh/year), reference/combine
electricity and other energy products. To identify
efficient technical solutions an integrative computerized 1 2 3 4
process was applied, including both the district heating Electricity 125/0 218/209 145/55 99/78
simulation software MARTES [2], and detailed spread
Biofuel 0/90 0/444 0/294 0/1336
sheet calculations. In order to guarantee high quality
4
input data, representatives from these four companies Others - - 0/384 -
1
gave access to technical, environmental as well as 2
Besides ethanol also biogas and pellets is produced.
Also kerosene and nafta is produced.
economic data. 3
Fuel oil (21/15) and industrial waste heat (53/120).
4
Biogas (0/114) and Pellets (0/270)
144
th th
System 1 benefit in accordance with marginal electricity

In the current configuration of this system 15-20% of production.
the energy demand is covered with fuel oil, which
needs to be reduced. One interesting option could be GHG GHG
to convert biomass into bio oil by pyrolysis and then
Production,
use the bio oil in the existing oil boilers. Bio oil that is Power
distribution and
not used within the system can be sold (e.g. summer use of biomass system
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. DH system with
or without bioenergy GHG
System 2 production
In this system, there is no need for new production
units, rather there is a high production capacity,
Production,
allowing for integration of a bioenergy production unit. Direct GHG emissions distribution and use
System 2 has good access to biomass, but might have of transportation
difficulties to find a market for large quantities of by- f uel
products. Based on these prerequisites, a suitable
combine technology could be cellulose ethanol Fig.1. Illustration of the applied system approach for
assessing the changes of GHGs.
production with enzymatic hydrolysis aiming at high
yield and in-house use of energy by-products. In the assessment, all GHGs of significance are
Regarding the O&M cost for the enzymatic process in included [3]: carbon dioxide (CO2), dinitrogen oxide
Table I, future enzyme price are assumed [4], With (N2O) and methane (CH4). For all energy carriers, life
todays prices, the enzymatic process will not be cycle emissions are considered, i.e. both combustion
profitable. emissions and well-to-gate emissions such as
emissions from fuel extraction, processing and
System 3
transportation. Also leakages are considered when
In System 3 there is a need for new production
applicable. How the GHGs for the relevant energy
capacity, which is represented by a bio CHP in the
carriers are assessed are described in brief below, a
reference case. This system has good access to a
more thorough description can be found in [3].The
large energy market, which enables output of other
adopted life cycle GHG emissions associated with
energy products. Hence, a cellulose ethanol plant
changes in consumption/production of the energy
based on acid hydrolysis can complement the
carriers are summarized in Table II.
reference case investment to build up the combine
case. Table II. Emission factors for included energy carriers.
System 4 ENERGY CARRIER LIFE CYCLE EMISSION

This system is in many aspects similar to System 3, but (kg CO2 eq./MWh)
ethanol production is not in line with company strategy. Biomass 14-17
1
Moreover, System 3 has good access to peat, which

High emission elec. (E1) 800
could supplement biomass for a large scale production
Low emission electricity (E2) 260
unit. Hence, gasification of biomass for production of
synthetic biofuel is evaluated for this system. Pyrolysis oil 292
Ethanol 307
Environmental evaluation FT diesel 277
The assessment of the environmental implication of Fuel oil 312
introducing a bioenergy production in an existing Biogas 207
district heating system focuses on changes in
Pellets 286
emissions of green house gases (GHG). A system 1
approach for analysing the changes of GHGs is The lifecycle emission of biomass is dependent on how
applied. This means that besides changes of the direct the biomass is used in the energy combines (e.g.
hydrolysis for fermentation or gasification)
emissions on site, also the changes of emissions in
affected parts of the energy systems are included; see Biomass
Figure 1. For instance, production of biofuel in the The energy input in all four combines is in the form of
combines ads to the environmental benefit since fossil biomass. Production, distribution and use of biomass is
fuels can be replaced, while reduced electricity related to GHG emissions. The GHG emission from the
production has a negative impact to the environmental use of biomass differs depending on how the biomass
is used. Combustion raises emissions of both methane
145
th th
and N2O (the CO2 emission are assumed to be neutral Biogas and pellets
from a climate perspective), while hydrolysis and In the energy combine of System 3, also biogas and
fermentation is not assumed to raise these emissions. pellets are produced. The biogas is assumed to be
Hence, the net lifecycle emission of biomass differs used as a transportation fuel to replace both petrol and
between 14-17 kg CO2 eq./MWh fuel. diesel. The net GHG reduction for replacing fossil
transportation fuel with biogas is set to 207 kg/MWh
Electricity
including life cycle emission and gas leakage in the
In all district heating systems, the electricity production
production. The pellets are also assumed to replace
decreases as a consequence of introducing the
fossil fuel, in this case oil with a net GHG reduction of
combine (see Description of the cases). Any change in
286 kg/MWh pellets.
electricity production is assumed to be compensated by
changes in marginal electricity production. For Resource efficiency
instance, if the electricity production decreases by 85 With the emission factors in Table II and the energy
GWh/year, it is assumed that other producers will flows of the reference and combine case in Table I, the
increase their production by 85 GWh/year. To assess environmental benefit of the energy combine can be
the environmental impact of this, the decrease has to assessed. However, if biomass is assumed to be a
be multiplied with a emission factor for marginal limited resource from a sustainability point of view, it
electricity. makes sense to evaluate the use of biomass from an
efficiency perspective. Hence, the resource efficiency is
There are many opinions regarding the emissions of
assessed as the net GHG reduction potential (in kg
marginal electricity. Here we have used a high and a
CO2 eq.) per used quantity of biomass (in MWh). By
low level, based on dynamic response for electricity
comparing this key figure for the reference case with
production with two different developments over a long
the combine case for each system, the resource
time period [5]. By using a high and low figure, the
efficiency of the combines can be evaluated.
impact and importance of changes in electricity can be
illustrated in a clear way. For the high figure, the
Economic evaluation
reference case in [5] is used where lifecycle emissions
of marginal electricity are about 800 kg/MWhel. This In order to analyze whether an investment adds
marginal electricity is denoted E1 hereon. With more financial value we rely on a standard discounted cash
stringent environmental targets the electricity flow (DCF) model estimating the net present value
production can be carbon lean [5] implying that the long (NPV) for each project so that:
term lifecycle emissions would be about 260 kg/MWhel,

n
NPV CFt / 1 r
t
denoted E2 hereon. (1)
t 0
Biofuel
As seen in Table I, the evaluated bioenergy combines where CFt denotes the net cash flow in year t, r is the
have various biofuel products as output. In System 1 future weighted cost of capital and n is the number of
pyrolysis oil is produced. The pyrolysis oil is assumed years included in the cost-/benefit analysis. The cash
to replace fossil fuel oil (but is categorized as an biofuel flow at year 0 indicates the initial outlay. Concerning r,
herein). If lifecycle emissions are regarded according to the weighted cost of capital (WACC), we do not
the approach in ref. [6] for both pyrolysis oil and fossil predetermine a specific hurdle rate; instead we analyze
fuel oil, the net GHG reduction for replacing fuel oil with value added for three different levels of discount rates.
pyrolysis oil is 292 kg per MWh of pyrolysis oil exported We do so because any statements on the actual
from the combine. Also the amount of fuel oil used riskiness of the project or an estimation of the WACC
differs in the combine case from the reference case in for the companies are outside the reach of this study.
System 1 (see Table I). The net life cycle GHG of this As stated before, when estimating cash flows the point
fuel oil is set to 312 kg/MWh. of departure is a reference object. That is, our NPV
calculations only address the differences in cash flows
In systems 2 and 3 ethanol is produced, which is
between the reference and the bioenergy combine; this
assumed to replace gasoline with net GHG reduction of
for two reasons. First, only the incremental cash flows
307 kg per MWh of ethanol reaching the market.
are relevant in a DCF analysis. For instance, in the
In System 4, three biofuels are produced: Fischer case of System 3 they already decided that they would
Tropsch (FT) diesel, nafta and kerosene. All three at least build a combined heat and power (CHP)
products are assumed to replace fossil transportation facility, and the question is if they gain from making
fuel with the net GHG reduction of 277 kg/MWh. The additional investments in a bioenergy production unit.
possible leakage of methane from the gasification Second, by focusing on the differences we do not need
process is assumed to be negligible. to consider the cost structure in the reference case, it is
treated as a given. Besides simplifying the analysis,
146
th th
academic access is facilitated as there is no need to can see for what potential price changes extra concern
reveal sensitive information. is warranted. Certainly, a drawback with the sensitivity
analysis is that it is just a ceteris paribus analysis and
Table III. Assumptions made for non-site idiosyncratic input
and output prices (/MWh). does not take into consideration the potential
covariance of variables, for instance between ingoing
Ethanol 78 Biomass 19 biomass and outgoing biofuel.
FT-diesel 78 Fuel oil 57
Kersone 78 Pellets 25 Business context evaluation
Nafta 52 Electricity 47 The environmental and economic analyses of a joint
Biooil 47 Electricity excise 0.5 production operation act as a starting point for the
1 business context analysis. A wide-spread adoption
Biogas 68 Electricity certificate 21
demands not only indications of environmental benefits
1
Premium paid to producers of renewable electricity. and economic profits, but must also offer a fit with the
existing business context. Even though the degree of fit
Cash flows
is defined on company level we will not analyze it as
The initial outlay is assumed to take place in full at year
such. Rather we use the business context of the
0. Yearly operational cash flows are projected by first
studied systems in order to put together a compilation
estimating an operational cash flow for the first year. As
of restrictions and barriers to a wide-spread adoption.
cash flows are the products of price and quantity, this
The magnitude and importance of these will give
estimation is based on the technical analysis in order to
important indications of the short term possibilities of
obtain energy flow estimates (see Table I), and then
realizing environmental benefits and economic profits
multiply them with price estimates, to which we add
in making bioenergy combines a future growth industry.
out-payments for operation and maintenance. We
The restrictions and barriers are identified through the
extrapolate this operational cash flow over the 20 year
fit with existing input/output market situation, production
long investment horizon with a three percent yearly
and system configuration and general business
growth rate (adjusted for the fact that green certificates
conditions, (i.e. strategic focus and capacity to absorb
are obtained for fifteen years only). All cash flows are
additional risk) dominant in the host company.
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 ENVIRONMENTAL BENEFITS
depreciation (according to Swedish tax code), changes As already stated in the Research design, the
in working capital (approximated by dividing the environmental benefit from integrating bioenergy
difference between in-payments and out-payments of production into an existing district heating system is
year t by 12 and subtracting the corresponding value assessed as the reduction of GHGs from a system
from year t-1, save for the last year where the perspective. As also explained, the net difference
difference is set to zero) and a terminal value (5% of depends on the reference case as well as the
the initial outlay). Initial outlays are determined by composition of the energy combine. In Figure 2, the
consulting [7] [19]. Our price assumptions for non-site GHG reduction for the included parts of the reference
idiosyncratic inputs and outputs are presented in case and energy combine case of System 3 is
Table III. For translation between different currencies displayed. In the reference case (left bar in Figure 2)
the following exchange rates were used: 9.6 SEK/ and a combined heat and power (CHP) plant biomass is
6.5SEK/USD. converted into heat (for district heating) and electricity.
Sensitivity analysis The amount of heat is the same in both the reference
We then control the robustness of the NPV estimates and combine cases and, hence, not considered in the
through sensitivity analysis; that is, we examine how evaluation of GHG reduction. However, the production
the cost-/benefit analysis is affected when changing a of electricity will change and the system consequences
variable at the time, holding all else equal. We do this of that is, as stated, considered by including two
in two steps for each system. First, we illustrate the different assumptions for marginal electricity. Assuming
changes in estimated NPV by changing yearly in- that marginal electricity is related to about 260 kg CO2
payments, yearly out-payments, initial outlay and eq./MWhel (E2), the electricity produced in the
terminal value respectively. Second, we show how reference case results in a yearly reduction of 38
yearly in-payments and out-payments respond to price Mtonne (dark blue bar to the left in Figure 2). If the
changes. emissions of marginal electricity instead is assumed to
be 800 kg/MWhel (E1), the emission reduction would
By this sensitivity analysis, we can to some degree
increase by 78 Mtonne/year (light blue bar) to be in
compensate for the uncertainty that surrounds our
total 116 Mtonne (dark + light blue bar = E1). The
estimates of initial outlays and terminal value, and we
handling of the biomass is related to GHG emissions
147
th th
(see Environmental evaluation) and, hence, there is a reduction from introducing an energy combine in
negative bar of 8 Mtonne for biomass. To sum up, the System 3 is 158 or 109 Mtonne/year depending on the
net GHG reduction in the reference case is 30 or 108 assumption for marginal electricity (E2 and E1,
Mtonne CO2 equivalents depending on assumptions for respectively).
the marginal electricity.
The equivalents to the right hand bar in Figure 2 for all
The combine case of System 3 has lower electricity four systems are shown in Figure 3. As can be seen,
production than in the reference case (see Description the reductions of GHGs are significant in systems 2-4,
of the cases). Consequently, the GHG reduction from especially if the electricity is associated with low
the electricity production is also lower, which is seen as emissions (E2, dark blue bar only). In System 1, the
lower dark and light blue bars for the combine case; environmental benefit is negative, even if the marginal
middle stacked bar in Fig. 2. Moreover, the negative electricity is CO2 lean.
bar for biomass is larger for the combine since more
Significant environmental benefits, as displayed for
biomass is used in this case. In the energy combine,
systems 2-4, are expected since the combines in these
however, bioenergy products such as biofuel (ethanol
systems use more biomass, which eventually replaces
in this system), biogas and pellets are produced. As
fossil fuel in the system approach applied (in system 1
already explained, these energy products are assumed
less biomass is used which explains the negative
to replace fossil fuels and the resulting GHG reduction
results for this system). However, if biomass is
from the combine is significant: 188 or 217 Mtonne CO2
assumed to be a limited resource from sustainability
eq. with carbon lean (E2) and carbon intense (E1)
point of view, the use of biomass should also be
electricity production, respectively.
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
GHG reduction (Mtonne CO2 eq./yr)
Net reduction (E2/E1): Electricity, E1-E2*

300
30/108 188/217 158/109
Electricity, E2
in Figure 4 for both the reference case and the
200 combine case for the four district heating systems
Pellets
evaluated.
100 Biogas
Ethanol 400
0 350 Net reduction (E2/E1): Others*
Biomass
GHG reduction (Mtonne)
300
-2/-69 124/119 158/109 321/309 Biofuel
* additonal emission 250
-100
reduction/change if 200
Elec., E1-E2
electricity is related 150
-200 to high CO 2 emissions 100 Elec., E2
Reference Combine Difference 50
0 Biomass
Fig. 2. GHG reduction in System 3 for the reference case, -50
-100 * biogas
combine case and the net difference for converting to the and pellets
-150
combine.
System 1 System 2 System 3 System 4
The dark blue bars are related to marginal electricity Fig. 3. Environmental benefit from introduction of energy
associated to low GHG emission (E2). The additional combines.
emission reduction/change if electricity is related to
high GHG emissions (E1E2) is indicated by the light As seen in Figure 4, the energy combines are less
blue bars. The total emission/change for E2 is given by resource efficient than the reference cases (generally a
the sum of light blue and dark blue bar. biomass fired CHP plant) if the marginal electricity is
associated with high CO2 emissions (E1, dark + light
The implication in terms of GHGs of integrating blue bar). However, if the marginal electricity is
bioenergy production in System 3 can be visualised by associated with low CO2 emissions (E2, dark blue bar
moving from the left bar in Figure 2 to the middle bar. only), the combines are more resource efficient than
Consequently, the difference of the two bars shows the the reference cases. As also can be seen, the resource
GHG implication of converting to an energy combine in efficiencies do not differ dramatically between
System 3, which is presented in the right hand bar in systems 24. System 1, however, shows lower
the figure. The change from the reference to the resource efficiency, which can be explained by the fact
combine case gives rise to GHG reduction from the fuel that a major part of the produced pyrolysis oil is
products (green bars) However, the electricity consumed internally in the system instead of replacing
production decreases, implying decreased reduction fossil fuel off site.
(emission increase) and, hence, negative bars for
electricity. As can be seen in the figure, the net GHG
148
th th
250 For robustness control purposes, sensitivity analyses

E1-E2
E2 are performed, here presented for System 3. Figure 5
200 illustrates the estimated NPV consequences from
(kg CO2 eq./MWh biomass)
changes in marginal cash flows, disaggregated into in-

Resource efficiency
150
payments, out-payments, initial outlays and terminal
value.
100
50
100
In-payments Out-payments
Change in NPV (M, 10% disc. rate)

50
0 Initial outlay Terminal value
Comb.
Comb.
Comb.
Comb.
0
Ref.
Ref.
Ref.
Ref.
-50
System 1 System 2 System 3 System 4 -100
-150
Fig. 4. Resource efficiency of biomass quantified as -200
GHG reduction per used quantity of biomass.
-250
ECONOMIC VALUE -300

-30% -20% -10% 0% 10% 20% 30%
Whether the cost/benefit analyses return positive NPVs
Change in cash flows
depend largely on the hurdle rates assigned to them. In
Table IV a summary of the economic results are Fig. 5. Estimated changes in NPV (M) for System 3 as
presented including the initial outlay, the expected free a result of percentage changes in cash flows assuming a
cash flow for the first year and estimated NPVs for 4, 7 10% discount rate.
and 10% discount rates, respectively. With the
exception of System 1, where the bioenergy combine is A percent change in either of these, results (ceteris
actually cheaper than the reference plant, marginal paribus) in a NPV change, as indicated in the figure. It
initial outlays vary between M 140 and 330, and is clear that the project is most vulnerable for changes
expected cash flows for the first year of operations in in-payments followed by out-payments. Assuming a
between M -3 and 57. The largest addition to existing hurdle rat of ten percent, a 20% average increase in
cash flow (both in absolute and relative terms) comes yearly in-payments would result in an increase in NPV
from the bioenergy combine investment in System 4. of about 100 million. Correspondingly, a 20%
increase in yearly out-payments result in a NPV
Table IV. Summary of cost/benefit analyses for adding a reduction of 84 millions. Fig. 5 also show that the
bioenergy combine to the reference investment in the cost/benefit analysis is not very sensitive to changes in
studied systems. initial outlay and leave no visible mark for changes in
1 2 3 4 terminal value. The order of importance of NPV impact
of cash flow changes are similar in the other three
Initial outlay (M) - 13.9 144 194 327 systems, where in-payments being the most important
Cash flow (My) -3.4 18.8 15.7 57 ones.
NPV (M) for different discount rates
4% -40 76 -62 362 25%
Change in marginal in-payments
20%
Ethanol Biogas Pellet
7% -27 29 -89 207 15%
10%
10% -19 -4 -108 101 5%
0%
-5%
As also can be seen in Table IV, only two projects are -10%
value adding at a 4% discount rate, and System 4 is -15%

-20%
the only one that can bear a 10% discount rate. The -25%
results for System 1 are a bit upside down, since -30% -20% -10% 0% 10% 20% 30%
compared to the reference case the investment cost Price change
and net cash flows are negative for the combine. Fig. 6. Estimated percentage changes in in-payments for
System 3, perhaps being the weakest of cases System 3 as a result of percentage changes in input
analyzed, will not show positive figures for any positive prices.
discount rate.
149
th th
Having established the sensitivity to changes in cash biomass, since increased use of biomass implies
flows it follows naturally to examine also to what increased output of CO2 neutral energy products.
degree different cash flows changes with respect to However, from a resource efficiency point of view,
changes in underlying prices. In Figure 6, the relation biomass should not be used to replace transportation
between marginal in-payments and prices of ethanol, fuel as long as the marginal electricity is related to high
biogas and pellets are shown for System 3. It is clear CO2 emissions. One important explanation to the
that ethanol is by far the most important bioenergy coherent environmental profiles of the different
product, where a 20% increase in prices renders a 12% bioenergy combine solutions is similar resource
increase in in-payment. efficiency for the four technologies evaluated. Hence,
our results suggest that it is possible to find different
energy combine with similar resource efficiency.
20%
Change in marginal out-payments
15%
Biomass Electricity O&M
However, these similarities in resource efficiency do
10% not indicate similarities in economic attractiveness. In
5% fact, the economic evaluation seems to suggest that
0% some bioenergy production technologies are not
-5% currently economic viable for integration with district
-10% heating system. Furthermore, the results indicate that
-15%
not all district heating systems are suitable for
-20%
integration with a biofuel production unit. Despite being
-30% -20% -10% 0% 10% 20% 30%
Price/unit cost change of the same size, use the same raw material and being
evaluated only on marginal effects on the economic
Fig. 7. Estimated percentage changes in out-payments for situation, differences in district heating system
System 3 as a result of percentage changes in input
characteristics have a profound impact on the
prices/unit costs.
economic possibilities of energy combine integration. In
this study we have matched every system with a
Similarly, Figure 7 shows how out-payments vary with combine solution in order to maximize the site-specific
input prices. Inputs included in the figure are biofuel, opportunities in each system. This opens of course the
operations and maintenance (O&M) and electricity7. possibility that there exist other matches with less
Not surprisingly, biofuel is the key input, where a 20% resource efficiency but higher economic profitability.
price change results in a 10% change in out-payments, Even if this can be the case, we would like to point out
which in Figure 5 translates to a 42 million change in that one of the starting points of this study was to base
NPV. in-data on the conditions of real systems. This includes
The sensitivity analyses of System 3 show that minor taking various kinds of restrictions into consideration.
changes in underlying factors can result in significant Even though these restrictions vary, the ones
changes in the NPV estimates. However, a not prominent in this study can be grouped into four
insignificant part of the indicated variability in cash different categories:
flows should be hampered by the offsetting effects Proximity to input resources
driven by the probable covariance between prices for Proximity to customers or infrastructure for
biomass and bioenergy products. To be noticed is that transporting the finished products
the order of importance of the inputs in the other three Existing production and system configuration
systems show a similar ranking, where biofuel and
Dominant business conditions
biomass price being the two most important ones.
Proximity to input resources
FIT WITH EXISTING BUSINESS CONTEXT Some combine solutions (such as the one for System
4) demand huge amounts of biomass. This requires
The environmental and economic evaluations indicate large areas of regional biomass recourses and little or
that the integration of bioenergy production into no competition over it. Import by sea is an alternative
medium sized district heating systems can be but it requires production sites close to a harbour.
associated with both environmental and economic
benefits, but the picture is mixed and ambiguous. From Proximity to market for the finished product
The production of biogas is one example of both the
an environmental point of view, the results are coherent
importance of proximity to customers and to
across all systems: the absolute environmental benefit
of bioenergy production is in proportion to the use of
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.
150
th th
infrastructure. Only relying on local demand for biogas Yearly Cash Flows (M)
is considered too challenging at present time.

100
Existing production and system configuration 80

Biofuel
Investments in bioenergy combines are seldom green Free cash flow

Biogas
field but, as we have shown earlier, have to be adapted 60 Pellets
to suit existing heat volumes, demand curves, system In-payments Out-payments

Biomass
configurations and also production site layout. In one of 40
Industrial waste heat
the systems, the production site was too small to house
the large amounts of biomass necessary for achieving 20
Electricity
an economic profitable size of an ethanol operation. Electricity certificate
0
Dominant business conditions O&M
The results of the study show that two business areas System 1 System 2 System 3 System 4
-20
have an evident influence on the type of bioenergy
combine investments the companies carry out: 1) the Fig. 8. Marginal cash flows (in-payments/out-payments)
strategic framing of the district heating company and 2) for each system in comparison to free cash flow from
the risk that these investments innate. Concerning the existing operations in 2007 (shaded bar).
first, many of the municipally owners use the utilities to
enhance and to some extent even realize the The considerable positive free cash flow of system 2
environmental visions that are formed and expressed from its existing operations is explained by the
on the political level. Examples of these found among companys sell of hydropower. Although irrelevant for
the companies represented in this study include; the value of this investment, it could function as a
phasing out fossil fuels, use of local waste resources general safeguard against negative results, due to
and visions of a fossil free cities based around locally unfavourable relation between biofuel and biomass
produced bioenergy fuels. When present, strategic prices.
framing has a visible effect on limiting the number of
The investment in system 1 was not profitable
available alternatives for integrates production.
according to the valuation earlier. Despite this, it is
As stated, the second area that has an significant worth pointing out that the risk of this investment
influence on the type of bioenergy combine that these should be low since it uses its own products as input. It
companies consider is the risk that these investments too has, relatively speaking, a strong free cash flow
innate. Due to the municipal ownership, these from its current operation that will decrease the risk of
companies are inherently dependent on stable ending up in the red.
business conditions. The ability to absorb negative
results is strongly limited. The added business risk of CONCLUSIONS
bioenergy production must, if needed, be able to be
absorbed by cash flows from existing operations or a The results of the bioenergy combine analyses show
strong capital base. In principle, this can be done in two that there are indications for both environmental gains
ways, either by keeping the investment relatively small, and added economic value of such investments.
or by only accepting business propositions with cash However, these benefits seem to be limited by several
flows that can be made relatively stable. operational, environmental and economic
circumstances present in these systems. First, these
In Fig. 8, the operational risk of the investment can to investments are dependent on the need for making
some extent be visualized by the size of the marginal major changes in current production layout, typically
cash flows of the different investments. The investment the need for new or altered production plants. This
in system 4 stands out not only because it is the largest limits the available window of opportunity. There are
one but also because its in-payment comes from one also several limitations related to operational
source only. If the price correlation with biomass is characteristics, availability of input resources and
high, this might not be a large problem. However, it is suitable product markets. A closer investigation of
interesting to note the relatively small positive cash flow existing governance situation also shows that these
available from existing operations in Systems 4, and investments often are made to fit owner strategies
also for System 3. If the company carries through with regarding environmental goals of the local energy
the evaluated investment, it will dramatically change its system. Finally, the municipally ownership typically
operational risk profile and over-all business focus. 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.
151
th th
Never the less, it will hamper the potential for wide- [6] IVL, Miljfaktabok fr brnslen, IVL Rapport B
spread adoption of bioenergy combines. 1334B-2 (2001).
These circumstances lead us to conclude that not all [7] Svebio, Kraftvrmeutbyggnad 2007-2015, Svebio
biofuel production technologies are suitable for all repport 2008-03-31.
district heating system. Our economic analyses also
indicate that not all district heating systems are suitable [8] H. Hansson, S-E. Larsson, O. Nystrm, F. Olsson
for bioenergy combine production. In fact the barriers and B. Ridell, El frn nya anlggningar - 2007,
are so many that it is reasonable to assume they will Elforsk repport no 07:50 (2007).
effectively reduce the number of systems adopting this [9] M. Zakrisson, Internationell jmfrelse av
operational design in the near future. produktionskostnader vid pelletstillverkning,
Masters thesis no 29 2002, SLU.
ACKNOWLEDGEMENTS
[10] A. Hang and S. Ilic, En frstudie fr bioetanol
The main funding for this project is provided by produktion i Bors, Masters thesis at Institutionen
Fjrrsyn, which is a research program organized by the Ingenjrshgskolan, Hgskolan i Bors (2008).
Swedish district heating branch agency. Additional
[11] M. Lantz, Drivmedelsproducenters
funding is also received from the project Pathways to
betalningsfrmga fr energigrdor, Milj- och
Sustainable Energy Systems.
energisystem, LTH (2006).
We kindly thank the representatives from each district
[12] J. Benjaminsson and A. Dahl, Uppgradering av
heating system for a good cooperation and for
biogas, Presentation at Temadag uppgradering
providing us with technical and economic data of their
av biogas, Gteborg (2008).
systems. Without these inputs, the work would not
have been as solid as it is. [13] I. Granberg, Project leader at Jnkping Energi,
Personal commication (2008).
We also thank Karolina Nilsson and John Jonsson
(both at Profu) for their valuable contribution to the [14] M. Tijmensen, A. Faaij, C. Hamelinck, and M. van
work. Hardeveld, Exploration of the possibilities for
production of Fischer Tropsch liquids and power
via biomass gasification, Biomass and Bioenergy
REFERENCES
2002, Vol. 23.
[1] M. Odenberger, F. Johnsson, Pathways for the [15] I. Johansson, S. Larsson and O. Wennberg,
European electricity supply system to 2050, Int. J. Torkning av biobrnslen med spillvrme,
of Greenhouse Gas Control, 2010, Vol. 4:2, pp Vrmeforskrapport 881 (2007).
327-340
[16] E. Sandvig, G. Walling, R. Brown, R. Pletka, D.
[2] J.Sjdin and D. Henning, Calculating the marginal Radlein, and W. Johnsson, Integrated Pyrolysis
costs of a district-heating utility, Applied Energy, Combined Cycle Biomass Power Systems,
2004, Vol. 78:1, pp 1-18. Repport of Alliant Energy, Iowa, USA (2003).
[3] E. Axelsson, C. Overland, K. Nilsson, and A. [17] H. Thunman, F. Lind, and F. Johnsson Delstudie
Sandoff, Bioenergikombinat i fjrrvrmesystem, energikombinat, Elforskrapport, 2008.
Fjrrsynsrapport 2009:11.
[18] NREL, Research Advances Cellulosic Ethanol,
[4] T. Brandberg, Senior researcher at SEKAB E- NREL (2007).
technology, Personal communication, 2009.
[19] P. Sassner, M. Galbe, and G. Zacchi, Techno-
[5] H. Skldberg and T. Unger, Effekter av frndrad economic evaluation of bioethanol production from
elanvndning/elproduktion. Elforsk report (2008). three different lignocellolosic materials, Biomass
and bioenergy 2008, Vol 32.
152
th th
SEA WATER DISTRTICT COOLING FEASIBILITY ANALYSIS FOR TALLINN

1 1 1 1
A. Hani , I. Britikovski , H. Voll and T.-A. Kiv
1
Tallinn University of Technology, Department of Environmental Engineering, Estonia
ABSTRACT also important to locate the district cooling station near

to energy source.
In this paper sea water district cooling feasibility
analysis for Tallinn is presented. It has become more Typical SW district cooling system principle is indicated
and more interesting to study alternative solutions for in Figure 1. The system consists of three main
public buildings A/C cooling due to relatively high sections:
electrical energy prices. Besides economical aspects
Cold sea water pumping;
technical and environmental sides must be considered.
Cooling plant with heat exchangers;
Standard cooling distribution network.
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. Fig. 1 SW district cooling principle schematic
Operation started in 1989 in Baerum, near Oslo. The
first system in Sweden was built in 1992 in Vsters [2] Heat exchangers allow usage of the soft water in
and since then the district cooling in Sweden has distribution network while problematic salty sea water
developed rapidly. Since the 1990s the establishment handling will be done in open central circuit.
of commercial district cooling systems has increased
Environmental impact study is required before any of
rapidly worldwide. Nowadays, more than 20 countries
the projects will be executed. Large sea water
have a commercial district cooling system and this is
quantities have to be available to minimize pumping
expected to increase rapidly [4].
impacts. In addition to evaluation of the deep zone cold
The sea water (SW) district cooling is based on large water pumping, the analysis of recycling the sea water
natural cold water source. Enough cold water is back to lower sea water zone with higher temperatures
accumulated in lakes, seas, oceans, rivers, etc [8]. should be made.
Lowering the coolant temperature with sea water is an
Following factors shall be considered before system
alternative to conventional electrical energy consuming
design [7]:
chillers [5]. The system working principle is quite similar
Minimum altitudes between heat exchangers and
to geothermal energy production which is used in
water resource level should be designed;
heating systems [6]. Until now the sea water district
Centralized district cooling plant (heat exchangers,
cooling is quite conservatively expanded around the
pumping station and chillers) is less expensive
World.
than decentralised system;
Centralized system has less maintenance
SW DISTRICT COOLING PRINCIPLE problems.
The temperature in conventional cooling water network
is between +4+7 o C so applicable the sea water DESIGN PARAMETERS
temperature should be below +5 oC. Despite that Temperature of the sea water varies during different
compressor based cooling can be used in case cooling seasons and distance from the coast.
water temperature is higher than mentioned [9]. It is
153
th th
In following Table 1 and Figure 2 the relations of the cooling network with total capacity of 19 MW. Project is
SW parameter can be found. 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/m2 (building no 17
cooling load 60 W/m2). In calculations was not
considered residential area cooling load due to different
usage profile compared to public areas.
Tab. 2. Preliminary cooling demand calculation
Fig. 2 Temperature and SW depth relation
Build. Storeys Cooling
Build. Public
height, above 2 demand
Tab. 1. SW parameters no
m ground
area m
kW
Dist. Depth Annual Min Max 1 24 6 8764 1052
from (sea), aver. temp, temp,
o o o 2 24 6 18870 2264
coast, m m temp, C C C
3 24 6 1458 175
500 20 7,5-8,5 2,5 17,5
4 24 6 3564 428
1500 25 5,5-6,5 1,5 17
5 24 6 5780 694
3200 30 4-5 1 14-16
6 18 5 5198 624
4000 35 3,5-4 <1 8-15
7 11 2 2340 281
5500 40 3,5 <1 6-8
8 18 5 8775 1053
From previous studies has been found that cooling 9 24 6 2268 272
demand exceeds significantly when the outdoor
10 24 6 2430 292
temperature exceeds 16 C (see Figure 3).
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
Fig. 3 Ambient temperature and cooling power relation 21 22 5 3700 444

100% 19,2 MW
CASE STUDY TALLINN COSTAL AREA
85% 16,0 MW
The study was carried out to research feasibility to
build to the Tallinn costal area cooling plant and district
154
th th
Tab. 3 Main technical parameters for design the system case sea water temperature is below 5 oC. Maximal
pressure drop in both circuits is selected 0,85 bar. Heat
Max cooling demand 19,2 MW
exchanger parameters are indicated in Table 5.
o
Ambient temp.calc. 27 C
Tab. 5 Free-cooling heat exchanger parameters
Simultaneous factor 0,85
Heat exchangers capacity 5x3600 kW
Cooling station capacity 18 MW o
Sea water (SW) supply temp. 4,5 C
Annual average cooling consumption 21600 MWh o
SW return temp. 10 C
o
Supply water temp 6 C
SW flow 130 l/s
o
Return water temp (max consumption) 16 C o
District cooling supply temp. 6 C
o
Return water temp (min consumption) 13 C o
District cooling return temp. 16 C
District cooling flow 72 l/s

Water chiller cooling
Max pressure drop 0,85 bar
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
Tab. 6 Coolant parameters
consumers interest and cooling energy demand.
System contains four 4500 kW water chillers with Sea water (SW)
centrifugal compressors. It is possible to adjust the SW temp 1,5-18 C
o
cooling power of the unit between 3004500 kW which

makes the system more energy (el) efficient during the Max pressure 6 bar
partial load period. The condenser has to be produced District cooling liquid
from titan or similar resistant material due to fact that it Temperature 10-18 C
o
is being cooled with sea water. In the following Table 4

are indicated technical parameters for water chillers. Max pressure 10 bar
Tab. 4 Water chiller parameters Sea water cooling

Cooling power 4x4500 kW The sea water is supplied through insulated 800mm
pipes to pumping station using sea water gravity. Three
Refrigerant R-134a
pumps (max 1080 m3/h) with frequency converters are
o
Condenser temp. 28 C installed using parallel scheme to suction pipe. Sea
o water pressure is ca 1,5 m and pumps will add 2 bars
Seawater (SW) supply temp. 18 C
to overcome self-cleaning filters, heat exchangers and
o
SW return temp. 24 C condensers pressure drop. Frequency converters are
SW flow (each unit) 215 l/s used to lower energy consumption during partial load.
o
Evaporator temp. 3 C Cooling plant operation modes
COP full load 7 Cooling plants are designed to have three different
operation modes:
COP partial load 12
SW temperature < 5 oC. Completely free-cooling;
o
District cooling supply temp. 6 C
SW temperature 512 oC. Pre-cooling with SW +
o
District cooling return temp. 16 C compressor cooling;
District cooling flow (each unit) 115 l/s SW temperature > 12 oC Only compressor cooling
(free cooling heat exchangers are equipped with
bypasses).
Free-cooling
When sea water temperature is lower than return District cooling network
temperature from the network free-cooling through heat Supply (forward) water temperature is designed 6 oC.
exchangers can be used. Optimum logarithmic
temperature difference shall be app. 1,5 oC. Five heat Return water temperature between 1316 oC (see
exchangers with capacity of 3600 kW are selected, Figure 4).
which assure whole cooling plant capacity (18 MW) in
155
th th
Due to fact that summer period soil temperature in The cooling plant shall have three operational modes:
1,5 m depth is 10 oC it is not necessary to insulate the
Free-cooling;
return pipe of the district cooling network. Supply pipe
Pre-cooling + compressor cooling;
is insulated with 10 cm nowadays heat insulation
Compressor cooling.
material.
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.
Fig. 4 District cooling network temp [3] Euroheat and Power, (2003). District Heat in
Europe Country by Country/2003 Survey. Brussel
Belgium.
CONCLUSION
[4] Mildenstein, B. S. P, (1999). District Heating and
The sea water (SW) district cooling has until year 2000 Cooling Connection Handbook.
quite modestly developed among different countries
[5] Gosney. W.B, (1982). Principles of Refrigeration.
around the World. Due to the fact that energy prices
Cambringe University Press. Published by the
have raised rapidly more and more researches for free
press syndicate of the University of Cambridge.
energy resources are carried out. Wind power, heat
pumps, solar energy and sea water have obtained [6] Westin, P. E. H., (1999). Production Technologies
huge attention. in District Cooling Systems and the Importance of
Local Factors. New Energy Systems and
SW district cooling is centralized and will have
Conversion-NESC 99.). pp 6.Osaka.
advantages like less pollution, less maintenance
problems and in perspective also economic benefits. [7] Westin, P. E. H., Karlson, B., and Lundqvist, P,
(1999). Straategies and Methods For Increasing
Current feasibility analysis was done in Tallinn costal
the Capacity of District Cooling Systems.20th
area to define possible cooling plant load, potential
International Conferenss of Refrigeration, IIR/IIF.).
consumers and technical possibilities.
pp 1-8. Sydney.
Due low costal area it is possible to locate the cooling
[8] Nordell, B., and Skogsberg, K, (2002). Snow and
plant near to sea water. Further studies should add
ice storage for cooling applications.Winter Cities
some more economic aspects to the technical solution.
2002.Japan Aomori. Lule University of
Problematic is to develop the district cooling network in
Technology
Tallinn area (existing tunnels and subways will ease
the process). [9] Eliadis, C, (2003). Deep Lake Water Cooling A
Renewable Technology. Number of pages 3.
Most of the new built or renovated public buildings
have high cooling demand due to glass walls and high [10] Morris, A.P, (1995). The Road to Lockport:
internal heat loads. In present research 21 buildings Historical Background of District Heating and
with only public area were included (total cooling Cooling. Ashrae Transactions: Symposia.
demand 19,2 MW). The cooling demand rises
[11] Arvidson, J, Asplund, A-L, Birgerrson, E, (1997),
considerably when ambient air temperature exceeds
Cold production uning low temperature waste
16oC. Sea water temperature 5 oC can be found in
heat,. Kungl tekniska hgskolan Kemisk
depth of 3540 m.
apparatteknik. Pp 54, Stockholm
156
th th
ANALYSIS FOR THE OPERATION BEHAVIOR AND OPTIMIZATION OF CHP

SYSTEM IN DISTRICT HEATING AND COOLING NETWORK
1 1 1 2
Yong Hoon Im , Hwa-Choon Park , Byung-Sik Park and Mo Chung
1
Cogen. & Boiler Research Group, Building Energy Research Center,
Korea Institute of Energy Research, Korea
3 Mechanical Eng. Dept., Yeungnam Univ., Korea
ABSTRACT 12.3% on the basis of the total number of households

at the end of 2008 [1]. The annual heat sales, via DHC
A simulation program for analyzing the effects of the
network, in 2008 have reached 16,676 thousand Gcal
networking operation of existing DHC system in
and it increased by about 5% on average after 2001.
connection with CHP system on-site is to be discussed
Considering the trend of new-town development in
in this study. The practical simulation for arbitrary areas
metropolitan areas and newly developing residential
with various building compositions is carried out for the
areas on a large scale, it is generally expected to show
analysis of operational features in both systems, and
a clear increasing trend of DHC systems on the market
the various aspects of thermal network operation are
for the time being. Furthermore, the relevant changes
highlighted through the detailed assessment of
of circumstances such as the long-term expectation for
predicted results. The intrinsic operational features of
high prices of fossil fuels and the imminent realization
CHP prime movers, gas engine, gas turbine etc., are
of UNFCCC around the world will help the CHP and
effectively implemented by realizing the performance
DHC system tighten its grips on the forthcoming
data, i.e. actual operation efficiency in the full and part
heating and cooling market [2]-[3]. Among the several
loads range.
merits of DHC systems against separate heat & power
For the sake of simplicity, a simple mathematical (SHP) or central heating system, the distinctive feature
correlation model is proposed for simulating various of being able to construct the networking system with
aspects of change effectively on the existing DHC the neighbouring DHC systems certainly deserves to
system side due to the networking operation, instead of receive attention from the view point of efficient use of
performing cycle simulations separately. The empirical energy resources and operation costs reduction [4].
correlations are developed using the hourly based However, the effectiveness of networking operation of
annual operation data for a branch of the Korean CHP and neighboring DHC systems is strongly
District Heating Corporation (KDHC) and are implicit in influenced by the conditions of energy consumption
relation between main operation parameters such as behaviours and corresponding operation scenarios on
fuel consumption by use, heat and power production. In both sides. The different pattern of energy consumption
the simulation, a variety of system configurations are in new demand areas is highly desirable for creating
able to be considered according to any combination of synergy effects by networking operation. In addition,
the probable CHP prime-movers, absorption or turbo the different operation strategy of CHP system with that
type cooling chillers of every kind and capacity. From of DHC network can also improve the effectiveness of
the analysis of the thermal network operation networking operation. The optimal system configuration
simulations, it is found that the newly proposed of the CHP system with networking operation certainly
methodology of mathematical correlation for modelling differs from that of stand-alone CHP system not to
of the existing DHC system functions effectively in mention the operation characteristics. Since the heat
reflecting the operational variations due to thermal flows in the network are bi-directional, the appropriate
network operation. The effects of intrinsic features of modelling for the mutual effects on each system is
CHP prime-movers, e.g. the different ratio of heat and highly required for the accurate estimation of the
power production, various combinations of different networking operation.
types of chillers (i.e. absorption and turbo types) on the
The main purpose of this study is to examine the
overall system operation are discussed in detail with
feasibility of the network operation of the CHP system
the consideration of operation schemes and
on-site with the existing DHC system in terms of
corresponding simulation algorithms. The various
efficient use of primary energy and reduction of the
aspects of system configuration in terms of CHP
operation cost. In this study, a simulation program is
system optimization are also discussed.
developed for analysing the thermal networking
process between the existing DHC system and the
INTRODUCTION CHP system for the newly developing area. The effects
In Korea, the district heating and cooling (DHC) system of thermal networking on the existing DHC system
gains share of the market steadily and it amounts to operation are implemented using mathematical
157
th th
modelling with empirical correlations for main operative Fig. 2 and Fig. 3 show examples of the daily unit
parameters. The intrinsic features for the CHP prime energy load model of heating for the apartment and
movers is modeled using the actual performance data hourly unit energy load model of electricity for the office
of operation efficiency in full or part load conditions. building respectively.
The specific features of the newly developed program The annual hourly unit energy model can be obtained
in simulation of thermal networking process in district by synthesizing the daily and hourly unit energy load
heating is described in terms of the energy load models [5]. The final annual hourly energy consumption
prediction and operation simulation of various system for given building compositions and corresponding
configurations with CHP prime movers and types of scale is to be predicted with the input of the total areas
cooling chillers. The unit energy load model for various for respective buildings since the unit energy load
buildings by use, e.g. apartment, hotel, hospital, models have been developed by normalizing the
buildings for business and commercial use etc, is statistical energy consumption measurement data with
introduced for the accurate prediction of energy loads the corresponding building areas. The example of
for newly developing area. The effects of intrinsic annual hourly energy consumption for the apartment is
features of CHP prime movers, e.g. the different ratio of shown in Fig. 4.
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
Fig. 2. Daily unit energy load model for the apartment
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. 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
Fig.1. Relationship between load, operation and economic facility instead of performing thermodynamic cycle
analysis modules [9] simulations for respective facilities separately. In order
158
th th
to implement the schemes, the performance data for the number of units, and the load factor in terms of unit
the commercial products, operation efficiency in full capacity. When an option is selected by the user as
and part load condition, has been extensively described above, its corresponding technical data for
investigated and the database has been realized on the CHP product will be linked automatically in the
simulation program. subsequent operation simulation procedures. The
settlement of the system configuration for the cooling
One can consider a variety of CHP system
system can also be performed in a similar manner by
configurations with various CHP prime movers and
providing the data for the ratio of being in charge of
types of cooling chillers. If the type of CHP prime
turbo or absorption type chillers.
movers is being selected, the capacity of it is to be
determined in the form of any percentage on the basis
2. Modelling of DHC system for networking
of the maximum value of annual hourly electricity
operation
demand. Then, the feasible options, which can match
the condition entered by the user, are compiled In contrast with small cogeneration or CES system, the
according to the relevant algorithm as shown in Fig. 5. 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.
(a) Heating load
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
(b) Electricity load
Fig. 4. Prediction of annual hourly energy consumption for production in terms of energy consumption are given
the apartment 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)
Fig. 5. Parametric entry of option for CHP system product Annual, fuel consumption per day according to the
facilities of heat production (CHP, HOB)
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,
159
th th
The functional form of the mathematical correlation is correlations is a certain time, not a specific time during
given as follows, 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
F f ( H , P) (1) amount of Q , the heat load of DHC system can be
regarded to be changed from Q1 to Q2, i.e. Q2=Q1+ Q .
Where,
F: Fuel consumption Then, the operation behaviour for DHC system at the
H: Heating load moment can be estimated simply from the
P: Electricity load mathematical correlations by simply referring the value
of F2*, corresponding to Q2* and P* corresponding to
F2*. 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) Time vs. events
(a) Electricity production vs. fuel consumption
(b) Events vs. events

Fig. 6. Illustrative diagram for the correlation between
energy production and fuel consumption
(b) Fuel consumption vs. Heat production
Fig. 6 shows the illustrating diagram for the Fig. 7. Developed correlations for the energy productions
mathematical correlation between energy production vs. fuel consumption
and consumptions. For any time t1, an optimized
operation scenario already exists and corresponding
heat and electricity production, and fuel consumption SIMULATION OF THE THERMAL NETWORKING
has been fixed according to the operation scenario and OPERATION
for any time t2, it is the same as above. On the basis of
1. Operation Conditions and Schemes
the operation data for a year, the behaviour of system
operation can also be described between dependent The operation of the overall system should be carried
variables (e.g. F: Fuel consumption, H: Heat out by the order of priority of operation for the various
production, P: Electricity production). In the correlations heat sources. In this study, the basic schemes in order
between dependent variables, the time t is reflected of priority for supplying the energy demands in newly
with implicit manner and the meaning of time t in the developed area are established as shown in Fig. 8,
160
th th
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
(a) Case A
(b) Case B
Fig. 10. Comparison of energy load prediction: elec. Load
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 Fig.11. Comparison of energy load prediction: cooling load
system is performed for two distinct test cases of (Case B)
residential buildings only, and a group of non-
residential ones. 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.
(a) Case A
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
(b) Case B system.
Fig. 9. Comparison of energy load prediction: heating load
161
th th
(a) Gas engine (a) Gas engine
(b) Gas turbine

Fig. 12. Typical pattern of heating load and recovered (b) Gas turbine
waste heat for CHP prime-movers
Fig. 13. Annual thermal energy supply and demand
operating condition for newly developing area
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. (a) 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 (b) Gas turbine
produce the required amount of heat energy for newly
Fig. 14. Annual electricity supply and demand operating
developing area. However, it does not actually happen condition for newly developing area
because the operation of the HOB on -site has priority
over that of the HOB in existing DHC system according
to the operation schemes. 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
162
th th
the rate of operation for HOB on-site considerably. Fig. 15 shows the variations of electricity production on
From the view point of system operation efficiency, it existing DHC system side due to thermal networking
has a very positive impact in that the rate of operation operation. It is interesting to note that a minor increase
of CHP in DHC system increases to some extent. of the electricity production for existing DHC system is
However, in case of supply of surplus heat to existing observed during the intermediate seasons. This is
DHC system as shown for gas turbine, it is vice versa. caused by the increased rate of operation of CHP in
It is noted that the heat flow of thermal network can be existing DHC system due to thermal networking
bi-directional for the gas turbine as shown in Fig. 13. operation.
The annual supply and demand operating conditions The detailed variation of electricity production on the
for electricity are shown in Fig. 14. A comparatively existing DHC system side is given as shown in Fig. 16.
good electricity-tracking operation is observed for both The net increase of electricity production for gas
CHP prime-movers and the supply from the grid tends engines is larger than that of gas turbines. This is
to increase during the summer due to the peak of the because of the intrinsic feature for gas engine CHP
electricity demand. 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.
Fig. 15. Variation of electricity production on existing DHC

system side due to thermal networking operation
(a) Gas engine
(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
(b) Gas turbine schemes for LNG price in promotion of energy efficient
facilities such as CHP, cooling chillers based on
Fig. 16. Detailed variation of electricity production on the cogeneration system etc. the reliable estimation of LNG
existing DHC system side according to CHP prime mover consumption according to their usage is crucial for the
assessment of economics for the scenarios.
163
th th
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.
(a) Heat
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 (b) Electricity
means that a large amount of heat is flowing toward the Fig. 20. Variation of operating conditions due to thermal
existing DHC system side as shown in Fig. 19, and networking operation on the existing DHC system side
there will be serious effects on the operation of existing
DHC system.
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. 18. Heating load and recovered waste heat for gas
turbine CHP
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
Fig. 19. Annual thermal energy supply and demand production as shown in Fig. 20 (b), there is a minor
operating condition for newly developing area 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
164
th th
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.
(a) Absorption type 80%
Fig. 22. Heat balance of operating the absorption chillers

(b) Absorption type 20%
The operating characteristics for cooling load are Fig. 24. Heat balance of operating the absorption chillers
for different responsibility by absorption type cooling
described in the following with Fig. 22. It shows the
heat balance of operating the absorption chillers. The The heat balance of absorption chillers for different
cooling load exceeding the supply capacity from ratio of responsibility by absorption type cooling is
recovered waste heat is modelled to be covered by shown in Fig. 24. In case of 80% absorption type
providing auxiliary heat for absorption chillers by direct cooling, the recovered waste heat is not sufficient
gas combustion. The cooling load assigned to turbo enough to handle the assigned cooling load, so an
type chillers is dealt with as an electricity load auxiliary heat source, such as direct gas combustion, is
converted according to the COP of the corresponding needed to cope with the full absorption cooling load.
product of turbo chillers. 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. 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:
(b) Absorption type 20%
Fig. 25. Annual electricity supply and demand operating
condition for newly developing area for different
responsibility by absorption type cooling
165
th th
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 (b) Absorption type 80%
CHP system is given in the form of any percentage of
the peak value of electricity, i.e. the maximum value of Fig. 26. Annual LNG consumption rate by use for different
responsibility by absorption type cooling
the annual electricity consumption rate per hour.
Therefore, the criteria for defining the CHP capacity is
CONCLUSION
to be varied depending on the amount of cooling load
assigned to turbo type chillers. The respective LNG A simulation program that predicts the energy loads for
consumption patterns depending on the ratio of a mix of buildings and estimate the operational
absorption cooling load are compared in Fig. 26. It is characteristics for networking operation of existing
interesting to note that the composition of fuel DHC system with CHP system on-site is developed.
consumed by use is substantially changed according to The distinctive features of this simulation approach can
cooling load treatment during the summer. The results be summarized as follows,
confirm that the effects of various aspects of
The unit energy load models are developed for
configuration for CHP and cooling system on the
accurate prediction of energy consumption by use
prediction of operational parameters (e.g. fuel
accroding to any combiation of building type and scale.
consumption rate by use) are properly realized in the
simulation program. A simple mathematical correlation for reflecting the
variations of the network operation on an existing DHC
By using the simulation approach as presented in this
system side is newly proposed for the sake of simplicity
study, the optimal design of the CHP system in
and efficient simulation process.
networking operation with DHC system can be carried
out since one can access the detailed physical data The performance data for the commercial products,
regarding the whole operation of the network system operation efficiency in a full and part load condition,
such as annual rate of fuel consumption for respective has been extensively investigated and the database
systems (e.g. CHP, HOB, Chiller etc), annual has been realized successfully on the simulation
production of electricity, heat, and the amount of heat program.
exchange etc. Along with the appropriate cost
The operational characteristics of thermal networking
structures for fuel, product sales (heat and electricity)
operation has been assessed in terms of system
and the estimation of capital cost, civil construction,
configurations for the CHP and the cooling facility as
and O&M costs etc, one can also make the
follows.
assessement for the economic feasibility of various
scenarios. However, the detailed economic analysis for According to the intrinsic features of the CHP prime
the test cases and the procedures to determine the movers such as gas engine and gas turbine etc, the
optimized CHP system configuration based on it will aspects for the supply of surplus heat is progressing in
not be described in this paper due to the page different manners by and large. For a gas engine, the
constraints. These tems will be discussed in further on-site is almost short of heat so that the prediction
studies. 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
166
th th
side in the winter the rate of operation of HOB will be [4] H. Lund, F. Hvelplund, I. Kass, E. Dukalskis, D.
decreased. Blumberga, District heating and market economy
The thermal energy exchanges via the network and in Latvia, Energy, 1999, Vol. 24, pp. 549-559.
the corresponding changes in operation on both sides [5] H. C. Park, M. Chung, S. H. Kim, Development of
are prevailing in intermediate seasons in case of similar system simulator for community energy system,
heat consumption patterns on both sides. Report to Ministry of Industry, 2003.
The operation of cooling system on the newly [6] Y. H. Im, H. C. Park, M. Chung, A study of optimal
developing area is verified not to have much effects in heating supply systems for the newly developing
terms of thermal networking operation. However, the area in the vicinity of DHC system supplying area,
significant changes in the LNG consumption patterns Report to Korea District Heating Corporation, 2006
by use are observed according to the ratio of
responsibility by absorption chillers for the cooling load. [7] Y. H. Im, M. Chung, H. C. Park, Feasibility study
for small size cogeneration systems in the
The various aspects of system configuration in terms of metropolitan areas of Seoul, Final Report to SH
CHP system optimization are discussed with the (Seoul Housing) Corporation, 2008.
development of a simulation program in this study. It is
verified that the physical and mechanical mechanisms [8] M. Chung, H. C. Park, Development of a energy
concerned with the thermal networking operation has demand estimator for community energy systems,
been appropriately modeled from the assessment of Journal of the Korean Solar Energy Society, 2009,
operational behavior for test cases. Vol 29, pp. 37-44.
[9] M. Chung, H. C. Park, Development of a software
ACKNOWLEDGEMENT package for community energy system assessment
Part I: Building a load estimator, Energy, in
The author gratefully acknowledges the financial and
press.
technical supports for the research from the Korea
District Heating Corporation (KDHC). [10] H. C. Park, S. S. Lee, D. J. Kim, Development of
energy models for department stores, Korean
REFERENCES Journal of Air-Conditioning and Refrigeration
Engineering, 2003, Vol. 15, pp. 1088-94.
[1] Korea Energy Management Corporation, Statistics
[11] H. C. Park, M. Chung, Building load models for
for district heating and cooling enterprise in Korea,
hotels in Korea, Journal of the Korean Solar
2009.
Energy Society, 2009, Vol. 29, pp. 48-57.
[2] A. Marbe, S. Harvey, Opportunities for integration
[12] H. C. Park, Development of weighting factors for
of biofuel gasifiers in natural-gas combined heat-
variables associated with hourly energy
and-power plants in district-heating systems,
consumption pattern for hotels in Korea, SAREK
Applied Energy, 2006, Vol.83, pp. 723-748.
(Soc. Air-conditioning, Ref., Engineers of Korea)
[3] C. Weber, I. Heckl, F. Friedler, F. Marechal, D. Winter Annual meeting, 2002, pp. 76-82
Favrat, Network synthesis for a district energy
[13] H. C. Park, Analysis of energy loads for hospital
system: a step towards sustainability, Computer
buildings, SAREK journal, 2002, pp. 1088-93.
Aided Chemical engineering, 2006, Vol. 21, pp.
1869-1874.
167
th th
IMPROVED PRIMARY ENERGY EFFICIENCY OF DISTRICT HEATING NETWORKS BY

INTEGRATION OF COMMUNAL BIOMASS-FIRED COMBINED HEAT AND POWER
PLANTS WITH BIOMASS PYROLYSIS
1 2 1 1
T. Kohl , N.A. Pambudi , T. Laukkanen and C.-J. Fogelholm
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 Furthermore the scarcity of the biomass available

demands most efficient use of this resource.
This paper investigates the influence of the integration
of communal biomass-fired combined heat and power As shown in a previous study [1] it looks promising to
plants with wood-pyrolysis on the plants energy integrate biorefinery processes, that are linked to
balance and product distribution. Further the proposed transportation fuel production, with CHP plants, since
integration concepts influence on the environmental CHP plants can provide both a source for high
performance of the connected district heating network is temperature heat needed for thermal conversion of
pointed out. The environmental performance is biomass as well as the district heating network (DHN)
evaluated by means of the primary energy factor and as a sink for sensible heat that would usually be
the CO2 emission coefficient. For this evaluation, the rejected in stand-alone biofuel refineries. It has been
European standards EN 15603 and EN 15613-4-5 are further worked out that the integrated production of
applied and modified. interstage products, such as liquid fast pyrolysis product
(often referred to as woodoil) and wood pellets, have
The concept comprises the integration of a simple
several advantages: Firstly, the products are
pyrolysis model and of a steam dryer with a base case
independent from the transportation fuel market
combined heat and power plant. The yearly plant output
developments since they can be seen as a universal
is calculated by applying a multiperiod model of the heat
input for different upgrading processes to e.g. biodiesel,
duration curve. The work shows that, by co-generation
ethanol, methanol, hydrogen or other chemicals
of valuable pyrolysis product, operation hours and
production but they can also be directly combusted for
electricity production can be considerably improved.
power and heat generation. Secondly, they increase the
The integration also clearly improves the district heating
biomass energy density making it more sustainable for
networks primary energy efficiency and lowers its
transportation to central plants required for economic
carbon dioxide emissions significantly.
fuel production. Thirdly, technologies applied for such
pre-processing are relatively simple and robust, thus
INTRODUCTION keeping investment cost and system complexity on a
The European Unions carbon dioxide mitigation goals reasonable level and making it therefore also interesting
and plans to reduce energy import dependency require for local small-scale solutions.
action towards a more sustainable energy supply that is In this paper, outgoing from a base case, we simulate
based on renewable energy sources available in the the retrofit integration of wood fast pyrolysis with an
member states. Biomass is discussed controversially existing wood-fired CHP plant. The aim is highest
due to its wide range of upgrade possibilities from possible pyrolysis product generation using the free
power, heat, cooling to chemicals and transportation boiler capacity in part loads under the condition that the
fuels. Among others, EU directives 2001/77/EC district heat (DH) demand is still fulfilled.
(promotion of electricity produced from renewable
energy), 2004/8/EC (promotion of With help of a multiperiod model of the DHNs heat
cogeneration) and 2003/30/EC ( promotion of the duration curve, the work shows the influence of the
use of biofuels) state that the use of biomass for integration on plant operating hours, electricity
energy purposes should be expanded on a sustainable production and biomass throughput. In addition the
base. effects on the DHNs primary energy factor and CO2
emission coefficients are studied as well. The primary
However, the increased use of biomass is expected to energy factor and the CO2 emission coefficient are
raise prices for biomass which will negatively influence, calculated according to European standards EN 15603
among others, the economy of communal biomass-fired [2] and 15316-4-5 [3], applying a modified power bonus
combined heat and power (CHP) plants a technology method. However, no cost estimation is given, since the
that is currently competitive to fossil energy production. focus of the work was to find out if this integration
168
th th
concept is possible within the operational limits of the Heat Duration Curve - Multiperiod Model
CHP plant. 25
Integrated Case - Lower Loads
In the following, first the used multiperiod load model is 20
District Heat Load [MW]

described. Further a brief introduction to biomass fast
pyrolysis is given and it is shown how the process has 15
been simulated and integrated. Then the modification of

the European standards is explained, results are
10
presented and finally restrictions of the work and 5
options for further improvement are discussed.

0
0 50 100 150 200 250 300 350
DISTRICT HEATING LOAD Time [d]
The CHP plant chosen has been integrated into a virtual District Heat Load Real CHP DH Load
DHN. Therefore yearly data of a real DHN has been Multiperiod Model DH Load
scaled so that the CHP plant provides 60% of the hourly

peak demand of the DHN when on full load. The CHP Fig. 1b: DH Load Multiperiod Model Integrated Case
plant is assumed to be shut off at 50% load which
corresponds with 30% demand in the DHN. As stated in
CHP PLANT INTEGRATED WITH WOOD PYROLYIS
[4], those are common operating parameters for
communal solid fuel-fired CHP plants. Wood Pyrolysis Model
In order to represent the yearly production of the base Biomass fast pyrolysis is the thermal conversion of
case plant a multiperiod load model was developed. biomass in the absence of oxygen at temperatures of
One full load and five part load levels have been chosen approximately 500 C and pressures close to
to represent the heat duration curve. The pyrolysis atmospheric [5]. The basic idea of the pyrolysis unit is
integrated CHP plant is represented by 7 part load derived from the bioliq process developed by the
levels since lower DH loads can be supplied, as Forschungszentrum Karlsruhe (FZK). There, fast
explained later. Operating time periods per part load pyrolysis is applied in order to yield a high share of
level are set of equal length and -together with the full liquid pyrolysis product. Biomass is indirectly heated
load period- match the total operation hours and yearly and pyrolysed with sand in an inert atmosphere at a
DH generation of 94.5 GWh as shown in figures 1a and temperature of about 500 C. Subsequently, the
1b. For each load level, fuel input and pyrolysis yield are pyrolysis gases are condensed and the liquid fraction
then iterated matching the required DH output. DH (also referred to as wood oil) is mixed with the coke and
demand not provided by the CHP plant is assumed to forms the so-called bioslurry which leaves the plant as
be generated in oil-fired heat-only boilers with a thermal the final product. In this work we use data published by
efficiency of 0.85 FZK [6] and hence assume that 90% of the biomass
energy is converted into bioslurry whereas 10% accrues
Heat Duration Curve - Multiperiod Model in gaseous form. The pyrolysis gas is thought to be co-
Base Case fired in the boiler and hence its energy is subtracted
25
from the fuel input into the boiler.
20
As pyrolysis requires a low fuel moisture content of
District Heat Load [MW]
15
approximately 10% [5] a dryer must be integrated as
well. Indirect steam drying is applied, since this also
10 allows the regulation of the DH load. As explained later,
regulation is necessary since the enthalpy of the steam
5
flow after the modification exceeds the demand of the
DHN and hence must be adjusted.
0
The wood pyrolysis process is modelled as follows: The

0 50 100 150 200 250 300 350
Time [d]
heat of pyrolysis of wood is set to 1.87 MJ/kg (moisture
District Heat Load Real CHP DH Load
Multiperiod Model DH Load content 10%) using data for pine derived from [7].
Therewith the pyrolysis yield is calculated from the heat
Fig. 1a: DH Load Multiperiod Model - Base Case 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
169
th th
part load using the thermal power plant simulator Table 1: Base Plant Input Specification and Performance
Prosim. Performance data of the CHP plant and fuel at Design Load
input specification is derived from [4] and [8], Simulation model input data - design load
respectively and given in table 1. The higher heating Wood Fuel

Ultimate analysis C 50.64 O 42.22
value (HHV) is calculated by the simulation software.
H 6.10 N 0.16
Ash 0.8 S 0.08
Moisture 50 % HHV 18.8 MJ/kg
Fuel input 26 MW
Steam Cycle
High pressure steam 60 bars Condenser pressure 0.69 bar
510C
Plant Performance
District heat output 16.5 MW Electrical efficiency el 0.243
Power output 6.3 MW Power to heat ration 0.381
Fig. 1: CHP Plant Base Case

CHP Plant Integrated Wood Pyrolysis CHP plant Integrated Steam Drying
The modified CHP plant is illustrated in Fig. 2. In order The dryer is modelled as a steam tube dryer. Paying
to provide heat for the pyrolysis process, the heat must attention to the retrofit situation, live steam is extracted,
be extracted from the flue gases leaving the fluidized throttled to 10 bars and further cooled to 190 C by
bed reactor (numbered 3 in Fig.2) boiler at 850 C. The spraying in the saturated water leaving the dryer. Drying
required amount of flue gas is split off (18) after the of biomass to low moisture contents requires
fluidized bed reactor. As in the FZK process, those flue temperatures far above the saturation temperature at a
gases are thought to heat up sand to 550 C (which given pressure due to the hygroscopic properties of
would provide the heat for the pyrolysis process by biomass. Heat consumption of the dryer has been
cooling down to 450 C) [6]. The flue gas thereby is estimated to 2750 kJ/kg water evaporated [9]. Wood
estimated to cool down to 480 C. The flue gas is then and hot flue gases are led in the dryer (24). If heat is
mixed back (20) into the main flow before the available from the flue gases, those are cooled down to
economizer. The heat extraction needed for biomass 120 C and together with the fully condensing steam
fast pyrolysis process is modelled by help of an provide the heat needed for the drying process. Dried
additional evaporator (19). 90% of the biomass energy wood leaves the dryer at wet bulb temperature. For the
on a lower heating value base will form pyrolysis slurry drying process live steam is extracted (21), throttled
whereas 10% accrues as pyrolysis gas. The energy (22) to 10 bars and further cooled to 190 C by spraying
carried by the pyrolysis gas reduces the biomass fuel in condensate (23) leaving the dryer. The dryer
input as explained above (Wood Pyrolysis Model). condensate is throttled to 2 bars (26) and send to the
feedwater tank (15). Flue gas temperatures of 120 C
170
th th
are considered not to cause sulphur corrosion pyrolysis yield constantly increases with the decrease of
problems, especially not with low-sulphur wood fuels. the DH levels down to 60%. The maximum flow off the
Certainly this design specification must be reconsidered dryer (and thus its capacity) is be restricted by the
in case of changed fuel properties. pressure prevailing in the feedwater tank, which in turn
is given by the extraction pressure after the turbine
The maximum pyrolysis production for each load point
stage (11). The pressure decreases with falling live
is restricted by the maximum steam extraction rate and
steam parameters and steam massflow. Hence, there is
by the boilers maximum burning power. The maximum
a pressure dependant maximum enthalpy flow that can
possible pyrolysis yield logically requires highest
be fed into the feedwater tank until saturation state is
possible fuel input since heat must be provided both, for
reached for the mixture of the condensates from the DH
drying and pyrolysis. Conversely, this means that the
exchanger (13) and the dryer (24). In order to overcome
steam enthalpy exceeds the demand of the DHN. This
this restriction the feedwater tank pressure has been
is because the boiler temperature is controlled by
increased load-dependently to a maximum of 2 bars
means of the evaporator- and superheater tubes in the
matching its design pressure. However, due to the
boiler walls. If now, the heat input in the boiler is kept on
reason mentioned above, for loads below 60% the heat
a higher level as usual the water amount needed to
that would need to be dissipated in the dryer (in order
dissipate the heat from the boiler walls is only
to match the DH load) would result in such a high dryer
decreasing to a certain amount (resulting from a
condensate heat flow which again would bring the
reduced temperature after the economizer).
feedwater beyond saturation state. Hence for those
Consequently, in order to match the DH load, this heat
cases the boiler load is gradually decreased, resulting in
must now be dissipated in the pyrolysis heat
lower pyrolysis yields. The lowest DH load level that can
exchanger (19) or in the dryer (24). By iteration the DH
be represented is 28.6% of the plants full load.
load is matched by adjusting dryer load, correlated split-
Compared to a minimum load of 50% in the base case
off to the pyrolysis heat exchanger and fuel input. In all
which is given by the minimum fuel input required for
cases the boiler load (characterised by the fuel heat
stable combustion conditions in the boiler-, the
input) is restricted to 100%. So, the overload back-up
integrated process offers possibilities to increase the
capacity of the boiler is maintained. With this setup the
operating hours of the CHP plant considerably.
Fig. 2: CHP Plant with integrated pyrolysis and steam drying
171
th th
APPLICATION OF THE PRIMARY ENERGY In the power bonus method fEl. is defined as the PEF of
CONCEPT ACCORDING TO EN 15603 the electricity that is thought to be replaced by the
power generated in the CHP plant (for instance, in this
Primary Energy Concept
study the average power generation efficiency in
The EU standard EN 15603 [2] handles the energy Finland is used). This allocation pays attention to the
performance of a building as a whole and gives fact that the co-generated electricity is more sustainable
guidelines how energy use and production of a building due the CHP process high overall efficiency. The PEF
shall be calculated. In order to aggregate the different of the DHN can thus be determined according to;
forms of energy produced and used within the building,
primary energy (PE) and CO2 emissions are
accumulated and expressed by means of primary f F ,i E F P f El .
energy factors (PEF) and CO2 emission coefficients, f DH i
respectively. PE is energy that has not been subjected QDH
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 As production of products other than electricity is not
involved in the generation process are retraced to their defined in EN 15316-4-5 the power bonus method has
sources and all energy needed to deliver the final been extended by regarding the produced pyrolysis
energy product are aggregated to the total PE slurry as a bonus as well. The PEF of the pyrolysis
consumption and CO2 emissions. Thus the PE integrated CHP plant is thus calculated as:
approach applies the holistic principles of life cycle
assessment to an energy rating procedure. By retracing
energy consumption to the source, the system f F ,i E F ,i P f El . E Pyro f Pyro
boundaries automatically include the whole world, and f DH i
thus depict the real impact of the system concerning QDH

energy consumption and CO2 emissions.
Primary Energy Factor In this study PEFs as shown in table 2 have been used:
The total primary energy factor is the sum of all PE input

to the energy system divided by the useful energy Table 2: Primary energy factors and CO2 emission
delivered at the system border. It thus describes how coefficients for fuels and products
much PE input is needed in order to obtain one unit of
energy used and can hence be seen as an inverted kg/MWh
efficiency. fBM
2
1.09 cCO2/BM
2
14
2 2
In standard EN 15316-4-5 [3] more detailed guidelines fOil 1.35 cCO2/Oil 330
for the calculation of PEFs of DH systems are defined. fEl.
1
3.11 cCO2/El.
1
270
According to EN 15316-4-5 PEFs can be calculated for 1 1
fPyro 1.28 cCO2/Pyro 14
a certain part of the energy system. In this study the
system boundary comprises the power plants and the 1 2
: value is calculated, : value is taken from EN 15603, Annex E
DHN.
The PEF of the DHN has been calculated applying the Fuels assumed to be used are wood logs for the CHP
power bonus method. If yearly demand data of the DHN plant and fuel oil for the heat-only/backup boiler(s) and
and the generation data are known, the PEF of the DHN their PEFs are taken form annex E of EN 15603. The
can be calculated by applying the so-called power PEF of electricity production in Finland has been
bonus method. The power-bonus method is derived derived from [10]. The PEF of pyrolysis slurry in a
from the energy balance of the building which can be stand-alone unit has been calculated assuming a flue
written as: gas dryer (which is considered as the drying technology
most likely to be applied) with an energy consumption of
f F ,i EF f DH QDH P f El . , 3300 kJ/kg water evaporated [9] and a heat of pyrolysis
i of 1.87 kJ/kg [7]. Although the standard asks for more
where EF, QDH and P are the heat of the fuels used, DH detailed analysis of the energy chain as e.g.
and power co-generated respectively. fF,i, fDH and fEl. are consideration of transport, transmission and other
the PEFs of the fuels used, the DHN and of the co- processing should be included, this has not been
generated power. implemented into this study since those factors are
assumed not to differ between integrated and separated
production of pyrolysis oil.
172
th th
CO2 Emission Coefficient hours and a DH load as low as 30% (matching 18% of
The CO2 rating is done by calculating CO2 emission the total DH load). It can be seen from the table that for
coefficients (cCO2) that quantify the total amount of fossil all cases the DH output is the same for the 100-50%
fuel derived CO2, emitted to the atmosphere, per unit operating points. This results, in the first two cases, in
delivered energy. As for the primary energy factor the an identical total DH output of 70.85 GWh. This
system boundary comprises of power plants and DHN. corresponds with 75% of the total yearly DH load. Due
Also the power bonus method is applied for calculating to steam extracted to the dryer, the enthalpy flow
the DHNs specific CO2 emissions. For the sake of through the turbine in part load is decreased, which
completeness it must be mentioned that CO2-equivalent results in a lower electricity production in part load for
emissions of other greenhouse gases can optionally be the cases 2 and 3. Already for the second case
included. However this has not been implemented into pyrolysis slurry with an energy content in the same
this study, due to a lack of data. Similarly to the PEF the range as the DH load can be produced. Fuel input,
CO2 emission coefficients for the base case are which is defined as wood burned in the boiler and wood
calculated as: 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
E F ,i cCO2 , F ,i P cCO2 , El . decreased for lower load levels as explained above. If
cCO2 , DH i
. operation hours are extended by supplying lower DH
QDH 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
And for the modified plant as:
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
E F ,i cCO2 , F ,i P cCO2 , El . EPyro cCO2 , Pyro heat is almost cut in half. Together with the additionally
cCO2 , DH i . produced electricity this substantially improves the
QDH
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 CO2 emission coefficient changes somewhat
EF,i, EPyro, P and QDH represent heat in fuels, heat in
controversially by increasing in the 2nd case. This is
pyrolysis slurry and co-generated electricity and DH because the loss in electricity bonus cannot be
respectively. Accordingly, cCO2,F,i, cCO2,Pyro, cCO2,El. and compensated by the produced pyrolysis slurry, since the
cCO2,DH are the related CO2 emission coefficients. The CO2 emission coefficients differ widely. However for
corresponding values are given in table 2. case 3 specific CO2 emissions become even negative.
The negative value is very unlikely to reach and can be
RESULTS explained with the not fully accounted fuel production
chain. Nevertheless, it is obvious that the DHNs CO2
In table 3, three simulation cases are presented: the
emission factors can be considerably reduced with the
base case (case 1), pyrolysis integration with the same
presented integration concept.
operation hours (case 2) and the maximum pyrolysis
slurry production (case 3) with prolonged operation
173
th th
Table 3: Results Multiperiod Model
Base Case - Case 1

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 + Pyrolysis - Case 2

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 + Pyrolysis - Prolonged Operation Hours - Case 3

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
Table 4: Results - PEF and CO2 Coefficient

Base Case CHP + Pyrolysis CHP + Pyrolysis -
Case 1 Case2 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 Another open question is the influence of the real
pyrolysis gas on the combustion temperature and flue
The work shows that by integration of a CHP plant with
gas properties. In order to gather more details of the
wood pyrolysis operation hours can be increases by
pyrolysis process a simple pyrolysis model is currently
30%, a valuable product can be co-produced and PEE
under development. Together with the power plant
as well as the CO2 emission coefficient of the DHN can
model the integration can be further optimised aiming
be substantially improved. As next steps more
for highest PEE along with low CO2 emission
comprehensive data of the fuel supply chain should be
coefficients.
implemented to get more realistic values that will
approve the trend shown with this work. The process Further an economic analysis should be carried out in
can be further improved by integrating heat that is set order to show potential economic benefits. The
free during the condensation of the pyrolysis liquid and integration itself seems to be viable a statement that
gaseous product. The heat is available in a is supported by a press release from June 2009 where
temperature range from approximately 500 C to 25 C boiler manufacturer Metso and forestry company UPM
and could hence be used for steam superheating, announced the development of a new viable fast
feedwater preheating, but also for DH generation. This pyrolysis process benefitting from the integration with a
integration is not a simple task since many plant CHP plant [11].
parameters influence each other. The heat integration
Concerning the European standards used for
must be carried out together with a pinch analysis to
evaluation, it can be said that the power bonus method
assure an energy efficient integration.
can be easily adapted to a polygeneration concept
174
th th
yielding heat, electricity and pyrolysis slurry. It very [4] Savola, T., Modelling biomass-fuelled small-scale
likely can also be extended to other possible biorefinery CHP plants for process synthesis optimisation,
products as long as those are energy products. This Doctoral Dissertation, Helsinki University of
expansion option could be implemented into the Technology, Espoo 2007
standard.
[5] Bridgwater, A.V., 2000, Fast pyrolysis processes
However the most difficult question remains how the for biomass, Renewable and Sustainable Energy
PEF of other, less common co-products should be Rev., 4(1), pp. 1-73.
determined. In the case of pyrolysis slurry it is not
possible to find good average production efficiencies [6] Henrich, E., 2007, The status of the FZK concept
since the technology is not yet on the market. But, how of biomass gasification, 2nd European Summer
is co-generation of cooling evaluated? School on Renewable Motor Fuels, Warsaw.
In general it can be said that the implementation of the [7] Daugaard, D., Brown, R., Enthalpy for Pyrolysis
process will be strongly dependant on investment cost for Several Types of Biomass, Energy & Fuels
and on the market value of the product. The product 2003, 17, 934-939
value is currently difficult to predict and also its future [8] http://www.ecn.nl/phyllis: PHYLLIS is a service
price development will be strongly dependant on the provided by the Energy Research Centre of the
use of biomass in the future. Netherlands ECN, 17.9.2009 [selected
Summarising it can be said that even though many subgroups: untreated wood birch and
questions still need to be answered, this works shows fir/pine/spruce].
clearly that the integration of communal CHP plants [9] Brammer, J., Bridgwater, A., Drying Technologies
with wood pyrolysis is beneficial concerning the for an integrated gasification bio-energy plant,
connected DHNs PEF and CO2 emission coefficient. Renewable and Sustainable Energy Reviews 3
Vice versa it also shows that CHP plants can play an (1999) 243-289
important role in the sustainable bio-refineries of the
future. [10] Dones, R. et al, 2004, Life Cycle Inventories of
Energy Systems: Results for Current Systems in
Switzerland and other UCTE Countries, ecoinvent
ACKNOWLEDGEMENTS
report No. 5, Paul Scherrer Institute Villigen, Swiss
This work is part of the Primary Energy Efficiency Centre for Life Cycle Inventories, Dbendorf, CH,
project of Nordic Energy Research. p.170.
The funding of the Graduate School of Energy Science [11] N.N., press release on
and Technology (EST) is gratefully acknowledged. http://www.metso.com/news/newsdocuments.nsf/w
eb3newsdoc/C89A8AC3F77ABD29C22575CF003
REFERENCES 111F5?OpenDocument&ch=ChMetsoWebEng
[12] H. Lund, F. Hvelplund, I. Kass, E. Dukalskis, D.
[1] Kohl, T., Jrvinen, M., Fogelholm, C.J.,
Blumberga, District heating and market economy
Gasification and biorefinery in combined heat and
in Latvia, Energy, 1999, Vol. 24, pp. 549-559.
power plants, Proceedings of the 11th
International Symposium on District Heating and [13] H. C. Park, M. Chung, S. H. Kim, Development of
Cooling, Reykjavik, Iceland, 2008 system simulator for community energy system,
Report to Ministry of Industry, 2003.
[2] EN 15603:2008, Energy performance of buildings.
Overall energy use and definition of energy [14] Y. H. Im, H. C. Park, M. Chung, A study of optimal
ratings, European Committee for Standardization, heating supply systems for the newly developing
CEN, Brussels. area in the vicinity of DHC system supplying area,
Report to Korea District Heating Corporation, 2006
[3] EN 15316-4-5:2007, Heating systems in buildings.
Method for calculation of system energy [15] Y. H. Im, M. Chung, H. C. Park, Feasibility study
requirements and system efficiencies. Part 4-5: for small size cogeneration systems in the
Space heating systems, the performance and metropolitan areas of Seoul, Final Report to SH
quality of district heating and large volume (Seoul Housing) Corporation, 2008.
Systems, European Committee for [16] M. Chung, H. C. Park, Development of a energy
Standardization, CEN, Brussels, 2007 demand estimator for community energy systems,
Journal of the Korean Solar Energy Society, 2009,
Vol 29, pp. 37-44.
175
th th
[17] M. Chung, H. C. Park, Development of a software [20] H. C. Park, Development of weighting factors for
package for community energy system assessment variables associated with hourly energy
Part I: Building a load estimator, Energy, in consumption pattern for hotels in Korea, SAREK
press. (Soc. Air-conditioning, Ref., Engineers of Korea)
[18] H. C. Park, S. S. Lee, D. J. Kim, Development of Winter Annual meeting, 2002, pp. 76-82
energy models for department stores, Korean [21] H. C. Park, Analysis of energy loads for hospital
Journal of Air-Conditioning and Refrigeration buildings, SAREK journal, 2002, pp. 1088-93.
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.
176
th th
CHP OR POWER STATION? QUESTION FOR LATVIA

1 1 1 1
D. Blumberga , G. Kuplais , F. Romagnoli and E. Vigants
1
Riga Technical University Institute of Energy Systems and Environment
Kronvalda boulv. 1, LV-1010 Riga, Latvia
ABSTRACT In CHP station this dominance is almost total and very

high in district heat supply boiler houses. As it well
This paper presents aspects and problems of the
known Latvia is a great consumer of imported fossil
Latvian energy-system connected to the choice of the
from one side but in the same the share of renewable
CHP and/or power stations for the future national
energy resources is one of the highest of Europe.
energy strategies. In the light of the last EU directive in
the subjects of Renewable Energy Sources (RES) the The use of specific energy resource depends on
share of electricity produced from RES at the moment energy supply policy, and total consumption of energy
is attested on the value of 42.4 % but should be resources depends on development of every type of
increase to 49.3%. In the same time the share of energy resources in regions. Now there is unjustified
renewable energy resources in the final energy high proportion of fossil fuel in state energy balance
consumption for 2020 should reach the level of 40% which is possible to reduce by a beginning of active
from 30%. use of local fuel in regions. The EU directive also
requires that Latvia in the year 2015 would generate
Dependence on imported energy sources, growth of
49% of electric power from renewable resources
electricity prices, the need to support local producers
(currently it is 45%) [3]. This is supportable, but the
are the main reasons for the use of new renewable
power supply of Latvia cannot be let out of the sight
energy technologies in the Latvian energy sector to
and this issue is already problematic.
implemented in refurbished energy supply system.
Latvia has some electricity production from co-
Several methods fro the evaluation of the best strategy
generation plants and some from hydro-power plants.
are explained.
However, the production of electricity from the hydro-
This apaper summarizes the application of the Energy power plants fluctuates a great deal from year-to-year.
Indicators for Sustainable Development (EISD) as good The rest of the electricity for consumption is imported
tool for analyzing trends, setting energy policy goals from the neighbouring countries.
and monitoring progress. The results from the
In order to understand the role played by CHP and
application of a multi-objective optimization regarding
power plant it is fundamental to understand the actual
the implementation of the landfill biogas in the biogas
situation in Latvia for thermal energy where more than
treatement plant Daibe are reported.
a half of Latvia district thermal energy is distributed and
consumed mainly in Riga.
1. INTRODUCTION
Latvian heating primarily is performed on a centralized
The structure of energy user in Latvia is characterized basis and after the used of wood energy the natural
by high energy consumption in households, public and gas imported from Russia is the main source (see
service sectors, comparing with relatively low Fig. 1).
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 Fig. 1. Main resources used for local and individual heat
as a convenient way of exerting economic pressure [2]. supply [3]
177
th th
This type of heating supply scheme means that of the gross energy-consumption from renewable-
consumers are grouped and heating is performed from energy sources in Latvia in the year 2007.
heat source which is intended for the consumer group.
The most important domestic renewable-energy
About 70 % out of this thermal energy volume is resource in Latvia is biomass in the form of fuelwood:
produced in the cogeneration cycle (only in Latvenergo in fact approximately 45% of Latvia is covered with
owned CHP, and Rigas siltums) and around 30% of woods and this substantial area makes wood a
centrally supplied heat energy is produced in Riga CHP significant potential as a resource for energy supplies.
plants and boiler houses. Of course as main fuel in
Even though the share of renewable is one of the most
Riga natural gas is used approximately 98% of thermal
large Europe the EU directive fixes the target of 40%
energy is produced from natural gas (CHP plants and
share of renewable energy resources in the final
boiler houses together) [3].
consumption in 2020.
As for heat supply outside of Riga, the dominant
This means that the increase is not feasible without the
thermal energy is produced in boiler houses with
need of refurbishment and/or construction of energetic
relatively high proportion of local fuel usage (as shown
infrastructures.
in Fig.1). Outside of Riga CHP heat production rate
does not exceed 5%. The fact that Latvia has domestic renewable-energy
resources makes it interesting because the utilisation of
2. EXISTING ENERGY SITUATION IN LATVIA: the domestic fuels would be sustainable both from an
SHORT OVERVIEW environmental and an economic point of view.
During the recent past central (large) power plants in Latvia has comparatively well developed power, natural
Latvia supplied roughly 65% of the total annual power gas supply and district heating systems, and as a
demand - distributed energy resources (DERs) covered consequence the electricity is basically produced by
36%, but the rest were received as import supplies hydro power plants and by cogeneration plants, which
from Estonia, Lithuania and Russia (mainly) [4]. are operated according to district heating demand, and
part of electricity is imported (fig. 2). Consequently the
Regarding fuel sources Latvia has no real fossil-fuels of main objectives of the Latvian energy policy now are to
its own and the consumption must be imported. ensure sustainable accessibility to necessary energy
However Latvia uses the domestic renewable-energy resources and security of supply in order to favourite
resources hydro-power and biomass. the economic growth and improve quality of life, to
Table I. primary energy-consumption in Latvia in the ensure environmental quality retention and meet the
year 2007 [1] objectives set in the Kyoto protocol of UN FCCC and
Latvian Climate Change Program.
PJ %
Electricity Supply in Latvia
Natural gas 56.92 27.8
Electricity amount, billion kWh
8
biogas 0.32 0.16 7
imported electicity
6
Biodiesel 0.07 0.03 5
wind generators
small HPS
4
Oil products 73.33 35.8 small CHP
3 CHP
Fuelwood 48.47 23.7 2 HPP
1
Hydroenergy 9.84 4.8 0
2000 2001 2002 2003 2004 2005
Import of Electricity 10.80 5.3 Year
Import Coal and coke 4.36 2.1

Fig. 2. Electricity supply in Latvia (Source: state JSC
Wind 0.19 0.09 Latvenergo, Ministry of Economics, Central Statistical
Bureau)
Biodiesel 0.07 0.03
2.1 Lack of energy sources for electricity
Total 204.6
The main domestic electricity capacity consists of 1517
MW of hydro and 520 MW [5] of thermal (CHP units in
The use of primary energy for the gross energy-
Riga) all of which is controlled by the state company,
consumption in Latvia can be seen in Table 1. The
Latvenergo. The generating potential mainly consists of
share of renewable energy in the gross energy-
three hydro power plants (HPP) on the Daugava River,
consumption is made up of fuelwood and hydro energy.
hence directly dependent on the rivers water flow. Due
That means that there is approximately a total of 30%
to small reservoirs, utilization rates are low and the
178
th th
production is quite seasonal following the water flows. equipment with this capacity heat production by
The amount of power produced by the Daugava river cogeneration would be a maximum.
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 Fig. 3. Heating energy distribution by cities in Latvia
produced by wind energy. The net electricity import
(including the amount of energy exported) is around If we are looking at the district heating division of Latvia
10.8 PJ approximately the 50% of the national supply. a huge difference can be seen in quantity of heat
These figures shows the lacks of energy sources in the supply in Riga and the rest of Latvia (see fig. 3)
national system and seems reasonable to foreseen a Two large CHP plants, Riga TPP-1 with an installed
more large fraction of other energy sources for the electric capacity of 144 MW and Riga TPP-2 (390 MW),
production of electricity, the main question is on which are located in Riga [5]. CHP plants are the main heat-
methodology base this strategy . generating sources of heating networks of Latvian
capital. Power is produced mainly in cogeneration
2.2 Well organized and developed DH system mode, according to the heatload curve.
Latvian heating primarily is performed on a centralized
During the heating season, when there is a substantial
basis consequently consumers are grouped and the
demand for heating and hot water, Riga CHP plants
heat is supply from heat source which is established for
produce approximately 80% of the total annual
a certain consumer group. The heat source power,
production volume, while during summer the volume of
depending on type of consumer group, varies from the
production reduces [5].
range of kW to several hundred of MW. In general
lower power can correspond to building groups, Nowadays Riga CHP plants cover about 20% of the
individual houses or even apartments heating. total annual power demand of Latvia [5] .
Residential and separate heating of individual houses
The main fuel used in Latvia biggest cities is natural
belongs mainly from the decentralized heating. One of
gas and the rates of thermal energy are 75% - 85% [3].
the benefits of district heating is centralization of heat
In Riga and other cities where most part of the heat is
load, which gives a possibility to increase the heat
produced in cogeneration cycle, the increase of rates
source power and to form basis for the development of
was not so high and currently (in the autumn of 2009)
cogeneration power. For large heat consumers in
heat rates are lower that in the cities where wood chips
Latvia (mainly heating systems in large cities like Riga)
are used.
large cogeneration plants are installed. The customers
who are not connected to a district heating cannot be From the thermal energy point of view seventy percent
provided from this system. In the other regions far from of the heat in Latvia is supplied from district-heating
the big cities the heat supply system is mainly based systems either from boiler houses or co-generation:
on district heating, consequently it means that that 37% of the district heating in Latvia was produced by
there possibility for a CHP development. means of co-generation plants [6]. This means that
63% of the district heating is produced in boiler houses
CHP plants cover only a part of the total heat load. The
[6]. This means that there is potential to replace some
rest of the load is covered by the peak load boilers.
of the heat plants with co-generation units (Eighty
This means that following the total heat capacity of the
percent of the district heating in Denmark is supplied
source, the potential heat capacity of cogeneration
from CHP [6]).
should be assessed quantitatively. Heat capacity of
cogeneration plant has to be selected so that operating As for heat supply outside of Riga, the dominant
thermal energy is produced in boiler houses with
179
th th
relatively high proportion of local fuel usage. Outside of gas), constructive parameters of cogeneration plant,
Riga CHP heat production rate does not exceed 5% [3] parameters of heat energy consumers, heat load
(combined heat and power plant up to 4 MW of power duration curve, duration of heat energy consumption
operating in Bauska, Valmiera, Ogre, Vangazi, levels, behaviour of energy end users, installed
Daugavpils, Jelgava, Dobele, Grobia, Saldus, capacity, energy efficiency of technologies,
Ventspils, Ozolnieki, dai, Lielvrde and Cesis). development of demand side management factor, and
other factors.
3. METHOD FOR EVALUATION
3.2. Methodologies: EISD method and MOO method
In connection to achieving sustainable development on
In the following paragraph the algorithm of ISED core
global scale the correct and judicious use of resources,
set tool, included in the conceptual framework used by
technology, appropriate economic incentives and
United Nations Commission on sustainable
strategic planning at the local and national levels is
development (CED), is shown. After is also shortly
required. Therefore, choosing energy fuels and
reported the MOO methodology
associated technologies for the production, delivery
and use of energy services, it is essential to take into The EISD is an analytical tool developed which can
account economic, social and environmental help energy decision and policymakers at all levels to
consequences. The research on criteria and/or incorporate the concept of sustainable development
indicators in order to understand the best energetic into energy policy. EISD core set is organized following
choice for Latvia is the first step for a correct energy the conceptual framework used by United Nations
planning. Commission on sustainable development (CSD).
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) Fig. 4. set of core EISD [8]
[8], Life Cycle assessment (LCA) [9, 10]. Each of these
methodology start from different point of views and There are 30 indicators, classified into three
bases: MOO methodology is connected to best dimensions (social, economic and environmental) and
optimization choice of a certain number of variables grouped in 7 big themes. There are four social
that optimize certain objectives, EISD methodology dimension indicators: three of them represent equity
aims to evaluate (and consequently increase) the (accessibility, affordability, disparities) and one health
concept of sustainability based on social, economical theme (safety). The set of energy indicators of
and environmental indicators, LCA aims to figure out economic dimension consists of 16 indicators. There
the global environmental load of a process and/or are nine environmental dimension indicators in the
product taking into account the entire outflows and EISD core list. The scheme of core EISD indicators is
inflows connected (in terms of energy, substances and presented in Fig. 4. The priority areas for energy sector
emissions), in this last case the indicators change analysis in Latvia can be were selected based on the
depending on type of Life cycle assessment methods main EU energy policy directions. These priority areas
choosen. are as follows:
A summary of the factors that can influence CHP Energy use.

development in Latvia has been proposed in previous Energy intensities.
papers. A. Volkova et al. [2] identify four main factors: End-use intensities of economic branches.
political, geographical-climatological, legislative and Energy security.
technological. Environmental energy impacts.
In general the total amount of electricity produced in a The next Fig. 5 shows the linkages among the
cogeneration regime and condensing mode depends indicators selected for energy policy analysis in Baltic
on constructive solutions (e.g. technical solution for the States. Relevant policy actions based on analysis
biogas collectors), availability of source used (mainly
180
th th
conducted in the previous sections are defined based of produced biogas. The main barriers for improved
on targeted indicators. 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
Fig. 5. Linkages between indicators and relevant policy
actions based on the targeted indicators [8] 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
Multi-objective optimization (MOO), also known as need for waste sorting in landfills, but in the other it is
multi-criteria optimization, particularly outside important to sort both before waste collection and in
engineering, refers to finding values of decision landfills.
variables which correspond to and provide the optimum
of more than one objective. Unlike in single objective Utilization of landfill biogas in Latvia is based on energy
optimization (SOO), which gives a unique solution, production in power station placed close to landfill for
there will be many optimal solutions for a multi- different reasons. One of the most important reasons is
objective problem. Multi-objective optimization involves financial state support of small scale power stations (4
special methods for considering more than one MWe) from renewable energy resources. Such kind of
objective and analyzing the results obtained [7]. support prevents both, development of waste sorting
and utilization of refuse derived fuel in cement
Often, the various objective functions conflict with each production, and biogas improvement to cover needs
other (i.e., optimizing one of them usually tends to transportation sector or to connect to natural gas grid.
move another towards undesirable values), for solving
such models one needs to know how many units of one In the following is shortly reported the methodology
function can be sacrificed to gain one unit of another, regarding the optimization model of biogas use in
but this trade-off information is not available. In other landfills in Latvia in connection to the data collected
words, one is forced to determine the best compromise from landfill Daibe. After the analysis only two of the
that can be achieved. independent parameters have been chosen: quality of
biogas (characterized by heat value), and technological
In the following paragraph an example of MOO applied equipment (characterized by electrical capacity).
to the evaluation of possibilities to utilize landfill biogas
for electricity production in one of Latvias landfills. This optimization model for biogas utilization in landfills
includes four modules and is based on technological,
climate and economical sub models.
4. TESTING OF LANDFILL GAS PRODUCTION
Results of economical optimization show that in case of
The improved biogas is one of the cleanest fuels with a low biogas quality (4 kWh/m3) the optimal installed
little impact on the environment and human health [11]. capacity is 2.2MW. In case of biogas quality of 5
One of the advantages of biogas injection into natural kWh/m3, optimal installed capacity is 2.8MW, and
gas grid is the fact that natural gas grid connects a 3.4 MW in case of high biogas quality (6 kWh/m3).
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
181
th th
5. COMMENT AND DISCUSSION

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 (comparing with Denmark where approximately
80% of the district heating in Denmark is supplied from
CHP [6].
Hence promotion of high-efficiency cogeneration (CHP)
Fig. 6. Diagram of the economical optimization[12]
based on a useful heat demand is a priority with regard
Results of technological optimization show that, the to saving primary energy, avoiding network losses and
higher the installed capacity, the shorter the operation reducing emissions, in particular of greenhouse gases
time of equipment. If assumed that operation time of [2].
the equipment could be 5 up to 10 years, then the Of course the choice of the fuel is fundamental in order
installed capacity can be 0,5MW and higher. 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
Fig. 7. Equipment operation time vs installed capacity [12] supply system, but it must also become part of
strategies for economic development and for the
Results of the climate sub model show that the higher improvement of the import/export balance of the
is installed capacity, the greater the reduction of country. These measures can succeed in not only
greenhouse gas emissions. Besides that, it is not developing local production and job creation, but if can
possible to reach extremis by using two objective also stimulate and increase the potential export.
functions (heat value of biogas and installed capacity),
It is particularly important to conduct engineer-technical
which have been used in case of economical and
and economic analysis of the various technological
technological sub models, and it is necessary to
solutions possible to implement wood use in the
introduce another objective functions.
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
Fig. 8. Diagram of the environmental optimization [12] mix.
Improve maintenance of existing energy
infrastructure.
Model of power production in landfill shows that feed-in
Eliminate constraints and investment in new
tariff stated as financial support today in Latvia allows
facilities.
to reach economically feasible projects even in case if
Increase the efficiency of energy supply in
cogeneration unit is operated in power station regime
electricity generation.
(generates only electricity). Results show that state
Increase the share of electricity produced by
policy needs corrections to improve energy efficiency of
combined heat and power (CHP) plants.
biogas utilization for energy production.
Increase the share of renewable and domestic
energy sources in the energy mix.
182
th th
6. CONCLUSIONS 7. REFERENCES
1. In the paper has been presented aspects and [1] Construction, Energy and Housing State Agency
problems of the Latvian energy-system connected to Energy Department, Latvian energy in figures,
the choice of the CHP and/or power stations for the Riga, 2008.
future national energy strategies in the light of the last
EU directive in the subjects of RES. The dependence [2] A. Volkova, E.Latev, A. Siirde, Small-scale CHP
on imported energy sources, the growth of electricity potential in Latvia and Estonia, Scientific Journal of
prices, and the need to support to local producers are RTU Environmental and climate technologies, Ser.
the main reasons for use of new renewable energy 13, n. 2, Riga, 2009.
technologies in the Latvian energy sector. [3] Latvias district heating association , Heat supply in
2. In this paper has been summarized the results from Latvia,
the application of the Energy Indicators for Sustainable http://www.lsua.lv/en/index.php?option=com_conte
Development (EISD), a good tool for analyzing trends, nt&task=view&id=4&Itemid=5.
setting energy policy goals and monitoring progress in
order to indentify good policy indicators. Also a testing [4] D. Streimikiene, I. Roos, J. Rekis, External cost of
of landfill gas using MOO method has been reported electricity generation in Baltic States, Renewable
where only two of the independent parameters have and Sustainable Energy Reviews n. 13, 2009, pp.
been chosen: quality of biogas (characterized by heat 863870.
value), and technological equipment (characterized by [5] D. Streimikiene, I. Roos, J. Rekis, External cost of
electrical capacity). Model of power production in electricity generation in Baltic States, Renewable
landfill shows that feed-in tariff stated as financial and Sustainable Energy Reviews n. 13, 2009, pp.
support today in Latvia allows to reach economically 863870.
feasible projects even in case if cogeneration unit is
operated in power station regime (generates only [6] L.H. Rasmussen, A sustainable energy-system in
electricity), but if is feasible from an economical point Latvia, Applied Energy n. 76, 2003, pp. 18.
view is not the same if reference to environmental [7] G.P. Rangaiah, Multi-Objective Optimization:
impact. Techniques and Applications in Chemical
3. In the paper has been discussed how LCA can be a Engineering, World Scientific, 2008, p. 454.
good approach that enables the energy requirements,
[8] D. Streimikiene, R. Ciegis, D. Grundey, Energy
GHG balance and other environmental impacts of
indicators for sustainable development in Baltic
bioenergy production chains to be accounted and
States, Renewable and Sustainable Energy
accurately compared. Hence LCA is good tool in order
Reviews, 2007, Vol. 11, pp. 877893.
to give the possibility to compare different RES usage
strategies. [9] G. Rebitzera et al., Life cycle assessment - Part 1:
4. Due to high electricity feed in there is an economical Framework, goal and scope definition, inventory
motivation for power plant operation with low efficiency. analysis, and applications, Environment
For electricity produced in renewable energy power International n. 30, 2004, pp. 701 720.
plants with nominal capacity of up to 4MW high feed in
[10] D.W. Pennington et al., Life cycle assessment
tariff has been transposed in Latvias legislative acts.
Part 2: Current impact assessment practice,
OF course this is not good from environmental point of
Environment International n. 30, 2004, pp. 721
view.
739.
5. The use of CHP instead of conventional plant will
always improve energy efficiency and will reduce CO2 [11] D.Blumberga, . Kuplais, I. Veidenbergs, E.Dace,
emissions significantly, in Latvia there is potential to The benchmarking method for an evaluation of
replace some of the heat plants with co-generation biogas improvement methods, Scientific Journal of
units. RTU Environmental and climate technologies, Ser.
6. Only crucial measures such as the reconstruction of 13, n. 2, Riga, 2009.
energy sources in the larger cities (including Riga
[12] G. Kuplais, D. Blumberga, E. Dace, F. Romagnoli,
TEC 1 and Riga TEC 2) adjusting the use of fossil fuels
Optimisation model of biogas use in landfills in
to biomass and conversion to non-natural gas sources,
Latvia, 7th International conference ORBIT2010:
will produce results. Biogas and landfill gas favorite the
Organic resources in the carbon economy, June
environmental impact displacing usage of natural gas,
29-July 3, 2010, Heraklion, Greece.
the possibility of the feasibility solution for connected
CHP in out-of-city region to heat consumer must be [13] A. Blumberga et al., Assessment on the use of
evaluated. renewable energy resources in Latvia until 2020:
report, LVAF, December 2008, Riga.
183
th th
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 energy become more viable in various locations

around the world. It is thus important to investigate
The aim of the study is to calculate primary energy
their primary energy efficiency and environmental
factors, fp, stating the primary energy efficiency as well
impact for comparison with other energy conversion
as factors for CO2 emission, K, for geothermal
technologies. These energy performance indicators
combined heat and power production at the Hellisheidi
can be used to help decision making of future
CHP plant in South-West Iceland. These factors state
developments, policy making and energy rating of
how much primary energy consumption and CO2
buildings.
emissions result from the production of 1 MWh of heat
and electricity due to geothermal utilization. Methods of Countries that have access to geothermal areas and
life cycle assessment (LCA) are used to calculate these produce power by geothermal utilization within the
factors by taking into account all energy and material European Union (EU) are: Austria, France, Germany,
streams to and from the CHP plant during construction Greece, Hungary, Italy, Netherlands, Portugal,
and operation. The results show that producing heat Romania, Slovakia and Spain. Other European
and electricity in a combined heat and power plant countries such as Iceland and Turkey, which are not
minimizes the primary energy factor for the electricity current member states of the EU, also utilize
generation and produces a relatively low primary geothermal energy extensively [2]. Also, 32 European
energy factor and CO2 production factor for the heat countries use geothermal energy directly for various
generation process. From the results, it can also be purposes such as district heating [3]. Thus, electricity
seen that life cycle assessment is a useful method to and heat based on geothermal energy are a part of
evaluate the total impacts of the geothermal energy Europes energy mix. For countries using geothermal
conversion process, especially for the emission of based power and/or heat and complying to EU
greenhouse gasses during the lifetime of the legislation, it is therefore important to have easy access
production facilities. The experience in this study also to standardized factors accounting for the primary
demonstrates that the method can equally be used for energy efficiency and CO2 emissions from geothermal
processes as it is commonly used for the analysis of based heat and power.
total impact of products.
The aim of this study is to produce standardized factors
for primary energy efficiency (fp) and CO2 emission (K)
INTRODUCTION for geothermal heat and power production.
The calculation of primary energy and CO2 production
factors for geothermal power production has had little ENERGY PERFORMANCE INDICATORS FOR
attention while factors for some other types of energy PRIMARY ENERGY CONSUMPTION AND CO2
technologies such as hydropower, nuclear and coal EMISSIONS
fired power plants have been developed during the
The primary energy factor is defined as the ratio
recent years. The importance of these factors is stated
between the total primary energy inputs involving
mainly in the new recast of Directive 2002/91/EC of the
energy production to the actual energy delivered to the
European Parliament and of the Council on the energy
consumer. According to [4], it should always account
performance of buildings [1]. There it is stated that
for the extraction of the energy carrier and its transport
before the end of year 2010, all new building occupied
to the utilization site, as well as for processing, storage,
by public authority should be issued energy
generation, transmission, distribution and delivery.
performance certificates showing these factors, based
There are two primary energy factors defined:
on the energy mix used by the building and the
buildings energy performance. Total primary energy factor, accounting for
primary energy use of both renewable energy
At present time, geothermal power plants are situated
sources and non-renewable sources.
in 24 countries [2] and a total of 78 countries have
Non-renewable primary energy factor,
reported direct use of geothermal energy [3]. With
accounting only for the primary energy
increasing fossil fuel prices and focus on renewable
consumption of non-renewable energy sources.
energy sources, these power plants producing green
This factor is used when expressing only the use of
184
th th
fossil or other non-renewable or polluting energy the power generation to 120 MW. A year later, another
sources in the energy conversion system. 90 MW were added, resulting in a power generation
capacity of about 210 MW (213 MW in February 2009).
The CO2 production coefficient, K, shall include all CO2-
Further developments of the power plant include
emissions associated with the primary energy used.
adding heat production in 2010 for district heating and
Furthermore, equivalent emissions of other greenhouse
also increasing the power production if possible.
gases, e.g., methane, may be included [4].
Estimated production capacity for the completed
According to Directive 2002/91/EC, indicators on the Hellisheidi Plant is 300 MW electricity and 400 MW
energy performance of buildings shall include the thermal energy [5].
consumption of primary energy and the CO2 emissions
The plant today is a double flash power plant with high-
resulting from the buildings energy usage. Factors for
and low-pressure turbines and separators as seen in
primary energy consumption and CO2 emissions have
Figure 1. The heat production facilities are currently
been calculated for various energy chains producing
under construction with a planned 133 MW thermal
electricity, and values for these factors are given in
capacity at the end of year 2010. The technical
Annex E of the standard EN15603 on the energy
complexity is moderate and the plant makes a good
performance of buildings. An overview of these factors
basis for a LCA study to evaluate the primary energy
is given in Table 1.
efficiency and CO2 emission of this type of geothermal
Table 1: Energy performance indicators for various power plant. Since it is fairly newly constructed, access
sources of electricity [4] to detailed background data for the inventory modelling
is possible, making the study more reliable and
Primary energy factors fp accurate. Environmental assessment for the
CO2
[MWh primary energy / production is available as well as measurements of
Source of production
MWh delivered energy] coeff. K various environmental impacts of the power plant,
electricity
providing data for the impact assessment of the LCA
[Kg/MWh]
Non- study.
Total
Renewable
In this study, a steady production of 213,6 MW
Hydraulic
0.50 1.10 7 electricity and 121 MW heat is used as a basis for the
power
LCA model. The reason for this choice is that the
Nuclear newest inventory data on the construction phase and
2.80 2.80 16
power
mass extraction are built on these production
Coal power 4.05 4.05 1340 capacities, and that the base thermal load is estimated
Electricity to be 121 MW and not the full capacity of 133 MW.
mix 3.14 3.31 617
UCPTE
PRIMARY ENERGY OF VARIOUS ENERGY
SOURCES
As seen in the standard EN15603:2008 [4] and
Table 1, no indicators are given for geothermal power. There is a matter of inconsistency in primary energy
The directive is under reconstruction and a recast has calculations of various energy sources as many
been released, as mentioned before. Also, the table different methods are in use and accepted by different
does not give factors for sources of thermal energy energy authorities [6]. As an example, the primary
used by buildings for space heating. Thus, there is energy factors for power produced from renewable
clearly a need to calculate these factors for energy energy sources such as hydro power, wind energy and
chains that involve geothermal energy, since they solar energy are sometimes calculated by assuming
produce both electricity and heat which is delivered to that the primary energy factor for the energy
buildings within the European Union and in countries conversion system is one, which is the same as
following EU legislation. assuming that the energy conversion process is 100%
efficient. The reason for this assumption is that the
GEOTHERMAL HEAT AND POWER PRODUCTION primary energy is defined as the first usable stage of
AT HELLISHEIDI CHP PLANT the energy flow, which in the case of wind, solar and
hydro is the electricity itself produced from these
Hellisheidi geothermal CHP plant is situated at the primary sources [7]. For electricity production from
Hengill geothermal area close to Reykjavik, the capital heat sources, the first usable stage of the energy
of Iceland. A 90 MW electricity production started in stream is defined as the steam input into the turbine,
2006 after several years of construction and research. according to an energy statistics manual from the
In 2007, a low pressure turbine was added, increasing International Energy Agency (IEA) [8]. The methods
185
th th
Cold water
tank
G
HPT W: Geothermal production well
HPS G: Generator
HPC HPS: High pressure steam separator
HPT: High pressure steam turbine
G
W TV HPC: Condenser for high pressure
LPT
CT turbine
LPS: Low pressure steam separator
LPS
LPT: Low pressure steam turbine
CP
LPC
Hot water LPC: Condenser for low pressure
tank turbine
HX1 HX2 CT: Cooling tower

CP: Cooling water pump
HX1: Heat exchanger 1 for DH system
HX2: Heat exchanger 2 for DH system
IW
IW: Reinjection well
Figure 1 A simple schematic of the Hellisheidi geothermal CHP plant
used to calculate primary energy demand of power system compared to the assumptions made for the
production from renewable energy sources tends to heat conversion processes such as coal, oil and also
underestimate the primary energy input from the geothermal.
original energy sources into the energy conversion
system compared to the assumptions made for the Definition of Primary Energy of Geothermal Fluid
heat conversion processes such as coal, oil and also There is no clear definition of primary energy from
geothermal. geothermal energy sources. Published methods of
determining the primary energy consumption in
PRIMARY ENERGY OF VARIOUS ENERGY geothermal power plants are the following [6]:
SOURCES
Working Group III (WG III) of the
There is a matter of inconsistency in primary energy Intergovernmental Panel on Climate Change
calculations of various energy sources as many (IPPC) records electricity from geothermal on a 1:1
different methods are in use and accepted by different basis. This results in a fp factor of 1.
energy authorities [6]. As an example, the primary The Engineering Information Administration (EIA)
energy factors for power produced from renewable uses a factor of 6.16 units of primary geothermal
energy sources such as hydro power, wind energy and energy for each unit of geothermal electricity.
solar energy are sometimes calculated by assuming International Energy Agency (IEA) records a fp
that the primary energy factor for the energy value of 10 by assuming 10% conversion efficiency
conversion system is one, which is the same as of geothermal power plants.
assuming that the energy conversion process is 100%
In this LCA study, where the main goal is to calculate
efficient. The reason for this assumption is that the
an accurate fp factor for a specific conversion
primary energy is defined as the first usable stage of
technology, the main issue is the primary energy
the energy flow, which in the case of wind, solar and
content of the geothermal fluid extracted from the
hydro is the electricity itself produced from these
production wells. The primary energy content of the
primary sources [7]. For electricity production from
geothermal fluid can be based on different
heat sources, the first usable stage of the energy
assumptions. The first one is the energy content of the
stream is defined as the steam input into the turbine,
geothermal fluid based on its enthalpy in kJ/kg.
according to an energy statistics manual from the
Second, the exergy content of the fluid can be used as
International Energy Agency (IEA) [8]. The methods
a basis. However, in this study, the primary energy
used to calculate primary energy demand of power
content of the geothermal fluid taken from the
production from renewable energy sources tends to
production wells (and utilized for both electricity and
underestimate the primary energy input from the
heat production) is chosen to be the enthalpy above
original energy sources into the energy conversion
186
th th
15 C, an International Standard Atmosphere (ISA) production based on geothermal energy will help
reference temperature [9], and calculated in the identify how much effect the construction, collection of
following manner: 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
(1) 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
Where is the specific primary energy content in the power plant itself and the primary energy extracted
from the geothermal reservoir and even the impacts of
kJ/kg, is the enthalpy of the fluid and is
constructing and operating the distribution facilities.
the saturated liquid enthalpy of the fluid at standard
reference temperature of 15 C. The different phases of performing LCA will be
described in the following sections. The main phases of
LCA include:
LIFE CYCLE ASSESSMENT
Defining the goal and scope of the study
The Directive 2002/91/EC defines the concept of
primary energy as energy that has not undergone any Performing inventory analysis
energy conversion process [1]. The primary energy Performing impact assessment
factor must thus represent all the primary energy
consumed in order to provide one unit of heat or power Goal and Scope of the Study
to the consumer. Primary energy consumption of The main goal of this LCA study is to analyze the two
energy chains is not only based on the consumption of energy performance indicators presenting the primary
fuel (or other energy source) in the power or heat energy efficiency and the CO2 emissions for both the
generation process, but also all the primary energy electricity and heat production at Hellisheidi power
needed for the construction, operation and possibly plant. The LCA calculations and impact assessment
demolition of the production facilities. Also, some where done by using the LCA software SimaPro 7 [12]
primary energy is needed for the distribution of the and using different databases such as the Ecoinvent
product. To calculate such accumulated primary database [13] for the inventory information on various
energy, the method of life cycle assessment is well raw materials and processes used in the geothermal
suited. LCA is a method that has been developing power plant.
since the earliest performance of such a study in 1969
and standards on the methodology where issued in the There are numerous geothermal power plants
late 1990s [10]. worldwide using similar technology as the Hellisheidi
power plant to produce electricity (double flash power
LCA has been considered a good tool to achieve a plants produced 23% of the electrical power from
holistic approach on evaluating the environmental geothermal resources in 2007 [14]), so the results for
impact of products. Today, it is widely used to the energy performance indicators for the power
investigate all kinds of production systems and has production at Hellisheidi could be used to represent
given valuable insight on the total impact of products these power plants. Other types of geothermal energy
and systems on the environment by not only focusing conversion systems, such as single flash and binary
on the operational aspect [11]. Many interesting results systems, should be treated individually when
have been achieved by using this methodology and calculating energy performance indicators for the
those results form a basis for evaluating and comparing electricity production.
different solutions for production of various products,
such as vehicles for transport, soft drink containers and Geothermal combined heat and power plants are not
power conversion technologies. On the other hand, common worldwide, but regarding Europe they can be
LCA in the process industry has had much less found in Iceland as well as Austria and Germany. By
attention than for manufacturing products, and producing heat as well as electricity in geothermal
research is needed before complete methods for applications, the utilization of the heat taken from the
processes are readily available [11]. The application of geothermal reservoir in the form of geothermal fluid is
LCA on geothermal energy utilization can be valuable maximized. The heat produced has a variety of useful
for LCA developers working on further improvements applications, such as for district heating, agriculture,
and adjustments on the LCA methodology for the fisheries, swimming pools, snow melting and heating
process industry. up greenhouses [3]. The calculations of the primary
energy factor of the heat production at Hellisheidi
Using LCA to calculate the total primary energy geothermal CHP plant will emphasize this increase in
consumption and CO2 emission for heat and power thermal efficiency of the power plant.
187
th th
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
1 kWh electricity 1 kWh heat
Geothermal
fluid, from
electricity
Electricity, production Heat,
geothermal, at geothermal, at
Hellisheidi CHP Hellisheidi CHP
plant plant
Heat, from
condenser
Geothermal Geothermal
Geothermal Collection
fluid, at power heat production
power plant unit pipelines
plant unit
Drilling of
Power plant Power plant Geothermal Heating station Heating station
geothermal
equipment structures fluid, in ground Structure equipment
wells
Figure 2: Flow model for the life cycle assessment of the Hellisheidi CHP plant
Functional Unit plant unit, the geothermal heat production unit and the
The primary energy and CO2 factors are defined as geothermal fluid. The geothermal power plant is
primary energy usage and CO2 emission per MWh and constructed from the power plant structures and
thus, the functional unit of the study is chosen to be equipment while the fluid is transported in collection
MWh of electricity or heat produced in the Hellisheidi pipelines from geothermal wells that need to be drilled
geothermal CHP plant. The functional unit is the for the production. The heat production unit consists of
reference flow to which all other modelled flows of the the heating station structure and equipment. The
system are related. energy input into the heating process is waste heat
from the power production process in form of heat
System Boundaries taken from the steam in the condenser for preheating
of district heating water, and the waste geothermal fluid
The processes included in this LCA study are mainly
from steam separators used for final heating of the
the operation and construction of the power plant. The
district heating water. Inventory data on all these
demolition or end-of-life phase is disregarded due to
different components in the flow model was collected
insufficient information at this time. Also, the energy
and used for the LCA study of the Hellisheidi CHP
and material flows due to maintenance in the
plant.
operational phase of the power plant are disregarded
but both the demolition and the maintenance will be
Impact Categories and Methods for Impact
included in further studies. The time horizon in this
Assessment
study is chosen to be 30 years, which is the technical
lifetime of the power plant capital goods. To calculate the two energy performance indicators,
the two main impact categories to be used are the
A flow model of the CHP plant as modelled in the LCA primary energy demand of the production process in
study is shown in Figure 2. The two outputs of the MWh and Global Warming Potential (GWP) given in
production system are 1 MWh of electricity and 1 MWh CO2 equivalents.
of heat. The main material and energy inputs into the
energy conversion system are the geothermal power
188
th th
Two different methods of impact assessment had to be Data Quality

used in the impact assessment calculations. For the To calculate the energy performance indicators by
primary energy factor, the Cumulative Energy Demand methods of LCA, reliable inventory information is
(CED) method [15] was used which is based on a needed on material and energy flows to and from the
method published by Ecoinvent 1.01 and available in geothermal power production facilities during their
SimaPro 7 impact assessment methods. For the lifetime.. The inventory in this study is constructed
calculation of the CO2 emission factor, the IPCC 2007 from data provided by Reykjavik Energy, the power
GWP 100a V1.01 [16] was used to get the CO2 company in ownership of the Hellisheidi plant. The
equivalent total global warming potential for the chosen data on the construction phase is retrieved from the
functional unit of 1 MWh electricity produced. conditions and specifications in a tender for the
construction of the power plant, where quantitative
Principles for Allocation information is collected on all major material flows
To allocate the impacts of the different products, required for the constructions and machinery. The
electricity and heat, produced at Hellisheidi CHP plant, inventory information for the fluid collection and drilling
several methods can be used. The method used is retrieved from a report done by Reykjavik Energy,
should reflect the physical relation between the two including the power and performance of the geothermal
products, such as how the different inputs and outputs wells drilled for the power and heat production [17].
of the process are dependent on the two different
For a LCA study, the following data quality indicators
products. Simple methods of allocation for an energy
must be presented:
conversion process can be:
Time period
Based on energy content of the products
Region
Based on exergy content of the products
Type of technology and representativeness
Based on the monetary value of the products
Allocation
The abovementioned methods can be used when the System boundaries
physical relation between the two products is unclear.
In this study, the time period of the data is from 2005 to
In the case of the Hellisheidi CHP plant, the physical
2009 and the region is Western Europe. The type of
relation between the two outputs (electricity and heat)
technology is modern and the representativeness is
is mainly the use of waste heat from condensers and
data from a specific company. The allocation, as
the geothermal fluid from the production wells, as
mentioned before, is by physical connections between
shown in Figure 2. The impacts of construction can
the two outputs. The system boundaries are described
easily be divided between the electricity and heat
by three different criteria. First, the cut-off criteria is in
production with the detail of inventory data provided.
general set to be less than 5% which means that all
Also, the geothermal fluid used in the heat production
inventory data that does not contribute more than 5%
is taken from steam separators in the electricity
to the overall impacts of the two products is
generation process and would otherwise be reinjected
disregarded. Also, the system boundary is chosen to
back into the geothermal reservoir via reinjection wells.
be of the first order, only to account for the materials
The disposed heat in the condenser is utilized to used in the construction and operation of the CHP plant
preheat the district heating water by using it as cooling but not the processing and transportation of these
water. The condenser pressure determines the materials. The third system boundary criterion is the
temperature of the steam output from the turbines and system boundary with nature, which in this study is
thus, also the final temperature of preheating of the described as unspecified at this stage of the LCA
district heating water. If the heat demand is high, the study.
condenser pressure must be higher than the optimum
for power production in order to supply high enough RESULTS FOR THE ENERGY PERFORMANCE
temperatures to the district heating water. This limits INDICATORS
the electrical power production and requires that more
geothermal wells have to be drilled in order to sustain Energy Performance Indicators for Electricity
the electrical production under high thermal loads of Production
the district heating system. These limitations on the The results for the impact assessment of the electricity
electrical production imply that the allocation of impacts production alone, focusing on the two energy
from the drilling of wells should be related to the performance indicators, is shown in Table 2. The
number of wells that have to be drilled to sustain both highest value of fp 6.33 MWh primary energy/MWh
the electricity production and the highest thermal load produced energy, is obtained when no heat production
designed for the district heating system. is present at the power plant and the effects of
reinjection of waste streams is not taken into account.
189
th th
The value of 5.33 for fp is obtained in the two latter The factor K for the CO2 emissions is the same for all
cases, where the waste heat is either reinjected back three cases of electricity production as reinjection and
into the reservoir or used for heating of DH water. In utilization of waste stream does not have significant
those cases, the primary energy content of the waste effects on the total emissions due to the power for
stream can be subtracted from the primary energy production. The origins of the CO2 emissions can be
content of the geothermal fluid used for the electricity seen in Figure 4. The largest contributor to the CO2
production, resulting in lower fp values. The share of emission from the electricity generation over 30 years
non-renewable primary energy sources such as oil and of production is the geothermal fluid, responsible 88%
gas used in the construction phase or in the of the CO2 emissions per kWh of electricity production.
manufacturing of various power plant components, only
account for about 0.01 of the total fp value in all cases.
Table 2 Results for the primary energy factor and CO2 emission factor for electricity based on geothermal energy
Primary energy factors fp

CO2 production
[MWh primary energy / MWh produced coeff. K
Source of electricity
energy]
[Kg/MWh]
Non-Renewable Total
Electricity from Hellisheidi geothermal power
0.01 6.33 29
plant
Electricity from Hellisheidi geothermal power
0.01 5.33 29
plant, with reinjection
Electricity from Hellisheidi CHP plant 0.01 5.33 29
A small share of 8% originates from the drilling of value reduces to 0.69. In both cases, the share of
geothermal wells while the construction of the power primary energy from non-renewable energy sources is
plant, along with the manufacturing of its main less than 0.01. In both cases, the CO2 production
components, is responsible for 4% of the CO2 coefficient is 0.98 kg CO2 equivalents per produced
emissions. MWh.
The origins of the CO2 emission from the heat
GWP 100a for Electricity Production generation process can be seen in Figure 4. The
in kg CO2 eq largest contributor to the total emissions is the drilling
of the geothermal production wells that were needed to
0.5%
4% Geothermal fluid sustain the electricity production while the heat
8% (87.5%) production is at maximum load of 133 MWth. The
Power plant and manufacturing of the district heating pipeline from the
components (4%)
production area to the rural area of Reykjavk city
Geothermal well
contributes to 15% of the total emission resulted from
87.5% drilling (8%)
the heat generation process.
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 fp is obtained in the case where reinjection is not Figure 4 Origins of CO2 emissions from the different
taken into account, with the value of 1.78 MWh primary processes of the heat generation
energy/MWh produced energy. With reinjection, the
190
th th
Table 3 Results for the primary energy factor and CO2 process. Reinjection of geothermal brine is recognized
emission factor for heat from a geothermal CHP plant to improve heat mining and stabilize the production
capacity of geothermal fields, if successfully carried
Primary energy factors fp out. It can also counteract pressure draw-down in the
CO2 reservoir by providing an artificial water recharge [18].
[MWh primary energy /
Source of production
MWh produced energy] In this study, reinjection of the waste stream is
heat coeff. K
[Kg/MWh] modelled, which decreases the use of primary energy
Non- in the energy conversion process, since a part of the
Total
Renewable
primary energy from the geothermal fluid it is returned
Heat,
Hellisheidi >0.01 1.78 0.98 back to the reservoir. Reinjection is present at the
CHP plant Hellisheidi geothermal CHP plant so the values of the
Heat, energy performance indicators with reinjection are valid
Hellisheidi for the power plant.
>0.01 0.69 0.98
CHP plant,
reinjection
4) Life cycle assessment is especially useful to
evaluate the total impact of geothermal power plants
DISCUSSION
with respect to their emission of greenhouse gasses.
The following discussion highlights the most significant Figure 3 and Figure 4 show how the different phases in
results from this study: the life cycle of the power plant significantly contribute
to the overall emission in CO2 equivalents. If LCA had
1) By comparing the energy performance indicators
not been carried out for the process, 12% of the CO2
calculated in this study and shown in Table 3 to the
emissions resulting from the electricity generation
indicators given in Table 3 it can be seen that electricity
would not have been accounted for and no emissions
from geothermal power plants has the highest total fp
would have been found for the heat production, since
factor while the share of non-renewable energy
the emissions from drilling, construction of buildings,
sources is the lowest. The main reason for the high fp
and manufacture of components had not been
factor is the low conversion efficiency of geothermal
accounted for.
power plants due to low working temperatures and
pressures. The CO2 production coefficient is relatively
low compared to the other energy conversion ACKNOWLEDGMENTS
technologies and could be lowered even further if Special thanks are given to the following partners:
measures are taken to control the emissions from the Nordic Energy Research (NER) for funding the study
power plant. The results for the Hellisheidi geothermal and the Energy Research Fund of Landsvirkjun for their
CHP plant cannot be used to represent all geothermal support. To Orkuveita Reykjavkur for providing data for
power plants producing either electricity alone or with a Hellisheidi Power plant, to Mannvit engineering for
combined production of electricity and heat. Further discussion and data provision and to Ragnar Gylfason
studies are needed on different types of geothermal for his contribution in the data gathering phase.
power plants, such as single flash and organic Rankine
cycles, to be able to produce specific or average
REFERENCES
factors representing geothermal utilization.
[1] EU. (2003, January 4). Directive 2002/91/EC of
2) The results for the heat production at the
the European Parliament and of the Council of 16
Hellisheidi geothermal CHP plant, given in Table 3,
December 2002 on the energy performance of
show that the energy performance indicators are
buildings. Official Journal of the European
relatively low and, in the case of reinjection, below
Communities .
unity. This is because the primary energy needed to
preheat the DH water is not accounted for in the heat [2] Bertani, R. (2010). Geothermal Power Generation
production but rather assigned to the electricity in the World 2005 2010 Update Report.
production. This is due to the fact that the preheating of Proceedings World Geothermal Congress 2010,
the DH water from 5 C to 41 C is done in the (April), 25-29.
condenser for the high pressure steam turbine as seen
[3] Lund, J. W., Freeston, D. H., & Boyd, T. L. (2010).
in Figure 1 and is a necessary step in the electricity
Direct Utilization of Geothermal Energy 2010
production, but a beneficial step in the heat production
Worldwide Review. Proceedings World Geothermal
for the DH system.
Congress 2010, (April), 25-29.
3) Values for the indicators for both electricity and [4] EN 15603:2008. Energy performance of buildings.
heat are calculated with and without reinjection of the Overall energy use and definition of energy ratings.
cooled geothermal brine from the energy conversion
191
th th
Geneva: International Organisation for [12] PRConsultants. (2009, September 6). SimaPro
Standardisation (ISO). LCA software. Retrieved October 14, 2009, from
[5] Helisheidi Geothermal Plant. (2009). Retrieved SimaPro LCA software: http://www.ecoinvent.ch/
May 2009, from http://www.or.is/English/Projects/ [13] Ecoinvent. (2009, August 13). Home. Retrieved
HellisheidiGeothermalPlant/ October 14, 2009, from Home:
[6] H. Douglas Lightfoot. (2007). Understand the three http://www.ecoinvent.ch/
different scales for measuring primary energy and [14] DiPippo, R. (2008). Geothermal Power Plants
avoid errors. Energy, 32, 1478-1483. Principles, Applications, Case Studies and
[7] Segers, R. (2008). Three options to calculate the Environmental Impact (2nd edition ed.). Oxford:
percentage renewable energy: An example for a Butterworth-Heinemann.
EU policy debate. Energy Policy , 36 (9), 3243- [15] Klpffer, W. (1997). In defense of the cumulative
3248. energy demand (editorial). International Journal of
[8] IEA. (2004). Energy Statistics Manual. Life Cycle Assessment , 2, 61.
International Energy Agency (IEA). Paris: [16] PRConsultants. (2009, September 6). Methods.
OECD/IEA. Retrieved October 14, 2009, from SimaPro LCA
[9] ISO 2533:1975. Standard atmosphere. software:
International Organization for Standardization, http://www.pre.nl/simapro/impact_assessment_met
Geneva, Switzerland. hods.htm#CML2
[10] Russell, A., Ekvall, T., & Baumann, H. (2005). Life [17] Gunnlaugsson, E., & Oddsdttir, A. L. (2009).
cycle assessment - introduction and overview. Helisheidi - Gufuborholur 2008 (Hellisheidi - Steam
Journal of Cleaner Production , 13 (13-14), 1207- wells 2008). Reykjavk: Orkuveita Reykjavkur.
1210. [18] Stefansson, V. Geothermal reinjection experience.
[11] Baumann, H., & Tillman, A.-M. (2004). The Hitch Geothermics, 26, (1997), 99130.
Hiker's Guide to LCA. Lund, Sweden:
Studentlitteratur AB.
192
th th
FLEXIBILITY FROM DISTRICT HEATING

TO DECREASE WIND POWER INTEGRATION COSTS
1 2
J. Kiviluoma and P. Meibom
1
VTT Technical Research Centre of Finland
2
Ris DTU
ABSTRACT INTRODUCTION
Variable power sources (e.g. wind, photovoltaics) Wind power is projected to be a large contributor to
increase the value of flexibility in the power system. fulfil electricity demand in several countries. This could
This paper investigates the benefits of combining take place due to relatively low cost of wind power
electric heat boilers, heat pumps, CHP plants and heat electricity or policy mechanisms promoting renewable
storages in a district heating network when the share of energy. In any case, power systems with a large
variable power increases considerably. The results are fraction of power coming from a variable power source
based on scenarios made with a generation planning will need to be flexible. Flexibility is used to cope with
model Balmorel [1]. Balmorel optimises investments the increased variation in residual load (electricity
and operation of heat and power plants, including heat demand minus variable power production) and with the
storages. It uses hourly resolution and enforces increased forecast uncertainty in the residual load. On
temporal continuity in the use of the heat storages. the other hand, lack of flexibility will cause larger costs
Scenarios with high amount of wind power were from increased variability and forecast errors.
investigated and the paper describes how the increase Therefore, it is prudent to investigate the cost optimal
in variability changes the profitability and operation of configurations for the combined power and heat
different district heating options in more detail than was generation portfolios.
described in the article by Kiviluoma and Meibom [2].
Heat generation could offer significant possibilities for
Results show that district heating systems could offer
increasing the flexibility of the power system. Currently,
significant and cost-effective flexibility to facilitate the
part of the inflexibility of the power system comes from
integration of variable power. Furthermore, the
CHP plants that are operated to serve the heat load
combination of different technologies offers the largest
while electricity is a side product. Installation of electric
advantage. The results imply that, if the share of
resistance heaters next to the CHP units or elsewhere
variable power becomes large, heat storages should
in the heat network could break this forced connection.
become an important part of district heating networks.
During periods of low power prices, which will become
more common with high share of wind power, CHP
NOMENCLATURE plants could be shut down and heat would be produced
with electricity. The dynamics can be made more
Indices economic with the use of heat storages. Further option
i, I Unit, set of units is to have heat pumps in the DH network, but they will
I
HeatSto
Heat storage units require large amount of full load hours to be profitable
t, T Time steps, set of time steps
and will compete with CHP plants for the operating
space.
a, A Area, set of areas
Variables In most countries heat demand is in the same order of
C New capacity magnitude as electricity demand. For example, in UK
the demand for primary energy due to heat is around
P Power generation
40% of total primary energy demand [3]. About 25% of
Q Heat generation
the primary energy demand is due to space and non-
Z Charging of heat storage industrial water heating. In the US all kind of heat use
Parameters accounts for about 30% of the primary energy
Inv
c Annualized investment cost consumption [estimated from 4].
Fix
c Fixed operation and maintenance costs Heat is inexpensive to store compared to electricity.
Operation
c Operation cost function of the unit Electricity storage has been seriously considered to
w Weight of time period alleviate the variability of wind power [5-6]. Therefore, it
h Heat demand 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
193
th th
optimal investments in new power plants and heat Loading of heat storage adds to the heat demand. Loss
storages. during the heat storage process is not considered. The
dynamics of heat networks were not taken into
The study has been restricted to residential and
account.
industrial district heating systems. Buildings not
connected to district heating systems were not
considered, although these also require heat. Cooling
Q
iI
i ,t hr ,t Z i ,t
t T ; a A (2)
iI aHeatSto
demand could also offer similar possibilities, but the
problem was not addressed here. Industrial heat Analysis is done for the year 2035. By this time, large
demand and water heating do not usually have strong portion of the existing power plants are retired. Three
seasonal variation and can therefore be more valuable district heating areas were considered. These have a
towards the integration of variable power. rather different existing heat generation portfolio by
2035. This helps to uncover some interesting dynamics
in the results section.
METHODS AND DATA
In this paper, scenarios without new nuclear power are
The model and assumptions used for the analysis are
compared (scenarios Base NoNuc and OnlyHeat
described in more detail in [2]. For convenience, most
NoNuc in article [2]). This meant that wind power had a
important sections are referenced below. The heat
very high share of electricity production. Accordingly,
sector of the model is described more thoroughly here.
there was more demand for flexibility in the system.
The Balmorel model is a linear optimization model of a
Urban area presents the heat demand in the capital
power system including district heating systems. It
region of Finland. The existing power plants in 2035
calculates investments in storage, production and
cover over half of the required heat capacity. Largest
transmission capacity and the operation of the units in
share comes from natural gas, which is a relatively
the system while satisfying the demand for power and
expensive fuel in these model runs. The annual heat
district heating in every time period. Investments and
demand is smallest of the considered areas: 6.2 TWh.
operation will be optimal under the input data
assumptions covering e.g. fuel prices, CO2 emission Industry area aggregates the known industrial district
permit prices, electricity and district heating demand, heating demand from several different locations. This is
technology costs and technical characteristics (eq. 1). a necessary simplification, since Finland has over
The model was developed by (Ravn et al. [1]) and has hundred separate DH areas and the model would not
been extended in several projects, e.g. (Jensen & be able to optimise all of these simultaneously. The
Meibom [10], Karlsson & Meibom [11], Kiviluoma & industrial heat demand in Finland is driven by paper
Meibom [2]). and pulp industry, which produces waste that can be
used as energy input. This capacity is assumed to be

min ciInvCi ciFix CiEx Ci wt ciOperation Pi ,t , Qi ,t (1) available in 2035 and as a consequence the model
iI iI tT iI does not need more industrial heat capacity. The
The optimization period in the model is one year annual heat demand is 46.8 TWh.
divided into time periods. This work uses 26 selected Rural area aggregates non-industrial heat demand
weeks, each divided into 168 hours. The yearly excluding the capital region considered in Urban. This
optimization period implies that an investment is carried is probably the most interesting example, as the
out if it reduces system costs including the annualized existing capacity covers only 20% of the heat capacity
investment cost of the unit. demand. Therefore, the model has to optimise almost
The geographical resolution is countries divided into the whole heat generation portfolio. There are wood
regions that are in turn subdivided into areas. Each resources (limited amount of forest residues and more
country is divided into several regions to represent its expensive solid wood) available unlike in the urban
main transmission grid constraints. Each region has area. The annual heat demand is 21.0 TWh.
time series of electricity demand and wind power
production. The transmission grid within a region is RESULTS
only represented as an average transmission and
Figures 13 give an example how heat production
distribution loss. Areas are used to represent district
meets heat demand in the different areas during the
heating grids, with each area having a time series of
same 4.5 days in January. Negative production
heat demand. There is no exchange of heat between
indicates charging of heat storage. Electricity price is
areas. In this article, Finland is used as the source for
on separate axis together with the cumulative content
most of the input data.
of heat storage. When electricity price is low, storage is
The hourly heat demand has to be fulfilled with the heat loaded with electricity using heat boilers and heat
generation units, including heat storages (eq. 2). pumps. When electricity price is high, CHP units
194
th th
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).
1200 450
NG_EX_UR
1000
Heat storage content (%)

Electricity price (/MWh)
375 NG_BP_UR
Heat production (MW)
800
MW_HB_UR
600 300
MW_BP_UR
400
225 EL_HP
200
EL_HB
0 150
Storage use
-200
75 Stor. content
-400
Elec. price
-600 0
Fig. 1. Example of operation in Urban heat area. Negative production indicates charging of heat storage.
4000 450
MW_HB_RU
WR_EX

3000 375
WW_EX
NG_BP_RU
2000 300
NG_CC_EX
PE_BP_RU
1000 225
WO_BP_RU
EL_HP
0 150
EL_HB
-1000 75 Storage use
Stor. content
-2000 0 Elec. price
Fig. 2. Example of operation in Rural heat area. Negative production indicates charging of heat storage.
6000 450
PE_BP_IN
5000 375
4000 WR_BP_IN
3000 300 WW_BP_IN
2000 225 EL_HB
1000 150 Storage use
0 Stor. content
-1000 75
Elec. price
-2000 0
Fig. 3. Example of operation in Industrial heat area. Negative production indicates charging of heat storage.
17500
15000
Wind
Electricity production (MW)
12500 Natural gas (NG)

10000 Hydro
Wood waste (WW)
Fig.7500
1. Example of operation in Urban heat area. Negative production indicates charging of heat storage.
Peat
5000 Solid wood (WO)
2500 Forest residues (WR)
Municipal waste (MW)
0
Nuclear
-2500
Electricity to heat
-5000
Fig. 4. Electricity production. Negative production indicates the use of electric heat boilers and/or heat pumps.
195
th th
Effects of heat measures in the three heat areas

3500 HEATSTOR
In the Industry heat area availability of heat measures
Heat production (GWh)

EL_HP
Heat capacity (MW)

(electric heat boilers, heat pumps, and heat storages) 3000
had relatively little effect (Fig. 2). The main reason is 2500 EL_HB
that the existing heat production capacity from 2000 NG_HB
industrial wood waste and the associated no-cost CO_EX
1500
1000 WO_HB
500 NG_EX_UR
25000
0 NG_BP_UR
Heat capacity (MW)
OnlyHeat
OnlyHeat
20000 HEATSTOR MW_HB_UR
Base
Base
EL_HB MW_BP_UR
15000
WW_BP_IN
10000 WR_BP_IN
Cap. Prod.
PE_BP_IN
5000 1
NG_BP_IN Fig. 6. Heat capacity and production in the Urban heat
0 area.
FO_BP_IN
OnlyHeat
OnlyHeat
Base
Base
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
Cap. Prod.
heat boilers were used to charge heat storage.
Fig. 5. Heat capacity and production in the Industrial
8 Accordingly, during average electricity prices heat was
heat area. 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
waste wood were not easily replaced. However, there of heat from the heat storages.
were some high wind situations with low power prices
where it was beneficial to use electric heat boilers to The most important difference between Urban and
produce heat and decrease heat production from wood Rural heat areas is the availability of wood residues in
waste in the Industry area. There was an annual the Rural heat area (Fig. 7). For the most part this
resource limit on wood waste on the country level and resource was able to outcompete heat pumps as
the wood waste use was transferred to the Rural heat means to produce heat. Heat measures still helped to
area. It was also profitable to install some heat storage replace coal CHP. The combination of electric heat
capacity. This enabled the full shut down of wood boilers and heat storages was again a large source of
waste back pressure power plants for the duration of additional flexibility to the system.
low electricity prices. This decreased electricity
production and gave more room for the upsurge in wind HEATSTOR
12000
power production. EL_HP

Heat capacity (MW)
10000
EL_HB
In the Urban heat area heat measures enabled the 8000 NG_HB
replacement of CHP coal units with production from
6000 NG_CC_EX
heat pumps and to smaller extent from electric heat
4000 CO_EX
boilers (Fig. 6). Also wood based heat boilers were
replaced. Investment in heat storage was relatively WR_EX
2000
smaller. However, they were cycled more due to faster 0
WW_EX
charging rate. WO_HB
OnlyHeat
OnlyHeat
Base
Base
WO_BP_RU
PE_BP_RU
NG_BP_RU
Cap. Prod.
MW_HB_RU
1
Fig. 7. Heat capacity and production in the Rural heat
area.
8
Heat production is from the modelled 26 weeks and should be
multiplied by 2 to get an estimate on annual production.
196
th th
Dynamics of heat storage In the Rural area during winter time, charging of heat
Most of the daily fluctuation in heat demand was storages is mostly based on the use of electric heat
smoothed with heat storages and electric heat boilers boilers. They create large amount of heat in relatively
in all heat areas. If CHP units were operated, they were short time during periods of low power prices. During
usually operated at maximum heat output. summer time, heat storages are charged by turning on
wood waste and forest residue CHP units. During
The investment cost for heat storage was assumed to spring and fall CHP units operate more often, since the
be 1840 /kWh. With the assumed ratio of 12 between heat load is larger, but still the heat storage helps to
storage capacity and heat capacity this translates to shut them down for periods of some hours.
153 /kW. In comparison the capacity cost of electric
heat boilers was assumed to be 40 /kW and 50 /kW Urban area has similar dynamics, but during summer
for natural gas heat boiler. This means that investment time the adjustment is made by heat pumps instead of
into heat storage capacity was not driven by need for CHP. In the winter during high power prices old natural
new capacity since heat boilers were cheaper. There gas CHP units are less expensive to operate than the
had to be operational benefits from the use of heat heat pumps.
storage to cover the additional investment costs.
CONCLUSIONS
Heat storages create operational benefits by moving
consumption from more expensive sources of heat to District heating systems offer good possibilities for
less expensive by shifting demand in time. In all increasing the flexibility of the power system, if the
heating areas whole operating ranges of heat storages penetration of variable power like wind power increases
were extensively utilized. During most 168 hour periods greatly in the future. According to the results, main
heat storage reached both the minimum and maximum vessels to increase flexibility are the use of heat
storage capacities. In the Rural area heat storage was storages, electric heat boilers and flexible operation of
2.1% of the time either full or empty. With a larger CHP units.
storage capacity this could have been reduced, but it
Investment in electric heat boilers in district heating
was not worth the investment.
systems is driven mainly by periods of very high wind
The size of the heat storage in Industry area was power production. The resulting cheap electricity is
larger than in other areas in relation to daily heat converted to heat and to some extent stored in heat
demand (Fig. 8). In Industry area charging of heat storages for later use. Investments in heat storage in
storages took place over several days during higher turn are driven by the same mechanisms, but also to
power prices, when wood waste CHP units were create flexibility in the electricity production when prices
producing extra electricity. Storing the extra heat are higher. To enable this, the operation of CHP units
required larger heat storage capacity. On the contrary, and heat pumps is altered with the help of heat
in Rural and Urban charging and discharging was storages. Heat pumps mainly compete against CHP as
more balanced and smaller heat storage was enough. 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
180 a considerably penalty due to CO2 cost. Heat pumps
Heat storage size
are not very important as a source of flexibility, since
160 Max daily heat
they require lot of full load hours due to their
140 Min daily heat investment cost.
Average
While the research has been conducted on district
Heat (GWh)
120
heating, similar dynamics could be achieved in
100
household heating not connected to district heating
80 networks. However, the costs are likely to be larger
60 unless there is an existing hot water tank. Flexibility
could also be gained from district cooling or air-
40 conditioning units with the addition of a cold storage.
20 Further research should also address some of the
0 shortcomings of current study. Sensitivity analysis
would be important, especially concerning the cost
Rural Urban Industry
estimates of the analysed heat measures. Heat storage
model was very simple and this should be improved.
Fig. 8. Heat storage size compared to maximum, minimum Heat grade, especially in the industrial environment,
and average daily heat demands. can vary and the model should take this into account.
197
th th
Heat pumps were assumed to work at constant COP [6] J.K. Kaldellis and D. Zafirakis, Optimum energy
and this is a crude approximation even if the heat storage techniques for the improvement of
source is groundwater or sea water. renewable energy sources-based electricity
generation economic efficiency, Energy, Vol. 32,
REFERENCES pp. 22952305. Elsevier.
[7] H. Lund and E. Mnster, Modelling of energy
[1] H. Ravn et al. Balmorel: A Model for Analyses of
systems with a high percentage of CHP and wind
the Electricity and CHP Markets in the Baltic Sea
power, Renewable Energy, Vol. 28, 2003, pp.
Region. Balmorel Project 2001. See also:
2179-2193. Elsevier. doi:10.1016/S0960-
http://www.balmorel.com/Doc/B-
1481(03)00125-3
MainReport0301.pdf
[8] H. Lund, Large-scale integration of wind power
[2] J Kiviluoma and P. Meibom, Influence of wind
into different energy systems, Energy, Volume 30,
power, plug-in electric vehicles, and heat storages
Issue 13, October 2005, pp. 2402-2412. Elsevier.
on power system investments, Energy, Volume
doi:10.1016/j.energy.2004.11.001
35, Issue 3, March 2010, pp. 1244-1255. Elsevier.
doi:10.1016/j.energy.2009.11.004 [9] H. Lund, B. Mller, B.V. Mathiesen and A.
Dyrelund, The role of district heating in future
[3] Energy consumption in the UK: overall data tables,
renewable energy systems, Energy, Vol. 35, 2010,
2009 update. Department of Energy and Climate
pp. 1381-1390. doi:10.1016/j.energy.2009.11.023
Change - secondary analysis of data from the
Digest of UK Energy Statistics, Office of National [10] K. Karlsson and P. Meibom, Optimal investment
Statistics and the Building Research paths for future renewable based energy systems
Establishment. Using the optimisation model Balmorel,
International Journal of Hydrogen Energy Vol. 33,
[4] Annual Energy Review 2008. U.S. Energy
2008, pp. 1777-1787.
Information Administration.
[11] S.G. Jensen and P. Meibom, Investments in
[5] H. Ibrahim, A. Ilinca and J. Perron, Energy storage
liberalised power markets. Gas turbine investment
systemsCharacteristics and comparisons,
opportunities in the Nordic power system, Int. J.
Renewable and Sustainable Energy Reviews,
Electr. Power Energy Syst. Vol. 30, 2008,
Volume 12, Issue 5, June 2008, pp. 1221-1250.
pp. 113124.
Elsevier. doi:10.1016/j.rser.2007.01.023
198
th th
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 Easier to optimize the operation that leads to

higher conversion efficiencies.
If daily heat load variations could be eliminated in
Less need for maintenance because of more
district heating-systems, it would make the operation of
smooth operation of the plants
the district heating system less costly and more
competitive . There would be several advantages in the To do this some questions need to be answered:
operation such as:
What input and output capacity to/from the heat
Less use of expensive peak load power where storage is needed?
often expensive fuels are used. What size of the heat storage is needed?
Less need for peak load power capacity. Are the daily heat variations in the specific system
Easier to optimize the operation that leads to large or small during a year?
higher conversion efficiencies.
Less need for maintenance because of more METHOD
smooth operation of the plants
Nomenclature
There are a number of ways to handle the daily
Ph= Present hour value [MWh/h]
variations of the heat load. Two often used are large
heat storages or using the district heating network as Pd= Mean hour value for the present day [MWh/h]
temporary storage. If it would be possible to centrally Pa= Mean hour value for the whole year [MWh/h]
control the customer substations, it would also be
possible to use heavy buildings connected to the Sh = Energy transfer capacity [MWh/h]
district heating system as heat storages. Sd = Size of heat storage [MWh/day]
To be able to find the best way to reduce or even Sa = Total annual daily heat load variation
eliminate the daily heat load variations, you need to h = Momentary daily variation [h/h]
understand the characteristics of the daily variations.
This paper will describe a way of characterizing daily d = Total daily variation [h/day]
heat load variations in some Swedish district heating-
systems.
a = Total annual relative daily variation [h/year]
INTRODUCTION Variables
Measured data has been collected from some district
For all heat generation/distribution systems, heat load
heating systems in Sweden. The collected data is the
variations leads to inefficiencies. You need to design
heat power that is generated and fed into the district
your system for the peak load even though you only
heating network. It is hour mean power that is used, i.e.
need the top capacity for a very short period of time of
8 760 data points per year. Only whole years is used
the year. This is of cause expensive. The solution to
from 1 of January to 31 of December. To describe the
this problem is heat storage. There are a number of
daily variation three variables is defined.
possibilities to store heat in DH systems:
Large heat storages at the heat generation plants 1. Momentary daily variation ( h )
Heat storage in district heating networks 2. Total daily variation. ( d )
Heat storage in heavy buildings in by allowing
small variation in indoor temperatures[1]. 3. Total annual relative daily variation. ( a )
If it would be possible to extinguish daily variations it Three system examples are presented in this paper to
would lead to several profitable advantages such as: exemplify the method to characterize district heating
daily heat load variation:
Less use of expensive peak load power where
often expensive fuels are used. System A: From a city in South of Sweden with an
Less need for peak load power capacity. annual heat generation of 200 GWh.
199
th th
System B: From a city in Southwest of Sweden with an 1 24

annual heat generation of 64 GWh. Ph Pd
2 h 1
System C: From a city in the middle of Sweden with an d
Pa
annual heat generation of 1550 GWh.
The total daily variation is presented in Fig. 2 for the
three example systems. The figure verifies that the
Momentary daily variation ( h ) variations are more pronounced in the two smaller
The momentary daily variation is proportional to the systems compared to the larger system. Another
amount of heat that needs to be fed in or out to the DH implication is that the highest day values are very few,
network to extinguish the daily variation. This variable giving an incentive to construct heat storages
describe the heat power capacity needed for in and out somewhat smaller than the peaks in the figure. Hence,
put from and to the heat storage. For each district the investment costs will be reduced more the lost
heating systems you will get 8 760 (8 784 during leap benefits from the storage, giving a more optimised heat
years) values per system and year. storage.
The momentary daily variation is defined as the Total daily variation
difference of each hourly measured value and the 5
mean value of heat per hour of the same day divided 4,5
Systen A
System B
by the mean heat per hour of the year. 4

System C
Ph Pd
3,5
h
Total daily variation, d
Pa
[h/day]
2,5
The momentary daily variation is presented in Fig. 1 for 2
the three example systems. The figure shows that the 1,5
variations are more pronounced in the two smaller 1
systems compared to the larger system.

0,5
0
- 50 100 150 200 250 300 350
Momentary daily variation Days of the year
0,7 Fig. 2 Total daily variation sorted by size day by day for
0,6 Systen A
System B
the three different district heating systems.
0,5
System C
Total annual relative daily variation ( a )

0,4
0,3
Momentary daily variation, h
0,2
0,1
Total annual daily variation is a variable that is
[h/h]
-0,1 proportional to the total amount of energy that at daily

-0,2
-0,3
basis divert from the mean value accumulated for a
-0,4 period of one year. It is used to compare different
systems between themselves. For each DH systems
-0,5
-0,6
-0,7
0 1000 2000 3000 4000 5000 6000 7000 8000
you will get 1 value per system and year.
Hour of the year
Total annual daily variation is defined as the sum over

Fig. 1 Momentary daily variation sorted by size hour by
the year of the difference between each hourly
hour for the three different district heating systems.
measuring value and the mean value of energy per
hour of the same day divided by two times the mean
Total daily variation ( d ) energy per hour of the year.
Total daily variation is defined for each day and is a
1 8760,365
variable that is proportional to the amount of heat that
divert from the daily mean heat load. If you want to
Ph Pd
2 h 1, d 1
extinguish the daily variation in a system this variable
a
Pa
describe the size of the heat storage. For each DH
systems you will get 365 (366 during leap years) values The annual daily variation is presented in Fig. 3 for 10
per system and year. different Swedish district heating systems. Since the
The total daily variation is defined as the sum over the annual daily load variation has a magnitude of
day of the difference of each hourly measuring value 250500 h, only 36% of the annual heat load is
and the mean value of energy per hour of the same generated above the daily average heat loads. Hence,
day divided by two times the mean energy per hour of it is the seasonal variations that dominate the heat load
the year. variations in the Swedish district heating systems.
200
th th
CONCLUSIONS
Total annual daily variation
500 An expected conclusion would be that large district

450
SYSTEM A
heating systems have smaller relative daily variations
( a ) than small district heating systems. There are two
400
Total annual daily variation, a
SYSTEM B
350
reasons for that:

300
[h/year]
250 SYSTEM C
1. In a large district heating system, the use of heat

200
150
100 power is spread on different distances from the heat

50 plant, i e the chilled water in the return pipe return back
0
1 10 100 1 000 10 000 to the heat generation at different time compared to
Annual heat supply [GWh]
when the return water left ach substation (geographical
Fig. 3 Total annual daily variation for 10 different district diversity)
heating systems in Sweden. 2. In large district heating networks, you would expect
that the operators have more active operation of the
RESULTS heat distribution network with respect to temporary heat
To characterize daily heat load variations in district storage.
heating systems three variables have been defined. But as can be observed in the Fig. 3 there does not
seem to be such a trend. One explanation could be that
h = Momentary daily variation
the heat users differ in different systems. e.g. in the
d = Total daily variation system in Fig 3 with an annual heat supply of 9 GWh,
mostly single and multi family houses are connected
a = Total annual daily variation and very few industry or office buildings are connected.
Together with the mean annual heat per hour (Pa) and Since there is a large diversity among the annual daily
the energy transfer capacity in and out of the heat variation more data need to be collected to be able to
storage, size of storage to extinguish the systems daily make any further conclusions.
variation and the total daily variation and can be
determined according to the expressions below. REFERENCES
Energy transfer capacity:
[1] Olsson L, Werner S: Building mass used as short
Sh = h Pa [MWh/h] term heat storage, The 11th International
Symposium on District Heating and Cooling
Size of heat storage:
Reykjavik 2008.
Sd = d Pa [MWh/day]
Total annual daily heat load variation:
Sa = a Pa [MWh/year]
201
th th
DISTRICT HEATING AS PART OF THE ENERGY SYSTEM:

AN ENVIRONMENTAL PERSPECTIVE ON PASSIVE HOUSES
AND HEAT REPLACING ELECTRICITY USE
1,2 3
Morgan Frling and Ingrid Nystrm
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 Increased energy efficiency is in itself a desirable goal

for a society it increases the robustness of the energy
Energy use for space heating, hot tap water and other
system and the possibilities for a resource efficient and
heat use at comparatively low temperature levels
more sustainable energy system in the long run.
represent a substantial part of the total energy use in
However, it is possible to create a system with higher
Sweden and countries with similar climate. It is thus of
environmental impacts with energy efficient buildings
importance to meet this demand in a way generating as
compared to less energy efficient buildings through
small environmental impact as possible. However, it is
choice of less good energy carriers. It is not enough
possible to create a system with higher environmental
that the individual parts of a system are good and
impacts with energy efficient buildings compared to
efficient to give a low environmental impact; the parts
less energy efficient buildings through choice of less
must be connected into the system in a good way.
good energy carriers. It is not enough that the
Thus it is important to identify system solutions that
individual parts of a system are good and efficient to
avoids sub optimization and gives us energy efficient
give a low environmental impact; the parts must be
buildings and an efficient energy system with good
connected into the system in a good way.
environmental performance.
From environmental perspective energy efficient In a synthesis studies within the framework of
buildings and district heating dont oppose each other Chalmers Energy Center [1] the role of district heating
good parts connected in a good system will give an in a future society with more energy efficient buildings
optimal. The results from the study of the three items of have been investigated. Here we report on general
household equipment show possibilities for district findings of this study with a special focus on the
heating to be an alternative with good environmental environmental performance of the possibility to convert
performance, but not under all heat generation some household electricity use into district heating - for
regimes. the use in dish washers, washing machines and tumble
driers [2]. The environmental performance is studied
INTRODUCTION using life cycle assessment methodology and different
assumptions regarding electricity and district heating
It is of importance to meet for space heating, hot tap
generation.
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 DISTRICT HEATING DEMAND SIDE
increasing the efficiency in the use phase and in the There are today several drivers in the direction of lower
heating systems of buildings as well as through heat total heat market for district heating in future [1].
generation systems with low environmental impact. Among possible such drivers in Sweden are:
During recent years there has been a focus on houses
with low need of space heating, low energy houses or Warmer climate (due to climate change)
passive houses. In such buildings the heat from the Higher energy prices
incoming sun radiation together with body heat from Increased environmental awareness
people living in the houses and different household Increased energy efficiency of existing building
equipment will cover the whole or at least substantial stock
parts of the space heating need over a year (extra Limited amounts of new housing
heating might be needed during the coldest days of a New housing more energy efficient
year). Hot tap water still need to be heated. For parts of
However, there are also possible drivers for a larger
the year this can be achieved by solar panels, but there
heat market in future, e.g.:
is a need for extra heating during winter. This might
result in the extra heating demand being covered by Increased wealth giving larger living space per
electricity, directly or indirectly. person and higher demands on comfort
202
th th
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 Fig. 2 Focus on the use of biomass e.g. for making optimal
amounts of high qualitative energy carriers with heat as a
A strategic role of district heating in the energy system residue (it could also e.g. be biomaterials production).
is the ability to utilize and deliver resources that
otherwise would have been lost. Among possible
At the same time we can also expect:
system drivers on the supply side in Sweden are [1]:
Increased competition for bio fuel resources
Increased utilization of industrial surplus heat
Higher prices on high quality energy carriers
Remaining large potential of waste incineration
(electricity and fuels) might drive towards smaller
Increase of CHP power production
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. 1 Possible change for district heating demand in

future decreasing demand but more even over the year.
a) b)
c) d)
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).
203
th th
THE OVERALL ENERGY SYSTEM If district heating should continue to be seen in general
as an environmentally preferable option it is important
The energy system of a country is complex, and it is
that district heating companies continue to develop
important to understand how changes in sub systems
district heating production in a favourable direction.
may affect the whole system. Sub optimizations might
easily occur. A simplified example of a situation where Heat for district heating should originate from resources
a more energy efficient building through sub that are otherwise wasted. In the long term that will
optimization of the total system gives a larger overall mean that bio fuelled district heating is not enough, but
primary energy need is illustrated in Fig. 3. Obviously it heat from other primary production like bio energy or
is possible to create a system with higher biomaterial combines producing transport fuels and/or
environmental impacts with energy efficient buildings bio based materials.
compared to a system with less energy efficient
buildings. It is not enough that the individual parts of a CONCLUSIONS
system are good and efficient to give a low
environmental impact; the parts must be connected into From environmental perspective energy efficient
the system in a good way. buildings and district heating dont oppose each other
good parts connected in a good system will give an
Thus it is important to identify system solutions that optimal. It is not enough that the individual parts of a
avoids sub optimization and gives us energy efficient system are good and efficient to give a low
buildings and an efficient energy system with a good environmental impact; the parts must be connected into
environmental performance. the system in a good way. The results from the study of
the three items of household equipment show
IMPLICATIONS OF NEW TYPES HEAT LOAD possibilities for district heating to be an alternative with
good environmental performance, but not under all
To better understand implications of different new types
heat generation regimes. Heat generation must
of heat load (as illustrated in the right hand side of
continuously be considered.
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 Fig. 4 Environmental impact from using district heat for
environmental impact category is considered. In Fig. 5 dishwasher, drier and washer. Case: Swedish av. district
this is exemplified with climate impact and acidification heating and European av. electricity.
impact.
204
th th
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.
Fig. 5 Environmental impact from using district heat for Gteborg Energi AB.
dishwasher, drier and washer. Case: bio based district [5] Morgan Frling (2004): Environmental limitations
heating production and Swedish av. electricity.
for the use of district heating when expanding
distribution into areas with low heat density. 9th
ACKNOWLEDGEMENT International Symposium on District Heating and
Financial support from the Knut and Alice Wallenberg Cooling, August 30-31, 2004, Espoo, Finland.
foundation and the Swedish District Heating
Association is gratefully acknowledged.
205
th th
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 oversized for all other heat loads. In addition, radiator

systems are generally also oversized for safety
The present paper describes how the control of a
reasons, as presented in both Swedish studies [3], [12]
radiator system connected to a district heating network
and international ones [5], [8] and [10], thus providing
via a heat exchanger can be optimised to provide the
further potential to reduce the return temperature.
lowest possible district heating return temperature. This
can be achieved for each operating point by employing
Objective
an optimal combination of radiator circuit supply
temperature and circulation flow rate. The control The objective of the study was to develop a control
algorithm gradually creates a modified control curve for algorithm for determining the optimal choice of supply
the radiator circuit, enabling it to consistently provide temperature and flow in an arbitrary radiator system for
an optimal cooling of the district heating water. Since every heat load in order to minimise the primary return
the heat exchanger is dimensioned for very low outdoor temperature.
temperatures, it is oversized for all other heat loads. In
addition, radiator systems are often oversized due to Limitation
safety margins. Such facts render it possible to reduce The present investigation has dealt with DH
the district heating return temperature. substations that were indirectly connected to the
DH network, i.e., hydraulically separated by HEXs.
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 OPTIMISED HEATING SYSTEM TEMPERATURES
that were carried out to practically verify it. The control There exist various ways to control the heat output in a
method could be implemented in any modern control heating system. Here, we have dealt with the prevailing
logics for adaptive control of a radiator circuit, and the control method used in Sweden; an outdoor
obtained results indicated that one can expect a temperature-compensated supply temperature,
lowering of the return temperature in line with previous ensuring that an adequate amount of heat is supplied
theoretical calculations. to the building at each outdoor temperature.
The benefits with regard to the primary return
INTRODUCTION
temperature from adjusting the flow according to the
The present paper demonstrates how the control of a heat load are known. The idea of using an optimal
radiator system connected to a district heating (DH) combination of flow and supply temperature was
network via a heat exchanger (HEX) can be optimised conceived by Frederiksen and Wollerstrand [2], and
to provide the lowest possible DH return temperature. this theory has been further studied [13] [11]. The
This is done by always choosing the optimal radiator guidelines from Euroheat & Power [1] state that the
supply temperature and flow rate. lowest return temperature is obtained by varying the
flow according to the consumption. If such a variable
Relevance of the topic flow is used, it is controlled by thermostatic radiator
Low return temperatures are beneficial for the valves (TRV) either in combination with a constant
production as well as the distribution of DH. A specific supply temperature or with an outdoor temperature-
advantage of the control method demonstrated in this compensated supply temperature. Langendries [4]
paper, as opposed to, for example, conventional low suggests a central control of the flow rate through the
flow balancing, is its robustness, enabling the lowest pumps rotating speed, but claims that it appears to be
possible return temperatures to be consistently a rather difficult and expensive system. Petitjean [9]
obtained. This is the case independently of the current proposes a lowering of the pump speed at low heat
outdoor temperature and heat load, even if the DH loads, when the TRVs are almost fully open, but finds it
supply temperature changes, the HEX becomes fouled, problematic to determine which parameter to use for
or the house heating requirements change. The idea is controlling the pump speed.
also to utilise the fact that, since a HEX is dimensioned It should be possible to implement the control algorithm
for an extremely low outdoor temperature, it is in fact presented in this paper in any modern, state-of-the-art
206
th th
control logics for building automation, which are today Table 1: A summary of flow-weighted mean primary return
often used for controlling DH substations. The control temperatures (bold) and resulting reduction for various
temperature programmes.
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).
Under the dashed line, results are shown for a system

100
Tp,s
that is oversized by 100 %. The first three temperature
90
Tp,r,opt programmes are 55/45, 60/40 and 80/30 C, whereas
80 Ts,s,opt the last two are optimised ones with variable flow.
Temperature [ C]
Ts,r,opt
The following conclusions could be drawn from the
70
Tp,r,55/45
60
table:
50
40
Primary return temperature reduction The oversizing of a radiator system leads, in itself,
to a significant reduction of the primary return
30
temperature, provided that some kind of
75 compensation has been made in order for the
Rel. flow [%]
ms
50
system to work properly, i.e., that an accurate
25
indoor temperature has been provided.
0
-15 -10 -5 0 5 10 15
Outdoor temperature [ C] By optimising the system (through the use of a
variable secondary flow), the primary return
Fig. 1 Temperatures with an optimised temperature curve temperature can be further reduced, especially if
and a variable flow in a 100% oversized system. The
the system is oversized.
primary return temperature from a 55/45 C programme is
shown for comparison. By extending the radiator HEX, the return
temperature can be further reduced with the
Flow-weighted, yearly mean primary return temperature programmes that employ a relatively
temperatures from the radiator HEX have been low flow.
calculated with regard to the outdoor temperature Regardless of the degree of oversizing, a
duration. Above the dashed line in Table 1, results are combination of an optimised temperature
shown for a correctly dimensioned system, with an programme and an extended HEX provides a
80/60C programme as well as with an optimised substantially reduced primary return temperature.
programme. The gain is estimated to just under two
degrees C. The last column shows how the primary The values presented in the table have been calculated
return temperature is affected when the length of the only for the radiator HEX. When considering the
HEX is doubled. This comparison can be justified by substations total return temperature, it can be said to
the fact that the primary return temperature is be smoothed by the DHW consumption. Calculations
significantly influenced by the lower secondary flow that corresponding to those in Table 1 for a parallel and a 2-
the optimisation entails, while the pressure drop and stage substation for 20 flats (based on the Swedish
heat transfer rate in the HEX can remain at a District Heating Associations recommendations for
magnitude close to the original ones. 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
207
th th
demonstrated [6] [3]. Euroheat and Power recommend i.e., a longer HEX. Furthermore, with optimised control,
that a 2-stage connection be used only in large multi- there exists a preparedness for future changes in
residential buildings if the primary radiator return system temperatures in the DH network. Should the
temperature is high. However, it should not be DH supply temperature be changed, an adaptive
employed if a low-flow heating system providing low control will ensure that the lowest possible return
return temperatures is used [1]. temperature is always achieved.
The advantage of extending the HEX when the In order to operate according to Fig. 1, the algorithm
secondary flow is low actually demonstrates the must combine a control of the radiator supply
optimisation problem: When the secondary flow is temperature with a control of the radiator flow as a
reduced, the secondary return temperature will function of the heat load and the DH supply
decrease. In the radiator HEX, the situation is different. temperature. In previous work [7], we have shown that
As the secondary flow decreases, the difference it is possible to manually determine the optimal radiator
between the primary and the secondary return supply temperature and flow. A natural continuation is
temperatures, increases as a result of the heat transfer to develop a method for automatic adjustment of
coefficient in the HEX being strongly flow dependent. parameter values for the optimal control algorithm.
Fig. 2 shows how the secondary return temperature is
lowered with a decreasing secondary flow while the THE TEST OBJECTS
difference between primary and secondary return
temperatures increases. This results in a primary return The tests have been carried out in four multi-residential
temperature that, at first, decreases and then increases buildings in the city of Karlshamn, Sweden. The houses
when the secondary flow is further reduced. The values were built in 1967-1968: three of them had three stories
in the figure have been taken from one of the test and a basement, and one had six stories and a
objects. For this heat load, the lowest primary return basement. The number of flats varied between 20 and
temperature was achieved for a secondary flow of 30 per house.
approximately 30 % of the original flow. The radiators in all houses were fitted with TRVs, but
these were at least ten years old. It was thus uncertain
40 7 whether they functioned properly. The circulation flow
Tp,r,rad was found not to vary significantly in any of the radiator
39
Ts,r
6 circuits, which may have been an indication that many
of the TRVs were not working. However, it should be
Return temperature, C
38 Grdigkeit 5
noted that the presented control algorithm is
Grdigkeit, C
Optimum, lowest Tp,r

37 4 independent of the use of TRVs in a system. Whatever
36 3
combination of optimal supply temperature and flow
that is identified for a given outdoor temperature, the
35 2 heat supply will be the same. The main task for TRVs
34 1
is to limit the heat supply in a room where additional
heat supply (solar radiation, bodily warmth or electrical
33 0 equipment) would result in an overheating of the room.
15 20 25 30 35 40 45 50
Flow, % The substations were of the 2-stage type and equipped
Fig. 2 Primary and secondary return temperatures, as well with control logics of the brand IQ Heat (Alfa Laval AB).
as the difference between them, as functions of the The equipment for the building automation was
radiator flow. manufactured by Siemens and furnished with a
separate communications module that could also be
Another reason for including the impact of an extended
used for executing minor computer programmes. There
HEX in the comparison in Table 1 is the opportunity of
was also an internet connection, rendering it possible
connecting to new installations. Large parts of the
to communicate in a number of ways, such as via the
housing stock in Sweden, built under strong political
software Saphir ScopeMeter (Siemens), or FTP. After
incentives during the 1960s and 1970s, are facing
a reconfiguration, the pump speed could be controlled,
substantial renovation needs. The results of this project
since all pumps were equipped with communication
can be considered consistent even if fewer radiator
modules.
systems be oversized in the future, whether
incorporated in older, renovated, or new buildings. The In order to monitor the circulation flow in the radiator
smaller potential for return temperature reductions circuits during the tests, clamp-on ultrasonic flow-
resulting from less oversized radiator systems may be meters were utilised. However, the objective was to
compensated by the ability to install a HEX that is develop a control algorithm based on modern, state-of-
dimensioned for of an optimised radiator programme, the-art equipment without using additional installations.
208
th th
To assure that the temperatures measured in the temperature around 0 C, which corresponded to a
substation corresponded to the average temperature load of approximately 50%. The actual flow rate was
levels in the various risers in the radiator circuits, about 1.1 l/s and the temperatures corresponded to
temperature sensors were installed in two of the 60/40 C, thus representing an oversizing around
houses. This enabled measurement errors or 100%.
disturbances in the radiator circuit to be identified. The
indoor temperature could be monitored thanks to six ADAPTIVE OPTIMISATION - METHOD
wireless sensors installed in each house in the area.
In the theoretical example, the system was assumed to
Modifications in the substations be 100 % oversized, while in an arbitrary system one
cannot be sure of the degree of oversizing. It is also
After some initial tests, the circulation pumps were
desirable to have a robust and adaptive control
found to be generally oversized to such an extent that
algorithm. The method found to function the best is
the flow rate could not be decreased as much as
described below. This approach consists in gradually
desired. There exists a predetermined minimum
modifying, by automatically performed tests, the control
rotational speed for this type of pump, implying that the
curve and determining the associated flow rate.
speed could be reduced by 6070%. Discussions with
the manufacturer revealed that the lowest pump speed
Online testing
could not be changed in this model, for which reason
the decision was made to throttle the flow with an By locking the control valve (CV), one can assume to
existing shut-off valve located after the pump, which have approximately the same primary flow through the
shifted the pumps operating range. The throttling was radiator HEX, and since the variations in the cooling of
conducted in order for the pump to give half the flow primary water is relatively small, the heat supply is also
rate at 100% rotational speed. The control curve was approximately constant. If the secondary flow is
modified accordingly, leading to the temperature drop reduced while the CV is maintained locked, the
in the radiator circuit becoming doubled and the heat temperature of the secondary flow leaving the HEX will
supply remaining unaltered. rise. When a new flow and its associated supply
temperature are tested, the current level of the primary
We were unable to receive a comprehensive reply from return temperature is compared to the level before the
the pump manufacturer with respect to the possible experiment. In this way, the new combination of flow
measures regarding the regulation of the pump. A and supply temperature can be either accepted or
discussion with another manufacturer implied that there rejected. This method renders it possible to implement
were no technical limitations for how far down the the adaptive algorithm in any arbitrary system, leading
pump speed could be controlled. However, such an to the control curve becoming gradually modified. This
extension of the manoeuvrable range has so far not method we suggested in [7].
been requested. After a simple modification of the
pumps frequency converter, the working range could One problem associated with this kind of optimisation is
be extended from todays 30100% to, in an extreme that the method is sensitive to disturbances. If the
case, 2100%. primary supply temperature, primary differential
pressure or the outdoor temperature changes during
Existing control of the radiator circuits the test, one cannot be sure that the heat supply is
constant. In that case, a reduced return temperature
Although the radiator circuits within the area were
could be the result of a heat supply that is too low.
designed by the same consultant, there is today a large
Such tests have to be rejected.
spread in the choice of control curve and resultant
temperature drop (1030 C). It is likely that the curves In order to render the tests less sensitive to
have been gradually adapted to the circuits hydraulic disturbances, the CV is locked only briefly, in order for
properties and balancing, and one can assume that this the HEX to stabilise. Subsequently, we return to
is a common situation. automatic control, but instead of using the control
curve, the control aims at maintaining a constant
When older houses are renovated and their radiator
temperature drop in the radiator system. If this is
circuits are modernised, there are no guarantees that
successful, the heat supply is also kept constant. One
oversizing is taken into consideration. For example, the
can assume that the secondary flow is relatively
radiator HEX in a substation that was installed in 2005
constant: as long as tests are conducted at night, no
in one of the houses was dimensioned for 185 kW heat
solar radiation is present and internally generated heat
output at DOT with temperatures corresponding to
is likely to be at a relatively steady level. If, for instance,
80/60 C at a flow of 2.25 l/s. However, when
the primary supply temperature or differential pressure
examining data for this substation, it turned out that the
rises during the course of a test, the CV will close
substation delivered less than 40 kW at an outdoor
somewhat causing the secondary supply temperature
209
th th
to decrease, and thereby also the temperature drop 80

Tp,s
and heat supply, to be detained at the same level. 70

Ts,r
Tp,r,rad
A test is started by keeping the CV locked for ten

Tp,r,tot
60
Ts,r
minutes. This leaves enough time for the HEX to Ts
Temperature [ C]
50
stabilise. The new level of the difference between the

To
40 To,dam p
primary and secondary return temperatures became Gr

30
stable already after about two minutes in the tested
objects. The CV was maintained locked for ten
20
minutes, which should be sufficient even for very low 10
flows and most types of HEXs. Subsequently, the 0
control was resumed in order to ensure a constant 00:30 01:00 01:30 02:00
Time
02:30 03:00 03:30 04:00
temperature drop on the secondary side.

20
CV,heat CV,DHW
The temperature drop was controlled by verifying the
%
10
current temperature drop, e.g., every five minutes, and 0

1
comparing it with the desired temperature drop, i.e., the 0.75
mp ms
temperature that was observed when the CV was
Flow [l/s]
0.5
locked. If the difference exceeded a certain value, 0.25
0.2 C has been used so far, the set-point for the 0
supply temperature was updated according to Heat supply [kW]

60 Qp Qs
Tsetpoint = Ts,r + Tsetpoint. 40
20
Fig. 3 displays a performed test: At 1:00 a.m., the CV 0

was locked and the radiator flow rate was reduced from
00:30 01:00 01:30 02:00 02:30 03:00 03:30 04:00
Time
0.59 to 0.36 l/s with the result that the secondary

Fig. 3 Results from a test. The flow was reduced at 1:00
supply temperature rose from 40 to 44 C. After ten and 2:30. The top graph shows temperatures in the
minutes, the temperature drop in the radiator circuit substation, the next graph presents the valve position for
was automatically controlled (in this case, the heat and DHW, and the last two display the primary
temperature drop was stable and it took more than 15 (including DHW) and secondary flow and the primary
(including DHW) and secondary heat supply, respectively.
minutes before the CV opening degree required
adjustment). After ninety minutes, the second flow An interesting aspect of this test was that the primary
reduction was carried out, to 0.24 l/s, and the supply temperature fluctuated a lot. Since the
secondary supply temperature increased to about
secondary temperature drop was kept constant, it had
48 C. no impact on the outcome of the test. One can see that
The total primary return temperature varied to a the CV generally demonstrated a lower opening degree
relatively large extent, partly because of tappings of later in the night, as opposed to before 1:00, when the
domestic hot water (DHW), but also due to the DHW primary supply temperature increased. Without the T
control in this substation being very unstable when no control, the heat supply would have been too high
tappings were made. However, the return temperature during the last part of the test.
from the radiator HEX was of interest for the tests. In
The radiator flow was altered by changing the set-point
this object, the difference between the primary and
for the pump speed, expressed as a percentage of the
secondary return temperatures was very small, and
maximum speed. It has been found that two flow
even for a low radiator flow, the grdigkeit was below
alterations of ninety minutes each are suitable per test,
one degree. One can see from the figure that the return
as this would allow the secondary return temperature to
temperature had fallen from just under 32 C to slightly
stabilise even at very low flows. The first test for any
over 28 C during the test. This resulted in, for a
outdoor temperature, as was the case in Fig. 3, means
current outdoor temperature of 8 C, the set-point for
that starting conditions include the original control
the secondary supply temperature being changed from
curve and flow rate. It is then desirable to perform two
40 to 48 C while the flow should be reduced from 0.59
fairly large flow reductions since, according to the
to 0.24 l/s.
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.
210
th th
point, one could expect a stable secondary return

11:50 PM < time < 0:00 AM temperature, e.g., during the last five minutes. In
Get To, use
modified control
Wait 10 min
addition to the secondary supply temperature, also the
Yes No curve
primary supply temperature is recorded. However, the
Wait 60 min with
constant pump speed
Save current
values
Set pump speed
for test
Test pump
speed 1
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
Set control Set control
Start timer Set T Wait 10 min
valve to Auto valve to Manual
Wait 5 min Test pump

subsequently kept constant at a level matching the
outdoor temperature (and heat load) at the time before
speed 2
< 80 min? No,

> 80 min
Pump speed 1? Yes the test was started.
Yes No, pump speed 2 The next step consists in using the information attained
from the test to modify the control curves. Initially, the
No T Tset-p > 0.2C Tset-p = Ts,r + Tset-p
original curve was used and the pump was, in our
Yes
Test done
case, controlled to give a constant differential pressure.

If the result of a test is that a lower primary return
Reject test
result
Not ok Check maximum
deviation for Q
Determine Tp,r,rad,min
(pump(0), pump(1)
temperature is obtained at a lower secondary flow rate,
the control curve is updated for that outdoor
(e.g., 5%) and To or pump(2))
Update Ok (e.g., 2C)
temperature. A reasonable resolution is 1 C. The

curves
Fig. 4 Flow chart describing the adaptive control algorithm. original control curve, generally based on 58 points,
was therefore initially extended to comprise values for
If a modified control curve is used before a test is about each outdoor temperature.
to start, the control should be interrupted and the pump
speed kept constant for an hour prior to the test. This If the experiment, as in Fig. 3 above, was performed at
way, one avoids the risk of the flow changing (due to 8 C, this point on the curve would be updated. Along
alterations in the outdoor temperature) too close to the with the new supply temperature there followed a new
test, which could result in unstable radiator system radiator flow, which in our case was expressed as a
temperatures. new set-point for the pump speed.
The supply and return temperatures were measured on The adaptive control continues in this manner night
four of the most remote risers from the substation, after night, and the control curves are continuously
during the tests. A continuous matching against updated. Outside the test periods of approximately
measurements on risers gives a good indication that three hours each night, the modified control curves are
the flow distribution in the system was not impaired by used for controlling the heating system.
the optimisation. The temperature profile was closely Fig. 5 shows an example of the gradual development of
matched to the profile at the substation. Both flow the modified control curve. The first graph shows a new
reductions resulted in increased temperature drops. point at 0 C (used for 0 0.5 C). In the second
(upper) graph, a point for 3 C has been added, while
Updating the control curves the range 0 to 3 C is complete in the third. The fourth
After the completion of a test, the obtained information graph shows a much more complete control curve
needs to be evaluated. The influence of the variation of (-5 to 10 C). Temperature curves corresponding to
the outdoor temperature is not entirely obvious; its constant flow systems with lower flows than the original
influence decreases with an increasing time constant system have been included as thinner lines. The value
for the building. Variations on the primary side normally for 10 C coincides with the curves of a system with a
have is compensated for since the heat supply is kept low flow, while the value of -5 C coincides with the
constant. As a result, it is sufficient to verify that the curves of a system with a moderately reduced flow
heat supply was maintained at a steady level during the (normal flow). The last graph clearly demonstrates that
test, avoiding any disruptions. the modified curves are based on a variable flow, i.e.,
they coincide with various constant flow curves at
If a test result is accepted, the primary return
different points.
temperatures for each tested flow are compared in
order to verify which flow resulted in the lowest return As shown in the second graph of Fig. 5, the modified
temperature. This flow also gave rise to a secondary curve could emerge in sections that subsequently are
supply temperature. It is however not obvious how to combined. One way to speed up the modification of the
read this temperature, given that it was regulated by control curves is to interpolate intermediate values
the controller and changed continuously. The most rather than wait for a flow optimisation at the missing
logical choice is to read the mean value at the end of outdoor temperature. Even the return temperatures
the test period, before the pump speed changes. At this could be interpolated, since it is possible to determine
211
th th
the required radiator flow for a known temperature drop where Ts,r,n is determined in analogy with Ts,s,n,
(and heat supply). according to:
70 70
Constant,
decreased flow
Ts ,r ,test Ts ,r ,n1
60 60
Constant, further
Ts ,r ,n
decreased flow
(3)
Original
Temperature
Temperature
50 control 50
curve
40
Modified
control curve
40
2
30 30
To ensure that the heat supply is kept constant, the

20 20
-10 -5 0 5 10 15 -10 -5 0 5 10 15
Outdoor temperature Outdoor temperature
70 70
required flow for the new temperature drop is
60 60 calculated. Since the flow is inversely proportional to
the temperature drop, it can be determined from the
Temperature
Temperature
50 50
40
last used flow and temperature drop, together with the
40
new temperature drop, according to:
30 30
20 20
-10 -5 0 5 10 15 -10 -5 0 5 10 15
(m Ts ) n1
Outdoor temperature Outdoor temperature
Fig. 5 A stepwise modification of the control curve. The

m s ,n (4)
supply temperatures are drawn in solid lines while the
Ts ,n
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 As mentioned earlier, the flow rate is set by changing
fully replace the original points on the curve. As more the set-point for the pump speed. According to the
tests are performed for the same outdoor temperature, affinity laws for fluid machines, the flow is proportional
one can proceed in several ways. Since the control to the rotational speed. The process of letting the last
should be adaptive and thus able to take into account modified supply temperature and the result of a new
changing circumstances both in the DH network and in test form a new modified supply temperature is
the building, the results of new tests should be illustrated in Fig. 6.
employed. However, one may expect that tests
60
performed close to one another in time, at equivalent Ts,s,test
outdoor temperatures, still provide slightly differing
results for varying reasons. A solution would therefore
50
be to use a forgetting factor, i.e., to gradually forget Ts,s,n
old values when the supply temperature curve is
Temperature
updated with new data. A possible approach for doing 40

Original
curves
so consists in calculating the new supply temperature, Modified
Ts,s,n, as a mean value of the obtained, Ts,s,test, and the curves
last used, Ts,s,n-1, supply temperature according to: 30
Ts,r,n
Ts ,s ,test Ts ,s ,n1 Ts,r,test
Ts ,s ,n (1) 20
2 2 4 6 8
Outdoor temperature
When a new test is performed at the same outdoor Fig. 6. An approach for modifying the control curve based
temperature, a new mean value is calculated, which on new test results.
means that older values will have less and less
The proposed method for updating the control curves
influence. To determine the secondary flow associated
indicates that if for instance the DH utility demonstrates
with the new supply temperature, i.e., the one providing
a long-term change in the supply temperature in the
the correct heat supply at the current outdoor
network, the control system gradually adapts to the
temperature, the expected temperature drop is
new temperature. However, there are always variations
calculated as:
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
Ts ,n Ts ,s ,n Ts ,r ,n (2) the outdoor temperature is expected to fall. Since the
primary supply temperature affects the primary return
212
th th
temperature, it is desirable for the adaptive control to the system decreases and all risers receive a more
also compensate for such short-term variations. One similar differential pressure. However, one must be on
way of doing so is to develop a number of parallel the look-out for errors (e.g., short circuits) in the
control curves for various intervals of the primary systems, a problem that is often emphasised in
supply temperature. If the temperature is greater than a connection with low-flow systems, as these tend to be
certain level, an alternative control curve is employed, more sensitive to hydraulic imperfections [12].
whereas if it is below a certain level, one utilises
another. This method has yet to be tested and there is Reduction of the primary return temperature
no basis for assessing how much impact one can To estimate a yearly mean return temperature
expect from normal variations in the supply reduction (as presented in Table 1) achieved by the
temperature or what would constitute reasonable adaptive control, an entire, or a major part of the,
intervals for parallel control curves in this case. Another heating season needs to be evaluated. The control
variant could be to perform a linear adjustment for the method presented in this paper was developed during
secondary supply temperature depending on the the winter and spring of 2009, and only a limited
primary supply temperature, according to: number of tests were performed during the spring.
However, Fig. 7 shows the obtained primary return
Ts ,s Ts ,s ,0 (1 a(Tp,s ,0 Tp,s )) (5)
temperature that was attained for the tests that were
performed in one of the houses. Note that these results
where a is a constant that can be determined from were first runs for each outdoor temperature (i.e., the
tests. flow was reduced to approximately 40%), signifying
that no further optimisations were undertaken. The
Regarding the measurement of temperatures and flows curve displaying the original return temperatures was
Regarding the temperature measurement in the based on the average return temperatures from the
substation, supply and return temperatures on both the radiator system prior to any of the modifications (i.e.,
primary and the secondary sides are required. One for the tests or the constant flow rate change, as
should keep in mind that, on the primary side, the described in section 3.1).
return temperature from the radiator HEX is needed
50
since the total return temperature is affected by the
DHW system. This temperature is normally available in
45 Tp,r,rad,orig
modern substation control equipment. Tp,r,rad,opt
Primary return temperature
It is desirable to avoid installation of a flow-meter in the 40

secondary circuit. On the primary side, where the
energy-meter is located, the total primary flow and the 35
total temperature drop in the substation are measured
and the energy required for DHW provision is thus 30
included. Since the tests are performed at night, DHW
tappings can be avoided to a large extent. By closing 25
the DHW CV for a short time, the primary flow passes
exclusively through the radiator HEX. By comparing the
20
average level of heat supply with a closed valve to the -10 -5 0 5 10 15
level prior to closing the valve, the flow required for Outdoor temperature
DHW re-circulation can be estimated.

Fig. 7. Primary return temperatures in the radiator system
In the test objects, indoor temperature measurements when the flow is reduced (dots), compared to the original
were used to verify that the adaptive control was able return temperatures (curve).
to give the correct indoor temperature. However, one
can in fact be sure that the correct amount of energy is CONCLUSIONS AND DISCUSSION
transferred to the system for each operating point, An adaptive control algorithm was developed in order
regardless of whether the original control curve or the to minimise the DH return temperature. The control
optimised curve is used. A possibility is that there is an algorithm can be implemented in any modern control
imbalance in the system. For example, the most distant logics for building automation. Some refinement may
riser may not receive the required flow because of a be done by compensating for short-term temperature
too low differential pressure when the pump speed is variations in the DH network. During the field studies,
decreased. It is, however, more likely that a better limitations in the speed control of the circulation pumps
balance in the system is achieved when the differential have presented a complication. A modification of the
pressure is lowered this since the pressure losses in frequency converter could increase the working range.
213
th th
There was not enough time to develop completely [5] Liao, Z., Swainson, M., Dexter, A.L., On the control
modified control curves for the test objects during the of heating systems in the UK, Building and
present heating season. On the other hand, a control Environment 40 (2005) 343-351.
curve with an adaptive controller is never definitive;
[6] Lindkvist, H., Walletun, H., Teknisk utvrdering av
rather it increases as more operational points (different
gamla och nya fjrrvrmecentraler i Slagsta
outdoor temperatures) are added and is then gradually
(Technical evaluation of old and new district
modified if outer conditions change. In order to receive
heating substations in Slagsta), Report 2005:120,
values for the primary return temperature on a yearly
Swedish District Heating Association, 2005.
basis using the adaptive control algorithm, the new
control curve needs to be modified for the entire [7] Ljunggren, P., Johansson, P.-O., Wollerstrand, J.,
temperature range. During the performed field studies, Optimised space heating system operation with the
the reduction of the primary return temperature was aim of lowering the primary return temperature,
about 3 C. Even though the test period limited the Proceedings from 11th International Symposium on
number of tests, the temperature range was still rather District Heating and Cooling, Reykjavik, 2008.
wide, including temperatures from -2 to 14 C.
[8] Peeters, L., Van der Veken, J., Hens, H., Helsen,
It is plausible that certain circuits are more suitable for L., Dhaeseleer, W., Control of heating systems in
a variable flow rate, e.g., depending on hydraulic residential buildings: Current practice, Energy and
balancing. It would also be possible to map out under Buildings 40 (2008) 1446-1455.
which circumstances other heat emitters than radiators,
[9] Petitjean, R., Total hydronic balancing, Tour &
such as fan coil heaters, can be included in a radiator
Andersson Hydronics AB, Ljung, Sweden, 1995.
circuit where the flow varies.
[10] Skagestad, B., Mildenstein, P., District Heating and
ACKNOWLEDGEMENT Cooling Connection Handbook, published by the
International Energy Agency (R & D Programme on
The Swedish District Heating Association, the Swedish District Heating and Cooling), 2002.
Energy Agency and Nordic Energy Research are
gratefully acknowledged for financing this work. [11] Snoek, C., Yang, L., Frederiksen, S., Korsman, H.,
Optimization of District Heating Systems by
Maximizing Building Heating System Temperature
REFERENCES Differences, Report 2002:S2, International Energy
[1] Euroheat & Power, Guidelines for District Heating Agency (R & D Programme on District Heating and
Substations, Downloaded from: Cooling) & NOVEM, Sittard, 2002.
http://www.euroheat.org/documents/Guidelines%2 [12] Trschel, A., Hydronic Heating Systems The
0District%20Heating%20Substations.pdf, Effect Of Design On System Sensitivity, Doctoral
20081117. Thesis, Chalmers University of Technology,
[2] Frederiksen, S., Wollerstrand, J., Performance of Gothenburg, Sweden, 2002.
district heating house station in altered operational [13] Volla, R., Ulseth, R., Stang, J., Frederiksen, S.,
modes, 23rd UNICHAL-Congress, Berlin, 1987. Johnson, A., Besant, R., Efficient substations and
[3] Gummrus, P., Petersson, S., Robust Fjrrvrme- installations, Report 1996:N5, International Energy
central (Robust District Heating Substation), Report Agency (R & D Programme on DHC) & NOVEM,
A 99-223, Dept. of Energy and Environment, Sittard, The Netherlands, 1996.
Chalmers Univ. of Technology, Gothenburg, 1999.
[4] Langendries, R., Low Return Temperature (LRT) in
District Heating, Energy and Buildings, 12 (1988)
191-200.
214
th th
POLICIES AND BARRIERS FOR DISTRICT HEATING AND COOLING

OUTSIDE EU COUNTRIES
1 2
A. Nuorkivi and B. Kalkum
1
Energy-AN Consulting
2
Energy & Utility Consulting
ABSTRACT PRELIMINARY COUNTRY SPECIFIC SURVEYS

The policies and barriers faced by DHC in the countries 1. Canada
outside the EU will be investigated during 20102011 1.1. Status of DHC
as a part of the Annex IX of the IEA Implementing
The old DH systems before 1985 are predominantly
Agreement on District Heating and Cooling (DHC),
with steam, whereas water systems have been built
including the integration of CHP.
since 1985. Both domestic hot water (DHW) and space
The countries to be covered are China, USA, Canada, heating (SH) have been included. Based on
South Korea, Russia and some other selected water/steam carrier, various combinations of heating
European countries outside the EU. The work is based and cooling are available in Canada.
on both interviews of the key officers and specialists
Historically, Canada has had the highest per capita
and the existing laws, regulations and policies of each
energy use of the developed countries, as a result of
selected country. The project will also provide
the harsh climate and relatively low-cost, abundant
examples of best practices useful for sustainable
energy. So the benefits of DHC would be particularly
development of DHC as well as offer recommendations
welcome to save energy. In Canada, there are records
to the countries to improve the institutional set up of the
of some 120-160 DHC systems in the country, and
DHC.
almost a half of them located in Ontario Province alone.
Regarding each country, the project will review, for About 27 Mm2 of residential, industrial and institutional
instance, the tariff setting, DHC related legislation, floor area are connected to the DHC systems. This
taxation rules, price regulation, customer definition and represents about 1,3% of all floor space in Canada.
points of delivery; ownership of fixed assets; allocation The largest DHC system is in Toronto with 522 MW
of CHP costs and environmental fees; social thermal capacity.[1]
considerations; municipal heat planning; and, heat
Natural gas distribution has spread everywhere, which
metering and control.
is a challenge for DHC expansion. Moreover, at
The project here is a twin project to EcoHeat4EU that is relatively low electricity prices, there is a little market
a thorough analysis of the barriers and opportunities of for CHP. No economic market for CHP exists in
DHC as well but in the selected EU member countries. Canada unless the feed-in tariff is in place or the
electricity is used in-house of producer. Power and gas
INTRODUCTION utilities have not been co-operating so far, because
there has not been any incentive to such co-operation.
There is no reliable statistics of DHC in most of the Because of the structure of the provincial utilities and
subject countries. The countries are in different stages low electricity prices, only a few CHP based DHC
of DHC development, as can be read out in the paper. systems are in operation.
The market drivers and barriers are different as well.
The aim of the study is to identify lessons learned from The utilities are empowered to provide the people with
all countries, including the EU that might be useful to gas and electricity at the lowest costs possible.
boost DHC development in the particular subject Economic drivers support the selection of the proper
country. Nevertheless, the lessons learned and technologies, and the provincial regulators ensure that
recommendations will be developed in fall 2010, after the system availability and safety are maintained at all
the Symposium, and the final and complete study will times. Provincial governments provide some directions
be available in May 2011. Therefore, all information to the energy industry, but limit themselves to setting
presented in the paper regarding four countries, overall goals only. The selection of the technologies is
Canada, China, Ukraine and USA is based on the left to the utilities. Natural gas is widely available
preliminary survey that will be finalized by throughout the country, which is a challenge for other
October 2010. heating modes to enter the market. Serious lack of gas
reserves is expected in the future, which means
alternative energy sources to become increasingly
215
th th
realistic. To substitute natural gas, DHC based on while fearing of intervening the private sector driven
biomass and possibly with CHP is a superior option. heating market.
For DH, two-tier tariffs are used in which energy fee is
1.4. Current Activities
pass-through of energy costs, and the fixed fee covers
the profit, the connection costs and all other cost The Integrated Community Energy Solutions (ICES)
except energy. The fixed fee can be adjusted Roundtables have been established to accelerate
annually/biannually with CPI (Consumer Price progress toward reducing GHG emissions by bringing
Index).The customer contracts are made for a long together senior-level stakeholders to exchange views
period, say 10-20 years, during which the capital cost on the best way forward from here. The Roundtables
have been discounted to the fixed fee. Municipal build upon ICES. The Roadmap for Action, which was
companies operate as non-profit but private companies released by the Canadian Council of Energy Ministers
with reasonable profit. at its annual meeting in September 2009, describes the
role that Canada's federal, provincial and territorial
1.2. Market Drivers governments can play in advancing ICES and it sets
out a broad strategy for action. It also includes a variety
In Canada, the federal government is committed to
of options from which the governments can choose,
reducing GHG emissions by 17% below 2005 levels
according to their priorities, to advance community
by 2020, being the main driver of DHC. The DHC
energy performance and complement existing energy
market is expanding smoothly to start creating a
efficiency activities in different sectors.
different infrastructure to substitute depleting resources
of natural gas. The ongoing collaboration of key energy actors and
enablers across Canada from the private and public
As mental drivers, there is strong interest in
sectors through the Quality Urban Energy Systems of
municipalities to consider DHC introduction and further
Tomorrow (QUEST) collaborative also informed the
expansion very much based on European practise.
Roundtable discussion. In particular, preliminary results
Many municipalities have set voluntarily targets to the
from a QUEST-led study suggest that ICES could
reduced GHG emissions. DHC systems are widely
reduce GHG emissions at the community level by as
recognized as a potential measure to achieve the
much as 40% to 50%, resulting in reduction of 65 Mt
targets. The DHC is considered a tool for the urban
by 2020, which is about 20% of Canada's official
planners but not an energy issue per se.
2020 target reductions. These results are very
As an example of investment support, Ontario Power promising and highlight how ICES could contribute
Authority (OPA) subsidizes investments in electricity significantly to improving Canadas energy and GHG
savings by paying up to $800/kW of the saved electric performance.
capacity. The subsidy used to be 400/kW, but was
doubled at the end of 2009. Customers can use that 2. P.R. China
money as the partial payment of the connection costs
2.1. Status of DHC
of DHC, thus DHC companies indirectly benefitting
from the subsidy system as well. In China, the DH development has been very strong,
more than 10% annually during the past decade on
1.3. Main Barriers average. By the end of 2005, DH supply (including
steam and hot water) was over 2 100 PJ; of which CHP
There is no formal DHC strategy or policy supporting
accounted for 47% and boilers accounted for 51%.In
DHC and CHP development in Canada. The
the supply of steam and hot water, steam supply is 715
Government does neither have the tradition nor the
PJ, of which CHP accounts for 81% and boilers
willingness to take strong position in DHC
account for 17%; the total hot water heating supply is
development. The private sector that could bring
1395 PJ, of which CHP accounts for 29% and boilers
investments and entrepreneurship cannot be much
account for 69%. The heating supplied by CHP units
interested, because starting the DHC is risky: long pay-
and boilers are respectively 992 PJ and 1086 PJ.
back times ranging beyond 10 years, limited access to
municipal property, challenging contracting of Apart from Europe, only SH is supplied by the DH
residential, municipal and federal buildings, overall systems, and the DHW by individual systems: solar
billing and collection of different types of customers. collectors, propane, electricity, etc. [2,3]
Nevertheless, the municipalities are rather weak,
During the few years to come, China will become the
because the municipal taxation only covers property
largest DH country in the world.
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
216
th th
2.2. Market Drivers charges and power grid balancing that need to be
The rational of strong DH development in China is addressed. At present, the State Power Grid Group
based on eliminating the small and polluting coal fired is responsible for the power grid operation. As
boilers in the northern, western and central provinces such, more communication and coordination
and to provide feasible living conditions to the activities could be conducted between the DHC
population massively moving in to the cities. industries and the State Power Grid Group.
DH has been encouraged by the Chinese government Centralized DHW would benefit CHP. Missing
for several decades. China's DH heating area has DHW load hampers economic development of
increased from over 276 Mm2 in 1991 to over CHP schemes. Without DHW, the CHP plants can
1100 Mm2 in 2000, and exceeded 2500 Mm2 in 2005, operate all year round only if there is industrial
with an annual growth rate of 17%. The growth in DH steam load existing nearby.
mainly came from the northern and the northeast
regions. In China, residential buildings account for POLICY BARRIERS
about 70% of the total DH area and commercial
There also exist barriers in the area of economic
buildings the balance of about 30%.
support and administrative policies related to
The urban communities are very densely built, which CHP/DHC, including:
effectively supports centralized heating and cooling
There is a lack of monitoring and enforcement of
solutions. The new buildings comprise about half of the
the governments policies related to the efficient
DH connections, whereas the balance for existing
operation of CHP projects. Currently, it appears
buildings, the latter previously having had been heated
that some newly- built CHP projects are operating
by small coal boilers.
only in thermal generation mode after they have
been approved, thereby reducing their energy
2.3. Main Barriers
efficiency.
The DHC sector is expanding fast but there are still
some barriers regarding economy, policy, financing and There is a lack of targeted policy for smaller CHP
technology as summarized below. units. In order to fulfill the energy conservation
target, China is attempting to increase the number
of more efficient large power generation plants and
ECONOMIC AND PRICING BARRIERS
to close down smaller, older units. While it is
In order to become cost-effective and an attractive important that the smaller, more inefficient units be
investment, power and heating reform policies will need closed down, some small CHP units with high
to be undertaken. Some of the key issues include: efficiency are also being targeted for phase-out.
Based on the goal of increasing energy supply
Energy price policy reform is a priority. At present,
efficiency, a different policy should be adopted. For
in China, the coal price is based on the market,
example, in regions with low heating loads, small
which has grown rapidly in recent years. However,
CHP units could provide most of their energy
electricity and heating prices are still controlled by
needs at a fraction of the cost of larger units.
the government, and have only slightly increased.
While the government has provided limited
FINANCING BARRIERS
subsidies to DH companies, most CHP enterprises
and DH companies are currently not making a There are promising energy conservation projects
profit as a result of the lack of energy price reform. particularly in the DH sector that could be realized if
there were sufficient funds or other means available to
In addition, heating reform needs to be further address the gap in investment capital. In particular:
developed. Currently, in most cases, heat tariffs
are based on the building area, rather than on the Some planned CHP/DHC projects are not operated
actual heat consumption, which has a negative efficiently because they lack sufficient resources to
influence on improving the energy efficiency in invest in expanded heat pipeline infrastructure.
district heat facilities and buildings. Further, at many existing DHC projects, the heat
loss in pipelines is high, reducing the overall
Power sector reform is also needed. At present, efficiency of the heating system. Additional
the electricity produced by most DHC (and some financing is needed to invest in cost-effective heat
CHP) projects cannot interconnect with the power
pipeline retrofit projects, which will generate
grid, which has strongly reduced development. The sizeable energy efficiency benefits and GHG
technical issues of grid connection can likely be reductions.
addressed. However, there are also administrative
interconnection issues, such as added-capacity
217
th th
While energy service companies are expanding in 3. Ukraine

the commercial building energy conservation
3.1. Status of DHC
arena, they have not yet entered the CHP/DHC
area. There is some room for these types of third- Ukraine is one of the largest DH countries in Europe.
party players to come up with innovative means to Currently almost 80% of urban housing is supplied with
finance projects. DH through extensive grids of hot water pipes.
The DH sector is rather saturated, but in some eastern
TECHNICAL BARRIERS (Donbas) cities the DH systems are deteriorating fast,
and customers are either adopting apartment level gas
While CHP/DHC are proven, existing technologies that
boilers or even remain without heating, thus enjoying
do not require major research and development, there
on the heat losses penetrating to them through walls
are some advanced technologies that could be
free of charge from their heated neighbours. Even the
introduced from IEA Member Countries to improve
municipalities are offering investment subsidies to the
efficiency and operational benefits. In addition, there is
apartment owners to purchase apartment level gas
currently some debate about the relative merits of DC
boilers while disconnecting the DH services.
technology. China-specific research studies could be
conducted to confirm the primary energy conservation Such practices have led to extremely poor quality of
performance of these technologies. DH services: low water and room temperatures a well
as periodical heating are used to minimize fuel costs.
ORGANIZATIONAL BARRIERS There are coal (and anthracite) mines in Ukraine, but
There are some organizational barriers for optimal little used for providing fuel for DH: Most DH is based
development as well. on natural gas imported from Russia. The costs of gas
comprise 5070% of the DH, which explains why the
Scattered organizations with several heat suppliers DH is vulnerable to gas price changes.
and distributors prevail in one city. In the same DH
system, the heat supplier is responsible for Ukrainian heat generating facilities are ineffective for
operation and maintenance until the group many reasons. The most important reasons are as
follows:
substations that serve several buildings through
the secondary network, and the distributors being technology used for heat generation is outdated
responsible from the substations to the indoor and inefficient;
heating elements. Therefore, the holistic key assets are heavily deteriorated;
optimization can be often compromised by partial equipment is being used in a switching mode on
optimizations. unspecified fuel;
The DHC companies are operation and delays and failures to carry out regular repairs.
maintenance companies only, whereas investment
According to the Ministry of Fuel and Energy, more
decisions and financing depends on the municipal
than 90% of energy units have worked out their
and provincial budgets. This is one more reason for
projected service life (100 000 hours), more than 60%
that there is little business minded atmosphere in
have been in service longer than 200 000 hours.
the extensively staffed DHC companies.
Heat tariff for final consumers is defined as a sum of
2.4. Current Activities tariffs for production, transportation and supply.
The DH systems are expanding fast in China, Tariffs for heat that is produced by CHPs, co-
simultaneously restricting coal consumption and
generation or alternative/renewable energy sources are
reducing overall GHG emissions of the heating set by the National Energy Regulatory Commission
services. (NERC) but they should not be higher than heat
The Ministry of Construction has issued the Housing produced by other sources.
and Building Reform on Energy Efficiency (HRBEE), Tariffs for heat production, transportation and supply
which requires more efficient buildings to be built as other than CHPs, co-generation or alternative
well as introduction of heat metering and consumption /renewable energy sources are approved by local
based billing. The first consumption based billing pilot governments. Due to that the tariffs differ much across
was initiated in Tianjin a few years ago with a two-tier the territory of Ukraine.
heating tariff. Such billing systems are slowly
expanding to other regions. 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
218
th th
production and provide for marginal profitability level directed to energy saving in heat, water supply and
that is not lower than the level defined by the Cabinet sewerage. But according to monitoring results, the
of Ministers on the base of calculations by the central funds are allocated to other purposes. Only four
body of executive power in heat supply. regions used the funds for energy saving. Other
regions used from 7% to 40% instead of the required
If heat tariffs do not cover the cost of heat and marginal
75% to energy saving.
profitability level, the body that has set the tariff should
provide for the compensation according to effective Other measures of energy saving that would be
legislation. That is, if the tariffs for heat from thermal appropriate include:
power station and boilers that are approved by the local
replacement or reconstruction of steam and gas
government on the basis of heat producer calculation,
boilers with efficiency that is lower than 89%;
and they are lower than economically sound cost
improvement of heat pipes insulation to decrease
including marginal profitability level, the local
losses in transmission pipelines;
governments must compensate the losses from the
installation of heat meters; and,
local budgets.
Installation of co-generation equipment.
Meanwhile, the Ministry of Economy elaborated the
Another stimulus for companies to introduce energy
draft that specifies binding of the household services
saving technologies is outlined in the Law of Ukraine
tariffs to energy prices. First of all, it means heat, hot
On Heat Supply. According to the Article 8, in case
water and gas supply to households. The current
heat supply or heat transportation companies introduce
system of tariff setting reduces the competitiveness of
energy saving measures that result in saving of heat
Ukrainian industry, since industry is forced to
losses, the body of executive power, that is entitled to
compensate for low households tariffs.
regulate heat tariffs according to the Law, may leave
The procedure to raise the heat tariffs is rather the tariffs unchanged for the three consecutive
complicated and time consuming, as follows: years.[4]
1) The district heat supply company receives official
3.3. Main Barriers
notification from NERC on gas price increase. Only
after that the company may start developing the In general, there are a number of decent laws and
proposal on the heat tariffs increase. regulations that would support DHC development, but
they are not implemented properly, as mentioned
2) The new heat tariffs have to be approved by the above already.
following authorities: Commissions of the Municipal
Council (miskrada) and regional council (oblrada). The Therefore, there is little if any incentives to business
tariff proposal has to be reviewed by several instances oriented development of the heating services, but the
as listed below: systems are run at minimum investments and reduced
technical performance. The DH companies are solely
Trade unions operation organizations, mainly departments of the
Antimonopoly Committee municipality. The municipalities take care of billing and
Department for Price Administration collecting based on subsidized lump sum tariffs, and on
Department for the Protection of Consumer Rights investment decisions.
Public hearings
There are many privileged customer categories that
3) Municipal Executive Committee (miskvykonkom) enjoy reduced costs of DH services. In Odessa, for
has to approve the new heat tariffs as well. instance, 25% of the customers in year 2006 enjoyed
4) The tariff changes shall be publicized via official such privileged heating prices. Their billings were
mass media of Municipal or Regional Council. If during decreased by 20, 30, 50, 75 or even 100%, which
a month there are no official protests from the Office of effectively destroys the business opportunities of DH.
Public Prosecutor, the company is entitled to apply the Individual and autonomous heating in every apartment
new tariffs. seems the most favourable option for consumers. In
The above steps clearly show how cumbersome any such a case they do not pay for heat and hot water but
tariff increase can be in practise. only for gas and cold water. In addition, they can
regulate temperature in their apartments and do not
3.2. Market Drivers suffer from overheating in spring and insufficient
heating in winter. But sometimes it is impossible to
Article 54 of the state budget of Ukraine for 2006 and
install autonomous boilers in every apartment, because
the Cabinet of Ministers Decree No.207 of 9 March,
there is not enough space for heating equipment and
2006 stipulate for subsidies from the state budget to
the vertical ventilation ducts are not designed for flue
local budgets. No less than 75% of the subsidy must be
gases. Therefore, it would be appropriate to install one
219
th th
boiler for the whole building (several apartments) or DHC/CHP system in New York City is the Worlds
several buildings. Another problem for individual and largest steam system with 1850+ customers.
autonomous heating is that in case of gas supply
DHC (primarily DH currently) delivers about 3,5 % of
interruption there is no reserve fuel resources to
the total final energy demand in the industrial,
continue heating. Reserve fuel can be provided only for
residential, public, and commercial sectors. In the past
centralized DH.
two decades, some 47 Mm2 has been connected to the
Frequent failures in the heating systems as a result of DHC systems, but the total customer base volume
outdated equipment and poor funding are still common number is not available.
throughout the country. Some service breaks in coldest
The DHC systems are predominantly (80%) with
winter times have caused serious impacts on human
steam, the consumption being a mixture of steam
life already.
heating, cooling and DHW depending on the particular
Legally, local authorities that establish tariffs for case. There is little residential heat load but the
population lower than the cost coverage level have to majority is public: offices, malls, universities and
compensate the difference to energygenerating military bases.
companies. In practice the compensation is not always Countrywide, the DH and DC markets are expanding at
paid in full which leads to arrears accumulation and 3-4%/a and up to 10%/a, respectively, but almost solely
aggravates financial state of heatgenerators. The on campuses, hospitals, military bases and in the
procedure of heat tariffs increase is rather complicated, downtown commercial and public buildings.[5]
as well as time consuming.
In general, however, DHC together with CHP has been
According to the Law of Ukraine adopted in April 2006, tragically underutilized as a tool to combat climate
heat producers such as CHPs and renewable sources change, to reduce life-cycle costs of energy supply and
power plants are not allowed to crosssubsidy heat to defend energy independence in U.S.A.
production to cover losses from heat production at the
cost of electricity production or other activity. 4.2. Market Drivers
Nevertheless, official sources say that due to low heat The U.S. Congress has acknowledged the benefits
tariffs for CHPs heat production is subsidized by the DHC/CHP by stating that:
cost of electricity production. But the unofficial sources approximately 30% of the total quantity of energy
assert that CHPs may charge heat tariffs that are even consumed in the United States is used to provide
higher than heat production cost to cover losses from thermal energy heating and cooling building
electricity production, because electricity tariffs are set space, DHW and industrial processes;
only by NERC while heat tariffs are set by heat thermal energy is an essential, but often
production companies with the approval of local bodies overlooked segment of the national energy mix;
of power. DHC systems provide sustainable thermal energy
infrastructure by producing and distributing thermal
3.4. Current Activities energy from CHP, sources of industrial or
The DH strategy is under preparation in Ukraine as a municipal surplus heat and from renewable
multi-ministerial approach and it should be ready in fall sources such as biomass, geothermal, and solar;
2010. CHP development is in the focus of the strategy.
DHC systems provide advantages that support
There has also been comprehensive framework
secure, affordable, renewable, and sustainable
support initiated by USAID, EBRD and EU to
energy for the U.S., including use of local fuels or
reformulate the national energy policy, including DHC
waste heat sources that keep jobs and energy
and CHP. It is uncertain now how much the political
dollars in local economies, stable, predictable
election of April 2010 will influence availability of such
energy costs for businesses and industry,
foreign technical assistance in the years to come.
reduction in reliance on fossil fuels, reduction in
emissions of GHG, and flexibility to modify fuel
4. U.S.A. sources in response to future changes in fuel
4.1. Status of DHC availabilities and prices and development of new
technologies;
The total DHC industry base comprises approximately
DHC helps cut peak power demand and reduce
2 500 systems, in which the number of customer
power transmission and distribution system
buildings served by a typical DHC system may range
constraints; and,
from as few as 3 or 4 in the early stages of new system
development to the largest system served by CHP systems increase energy efficiency of power
Consolidated Edison in Manhattan. The downtown plants by capturing thermal energy and using the
thermal energy to provide heating and cooling, more
220
th th
than doubling the efficiency of conventional power The definition of CHP is rather complicated. The
plants. Internal Revenue Code 26 USC and its 48 define
CHP as producer of:
The Department of Energy has estimated that
increasing CHP from its current 9% share of U.S. at least 20 % of its total useful energy in the form
electric power to 20% by 2030 would avoid 60% of the of thermal energy which is not used to produce
projected increase in U.S. carbon dioxide emissions electrical or mechanical power (or combination
(equivalent to taking half of all U.S. passenger vehicles thereof), and
off the road); and, generate $234 billion in new at least 20% of its total useful energy in the form of
investments. electrical or mechanical power (or combination
thereof), and
DHC would be a critical component of this CHP growth.
The local electric distribution companies (LDCs) are the energy efficiency percentage of which exceeds
interested in DHC as a means to reduce the summer 60%.
peak and to release transmission and distribution The Thermal Energy Efficiency Act of 2009 establishes
capacity to other electric applications that have more the Thermal Energy Efficiency Fund that would award
even consumption during the year. grants for DHC, CHP, and recoverable waste energy
The developers of the building sector are interested in projects. It includes biomass facilities. Under a federal
DHC as well, because it would leave more room space GHG emissions regulation program, 2% of emission
in the building for sale. allowances established for each calendar year from
20122050 would be allocated to the Fund.
At the municipal level, the market driver for DHC is the
reduction of the GHG emissions. Many municipalities This legislation would dedicate 2% of revenues from
have set voluntarily targets to the reduced GHG climate change legislation to fund CHP, waste energy
emissions. recovery, and DHC projects. Based on various
estimates, this could mean roughly between $1 billion
4.3. Main Barriers and $1,5 billion per year for clean energy infrastructure.
The Thermal Energy Efficiency Act would provide 40%
In general, the barriers are very much the same as
of its funding for institutional entities (defined as public
already discussed in Canada. Private sector as
or non-profit hospitals, local and state governments,
investor cannot be much interested, because starting
school districts and higher education facilities, tribal
the DHC is risky: long pay-back times ranging beyond
governments, municipal utilities, or their designees),
10 years, limited access to municipal property,
40% for commercial and industrial entities, and 20% to
challenging contracting of residential, municipal and
be used in the discretion of the Secretary of Energy to
federal buildings, overall billing and collection of
fund institutional entity projects, commercial and
different types of customers.
industrial projects, or federal facility projects. A match
Only little expansion on residential sector is is required of all non-federal applicants, starting at 25%
recognized, and that is because there is voting needed from 2012-2017, and rising to 50% from 2018 to 2050.
among the condominium owners. The centralized The breakdown of how the money would be used is
energy systems, that the condo owners are not familiar 75% for construction of infrastructure, 15% for
with and perhaps difficult for them to understand the planning, engineering, and feasibility studies, and the
benefits, have not been adopted on the residential remaining 10% to be used at the discretion of the
sector in a considerable scale so far. Secretary for either infrastructure or planning,
depending on the need.
4.4. Current Activities
In competition with grid power plants receiving
There are several laws and regulations that are generous allowances in ACES, CHP systems could be
expected to support DHC development in the shut down. Unless allowances are allocated to the
U.S.A.[6,7] DHC CHP system, it will have to purchase allowances
Rising interest on development and extension of for all gas consumed in the facility, resulting in an
renewable energy sources as well as improving overall additional cost equal to 15% of the average 2007
energy efficiency is to be converted to legislation at the wholesale power price ($57 per MWh) at the $16 per
moment. Unfortunately, DHC has not been successful metric ton allowance price projected by Environmental
in the legislation process so far, but both the Protection Agency (EPA) for the year 2020. In contrast,
Department of Energy as well as the DHC and CHP the merchant coal plant will only have a GHG
associations such as IDEA and USCHPA are working allowance cost of only 5% of the average 2007
on it. wholesale power price, because allowances will be
allocated for nearly all (83%) of its emissions.
221
th th
Faced with this huge competitive disadvantage in the CONCLUSION

marginal cost of power generation, some existing CHP
The survey work is still underway, and therefore, the
facilities will shut down and construction of new CHP
lessons learned and recommendations will be issued in
plants will be choked off.
the final report in spring 2011.
In the ACES, DHC systems are not directly covered
entities unless they qualify as electricity sources. ACKNOWLEDGEMENT
District systems would be covered indirectly through
the costs of allowances built into the prices of The authors express their gratitude to the interviewed
purchased fuel oil, or natural gas if purchased from the specialists, Mr. B. Gilmour (Canadian Urban Institute),
gas LDC. However, gas purchased on the wholesale Mr. K. Church (Natural Resources Canada), Mr. M.
market or coal users not qualifying as an electricity Wiggin (Public Works and Government Services
source would not be required to submit allowances. Canada), Mr. R. Thornton (International District Energy
Association IDEA), Mr. D. Kaempf and Ms. P.
This is a fundamentally good framework with the Garland (U.S. Department of Energy), Mr. B. Hedman
exception of the concerns about CHP systems to be (ICF International) and Messrs. G. Draugelis and P.
covered or not. However, if the upcoming legislative Salminen in the World Bank.
process results in modifications that make many DHC
systems covered entities, it is critical that changes in
REERENCES
allowance allocations be made as discussed below.
For example, if the final climate change bill regulates all [1] National DHC Survey, Canadian DHC Association
sources with emissions greater than 25 000 metric tons (CDEA), 2009.
CO2e (the threshold generally used in the ACES as [2] Ministry of Construction, China City Construction
well as a number of past bills), over 70% of DHC Statistic Annual. The DH data does not include
systems and over 95% of DHC output would be industrial steam and hot water.
capped. In such a way, more efficient systems will have
competitive advantage, because the quantity of [3] T. Kerr, IEA Collateral, Sustainable Energy in
allowances needed per unit of energy will be lower. China: The Role of CHP and District
Heating/Cooling, 2008.
American Clean Energy Leadership Act (ACELA) in
June 2009 and Federal Renewable/Energy Efficiency [4] A. Tsarenko, Overview of Heating Sector, CASE
Standard establishes a Renewable Electricity Standard Ukraine, 2007.
which includes provision for energy efficiency credits as [5] IDEA Report, The DHC Industry, 2005.
well as renewable energy credits that can benefit DHC
as well. [6] DHC Services, Commercial Data Analysis for EIAs
National Energy Modeling System, Energy and
Renewable Electricity and Energy Efficiency Standard Environmental Analysis, Inc. and International DHC
established by ACELA is applicable with the electric Association, 2007.
utilities selling >4 TWh a year. The utilities are required
to supply 20% of demand from combination of [7] M. Spurr, Climate Change Legislation in Dollars
renewable sources and increased energy efficiency and Cents, presentation in IEA DHC in
meaning 15% renewable together with 5% efficiency Gustavelund, Finland, in Aug 2009.
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.
222
th th
BARRIERS TO DISTRICT HEATING DEVELOPMENT

IN SOME EUROPEAN COUNTRIES
1 2
Dag Henning and Olle Mrdsj
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 are difficult to use for individual buildings, such as

unrefined biomass fuels, heat from waste incineration
District heating (DH) offers low primary energy
and industrial surplus heat. The latter may, for
demand, high security of supply and small CO2
example, be a by-product from production of
emissions. Barriers to DH in the UK, Ireland, France,
automotive biofuel. District heating can provide cheap
Romania and the Czech Republic have been compiled
energy to consumers by using low-cost energy
through publications and interviews.
sources, such as wood, waste and surplus heat. Many
DH systems require large investments, have negative of these resources can be of local origin and promote
initial cash flow and long payback time, which obstructs local business and industry.
financing. One actor should control DH from source to
The main advantages with district heating are high
consumption. If the value chain is fragmented,
security of supply through utilisation of domestic
contracts are required between the links. It increases
renewable energy resources, if available, low primary
risks and financing costs, like in the UK and Ireland,
energy demand due to high conversion efficiency, as
where DH is not established. There are few multi-family
well as small CO2 emissions thanks to low fossil fuel
houses with central heating and it is expensive to build
use and the high energy efficiency. Incineration of
DH networks in built areas.
waste with heat recovery to district heating may be
Most French DH systems are operated according to used at very low cost. District heating also gives
long-term concessions by companies that sell electricity opportunity for cogeneration of power and heat with
and gas. No strong actor provides unbiased DH very high efficiency. District heating enables profitable
support. In the Czech Republic, gas offers DH severe heat supply with outstanding environmental
competition. Much DH is produced at the expense of performance but there are in many places various
electricity that is considered more valuable, and waste barriers to a prosperous DH development.
incineration is not popular. In Romania, DH
Barriers to district heating in the United Kingdom (UK),
consumption was reduced by one-half. Distribution
Ireland, France, Romania and the Czech Republic, as
losses are enormous. New less polluting plants are
well as barriers to export of Swedish district heating
needed.
knowledge and products to these countries have been
Consortia from established DH countries could offer DH compiled from publications and through personal
systems from fuel to customer if local policies facilitate communication with people in public and private energy
DH development. bodies and companies in Sweden and abroad [1].
In the studied countries, there are large potentials for
INTRODUCTION
district-heating development and for Swedish sales of
This paper describes barriers to district heating (DH) in DH related goods and services. But for district heating
various parts of Europe and to Swedish involvement in and export to succeed, there are several barriers to
district-heating business abroad. The paper is based overcome in Sweden as well as in the other countries.
on a report called District Heating in Europe: Barriers It should be emphasised that this paper focuses
to overcome for Swedish export [1], which was barriers and does not give the full picture of the
prepared for The Swedish District Heating Association. conditions for district heating, which also includes many
possibilities.
The losses by energy conversion in Europe are of the
same magnitude as the European heat demand and
BARRIERS IN WELL-DEVELOPED DH COUNTRIES
consist mainly of heat that is wasted by electricity
generation [2]. District heating is a means to utilise In many countries with well-developed district-heating
such surplus heat to cover heat demand. industry, such as Sweden, much DH competence
resides in municipally owned energy companies. They
District heating can utilise the heat from electricity
have system knowledge, which could be applicable in
generation in combined heat and power (CHP) plants.
other countries. District-heating companies owned by
District heating can also use other heat sources that
223
th th
Swedish municipalities must, for judicial reasons, limit Another general barrier to district heating is the EU
their business abroad to sales of services, and to a emission trading scheme, which favours individual
very limited extent goods. For municipal district heating heating because individual CO2 emissions do not need
companies, domestic judicial restrictions are the first allowances.
barriers to overcome before operations in other
countries can commence. TYPES OF DH BARRIERS
Only certain components for production and distribution In the countries analysed in this project, the barriers
of district heating are manufactured in a single country, are of very diverse nature. The obstacles are
which calls for international cooperation. The Swedish dominated by difficulties for district heating itself rather
Government provides certain but limited support to than for foreign companies operations in the countries.
promotion of district heating business abroad. For In the British Isles, it is largely a question of
example, Swedish district heating consultants work establishing district heating as a natural element in
abroad but it is seldom followed by goods export. society. In France, it is about large domestic companies
that may offer superior competition to foreign firms. In
FINANCING THE DH VALUE CHAIN the Czech Republic, French and other companies from
abroad dominate the DH business but the technical
Financing is a large barrier to district heating
design of district-heating production may hamper DH
development. DH systems require large investments
development. In Romania, there are several problems
and may have long payback times. The cash flow is
with facilities in bad shape and public bodies that have
negative for a long time during the establishment of a
not addressed the issues properly.
new DH system. Time horizons are distant, which
stresses financers in our present situation of rapidly Table I is an attempt to assess how large the various
changing conditions. Private companies often focus on barriers are in the studied countries. The table starts
short-term profit and public involvement may be with some general conditions. Ownership and
necessary for the deployment, modernisation and long- organisation considers if district-heating companies are
term development of district heating systems. owned, or DH operations are organised, in ways that
make it more difficult for Swedish companies to do
District heating is a comprehensive concept for heat
business. Corruption may be a problem through, for
from source to consumption. Its strength lies in
example, indirect bribes by procurement. National and
maintaining the value chain (Fig. 1). This may fit badly
local control encompasses national laws and policy
in an exaggerated market context where every little link
instruments that are disadvantageous for district
of the value chain is organised separately with an
heating, DH price regulations, as well as municipalities
interface of costs and revenues to other links. A
not facilitating district heating by planning of new
fragmented value chain increases interface costs and
developments. But rules complicating combined heat
total risk. EU regulations have a tendency to promote
and power production are included in the CHP line in
such fragmentation. Between the links of a fragmented
Table I.
supply value chain, many complicated agreements are
required, which all include risks. It means a larger total Financing is one of the largest barriers to district
financing risk, which raises interest rates and shortens heating, primarily because DH schemes give a low rate
amortisation periods for loans. This implies a mismatch of return. A fragmented value chain cause contract
with the depreciation in the balance sheet due to the risks at several instances. Entrance barriers for foreign
long economical lifetime of district heating versus the companies in Table I consider additional difficulties for
short amortisation time. foreign firms besides the other parameters and the
general disadvantage of not being familiar with the
TWO GENERAL DH BARRIERS domestic business culture.
Two general district-heating barriers are related to CO2 Some parameters in Table I are related to district-
emissions and the attempts to reduce these through, heating sales. DH competitiveness includes the
for example, reduced energy use. Global warming and availability and price of other forms of heating, primarily
better insulated houses reduce heating demand and, natural gas. Customer relations concern customer
hence, the advantages of district heating because attitudes toward district heating, customers and
investment costs must be carried by less supplied heat. suppliers perceived insecurity whether they can
Fig. 1. District heating value chain with heat production, distribution and sales in focus [1]
224
th th
establish and maintain relations, as well as if Table I encompasses difficulties with building networks
disconnections have occurred or may occur. Built and deficiencies in existing distribution.
environment relates to how common multi-family
The assessments in Table I were primarily made within
buildings are and if these have a central heating
each country and secondly countries were compared but
system for the whole house. Table I ends with district-
mostly the ranking of countries for a parameter is
heating production and distribution issues. Biomass
appropriate. However, every grade has a certain width
considers domestic biomass supplies and
and two countries with the same digit may differ. As an
infrastructure for biomass fuel supply. Waste includes
example, district heating is assessed to be somewhat
current waste management and attitudes toward
less competitive in Romania than in the Czech Republic.
waste incineration. CHP concerns regulations
It follows a description of barriers in the individual
hampering CHP production as well as problems in
countries emphasising the largest barriers.
existing plants. Finally, district heating distribution in
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 is rated as a rather large barrier in Table I because

supplies are limited in the British Isles and fuel supply
In the United Kingdom (UK), and even more in Ireland,
systems are less developed.
district heating is not really an established
phenomenon. Figure 2 shows that residences mostly UK Government and municipalities have hitherto not
are heated with gas in the UK, often through a gas facilitated district-heating development sufficiently and
boiler for the individual household. Oil is the most strong incentives for deploying district heating systems
common fuel in Irish homes but gas is expanding. are lacking. Heating is generally not regarded as a
public concern, but as a concern for each individual.
The largest problem is district heating distribution
National and local control is therefore indicated as a
(Table I). It is expensive and complicated to build DH
rather large barrier in Table I. In Ireland, the situation
networks in already built areas and, at least in the UK,
seems to be slightly better but in both countries certain
it is not straightforward to obtain a licence for putting
regulations, designed with electricity and gas in mind,
district heating pipes into streets. The financing
are disadvantageous for district heating. CHP suffers
difficulties in the British Isles are primarily due to a
especially from rules on how produced heat and power
fragmented value chain with many contract issues that
may be supplied.
need to be solved before a larger district heating
scheme can be deployed. British thinking is based on Customer relations are complicated because district
competition and individual choices. A collective large heating is a rather unknown energy form and there is a
scale solution, such as district heating, may conflict certain resistance against collective solutions [3]. There
with principles and tradition. Another large barrier is the is a lack of standardised terms of contract for
built environment. Few people live in multi-family connection to and delivery of district heating. Potential
houses in the UK and even fewer in Ireland [3], and heat suppliers and customers feel insecure concerning
even these buildings often lack central heating, but how many users that will connect to a DH grid, for how
individual heating of apartments is common. Biomass long they will stay and if heat supply may be
225
th th
interrupted. The competitiveness of district heating considered as a medium severe barrier in the British
compared to gas concerning availability and price is 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.
100%
100%
90%
90%
80%
80%
70%
70%
60%
60%
50%
50%
40%
40%
30%
30%
20%
20%
10%
10%
0%
0%
UK
UK Ireland
Ireland France
France Czech
CzechRepublic
Republic Romania
Romania
District heating
District heating Gas Biomass BiomassPeat
Gas PeatElectricity
Electricity Oil Oil Coal Coal
Fig. 2. Heating of residences [1], [4][6]
FRANCE
Miscell-
Table I shows that one of the largest barriers in France aneous
Renew-
concerns the organisation of district-heating operations. Natural
ables
Most DH systems are managed by private French gas CHP
companies according to long-term concessions [7]. The
companies have successfully applied this DH
management model in several other countries. By such Coal
arrangements, it is important that operators have
incentives to make investments even if these have Natural
Oil
payback times longer than the concession period [8]. It gas heat
is unclear if the French DH management model is
disadvantageous for district heating development but it
Fig. 3. District heating production in France [9]
should anyway be a large barrier for foreign companies
wanting to enter the French market. In general, Financing is considered to be a smaller problem in
domestic solutions are preferred. There is no strong France. The market domination by a few actors may
actor who provides unbiased support for district present an indirect financial barrier. Quite a few people
heating. The dominating DH operators also sell live in apartments but most multi-family houses lack
electricity and gas, which both cover a large fraction of central heating. The large French nuclear power
the heat demand (Fig. 2) and offer district heating production is one reason for worse CHP conditions,
severe competition. Only ten percent of the apartments which is assessed as a medium-grade barrier (Table I).
and four percent of all residences have district heating
today, and DH expansion is slow [6]. THE CZECH REPUBLIC
Fig. 3 shows that one-half of the district heating in Fig. 2 shows that district heating covers a substantial
France is produced with natural gas, mostly in CHP part of residential heating in the Czech Republic, but
plants. The main part of the renewable energy used for electricity is used to the same extent and gas is the
district heating production is waste, which is used to a most common heat source. District heating covers one-
slowly growing extent [7]. But French waste half of the apartments and 60% of urban heating [7].
incineration plants are mostly built far away from towns,
A large barrier in the Czech Republic is, according to
which makes it difficult to utilise the heat [6].
Table I, the competitiveness of district heating.
226
th th
Gas prices make it difficult for gas-based district use. Many district heating users switched to gas due to
heating to compete with individual gas heating [7]. low gas prices and heavy, government-regulated DH
There are some disconnections from DH systems. price increases [7], whereas households and district
heating plants had the same gas price.
Miscell- Natural
aneous gas
Hard
Coal
coal
Natural
gas
Lignite Oil
Fig. 5. District heating production in Romania in 2005 [7]

Fig. 4. District heating production in the Czech Republic
[10]
Fig. 5 shows that Romanian district heating production
Domestic coal dominates Czech district heating is completely based on fossil fuels. One-half of the heat
production (Fig. 4). Most of the district heating is is produced in, normally coal-fired, CHP plants. Large
produced in CHP plants. The problem concerning CHP investments are required in the Romanian district
(Table I) is that a large share of Czech district heating heating systems. CHP plants and heat-only boilers
comes from coal-fired power plants with extraction must be replaced for environmental reasons.
turbines where the heat is produced at the expense of Distribution losses are enormous [7].
electricity [7], which is considered more valuable. The Organisation is a rather large obstacle for district
benefit of this CHP production is not allocated to the heating in Romania (Table I). The municipalities are
heat [8]. Some biomass is used to produce district now mostly in charge of the district heating systems [7]
heating, but biomass use is complicated due to but much lobbying is required to achieve improvements
deficient fuel supply systems [7] and government and it takes time to reach an investment decision.
scepticism toward renewable energy. There is also Corruption is common. Some politicians and
much resistance to waste incineration from the public employees try to make their own profit on DH business.
as well as from politicians.
Financing difficulties largely concern insecurity whether
Financing may be a rather large barrier, partly due to a customers will remain because many have
certain district heating disconnection tendency. The disconnected from district heating. National and local
many private foreign district-heating companies in the control is a certain barrier because DH companies
Czech Republic [7] may be a difficult target for Swedish partly get heat production costs covered by central
and other district heating companies from abroad that Government and City Councils [7]. Besides the
are not established in the country. There may also be mentioned problems, the entrance barrier for foreign
some reluctance toward foreign enterprises. A certain companies should be rather low. Waste collection and
barrier is the common corruption by public procurement sorting are now deficient but, on the other hand, new
(Table I). The value chain is sometimes fragmented possibilities should emerge when Romania wants to
into production and distribution run by different actors. introduce waste incineration, and waste is therefore
considered to be a medium-size barrier in Table I.
ROMANIA
In Romania, biomass covers the largest fraction of HOW TO OVERCOME BARRIERS
residential heat demand among the countries under This paper focuses barriers and omits more positive
study (Fig. 2). Individual boilers and stoves for wood circumstances for district heating. It may be depressing
and gas cover more than one-half of the heat use in but the message is not that district heating has no
households. Gas is the most widely used heating prospects. The report should rather be understood as a
source for residences and it is expanding at the realistic guide to DH development in the studied
expense of district heating [7]. countries.
Table I shows that district heating has large problems To have a chance to overcome the outlined barriers to
with competitiveness and customer relations. Today, any significant extent, powerful initiatives are required
the DH consumption is just one-half of the previous from countries with established district-heating
227
th th
industries, such as Sweden. Initiatives should comprise ACKNOWLEDGEMENT

many different players, for example, district-heating
The Swedish District Heating Association and The
companies, equipment manufacturers, consultants and
Swedish Energy Agency are gratefully acknowledged
governmental bodies. Such consortia could offer
for financing this study through the Fjrrsyn
district-heating systems from fuel supply, via heat
programme. We would also like to thank everybody
production plants and DH networks to customer
who has contributed to the study with facts and
contracts. Now, many foreign groups visit municipal
viewpoints.
district-heating systems in Sweden but these
opportunities are seldom utilised to sell a
REFERENCES
comprehensive DH solution.
[1] D. Henning and O. Mrdsj, Fjrrvrme i Europa:
Municipally owned district heating companies have
Hinder att vervinna fr svensk export, Rapport 2009:3,
system knowledge that can be applicable in other
Fjrrsyn, Svensk Fjrrvrme, Stockholm (2009)
countries. A competence transfer may be realised
http://www.svenskfjarrvarme.se/index.php3?use=biblo&
through deeper involvement that might include
cmd=detailed&id=1440
ownership of plants in other countries. Business
models should be developed, which allow utilisation of [2] S. Werner, Ecoheatcool work package 4:
municipal knowledge abroad and give municipalities Possibilities with more district heating in Europe,
reasonable returns. Euroheat, Brussels (2006)
www.euroheat.org/ecoheatcool
For a successful transfer of district-heating solutions
from established to emerging markets, private and [3] WS Atkins Consultants Ltd, Assessment of the
public companies must focus marketing on the Barriers and Opportunities Facing the Deployment of
countries, places, projects and forms of involvement District Heating in Ireland, Sustainable Energy Ireland,
that have the greatest expectations to succeed. At the Dublin (2002)
same time, national and local policies should reduce www.sei.ie/uploadedfiles/InfoCentre/DistrictHeatingRep
and remove described barriers and facilitate district ortatk.pdf
heating development as a means for increased [4] S. Werner, Ecoheatcool work package 1: The
efficiency of energy utilisation, higher security of supply European heat market, Euroheat, Brussels (2006)
and decreased environmental impact. www.euroheat.org/ecoheatcool
[5] SEI, Energy in Ireland: Key Statistics, Sustainable
CONCLUSION
Energy Ireland, Dublin (2008)
There are several barriers to district heating www.sei.ie/Publications/Statistics_Publications/EPSSU
development in the countries under study. In the UK, _Publications/Energy_in_Ireland_Key_Statistics/Energ
there are not many district heating systems. There are y_in_Ireland_Key_Statistics_Final.pdf
few multi-family buildings with central heating in
[6] P. Cousinat, District Heating: A Tool for Rational
Ireland. The long-term operating concessions of French
Heat Management, Master thesis 2006:21, Department
district heating systems might hamper their
of Civil and Environmental Engineering, Chalmers,
development. In the Czech Republic, much district
Gothenburg (2006).
heating is produced in extraction turbines at the
expense of more valuable electricity. Romanian district [7] Euroheat, District heating and Cooling country by
heating use was reduced by one-half by cheap gas. country 2007 survey, Euroheat, Brussels (2007).
In general, it should be advantageous that one actor [8] J. Zeman and S. Werner, District Heating System
controls the whole district-heating value chain from Ownership Guide, DHCAN project, BRE, Watford
source to consumption in order to utilise synergies and (2004) http://projects.bre.co.uk/DHCAN/guides.html
to avoid economic risks with contracts between the [9] SNCU, Les rseaux de chaleur et de froid: lnergie
separate entities of a fragmented value chain. Like for citoyenne, SNCU, Paris (2004).
other long-term large-scale infrastructure investments, www.fg3e.fr/public/federation/syndicats/plaquettes.php
public involvement may be necessary for district ?root_page=6
heating development.
[10] T. Zenaty, CHP/DH sector in the Czech Republic:
Through cooperation among various well-established situation / problems / wishes, Energy Policy EHP
players in the district heating industry, knowledge, meeting, Budapest, 11 September 2008,
products and services can be transferred to evolving www.lsta.lt/files/seminarai/080911_Budapestas
district heating markets, which promotes industrial /CZ.pdf
prosperity for all parties and helps building sustainable
energy systems in Europe.
228
th th
IMPACT OF THE PRICE OF CO2 CERTIFICATES ON CHP AND

DISTRICT HEAT IN THE EU27
1
Markus Blesl
1
Institute of Energy Economics and the Rational Use of Energy (IER)
University of Stuttgart
INTRODUCTION modelling language of GAMS due to reasons of being

better transferable. TIMES is a multi-periodic linear
In the current energy and climate policy debate, one of
optimization model based on a technical approach at
the key points is the discussion about emission
which single plants are aggregated. The purpose is the
reduction targets and how they are spread among
evaluation of the economically optimal energy supply
different world regions or countries and also among
structure at a given need of end use energy and energy
different sectors. To find a cost optimal burden sharing
services and also at given energy and climate policy
of an emission reduction target, the different reduction
requirements. For this, the discounted system costs are
potentials of the particular sectors or technologies have
minimized, whereas the single players (industry,
to be known. To reach a reduction target, emission
supply, households) could have different economic
certificates in a country or region (like EU-27) are
considerations. The main objective of the model
allocated among the different sectors or between
development of TIMES is the flexible structure to
different types of heat and power generation
ensure a simple mathematic adjustment to the
technologies. This allocation (for example, auctioning)
respective problem.
of emission certificates is an important issue to
negotiate since the costs of buying certificates could be The pan European TIMES energy system model
an important factor in technology choices for (abbreviated as TIMES PanEU) is a model of 30
investment. regions which contains all the countries of EU-27 as
well as Switzerland, Norway and Iceland. The objective
The significant advantage of this approach is that the
function of the model is a minimization of the total
analysis of the different competing pathways to achieve
discounted system costs over the time horizon from
emission reductions also assesses how they influence
2000 to 2050. A perfect competition among different
each other. In the context of efficiency improvement in
technologies and paths of energy conversion is
industrial CHP and district heating and cooling, the use
assumed in the model. The TIMES PanEU model
of waste heat becomes an interest field. Efficiency
covers on a country level all sectors connected to
improvements in the residential or commercial sector is
energy supply and demand such as the supply of
examined in the topic of energy saving. Without
resources, the public and industrial generation of
analysing the entire energy system the possible
electricity and heat and the industrial, commercial,
advantages of CHP and district heating and cooling
household and transport sectors. Both greenhouse gas
couldnt be taken into account. This shows the
emissions (CO2, CH4, N2O) and pollutant emissions
difference to a standard cost potential curve approach,
(CO, NOx, SO2, NMVOC, PM10, PM2.5) are covered
which has a fixed order of measures depending on
by TIMES PanEU.
their avoidance cost.
The transport sector is disaggregated into four areas:
This analysis will evaluate the reduction potential of
road transport, rail traffic, inland shipping and.aviation.
CHP plants or in general the production of district
The road traffic includes five demand categories for
heating and cooling in the EU-27 using the energy
passenger transportation (car short distance, car long
system model, TIMES PanEU /Blesl et al 2008; Blesl
distance, bus, coach, motor bikes) and one for freight
2008; Blesl et al 2008b, Kuder Blesl 2009; Blesl 2009/.
service (trucks). The rail traffic includes three
categories: rail passenger transportation (divided into
TIMES PAN-EU MODEL short and long distance) and rail freight transportation.
The energy system model, TIMES (The Integrated The transport modes of inland shipping and aviation
Markal Efom System), is a further development of the are represented by a non-specified general process
two model generators, MARKAL and EFOM-ENV, where the development of the transport demand is
written in GAMS. TIMES was developed in recent embodied by the final energy demand.
years within the Energy Technology Systems Analysis The household sector contains eleven demand
Programme(ETSAP) from the IEA with contribution categories (space heating, cooling, hot water, cooking,
from the IER. It is classified in one category with the refrigeration, lighting, washing machines, laundry dryer,
models MARKAL, EFOM or MESSAGE. The model dishwasher, other electrics, other energy use), whereof
generator, TIMES, was developed in the general
229
th th
the first three correlate to specific building types (single Industrial heat demand by temperature and sub-
family houses in urban and rural areas and multi-family sector in the EU27
houses each described as existing stock and new The particular sub-sectors of the industrial sector use
build). The commercial sector is represented by a different chemical and physical conversion processes.
similar reference energy system (RES) and consists of Therefore, they need heat on different temperature
nine demand categories (space heating, cooling, hot levels (Figure 1). Processes with a need for very high
water, cooking, refrigeration, lighting, public street temperatures (> 1400 C) are e.g. blast furnaces
lighting, other electrics, other energy use). The first (iron/steel industry) or kilns (cement or lime industry).
three of them are subdivided according to different Processes with lower temperature levels occur in the
building types (large/small). food/tobacco (sugar production, dairy) industry, other
The agricultural sector is described by a general industries or in general for the supply of space heating
process with a mix of several energy carriers as input and hot water. Also, the pulp/paper industry has a high
and an aggregated demand of end use energy as need for heat at a lower temperature level (< 100 C).
output. Most of the heat is produced by the combustion of
The industrial sector is subdivided into several fuels. Other heat is generated by the use of electricity.
branches (for example, iron and steel, cement, lime) Key processes using electricity for high temperature
and into energy intensive and non-intensive branches. heat are chlorine electrolysis, aluminium electrolysis,
While the intensive ones are modelled by a process electric arc processes (iron/steel) and copper
orientated approach, the other industries have a similar electrolysis.
structure but with five energy services (process heat, 2,500
Others
steam, machinery drive, electrochemical, others).. Food/Tabacco
Pulp/Paper
The generation of electricity and heat in power plants, 2,000

Final energy consumption [PJ]
Oth. non-metallic
minerals
CHPs and heating plants is differentiated into public Glass flat
and industrial production. The model contains three

1,500 Glass hollow
different voltage levels of electricity (high voltage,

Lime
Cement
medium voltage, low voltage) and two independent

1,000
Oth. chemicals
heat grids (district heat, local heat). Chlorine

500
Ammonia
In the supply sector, all primary energy resources Oth. non-ferrous
(crude oil, natural gas, hard coal, lignite) are modelled

metals
0 Copper
0-60
60-100
100-120
120-180
180-240
240-300
300-360
360-420
420-480
480-540
540-600
600-700
700-800
800-900
900-1000
> 1500
1000-1100
1100-1200
1200-1300
1300-1400
1400-1500
by supply curves with several cost steps. Three Aluminium
categories can be differentiated: discovered reserves

Iron/Steel
Figure 1: Final energy consumption for industrial heat

(or developed sources), growth of reserves (or
production by temperature and sub-sector in the EU-27
secondary and tertiary extraction) and new discoveries. in 2005
Additionally, seven bio energy carriers are
differentiated: matured forest, bio gas, household On the country level, the role of the different member
waste, industrial waste, as well as energy plants states concerning a particular temperature level
containing sugary, starchy and lignocelluloses.. depends on the structure of the industrial sector in that
country. In general, the final energy consumption for
Due to its regional resolution, TIMES PanEU allows the
heat production at a specific temperature level is
consideration of country specific features, for example
dominated by the bigger member states and members
different structures of the stock of power plants,
of the EU-15 like Germany, Italy, UK, France and
different extension potentials for renewables as well as
Spain. However, new member states like Poland,
potentials for storing CO2. An interregional electricity
Czech Republic or Romania also play an significant
trade is implemented in the model, so that exports and
role. Some countries only play a key role at single sub-
imports of electricity according to the existing border
sectors and thus only for some temperature levels.
capacities could be calculated endogenously in the
model. The lower temperature levels are dominated by the
industrial sub-sectors pulp/paper, food/tobacco and
The role of CHP and district heating will be influenced
others. Due to high activities in those areas, the
in the future by the heating demand for the heat, space
heating demand is clearly influenced by France (strong
heating and cooling processes. The following chapters
for food/tobacco), Sweden and Finland (strong for
describe the status and the assumed development for
pulp/paper) next to other big countries like Germany,
Europe.
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
230
th th
chemical and food & tobacco industries are the most occuring at this price level are analysed according to
important ones. Each country is clearly specialised in the role of the different reduction possibilities.
differing industrial sub-sectors.
The foundation for the CO2 price variation is set based
on the CO2 price outcomes from two scenario runs with
Space heating and cooling demand in Europe
a reduction target of 15% [scenario: 15% reduction
today and in future
(2020)] and 40% [scenario: 40% reduction (2020)] in
The demand for space heating and cooling differs 2020 compared to the Kyoto base year (Table 1). In the
among the countries in Europe due to the differences in long run (2050), both of these restricting scenarios
climatic conditions and in living standards (e.g. square have the same target which equals a 450ppm goal
meters per capita) and building standards. This is (-71% in 2050 compared to 1990).
especially applicable to the assessment of current and
near future energy demand. 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%
14000
12000
Table 1: CO2 reduction pathways for the two restricting
scenarios
10000
Heat demand in [PJ]
8000
The resulting CO2 prices of these two restriction
scenarios build the framework for the price variations.
6000
Within the range of the resulting CO2 prices, the carbon
4000
price varies between 10 /tCO2 and 110 /tCO2 in 2020
2000 in increments of 10 . In 2030, the price varies between
0
27 / tCO2 and 123 / tCO2. The price increases until it
2005 2010 2015 2020 2025 2030 2035 2040 2045 2050 reaches the level of a 450 ppm scenario in 2050
Commercial Urban MFH Urban SFH Rual New Houses
(Figure 3). The emission reductions are evaluated
using the results from the different scenarios in
1200
Residential SFH comparison to the case of the lowest CO2 prices (10 /t
1000 Residential MFH
in 2020, 27 /t in 2030). First, the total reductions over
Commercial
all sectors are presented and afterwards the focus will
Cooiling demand in PJ
800
be on the industrial sector. The drivers of the reduction
600 are shown separately. Looking at the industrial sector,
the reasons for the emission reductions could be split
400
up into more efficient production processes, more
200 efficient heat supply, fuel switch in heat generating
units or CCS technologies in production processes and
0
energy supply.
2000 2005 2010 2015 2020 2025 2030 2040 2050
year
450
Figure 2: Demand for space heating/hot water and for

15% reduction
(2020)
40% reduction
cooling in the EU27

400
(2020)
CO2_10
350
CO2_20
In the 2000, the useful demand for cooling was less

300 CO2_30
CO 2 price [ 2000 ]
CO2_40
than 5% lower than the useful demand for space

250
CO2_50
200
heating and hot water. In the long term, the cooling 150
CO2_60
CO2_70
demand will be dominated by the commercial sector. 100

CO2_80
The increase of cooling demand in the EU27 up to

CO2_90
50
CO2_100
2050 will reach approx. 1120 PJ in the residential and 0

2020 2025 2030 2035 2040 2045 2050
CO2_110
commercial sectors.
Figure 3: CO2 prices of the different scenarios
SCENARIO DEFINITION OVERVIEW OF THE DYNAMIC DEVELOPMENT OF

A parameter variation is used to evaluate the reduction THE ENERGY SYSTEM OVER TIME
potential and the role of CHP and district heat in the In the following analysis, the two scenarios with the
energy system of the EU27. By varying the CO2 price, lowest (10 /t CO2 in 2020, scenario CO2_010) and the
the reduction potential curves are constructed. highest (110 /t CO2 in 2020, scenario CO2_110)
Therefore, different scenarios with different CO2 prices prices are displayed to show the range in which the
(one common price for ETS and Non-ETS sectors) are results of the price variation occur. Therefore, the
calculated with TIMES PanEU and the reductions development over the whole modelling horizon
(20002050) is presented to rank the more detailed
231
th th
results from a point of time within these more general The electricity generation from CHP plants in the EU27
results over a period of time. Since the CO2 reduction increases by 79% from about 380 TWh in the year
target of the two bounding scenarios [scenario 15% 2000 to 640 TWh by the year 2020 (see Figure 4). The
reduction (2020) and scenario 40% reduction (2020)] extension of the electricity generation from CHP plants
clearly differ in the mid-term periods of 2020 and 2030 is essentially supported by gas-fired and biomass
(Table 1) and the corresponding prices are more based CHP plants. Additionally, existing public CHP
different in these periods (Figure 3), the energy system plants with an extraction condensing turbine are
shows the most variations during this time. substituted by CHP plants with a higher power-to-heat
ratio and there is also an extension of industrial CHP
To show the development over the modelled time
plants, which are often used in cooperation with
period, first of all the net electricity generation of EU-27
communal facilities. The intermediate growth of CHP
is displayed (Figure 4). The overall electricity
plants in the commercial sector between the years
generation remains almost constant at 2010 levels
2015 and 2035 are based on efficiency advantages of
(about 3 200 TWh) until 2030. In later periods, there is
CHP plants with a medium sized internal combustion
a clear increase in electricity generation up to
gas engine. In the long term, the limited possibilities of
4 255 TWh (2050, scenario CO2_110). The increase in
using CO2 free fuels in commercial CHPs will result in
the later periods is driven by stronger emission
these phasing out in the commercial sector. Until the
reduction targets. To fulfil the restrictions, more
year 2050 the electricity production by CHP plants in
electricity with low specific emissions and high end use
the scenarios further increases up to a level of 1055 to
efficiency in the demand sectors is used.
1100 TWh. CHP plants based on biomass as well as
According to the given CO2 prices of the two scenarios CCS CHP are an important option in the year 2050.
(CO2_010 and CO2_110), the main differences occur
in the mid term periods. While the total electricity 1200.00
demand in 2020 is lower in the scenario with higher

Net electricity generation CHP in [TWh]
1000.00
emission certificate prices (-22 TWh in 2020 between
CO2_110 and CO2_010), the demand is higher by 800.00
86 TWh in 2030. The increase is due to the use of 600.00

Public
Comercial
more efficient technologies in the end use sectors Industry
400.00
resulting in lower electricity demand in 2020, while by
2030 the switch to electricity based technologies to fulfil 200.00
the emission restrictions has already taken place. 0.00

CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
4500 Others / Waste
non-ren.
4000 Other
Renewables 2000 2005.0 2010 2015 2020 2025 2030 2035 2040 2045 2050
3500 Biomass /
Net electricity generation [TWh]
Waste ren.
3000 Solar Figure 5: Net electricity generation CHP by sector in the
2500 Wind EU-27
2000
Hydro
1500
Nuclear
In addition to the net electricity generation, the primary
1000
Natural gas
(Figure 6) and final energy (Figure 7) consumption of
500
Oil
the EU-27 are also analysed over the whole time
0
Lignite
period. Overall, the primary energy consumption (PEC)
Statistic
CO2_010
CO2_010
CO2_010
CO2_010
CO2_010
CO2_010
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_110
CO2_110
CO2_110
CO2_110
CO2_110
Coal does not show clear changes and remains at a level of

2000 2010 2015 2020 2025 2030 2035 2040 2045 2050
about 75 000 PJ. The lowest total PEC occurs in the
mid-term periods. The total consumption is influenced
Figure 4: Net electricity generation in the EU-27 by an increasing efficiency till 2030 and later on by a
Aside from the changes in the total electricity demand, higher share of renewables and also CCS which both
there is also a change in the structure of the electricity lead to a higher consumption due to the lower thermal
generation. At higher CO2 prices, less coal (-120 TWh efficiency in the combustion processes.
from coal fired power plants in 2030) and more gas
Looking at the impact of the single energy carriers,
(+44 TWh) and nuclear (+30 TWh) are used and more
there is a distinct change between the two scenarios
electricity from renewable energy sources (+35 TWh
than in the total sum of the PEC. In 2030 at a higher
from wind, +56 TWh from biomass and renewable
CO2 price, less coal (-1 675 PJ) and petroleum
waste) is generated. Furthermore, CCS is used more
products (-881 PJ) and more Hydro, wind, solar
widely under the conditions of the CO2_110 scenario in
(+338 PJ) and other renewables +4856 PJ) (mainly
2030 compared to CO2_010.
biomass) are used.
232
th th
90000 Electricity 60000 Others (Methanol,

import Hydrogen)
80000
Primary energy consumption [PJ]
Total final energy consumption [PJ]

Waste (non 50000 Waste
70000 renewable)
Other 40000
60000 renewables Renewables
50000 Hydro, wind,

solar 30000
Heat
40000
Nuclear
20000
30000 Electricity
Natural gas
20000 10000
Gas
10000 Oil
0
0 Petroleum
Statistic
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_010
CO2_010
CO2_100
CO2_010
CO2_100
CO2_100
Lignite
products
Statistic
CO2_010
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_110
CO2_010
CO2_110 Coal
Coal
2000 2010 2015 2020 2025 2030 2035 2040 2045 2050
2000 2010 2015 2020 2025 2030 2035 2040 2045 2050
Figure 7: Final energy consumption in the EU-27

Figure 6: Primary energy consumption in the EU-27
In contrast to the year 2000, the distribution of local
The final energy consumption (FEC) shows and district heat to the household, commercial and
comparable results (Figure 7). The use of petroleum industrial sectors changes by the year 2050 with an
products declines over time in both scenarios additional approx. 1000 PJ district consumed in the
(-9 052 PJ in scenario CO2_010 between 2000 and year 2050 (see Figure 8).
2050). The use of gas increases at lower CO2 prices in
the mid-term periods (up to more than 13 500 PJ in 3500
2020 at scenario CO2_010), but declines in both

Supply
3000 Residential
Final energy consumption heat in [PJ]
scenarios at the very end. This shows that one early

Industry
Comercial
2500
and cost-effective measure for emission reduction is
Agriculture
the fossil fuel switch from petroleum products and coal 2000
to gas in the end use sectors. 1500
As already shown with electricity generation, the use of 1000
electricity also increases in the end use sectors. 500
Especially in the long run at higher carbon prices, there

0
is a clear rise. The use of renewable energy sources
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
CO2_010
CO2_100
also increases constantly in both scenarios. In 2020
and 2030, clearly more renewables are used in the
2000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050
CO2_110 scenario due to the higher CO2 prices Figure 8: Final energy consumption district heat in the
(+3900 PJ in 2030). EU27
In contrast to the PEC, the total FEC decreases slightly In the long term, the CO2 contents of the heat supply for
in the long run. The reason for this different the end use sectors will be reduced from 130 kg
development is that the higher conversion losses CO2/MWh to 122 kg CO2/MWh in 2020 and from 113
arising from a higher electricity demand and the kg CO2/MWh to 36 kg CO2/MWh in the year 2050,
extended use of renewables and CCS at the public which is one explanation for achieving the CO2
electricity generation are balanced at PEC and do not reduction targets in this area. On the other hand, the
influence the FEC. possibility to use renewable energy or to install CCS,
Even though more renewables (mainly biomass) are increasingly influences the penetration of CHP. By
used, due to the higher use of electricity with its high 2050, fossil heat plants will also be substituted with
end use efficiency and other efficiency improvements, large heat pumps and solar thermal heat plants in
the total FEC declines to 49 482 PJ (in 2050 at combination with storages, biomass heat plants fuelled
scenario CO2_110). This efficiency improvement with wood or woody crops and biogas.
occurs in the industrial sector mainly at industrial The overall emissions decrease is based on the
production processes, but is also clearly driven by emission reductions of the single sectors leading to
efficiency improvements ain the residential and different CO2 abatement costs (Figure 9). The total
transport sectors.9 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
9
For a detailed discussion of the effects in the different end industrial sector and the residential/commercial sector
use sectors and its impact on the total final energy
consumption see /Blesl et al. (2010)/ also show clear reductions. The transport sector tends
233
th th
only to reduce its emissions with very strict reduction 900

Transport
targets connected to high carbon prices. 800

Commercial
Additional emission reduction [Mt]

700
Residential
600
4500 Industry
4000 500
Conversion/
Transport Production
3500 400
Emissions of CO 2 [Mt]
3000
300
2500 Households,
commercial, 200
2000 AGR
100
1500 Industry
1000 0
36 46 56 65 75 85 94 104 114 123
500 Carbon price [2000/tCO2]
Conversion,
production
0
Figure 10: Additional CO2 reduction in the EU-27 in 2030

Statistic
CO2_010
CO2_010
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
CO2_010
CO2_110
by sector compared to the scenario with the lowest CO2
2000 2010 2015 2020 2025 2030 2035 2040 2045 2050 price of 27 /t
Figure 9: CO2 emissions in the EU-27 One reason for the CO2 reduction in the residential,
commercial and industrial sectors is the increase in
ANALYSIS AT A SPECIFIC POINT OF TIME WITH final energy demand from district heat (see figure 11).
FOCUS ON 2030 The overall increase of the district heat demand
After the general effects are described and the influenced by the different CO2 prices is 14%. The
scenarios with the lowest and highest CO2 prices are biggest growth can be seen in the commercial sector,
analysed over the whole period of time, a more detailed where the total district heat demand for district heating
analysis shows the effects in the industrial sector grows by over 30% between the min the minimum and
during the mid-term periods with a particular focus on maximum CO2 certificate price.
2030.
Firstly, the reduction potential of the different sectors
2000
1800
should be analysed (Figure 10). Both conversion
Final energy demand heat in [PJ]
1600
/production and the other end use sectors are taken 1400
into account. As in the results of the emission reduction 1200

Residential
Industry
from 2000 to 2050 (Figure 9), the industrial reduction 1000 Comercial
Agriculture
potential plays the key role next to the 800
conversion/production sector. Looking at the year 2030 600
and comparing the additional CO2 reductions when the

400
200
CO2 price is increased from 27 /t to 123 /t, the 0
strongest additional reduction occurs at the conversion 27 36 46 56 65 75 85 94 104 114 123
Carbon price [/T CO2]
sector (+351 Mt at a price of 123 /t compared to
27 /t). An additional 301 Mt of CO2 are reduced by the
Figure 11: Final energy demand of district heat in the
industrial sector. EU-27 in 2030 by sector
Especially at higher prices above 94 /t, the reduction
However, the generation of district heat from the use of
potential of the industrial sector becomes more and
renewable sources and CCS will be one reason for the
more important. Its share of the total additional
growth of the reduction potential in the conversion
reduction increases from 33% (36 /t compared to
sector (Figure 12). The share of the use of renewables,
27 /t) to 37% (123 /t to 27 /t). The lowest reduction
especially biomass, will rise from 29% to 60%. More
occurs in the transport sector. Till a price of 85 /t, only
than 1300 PJ of additional biomass will be needed.
an additional 6.3 Mt are reduced, while at a price of
Due to this increase, the average heat to power ratio of
123 /t an additional 18.9 Mt are reduced. In the
all CHPs will fall from 0.9 to 0.66. In the cases where
residential and commercial sector, some reduction
CO2 prices exceed 56 /tCO2, the district heat
possibilities are cost-effective even without a price on
generation in CCS CHP plants grow more rapidly. The
CO2. The energy savings outweigh the additional
specific emissions of the district heat generation
investment costs. Those reduction measures are
decrease from approx. 380 kg / MWh to 84 kg / MWh.
especially connected to the building/heating sector.
234
th th
2000 4200 30%
SUM
1800 4180
25%
Change of Electricity supply by technology

1600 4160 public
generation
compared to CO 2 price of 27 /t
20%
District heat generation in [PJ]
Use of Electricity [PJ]

1400 4140 Condensing
industrial
1200 Heat Plant RES 15%
4120
Heat Plant CHP
1000 industrial
CHP RES 4100 10%
CHP FC
800
CHP CCS 4080
5%
600 CHP
4060
400 0%
4040
200
-5%
4020
0
27 36 46 56 65 75 85 94 104 114 123 4000 -10%
27 36 46 56 65 75 85 94 104 114 123
carbon price in [/t CO2]
Carbon price [2000/tCO2]
Figure 12: District heat generation in the EU-27 in 2030 by Figure 13: Use of electricity in the EU-27 in 2030 by
technology group technology
In the industrial sector, the share of CHP will grow. The The other part of the industrial supply processes is the
additional emission reductions by the industrial sector industrial heat generation. The drivers for the emission
of 301 Mt in the year 2030 could be split into industrial reduction in industrial heat production are a switch to
supply and industrial production processes. The supply biomass (from coal and clearly from gas) and the use
side covers the industrial generation of energy of CCS in industrial CHPs (Figure 14). Between a CO2
commodities or energy services. These are electricity price of 36 and 56 /t of CO2 in 2030, there is a clear
from industrial condensing power plants and CHPs, increase in the use of renewables in boilers. The share
heat and steam from CHPs and boilers, space heating of renewables in the total fuel use in industrial boilers
and heat for hot water as well as cooling. The supply increases from 33% to 51%. As a result, the thermal
activities play an important role in the industrial sub- efficiency of boilers has an overall decrease.
sectors with a high share of space heating (such as
In industrial CHPs, there is also a slight increase in the
food & tobacco or other industries) or low temperature
use of s. This switch takes place between CO2 prices
process heat (such as pulp & paper or food & tobacco).
of 27 /t to 65 /t. However, the main change
In total, from the additional reduced emissions, 147 Mt concerning CHPs is the increasing use of CCS. At a
are reduced by industrial supply processes and 154 Mt CO2 price above 94 /t, there is a clear rise in the use
by production processes in 2030. While at lower a CO2 of this technology. These CCS CHPs are mainly gas
price more emissions are reduced on the supply side fired10. This is why the share of renewables used in
(66% of the additional reduction based on supply industrial CHPs declines at a price over 75 /t again.
processes at 46 /t), at higher prices more and more
Like biomass, the extended CCS use also leads to
reductions take place on the production side (49 %
lower efficiencies resulting in both the efficiency of
based on supply processes at 123 /t).
boilers and CHPs to decline over time. Accordingly, the
The additional electricity needed at high CO2 prices is key driver is not efficiency improvements, but the use
mainly generated by industrial autoproducers. Within of renewables and CCS. The effects of renewables and
this industrial production, the additional electricity CCS compensate the trend to lower energy intensity
mainly comes from CHP power plants. The use of within one technology. Gas boilers become more
electricity in the industrial sector from public generation efficient and as do biomass boilers. However, the more
remains relatively constant even when the CO2 price efficient biomass boilers still use more fuel than the gas
increases. Accordingly, one key way to reduce the boilers.
emissions on the supply side is through the extended
Looking at the heat output by technology, there is also
use of CHP plants for industrial power generation. This
a shift (Figure 14). At lower emission prices, the heat
higher amount of electricity from industrial
output from industrial boilers stays almost constant.
autoproducers (Figure 13) leads to higher conversion
Within this range, the share of renewables used
losses in total when the fuel use is considered. As
increases (as illustrated in Figure 13). Afterwards, at a
described above, that is one reason for the difference
price above 65 /t, boilers are substituted with heat
between final energy consumption and fuel
from CHPs and district heat. Both heat commodities
consumption. Another reason is the lower efficiency of
electricity generation due to the higher use of CCS.
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)/
235
th th
are generated in combination with an increasing share that the climate conditions within Europe differ
of renewables, a higher CO2 price and from CCS. substantially.
90%
CHP industrial
Within the energy system of the EU-27, there are
different emission reduction pathways. The emissions
(total)
80%
CHP industrial
70%
share RES
could be reduced by a fuel switch in more efficient (or
better, less carbon intensive) energy supply or by a
Boiler industrial

60%
Efficiency and share [%]
change in production processes. Key drivers

Boiler industrial
share RES
50%
CCS CHP
industrial (share
Fuel input)
concerning the emission reduction in production
40%
processes and in heat demand side are efficiency
improvements due to new technologies and
30%
20%
technological improvements. The key driver concerning
10%
the supply side of electricity and heat generation is the
0% increased use of renewables, mainly biomass, for heat
27 36 46 56 65 75 85 94 104 114 123
Carbon price [2000/tCO2] generation. The CCS technology also plays an
Figure 14: Efficiency of heat supply technologies and important role in the reduction of emissions. Due to the
share of CCS at industrial CHP in the EU-27 in 2030 increased use of renewables in CHP and heat plants
and the use of CCS, the efficiency in the supply
150
District Heat processes decreases at higher CO2 prices.
CHP industrial
In the long run to a CO2-free world, the possibility to

100
Boiler
change in heat output [PJ]
50
generate district heat with renewable energy and the
use of CCS make the decarbonisation of the energy
0
consumption in the end use sectors possible.
-50 In general, the progression of district heat depends
crucially on the possibility of generating CO2 emission
free district heat and electricity.
-100
-150
36 46 56 65 75 85 94 104 114 123
Carbon price [2000/tCO2] REFERENCES
Figure 15: Heat supply by technology in the industrial
sector in the EU-27 in 2030 compared to the scenario with [1] Blesl, M.; Kober, T.; Bruchof, D.; Kuder, R.: Effects
the lowest CO2 price of 27 /t of climate and energy policy related measures and
targets on the future structure of the European
In total, all these described effects concerning the energy system in 2020 and beyond, Energy Policy,
industrial supply processes lead to the additional 2010 (forthcoming)
emission reduction in 2030 of 147 Mt at a price of
[2] Blesl, M.; Kober, T.; Bruchof, D.; Kuder, R.: Beitrag
between 123 /t and 27 /t. In general, more emissions
von technologischen und strukturellen
are reduced in boilers than in CHPs. The reasons are
Vernderungen im Energiesystem der EU 27 zur
the fuel switch from coal and mainly gas to renewables
Erreichung ambitionierter Klimaschutzziele,
at lower CO2 prices and later on the substitution of
Zeitschrift fr Energiewirtschaft 04/2008
boilers with CHPs (less boilers are used and therewith
produce less emissions). [3] Blesl, M.: CHP and district heat in the Europe
under an emission reduction regime, in: 11th
Due to a higher use of CHPs, there is no clear increase
in emissions during the mid-term ranges. When the
Cooling in Reykjavik, Island
output of heat stays constant and a higher share of
CCS is used, then clear emission reductions from [4] Blesl, M., Cosmi, C. ,Kypreos, S. , Salvia, M.:
CHPs (additional 48.7 Mt in 2030 at 123 /t compared Technical paper n Technical Report n T3.18
to 27 /t) occur. RS 2a Summary report of Pan European model
results BAU scenario EU Integrated Project
CONCLUSION AND OUTLOOK NEEDS New Energy Externalities Developments
for Sustainability October, 2008
District heating generation offers an economic potential
for expansion in the future. Depending on the regions [5] DEHSt (2010): Deutsche
or countries, the development will be different because Emissionshandelsstelle, Kohlendioxidemissionen
the starting point is economic growth and the existing der emissionshandelspflichtigen Anlagen im Jahr
national laws or cross-subsidies for competitors energy 2009 in Deutschland, Mai 2010
carriers. In addition, it is necessary to take into account
236
th th
[6] EEA (2010): European Environment Agency, Probabilistic Scenarios, 2010, www.feem-
European Union emission trading scheme (ETS) project.net/planets
data viewer, 2010
[9] Kuder, Blesl (2009): Kuder, R.; Blesl, M.: Effects of
[7] Kober, Blesl (2010a): Analysis of potentials a white certificate trading scheme on the energy
and costs of storage of CO2 in the Utsira aquifer in system of the EU-27, Fullpaper 10th IAEE
the North Sea; report work package 4: Regional European Conference in Vienna, Austria, 2009
analysis at North Sea level, 2010, www.fenco-
[10] UNFCCC (2009): GHG inventory reports for the
era.net
single member states of the EU-27, submission
[8] Kober, Blesl (2010b): Perspectives of CCS 2009 situation / problems / wishes, Energy Policy
in Europe considering technical and economic EHP meeting, Budapest, 11 September 2008,
power plant uncertainties; in PLANETS work www.lsta.lt/files/seminarai/080911_Budapestas/CZ.pdf
package 6 deliverable No. 15 Report on
237
th th
CONSIDERATIONS AND CALCULATIONS ON SYSTEM EFFICIENCIES OF HEATING

SYSTEMS IN BUILDINGS CONNECTED TO DISTRICT HEATING
1 2 1
Maria Justo Alonso , Rolf Ulseth and Jacob Stang
1
SINTEF Energy Research, Department of Energy Processes
2
NTNU, Faculty of Engineering Science and Technology,
Department of Energy and Process Engineering
ABSTRACT of EN-standards. The main goal of the Directive is to

promote the improvement of the energy performance of
In order to harmonize the implementation of the EC
buildings within the Community, taken into account
Directive on the energy performance of buildings
outdoor climatic and local conditions as well as indoor
(EPBD) [1], and to provide guidelines and common
climate requirements and cost-effectiveness. The main
calculation tools, several technical standards have
focus is on reducing the primary energy use and the
been worked out by CEN in accordance with a
associated CO2 emission of buildings.
mandate from the EC. This paper focuses on
calculating system efficiencies of hydronic heating Figure 1 shows how the Primary energy use is
systems by using the standards EN 15316-x-x [4], [5], calculated based on all the steps where the energy is
[6]. changing its nature from the source to the end use. In
the current case, the energy calculations are performed
The paper has been written in order to ease and
for the systems within the building to be able to
diminish the time consuming process of interpreting
calculate the delivered energy to the building. This
details in the standards such as the numbered EN
means that the building substation with the heat
15316-x-x, and with the goal to enlighten main parts of
exchangers and tap water storage are included.
these standards.
In the current scenario, all the losses before the heat is
To exemplify some results, an apartment building of
delivered to the building are included in the primary
1000 m2 floor area located in a climate like Oslo is
energy factor (PEF) for the delivered heat. In case of
chosen. In the base case, the design distribution
considering the complete scenario, the boundaries for
temperatures in the building are 80/60. The different
the energy performance indicators are the whole
efficiency figures applying for this case are calculated
energy chain from the source to the end use. In this
efficiency values for the production of the heat, for its
case, if a CHP plant is represented, the power bonus
distribution through the building and its emission in the
method (EN 15316-4-5) should be used. This method
room. The room efficiency is the one that has the
is giving the produced district heat a bonus for the
bigger influence on the total system efficiency.
electricity produced assuming that this electricity
replaces electricity production with a high PEF-value.
INTRODUCTION
According to the implementation of EPBD, it is crucial
The Directive on the energy performance of buildings is that the system borders are clearly defined so that the
carried out in order to be used together with a number delivered energy is doubtlessly defined.
Calculation of Primary Energy use according to EPBD and mandated EN-standards
Primary energy use = DEdh PEFdh (f (x,y,z)) + DEel PEFel (f (x,y,z)) = (Weighted delivered energy indicator (kWh/m2)) AC
Delivered Energy (DE) dh +el (PEF might be PEFR or PEFT depending on purpose) ( AC = conditioned floor Area )
Net energy demand

Primary energy use calculated by PEF(x,y,z)
Distribution and
Transmission (el)

Electricity Energy carrier (z)
Hot tap water Energy carrier (y) Waste coll.

Air and room + Delivered energy (el) CHP-plant Logging
heating system Heat boilers Energy carrier (x)
Delivered energy (dh) Extraction
DH substation Storage Processing
Storage
End use Heating Distribution and Generation Transportation
systems Transmission (dh) Transformation 2010 / 04 / RU
demand !
Calculating end use and losses by EN standards worked out according to mandate from the EU Commission
Heating systems efficiency Calculating direction
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
238
th th
METHOD Finally; once distributed, the heat is emitted according

to the demand. For the present case, the heat is
Definition of building and system build up
delivered either by radiators (80/60 C) in the base
In this exemplified case, the main chosen building is an case, floor heating (35/28 C), or domestic hot water at
apartment building. This building category seems to be
60 C. Figure 2 gives a further visual explanation. For
the most representative concerning heat use among
the present paper, the supply of heat is just done by a
the building categories defined in the EPBD [1].
hydronic heating system. The possible heat loss from
In the shown example, the size of the building is the distributed air is neglected since the temperature of
chosen to be 1000 m floor area since this size should the air is assumed to be slightly lower than the
be rather representative and be a good compromise temperatures in the rooms.
between the previous and the proposed new recast of
the EPBD. [2] Categories of building
A building of these features corresponds to a three The presented analysis shows results for five kinds of
storeys squared building with four flats of about 80 m buildings described in the EPBD which are: single
per storey. family house and apartment block, office buildings,
hotel and restaurants, educational buildings and
In this setting, the total heating system efficiency in the hospital buildings. When it comes to heat consumption
building is built up based on the differentiation between for these buildings, the measurements performed in
the three main parts of the system. It must be defined Linda Pedersens PhD thesis [3] show that the
where the substation is located in the building, i.e. consumption of the apartment buildings is about
where the heat is exchanged from the distribution
116 kWh/m, while hospitals use 150 and office
network DH stage in Figure 2. The heat supply to the
heating system within the building from the district buildings use 100 kWh/m. These measured values
heating system is assumed to be provided by two heat include the domestic hot water (DHW) and the space
exchangers and hot water storage defined as the heating (SH) consumption.
building substation part of the system.
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
2010/05/RU place to place. Owing to this, the outdoor climate
affects not only the heat consumption but also the
DH relative losses. In general the relative losses are

increasing with an increased ratio between the degree
Ventil-
ation + + CW
days and T between the dimensioning internal and
+
air external temperature.
Hot tap Room Substation The average outside temperature affects the heat
water heating system consumption and the temperature variations affect the
distrib. distrib. border regulation of the heat emitters. This means that during
Figure 2 Sketch of the system elements for production, cold periods, the temperature of the supply water tends
distribution and conditioning of the rooms to be increased imposing an increase in the losses
related to the transport of water with higher
From the substation, the hot water is distributed either temperatures. The design temperature for the radiators
for air and space heating or as domestic hot water. in the base case in this paper is 80/60, and in warmer
Both uses are provided by their own heat exchanger periods, this temperature is decreased in order to
and the necessary pipelines will now be referred to as reduce losses and adapt the supply temperature to the
distribution pipelines. outside temperature. This affects the efficiencies in a
positive way.
239
th th
Finally, the outdoor climate affects the length of the temperature is here constantly at the designed point of
heating season. Usually, the lower the average outside 60 oC.
temperature, the longer the heating season. This
When dealing with distribution of SH, the losses are
however, does not affect the DHW since this is more or
considered dependent on the kind of insulation material
less steady all the year along.
and the ambient and the mean water temperature in
the supply and return pipes.
Positioning of substation
Figure 2 shows the positioning of the substation. The The heat emission to the room from the DHW draw-off
heat is delivered from the district heating pipelines tap discharge cocks is considered to be negligible in
through two separate heat exchangers, one for heating comparison to the total heat consumption.
the water in the storage tank for DHW by a circulating Dealing with space heating a distinction is done with
loop, and the other for the air heating and space respect to the kind of emission. Two major groups are
heating system. The main reason for having two heat considered: the emission by floor heating and by
exchangers is due to the different needs of temperature radiators. The first has a low temperature distribution of
levels. In the calculations dealing with the production, 35/28 C. As for the radiator system, the analyzed base
the used heat demand used is the total demand, while case is 80/60 C for supply/return design values.
for the distribution the heat is divided into heat
distribution for SH and for DHW. Besides the temperature level, the placing in the room
affects the stratification efficiency and the loss through
In order to calculate losses related to the storage tank, the outside wall. Furthermore, another point related to
it is assumed that the tank is of a common type with a the temperature is the regulation of the room
common value for the stand by heat loss. The system temperature, which in our case, is assumed to be a
design consists of the coupling in series of 289 litres PI-regulator, even if in a lot of apartments this
storage tanks. There are considered two tanks for the regulation is quite often done by on/off regulation.
apartment building of 1000 m2 but one more tank if the
case is dealing with hospitals, educational and hotel
RESULTS
buildings. For other sizes the number of tanks is
adjusted according to the demand. Production efficiency
Efficiencies to be studied The production efficiency is shown in Figure 3. By using

the losses on the heat demand and the temperature
For the present paper, as written previously, the
difference as basis for calculation, the values in
system is divided in three smaller system parts which
Table 1 are obtained.
are independent. For every component, the efficiency
is calculated following different standards: Production efficiencies for the different types of buildings according to the
EN 15316-4-5 :2007 (Oslo climate) with a distribution temperature 80/60
Production; according to EN 15316-4-5:2007[4] 0,990
Distribution; according to EN 15316-2-3:2007 and

EN 15316-3-2:2007[5] 0,980
System efficiency
Room emission; according to EN 15316-2-1:2007[6]

The efficiency of the production includes the losses 0,970
Apartment block
depending on the thickness of the insulation material, Office building
the insulation material itself, the storage tank, the Hotel and restaurant
building
complete local piping system of the substation system 0,960 Educational building
and the temperature difference between the two media

Hospital building
Single family
and the ambient. It takes into account the thermal loss 0,950
of the total substation. For this case the substation is 0 250 500 750 1000 1250 1500 1750 2000 2250
considered to be in an unheated part and therefore the Building floor area [m2]
losses are considered as unrecoverable.
Figure 3 Production distribution of the 80/60 C district
In case of the distribution, the efficiency depends on heating for different buildings
the use of the heated water. In case of being a part of a
DHW system; the energy used for heating the water
As Figure 3 shows, the bigger the building, the higher
which is not drawn-off and which slowly gets cold in the
the efficiency. This effect is due to the reduction of the
pipelines, has to be considered as loss. Moreover, heat
relative losses when the size of the substation (kW)
is used to heat up the pipes and fittings. Since the
increases. The curve profile is decreased slightly from
building is large enough to need a circulation loop this
2000 m2 and downwards, and then decreasing rapidly
loop is considered to be a source of loss? The water
from about 1000 m2 down to 500 m2.
240
th th
In addition, it can also be observed that among all loop ranging up to 35 % of the total losses from the
types of buildings, apartment houses represent total distribution system. This loss is related to the fact
somehow the highest efficiencies which justify the main that the water that remains in the distribution pipelines
focus in this study. The displayed case applies for the represents 5% of the total losses per flat.
values where the design temperature level is 80/60 C.
For this calculation it is assumed that the pipelines
It can be concluded from other calculations that the
have insulation which a loss of 0.3 W/mK (the
higher the design distribution temperature level, the
pipelines are considered to be according to the
lower the production efficiency. This conclusion is what
category installed after 1995 in [5]). In this calculation
could be expected considering the difference between
the losses due refilling the pipes with hot water are
the average temperature and the ambient temperature;
included. This heat could be considered as recoverable
the larger this difference, the larger the losses.
loss for space heating during the heating season but in
As shown in Table 1, the efficiency varies only between lack of a special national annex all the losses related to
0.9784 and 0.9673. It can be concluded, compared to the distribution of DHW should be considered as non
the distribution loss values that the production recoverable. These losses are not related to the
efficiency is not changing significantly even if the demand for heat and will consequently be lost or result
temperature level is changed. As a conclusion it can be in increased room temperatures.
said that the losses in the production are relatively low
When it comes to SH, the losses are related to the
for bigger houses but increasing quite rapidly for
temperature difference in the non-heated areas where
smaller buildings.
the water goes through. These losses are relatively low
Table 1 Efficiency of DH production system for different compared with tap water since most of these losses
design temperature levels. are considered to be recoverable. The values used are
tabulated in the EN standard [5].
Kind of
building 80/60 70/55 55/45 35/28 The percentage of recoverable losses is the cause of
the higher efficiency for distribution of space heating
Apartment which ranges 0.99, whilst the efficiency for distribution
blocks 0,9776 0,9778 0,9780 0,9784
of DHW is in thee range of 0.60.
Office building 0,9729 0,9732 0,9735 0,9740
Emission efficiency in the rooms
Hotel and
In this case, domestic hot water is not considered to
restaurant
building 0,9701 0,9703 0,9706 0,9709 contribute to the room heating since the losses from
the discharge cocks are considered negligible.
Educational
building 0,9676 0,9678 0,9681 0,9685 In case of space heating a difference has to be made
between floor heating and radiator heating when it
Hospital comes to the efficiency calculations.
building 0,9773 0,9775 0,9777 0,9780
Floor heating is by its nature emitted at lower
The quality of the insulation of the storage tank will also temperature, which has an effect on the stratification
influence the production efficiency. Manufactures efficiency since the lower the temperature level, the
should follow the standard pr EN5044:2005 [7] in order higher this str. By definition in [6] the stratification
to calculate these losses. Losses from storage tanks efficiency of floor heating is 1 whilst this parameter for
should be considered closely in practice, and tanks radiators goes down to 0.91 on the 80/60 distribution
with relative high losses should be considered for system. This value is combined with the efficiency
replacement or to be replaced by direct heat value of 95 % due to the positioning of the radiators on
exchangers for DHW. a normal external wall. Together these values make a
total room efficiency of 0.93.
Distribution efficiency However, a regulation with PI controllers for the
First the system for the distribution of tap water is radiators delivers an efficiency of 0.97 while the same
analyzed. In this case, the building includes a controller remains at 0.95 for floor heating.
circulation loop (in small dots Figure 2) which goes
In case of the embedded floor heating efficiency the
from the storage tank, to the third floor and the
efficiency is 0.93. Since it is considered to be normal
distribution branches (in bigger dots in Figure 2) which
insulation layer according to EN 1264, it results in a
deliver DHW from the central loop to the consumer.
emb of 0.95, the combination results in emb of 0.94.
The water temperature in the circulation loop is
assumed to be at 60 C throughout the whole year. The Due to these three parameters, floor heating all in all
biggest share of the losses come from the circulation has a room efficiency of 0.90 and radiators of 0.88.
241
th th
Total system efficiency Table 2 Total efficiencies of the systems for space
heating with floor heating and hot tap water system
The total system efficiency is in this paper calculated
as the product of the efficiencies of the different pasts
Total system efficiency
of the heating system in the building.
Figure 4 shows the room efficiency and the total Floor heating 0.87
system efficiency related to the temperature level
supply/return. The delivering water temperature to the Hot tap water 0.59
radiators is the parameter which affects the efficiency
the most.
CONCLUSIONS
The total system efficiency follows the pattern of the
The design temperature level for the system is the
room efficiency since this parameter has far the largest
most important factor when referring to the efficiency of
influence.
a hydronic heating system in buildings supplied by
Another conclusion from Figure 4 is that the efficiency district heating. . Therefore the possibility of lowering
of the complete system varies significantly with the the design temperature level of the heating system
design temperature level of the heating system. This is should be considered closely. This increases the
due to the temperature difference between the heating emission efficiency in the room and reduces the losses
system components and the ambient. With a lower from the distribution pipelines. It saves energy and
distribution temperature the losses will be smaller. 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
System efficiency for an hydronic heating system in actually decreases the room emission efficiency.
an apartment building of 1000m2 floor area
The introduction of an energy performance certificate
for buildings according to EPBD requires a transparent
0.94 calculation model according to the standards in the EN
0.92 15316 series. This paper gives a picture of system
efficiencies for hydronic heating systems and also an
Efficiency
0.9
0.88 idea of the time consuming process that has to be

performed in order to calculate the efficiency of a
Room
0.86 Efficiency
0.84 Total system system in detail. Therefore, it is concluded that some
0.82
efficiency
user-friendly guiding material should be desirable in
=80/60 =70/50 =55/45 order to enlighten and facilitate the calculation process.
Temperature distribution
In the present paper the potential heat losses from the
Figure 4 Room efficiency and total system efficiency for ventilation system are neglected due to the fact that it
the different design distribution temperatures according to is assumed that the air temperature is distributed at
EN 15136-2-1:2007 (Oslo climate). Radiators with temperatures slightly below the room temperatures.
thermostatic valves mounted on normal external walls and
with heat supply from district heating
ACKNOWLEDGEMENT
Calculated efficiencies for a floor heating system and This work has been supported by SINTEF, NTNU and
the hot tap water system are presented in Table 2. has been related to the project Systemvirkninsgrader
Space heating with floor heating has a slightly lower (System efficiencies) which was initiated by Standard
efficiency than radiators due to the lower efficiency for Norge and paid by the Norwegian Water Resources
the emission of the heat in the room. and Energy Directorate. It has also been supported by
the Primary Energy Efficiency project which is paid by
Tap water systems have a lower efficiency since the Nordic Energy Research and companies in the heating
system is by its nature losing a considerable amount of field in Norway.
heat when leaving the hot water in the pipes between
the tapping cycles. This water is cooled down inside
REFERENCES
the pipelines and is then being tapped without being
useful. In the present case the distance from the [1] European Parliament and Council on energy
substation to the furthermost apartment forces an efficiency of buildings, Directive 2002/91/EC on
installation of a circulation loop in order to reduce the the energy performance of buildings (EPBD)
waiting time for hot tap water at the tapping cocks. This
[2] Proposal on a recast of Directive 2002/91/EC on
is a stand by source of loss. These two factors cause
the energy performance of buildings, 2009-11-25
the rather low efficiency of the hot tap water system.
242
th th
[3] PEDERSEN, L. (2007) Load Modelling of [6] EN 15316 Heating systems in buildings - Method
Buildings in Mixed Energy Distribution Systems, for calculation of system energy requirements and
Department of Energy and Process Engineering, system efficiencies Part 2-1: Space heating
NTNU,(Norwegian University of Science and emission systems., 2007
Technology), Trondheim
[7] CEN: Efficiency of domestic electrical storage
[4] EN 15316 Heating systems in buildings Method water-heater German version pr EN 50440,2005
for calculation of system energy requirements and
system efficiencies Part 4-5: Space heating [8] VVS-tekniske klimadata for Norge, Norges
generation systems, the performance and quality of byggforskningsinstitutt, Hndbok 33
district heating and large volumes, 2007
[5] EN 15316 Heating systems in buildings Method
for calculation of system energy requirements and
system efficiencies Part 2-3: Space heating
distribution systems., 2007
243
th th
HEAT LOAD REDUCTIONS AND THEIR EFFECT ON ENERGY CONSUMPTION

1 2
Christian Johansson and Fredrik Wernstedt
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 balance. This definition is based on the fact that the

heat load reduction will continue to exert an influence
In this paper we investigate the consequences of using
on the buildings thermal buffer for some time even after
temporary heat load reductions on consumer
the heat load reduction in itself is ended. The length of
substations, from the perspective of the individual
this interval is specific to each building and is related to
consumer as well as the district heating company. The
the thermal inertia of the building in question.
reason for using such reductions are normally to save
energy at the consumer side, but the ability to control In this paper we study the consequences of using
the heat load also lie at the core of more complex different types of heat load reductions, and try to
control processes such as Demand Side Management analyse the way the thermal buffer of the building is
(DMS) and Load Control (LC) within district heating affected along with the actual heat load and energy
systems. The purpose of this paper is to study the way usage from both a local and a global perspective. We
different types of heat load reductions impact on the study the performance of both long low-intensity heat
energy usage as well as on the indoor climate in the load reductions (e.g. night time set-back) as well as
individual buildings. We have performed a series of short high-intensity reductions (e.g. those frequently
experiments in which we have equipped multi- used in DMS schemes). The use of night time set-back
apartment buildings with wireless indoor temperature has received some attention in previous works, e.g [1],
sensors and a novel type of load control equipment, and the possibilities to use the building as a heat buffer
which gives us the ability to perform remotely has been evaluated [11], but heat load reductions such
supervised and coordinated heat load reductions as those used in DSM and LC have to the knowledge
among these buildings. The results show that a of the authors not been thoroughly investigated.
substantial lowering of the heat load and energy usage
during periods of reductions is possible without Night Time Set-back
jeopardizing the indoor climate, although we show that Night time set-back means to lower the wanted indoor
there are differences in the implications when temperature during night time, with the purpose of
considering different types of heat load reductions. saving energy through reduced heat losses due to
decreased difference between indoor and outdoor
INTRODUCTION temperature. This is the most common way to perform
temporary heat load reductions, and many commercial
The main purpose of this paper is to investigate the
control systems support this feature. This is normally
consequences of using temporary heat load reductions
done by a parallel displacement of the heat control
on consumer substations within a district heating
curve during night hours. During night time set-back the
network. The most common way to perform temporary
wanted indoor temperature will be set to one, or a few,
heat load reductions is to use night time set-back, i.e.
degrees lower than during normal operations. There is,
to lower the wanted indoor temperature during night
however, an ongoing debate on whether night time set-
time while social activity is expected to be low.
back actually gives an energy saving or not [4], and
Emerging technologies like Demand Side Management
most practical implementations of night time set-back
(DMS) and Load Control (LC) also use temporary heat
suffer from morning peak loads when the control
load reductions in order to accomplish system wide
system returns to the original operational level. Still,
control strategies, although the characteristic of these
almost all control equipment companies sell equipment
head load reductions differ significantly from night time
that facilitates the use of night time set-back, and the
set-back.
use of this technique is widespread.
In the context of this study we regard a heat load
reduction to be the whole process from the initial Demand Side Management and Load Control
change of heat load, through the return to normal heat While night time set-backs are a solely local energy
load, and until no evidence of the heat load reduction saving technique, DMS and LC are usually performed
can be noticed in the dynamics of the building energy with a system wide perspective in mind. A building
244
th th
owner is normally only interested in lowering the Most of the previous work done on the subject is based
energy consumption, while the district heating company on simulated results. This is expected since the
is more interested in being able to optimize the whole dynamic thermal processes within a building are
production and distribution process. Optimizing the extremely complex and it is not surprising that
production normally translates to avoiding expensive comparisons between measurements and calculations
and, more often than not, environmentally unsound sometimes show large discrepancies. It is noted that
peak load boilers or trying to move heat load demand most calculations are dependent on variables that
in time in order to maximize utility during combined cannot be measured and verified, and that the building
heat and power generation. Basically, from the time constant is really not a constant [6].
perspective of the district heating company it is a
question of finding a balance between lowering EXPERIMENTAL METHOD
expensive heat load demand while still selling as much
energy as possible. Implementing this on a system In order to study the effects of temporary heat load
wide scale requires complex coordination control reductions we equipped a building with several wireless
strategies that dynamically adapt to the state of the temperature sensors in order to measure the
district heating system [2]. On the local building level fluctuations in indoor temperature. The building in
this is implemented by performing temporary heat load questions is an office building with semi-light thermal
reductions. On a local level these reductions are characteristics (light construct with concrete slab) and a
normally very short, i.e. one or a few hours, but they time constant of about 150 hours [7]. The indoor
can be of high intensity, even sometimes completely temperature sensors were placed on different locations
shutting of the heat load during shorter periods of time. within the building in order to get a good overview of
This behaviour requires the control system to be highly the thermal behaviour of the indoor climate. In addition
adaptive in relation to the dynamics of the buildings to the existing outdoor temperature sensor an extra
thermal inertia in order to avoid jeopardizing the indoor wireless sensor was also placed on the outside of the
climate. By coordinating such local heat load building. Unlike the existing outdoor temperature
reductions among a large group of buildings it is sensor the wireless one was placed in a position were
possible to achieve system wide DMS and LC. it was fully exposed to any possible sunshine. This
gave us an extra indication of the impact of free heating
Previous work through window areas, even though we did not have
any ability to measure the actual solar irradiance.
Most previous work regarding temporary heat load
reductions deals with night time set-back. This is a In order to control the district heating consumer station
technique that has been around for a long time, and is we connected a load control platform for system wide
based on the general idea that if you decrease the LC and DSM [8]. This platform is based on a novel
difference between the outdoor and indoor temperature form of hardware and software which enables us to
in a building you will save energy. One of the first manage the heat load of the substation without any
large-scale evaluations of night time set-back was major alterations or any damage on the existing
performed in 1983 when buildings in Sweden, USA, hardware. The software system is based on the open
Belgium and Denmark were evaluated. This source Linux operating system and is equipped with an
experiment concluded that night time set-back did not application programming interface (API) for I/O. This
save as much energy as was expected, at most a few makes it easy to apply additional sensors, e.g. for
percent for multi-apartment buildings [3]. In hindsight it measuring the forward and return temperatures of the
is possible to see that these meagre results were a radiator system. The platform also features
consequence of several interacting factors. First of all connections to a database system which enables real-
the control systems of the time were not capable of time logging and analyse of sensor data. The actual
properly handling the transition from night time set- heat load reductions are implemented by supplying the
back to the original operation mode, which causes a existing control system with adjusted outdoor
considerable over-compensation of heat load when the temperatures, which gives us the ability to manage the
systems tries to find the new control level. This extra behaviour of the heat load without exchanging any
boost in heat load during the mornings counteracts existing hardware. This adjusted outdoor temperature
large portions of the energy saving done during the can be managed with a resolution of at most 60
night. The theoretical part of the experiment also had a seconds. The computer platform uses either Ethernet
few draw-backs, e.g. assuming optimally adjusted or GPRS modems to communicate with the database.
radiator systems and linear relations between indoor In our case we used the existing Internet access in the
temperature and energy savings. Other articles show building. In addition to this primary experimental
that there is indeed a substantial level of energy saving building we also collected and analysed data from
to be found by controlling the local heat load [5]. previously installed buildings using the same basic
computer platform.
245
th th
Energy and heat load usage was primarily evaluated by system performs a controlled heat load recovery in
studying the dynamic differences between the forward order to avoid unwanted heat load peaks after the
and return temperature of the radiator system in reduction.
relation to the flow. These readings were then verified
The same values are shown for a long heat load
by specifications from the district heating provider
reduction in Figure 2. The heat load reduction starts
regarding energy consumption and momentary heat
slightly before the 600 minute mark and continues for
load usage.
several hours until about the 900 minute mark. After
Using this set-up we scheduled different types of that the control system performs a controlled recovery
temporary heat load reductions and studied their in order to return to the original operational state.
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
Figure 2: dT in radiator circuit with long heat load
reduction except in those cases when then purpose reduction
was to explicitly study the interaction between
subsequent heat load reductions.
Figure 3 shows the same values for a series of
EXPERIMENTAL METHOD recurring heat loads.
Figure 1 shows the temperature difference between the

forward and return temperature in the radiator circuit
during a short heat load reduction.
Figure 3: dT in radiator circuit with recurring heat load

reduction
Figure 1: dT in radiator circuit with short heat load

reduction
Each of the heat load reductions in Figure 3 is one hour
The heat load reduction starts at about 60 minutes and long intersected by one hour long recovery periods.
continues until the 120 minute mark. Between the 120 The first reduction starts at the 60 minute mark and
minute mark and about the 160 mark the control continues until the 120 minute mark.
246
th th
Figure 4 shows the energy consumption in relation to Figure 6 shows recurring heat load reductions instead
the outdoor temperature during week long periods with of single long ones. It is clear that the building is able to
and without heat load reductions implemented as LC. respond to the control scheme in this example also.
The squares are from periods without LC and the The largest heat load reduction during the recurring
triangles are from periods with LC. LC in this regard scheme is about 25%.
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 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
Figure 4: Energy usage in relation to outdoor temperature. deviation during heat load reduction is about 0.29 while
The squares are values during periods without LC, and the average deviation during periods without reductions
triangles show periods with LC is about 0.19.
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 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

Figure 5: Heat load showing 24 hours without reductions
temperature sensors during a time period of two days.
(black), 24 hours with reductions (dark grey) and control
scheme for reductions (light grey) The graph shows the outdoor temperature sensor
which is connected to the actual consumer sub-station
Figure 5 clearly shows that the reduction in heat load
in the building (black line). Normally these sensors are
closely follows the control scheme. The largest heat
placed somewhat in the shadow to avoid large
load reduction is about 30% in this example.
fluctuations due to solar radiation. We added another
temperature sensor (grey line) in order to estimate the
247
th th
impact of this solar radiation. Hence this sensor was will lower the need of additional heating from the
placed in full view of the sun. The first day was sunny radiator system, by coordinating the thermal inertia of
during most of the morning until midday, while the the building with freely available heat, e.g. heat from
second day was cloudier. 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
Figure 8: Outdoor temperature sensors placed in the
shade (black line) and in full view of the sun (grey line) 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
DISCUSSION
from reduction to normal operation. The results
When dealing with temporary heat load reductions it is showing energy saving is evaluated in relation to the
important to include the whole process of the reduction. total energy usage which also includes tap-water
This also includes what happens after the actual heat usage. Normally this is estimated to about 30% of the
load reduction has been performed. For example, when total energy use in a multi-apartment building.
just restoring the wanted control level after a long
In prior studies of temporary heat load reductions the
reduction, e.g night time set-back, the forward flow
focus has been on the fluctuations in the indoor
temperature in the radiator system will rise much faster
temperature as a way of evaluating the energy saving
than the return flow temperature. This causes a
[3]. This idea is based on the widespread notion that
substantial, although temporary, heat load increase in
any energy saving is linearly proportional to the
the radiator system which negates large portions of the
temperature difference between the indoor and outdoor
energy saving done during the actual reduction. Apart
temperature. This model might be true in a steady state
from decreasing the local net energy saving this
simulation where the temperature difference is
behaviour is also less than desired from a system wide
assumed to have had time to permeate the air mass as
perspective, since it causes massive heat load peaks if
well as the entire building structure, but it is obviously
done in many buildings simultaneously, e.g.
inadequate in a dynamic situation. We have instead
contributing to morning peak loads. In order to avoid
focused on the heat load and energy usage directly, i.e.
this it is important to factor in the whole process of the
the difference between forward and return temperature
reduction, and make sure that the control system
in relation to the flow within the radiator circuit. In most
properly handles the transition from the reduction level
of the buildings evaluated there has been a
to the original level. The inability among most
considerable reduction of energy consumption without
commercially available control systems to properly
any noticeable change in indoor temperature. The
handle this over-compensation is most likely
reason that there does not need to be a measurable
contributing a great deal to the lingering controversy
change of the indoor temperature is due to the
whether night time set-back actually gives an energy
dynamics of the thermal inertia of the building, e.g. the
saving or not.
time constant of a building is not a constant [6]. This
It is important to realize that the definition of an aspect comes into play when using very short heat load
acceptable indoor temperature is not about having the reductions, at most one or a few hours long. During this
indoor temperature at a certain precise level at all time, first part of the reduction it is mainly the actual air mass
but rather to have it within a certain, socially that is influencing the indoor temperature drop since
acceptable, temperature interval at all time. This has this body has a low resistance to change, i.e. the short
been discussed at great length in previous work [6]. time constant [10]. If the heat load reduction is
The general idea is that a greater temperature interval prolonged, like during a night time set-back, the
248
th th
building mass will start to interact with the air mass and REFERENCES
thus stabilizing the continuing temperature drop, i.e. the
long time constant [10]. [1] N. Bjrsell, Control strategies for heating systems,
University-College of Gvle-Sandviken.
The influence of external and internal free heat is large
enough that when these heat sources interact with [2] F. Wernstedt, P. Davdisson and C. Johansson,
other parts of the thermal process it hides shorter heat Demand Side Management in District Heating
load reductions in the ambient temperature. This can Systems, in Proc. Of Sixth International
be seen in Figure 7 where it is shown that although the Conference on Autonomous Agents and Multiagent
average indoor temperature is not noticeably affected Systems, Honolulu, Hawaii, USA, 2007.
there is still a somewhat larger deviation in the indoor [3] L. Jensen. Nattsnkning av temperatur I flerbo-
temperature which implies that there is indeed a higher stadshus, R64:1983, Byggforskningsrdet, 1983
level of temperature flux within the air mass and that (In Swedish).
this is triggered by the heat load reductions. The
control policies used during this work obviously set a [4] H. Lindkvist and H. Wallentun. Utvrdering av nio
high bar for the control system to handle, but as the fjrrvrmecentraler i Slagsta Report ZW 04/05,
average hardware develops it should be possible to ZW Energiteknik, 2004 (In Swedish)
implement such techniques on a larger scale. [5] F.B. Morris, J.E. Braun and S.J. Treado Experi-
Figure 8 gives another clear indication of just how mental and simulated performance of optimal
substantial such sources of free energy can be. This control of building thermal storage, ASHRAE
extra heating due to solar radiation through the Transactions, Vol. 100, No. 1, 1994
windows directly interacts with the mass of air inside [6] E. Isflt and G. Brms. Effekt- och energi-
the building, thus raising the temperature. besparing genom frenklad styrning och drift av
In addition to being able to help save energy usage in a installationssystem I byggnader, ISRN KTH/IT/M--
building temporary heat load reductions also form the 22--E. Institutionen fr Installationsteknik. Kungliga
backbone of DSM and LC, in which the goal is to Tekniska Hgskolan, 1992. (In Swedish)
manage the heat load (kW) rather than the energy [7] S. Ruud. Energimyndighetens program fr
usage (kWh). passivhus och lgenergihus Remissversion 2009-
03-10. Forum fr Energieffektiva byggnader, 2009.
FUTURE WORK (In Swedish)
In the future we plan to further develop models in order [8] F. Wernstedt and C. Johansson. Demonstrations-
to dynamically estimate the temperature flux within projekt inom effekt och laststyrning. ISBN 978-91-
buildings and develop theoretical and practical 7381-041-8, The Swedish District Heating
interfaces for incorporating this data dynamically into Association, 2009. (In Swedish)
the control systems. [9] J. Skoog, PM avseende komfort, F-Infrastruktur
AB, 2005. (In Swedish)
ACKNOWLEDGEMENT
[10] C. Norberg. Direktverkande elradiatorers reglering
This work has been financed by Blekinge Institute of och konstruktion Vattenfall Utveckling AB, Rapport
Technology and NODA Intelligent Systems AB. 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.
249
th th
VERIFICATION OF HEAT LOSS MEASUREMENTS

1 2
J.T. van Wijnkoop , E. van der Ven
1 2
Liandon B.V, Thermaflex International Holding B.V.
ABSTRACT conditions, as defined by the European standard [1].

The ability to conduct equally based heat loss
Heat loss tests are performed on different samples of
measurement result in an objective comparison of
the Thermaflex Flexalen 600 series and one
different types of (semi) flexible piping systems,
ST-PUR-PE sample at the Thermaflex heatloss
providing the opportunity to highlight strengths and
equipment and two German test facilities. At these
weaknesses of (competitive) piping systems.
facilities two different testing methods are used. These
Furthermore, in contradiction to most heat loss tests,
methods are both described in the European standard
the test time in the Thermaflex test-rig is only a few
[1] but show significant differences in the results. In this
hours so the test can be performed during production.
paper the different methods of testing are described.
This provides the opportunity to optimize the production
Furthermore the Thermaflex heat loss equipment is
process real-time and measure the heat loss of the
verified with the test institute that uses the same testing
product several times during a production run. This
method.
guarantees the quality of the produced batch.
INTRODUCTION In addition the handling of the equipment is made easy,

so no specially trained staff is needed for testing,
Last year Liandon developed a test-rig for Thermaflex making it possibly for operators to carry out the tests.
to measure heat loss of insulated plastic piping
systems. With this test-rig it is possible for Thermaflex
EUROPEAN STANDARD METHOD DESCRIPTION
to test the in house produced pre-insulated, semi
flexible pipes in various diameters. The European standard EN 15632 [1] allows two
different methods of heat loss or thermal conduction
To verify the test results, the results of the Thermaflex
testing. These methods both state the same on internal
heat loss equipment are compared with the test results
heating of the service pipe but vary on the method of
of two acknowledged institutions. For this paper two
compensation for heat loss in axial direction.
German institutions are chosen, since they both
measure in compliance with the European standard EN The first method, the guarded end method, states no
15632 [1], however with different methods described in axial heat transfer is permitted. This should be
this paper. In order to give an appropriate comparison, accomplished by the use of end guards, an extra pair
knowledge of the testing methods of both systems is of heating elements at both ends of the service pipe as
required. In this paper the testing methods of all three shown in Fig 1. By heating the ends separately to the
systems is covered, together with the comparison of same temperature as the middle test section no heat
the test-results. Since the testing facilities use two transfer will take place to the ends of the service pipe.
different methods described in the standard, the In this case a theoretical compensation is not required
comparison refers to the test methods and the test since the test section only has losses in radial direction.
results. This method is used in the Thermaflex heat loss
equipment and at one of the institutes.
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.
Fig 1, Longitudinal section guarded end heating probe

NOVELTY AND MAIN CONTRIBUTION
The Thermaflex test-rig is newly developed for the The second method described for compensating for
research of heat loss of pre-insulated pipes. The axial heat loss is the calibrated or calculated end
novelty of this system is its ability to measure the method. The calibrated end method will not be covered
overall heat loss of different samples under similar in this paper since it is not used in our comparisons.
250
th th
The calculated end method states the ends of the X, L: distance to next measuring point from the middle,
service pipe shall be insulated with a known thermal sample length
conductivity as shown in Fig 2.
T0m,T0X,T0m,T2: 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
Fig 2, Configuration calculated end cap. overall heat loss. These samples consist of 2 or 3 m of
the pre-insulated piping system. More information
The service pipe is heated, using a heating element about the Flexalen 600 system can be found in Heat
with only one section. During the tests a thermal profile loss of flexible plastic pipe systems analysis and
is made of the outer casing of the sample, showing optimization (E. van der Ven et Al.) [4]. Furthermore,
lower values at the ends. After testing the heat loss is method comparison tests are performed on competitive
compensated for the end loss with the van Rinsum or pre-insulated piping systems, a comparison of the
Nukiyama theory. For this investigation only the van products themselves is given in Performance of pre
Rinsum theory is used and therefore described. insulated pipes (I. Smits et Al.) [6].
According to the van Rinsum theory, the axial heat loss
causes a decrease in temperature not only towards the The tests on the Flexalen 600 products are performed
ends of the service pipe, but in the test section as well. by Thermaflex and by one of the acknowledged
With the use of the equations (1), (2), (3) this institutes, using the different methods. To ensure the
temperature decrease in the test section can be effect of ageing in the Flexalen 600 system is the same
calculated and added to the measured value, during all tests, the Flexalen 600 samples are tested
compensating the end loss. This corrected temperature simultaneously. To exclude effects of the production
process both tested samples are half of a 6 meter stick.
is used in equation (4) to calculate the overall thermal
conductivity. This method is used by one of the An alternative method is used for Flexalen 50A25 and
German institutes. competitive products. Here the same sample is tested
at the different test facilities.
D2
ln The comparison of the results is based on the outcome
D0
calc 2 L T of heat loss per meter, calculated as described in the
0m T2 (1) European standard [1]. This loss per meter is only
conclusive on a very small part of the entire system.
2 calc
c Therefore the complete Flexalen 600 system will be
D covered in paper Heat loss system optimisation (J.
A1 1 A2 2 ln D2 Korsman et Al.) [3] and New economical connection
0 (2)
solutions for flexible piping systems (C. Engel et Al.)
T0m T0X [5].
T 0m
cosh X c (3) In this report the following diameters of the Flexalen
600 piping systems are used for comparison of the
D2 measurements:
ln Flexalen 600:
D0
50A25, two guarded end tests* and calculated end

2 L T0m T 0m T2 (4) test.
160A90, one guarded end test* and calculated end
calc: approximate value of thermal conductivity test.
200A110, one guarded end test* and calculated
D2/D0: outer/inner diameters of casing and service pipe end test.
A1, A2: areas of the heating probe, inner service pipe Competitive products:
Sample 1 two guarded end tests
1, 2, : thermal conductivity of heating probe, thermal Sample 2 two guarded end tests*
conductivity of medium in the service pipe, thermal *At the time of writing the second test results were not
conductivity total sample. yet available.
251
th th
THERMAFLEX HEAT LOSS EQUIPMENT 600 section. With this method it is possible to measure the
heat loss by measuring the power needed to maintain a
The Thermaflex heat loss equipment is specially
constant temperature of the test sample. In
designed for the Thermaflex Flexalen 600 series. One
contradiction to the measurements at the test institutes,
of the major design goals was to develop a fast and
the Thermaflex heating probes temperature is
easy to use test rig with the precision of a laboratorial
regulated by PID controlled power supplies. In the test
test. These goals have resulted in a test rig that is able
results Graph 2 the power consumption versus test
to measure heat loss in a few hours, allowing direct
time is shown. This variable power supply makes it
optimization during the production process, and is
possible to pre-heat the probes in a short period of
operable by the production staff without the loss of
time, shortening waiting times considerably.
accuracy
Furthermore the use of the actual pipe material as a
heating probe increases the accuracy. Moreover it
Physical test facility
eliminates all additional heat loss by convection that
The physical part of the Thermaflex heat loss will be present with the use of smaller, not inner service
equipment consists of three segments. pipe connecting heating probes.
The first is the water cooled compartment in which all For testing competitive products with different
tests are performed. This compartment is kept at a diameters these advantages are lost. However by the
constant temperature, (23 C), during each use of thermal compartments in the service pipe the
measurement. test results can be guaranteed.
The second is a heat source, for which heating probes
are used. These heating probes are custom made by Thermaflex method of testing
equipping a two meter Thermaflex piping segment, of For testing, the heating probe with the appropriate
all available diameters, with three heating coils. diameter is inserted in the insulation covered with outer
casing, and inserted in the cooled test section. After
The third part of the heat loss equipment is the control
connecting the probe to the control unit the
unit. Here the heating probe is powered and all thermal
measurement can be started. Different testing
readings are done. By applying custom made software
conditions can be entered at this point such as the
all desired readings can be done. The final output is the
inner pipe temperature, representing the internal
actual heat loss in W/m through the entire pre-insulated
medium. When the test is started the heating coils heat
Flexalen pipe, consisting of the service pipe, insulation
the inner side of the probe until the desired
and outer casing.
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
Fig 3, Thermaflex heat loss equipment 600. consumption. Since the middle/testing coil is exactly
one meter in length the required energy represents the
Measurement principle Thermaflex exact heat loss through one meter of piping and
The Thermaflex test rig is designed in compliance with insulation in W/m. Since the actual piping material is
the European standard [1] and also the tests are used during the measurement, there are no other
carried out according to ISO 8497 and EN 15632. In losses, nor advantages, than there will be in practise,
the design of the heating probes the most realistic ensuring an objective measurement. Furthermore a
method, the guarded end method, is used. According realistic fit of the insulation material is guaranteed. As
to this method the heating probes are equipped with stated in the foregoing paragraph these advantages
three heating coils with separate power supply. As are lost for divergent diameters. However during this
shown in Figure 2, two 400 mm heating coils located at investigation the probes have proven suitable for
each end of the 1000 mm test section. These two testing, as both testing institutes also use smaller
sections provide a thermal insulation at both ends of heating probes.
the test section since all three are kept under uniform
temperature, eliminating axial heat loss of the test
252
th th
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 Graph 2, Power and temperature of the Thermaflex
insulation, whereas the last 30 minutes is the actual heating probe.
test time. Since, as the figure shows, the temperature
Outcome Competitive products for comparison of
is constant, the power usage equals the heat loss
testing method:
through the piping system in radial direction during the
last 30 minutes. The results, as given in Table 1, are For the comparison with test institute two, two samples
calculated by using the mean of the power of competitive products are tested. As these samples
consumption during the last 30 minutes of the heat loss are ST-PUR-PE system, a correction has been made
test. The results in Table 1, are also displayed in for using the PB heating probe using the Wallentn [2]
Graph 1 for the three tested samples. method. First the thermal conductivity of the insulation
is determent by the use of equation (5), hereafter the
Table 1, Results heat loss equipment for the Flexalen 600 heat loss is recalculated without the heating probe,
products using the temperature of the inner service pipe in
equation (6). The results are presented in Table 2 and
Heat loss of the Flexalen 600 series in W/m tested on
Graph 3.
the Thermaflex heat loss equipment
Product 40 C 50 C 60 C 70 C 80 C d3
ln
d2 W
50A25 3.6 6.4 9.3 12.0 15.0 i

m K (5)
160A90 6.2 10.1 14.0 17.9 21.8 Tp Tc 1 d2 1 6 1 d4
d
2 ln ln ln
probe st d1 p d c d3
200A110 6.5 12.0 17.5 23.0 28.5 5
Heat loss results thermaflex heat loss equipment

2 T st T c W
30 corrected
d2 1 d3 1 d4 m (6)
ln
1
ln ln
st d1 i d2 c d3

Heat loss [W/m]
20
Where:
Tp,Tc, Tst=Probe, Casing and Steel pipe temperature
10
d1 to d6 = inner/outer diameters of service pipe, casing
and heatingprobe
0
40 50 60 70 80 st, i, c, p = heat coefficient of service pipe, insulation,
casing and probe
Temp erature inner service pipe [C]
Flexalen 50A25 probe, corrected =probe power and corrected heat loss.
Flexalen 160A90
Flexalen 200A110
Graph 1, Results heat loss equipment Flexalen 600

products.
253
th th
Table 2, Results heat loss equipment for the competitive made of metal. Furthermore no heat guards are used.
products This means the outer ends of the piping system are
insulated and the heat loss is corrected with a
Heat loss of competitive products in W/m tested on calculated value. In the paragraph European standard
the Thermaflex heat loss equipment
method description a more detailed description is
Sample 40 C 50 C 60 C 70 C 80 C given. As can be seen in Fig 4 the heat distribution in
this case is not uniform along the test specimen,
Sample 1 2.5 4.5 6.4 8.4 10.3 proving the need for the van Rinsum correction.
Sample 2 6.5 8.8 11.2 13.5 15.9
Sample 3
Heat loss results thermaflex heat loss equipment

20
15
Heat loss [W/m]
10
Fig 4, Thermal image of the sample at institute one

5
In contradiction to the Thermaflex test rig, no integrated

0 computer controlled power supply system is used. The
40 50 60 70 80
power for the heating probe is first theoretically
Temp erature inner service pipe [C] calculated and manually set to this value. For the
Competitive samp le 1 temperature measurement thermocouples and a data
Competitive samp le 2 logger with computer link are used.
Graph 3, Results heat loss equipment for the competitive Method of testing
products
The heating probe is positioned in the centre of the test
pipe with positioning foam in three sections of the pipe.
TEST INSTITUTE ONE
On these foam blocks four thermocouples are placed in
This institute is specialized in measuring heat loss in 0, 90, 180 and 270 degrees on the inner surface of the
different types of insulation. The test facility used for service pipe. For the outcome of the pipe inner
the Flexalen 600 system is specially designed for temperature the mean of the four values is used. To
measuring the heat loss of (pre-) insulated piping measure the temperature on the outside casing of the
systems. This means the facility is designed to insulation, five groups of four thermocouples are used
measure all different types and diameters. in the same configuration as the inner pipe. The
difference being that the thermocouples are placed
Measurement principle institute one both on and in between the corrugations of the casing.
The measurements are all based on the calculated end The test sample, with the heating probe, is placed in
apparatus, using the van Rinsum theory as correction, the conditioned container thereafter the test can be
as described in the paragraph European standard [1] started. The power supply of the heater is turned on by
method description of this paper. setting the voltage and current of the power unit to a
fixed value so the electrical power equals the
Physical test facility calculated heat loss.
The physical part of the test facility is similar to the Depending on the diameter of the test sample and the
Thermaflex test rig and also consists out of the three test temperature the waiting time for the heating of the
elements: A temperature controlled compartment sample is five to eight hours due to the low, fixed power
where the tests are carried out at a constant input. After a constant temperature of the outer casing
temperature of 23 C. FIW also uses heating probes as is achieved the actual test cycles start. Each test cycle
a heat source but, since it is not specially designed for consists of a measurement of 30 min in which the outer
the Flexalen 600 system, they are made to fit all casing temperature is to be constant. If not the cycle
systems. To ensure the fit of the probes in all different has to be restarted. In total ten cycles will be performed
systems the diameters are smaller, and for durability on each sample. After the test the values are corrected
254
th th
for axial heat loss and the thermal conductivity, thermal An update to this paper will be made as soon as the
resistance and overall heat loss are calculated. Flexalen 600 results will become available.
Test results institute one Testing method institute two

As the actual measurement data are not available due The method used by this institute is generally the same
to the correction factor, only the calculated values can as the method used by Thermaflex; however the test
be discussed in this paragraph. As soon as the actual facility itself is different.
measurements become available this section will be
updated. Furthermore the results are not given for Physical test facility
exactly 60, 70 and 80 C due to the fixed power supply The testing facility at institute two consists of a
with no temperature set point, the displayed results are temperature controlled room, kept at the prescribed
calculated heat loss values at the set temperatures to 23 C. As a heat source a heating probe, consisting of
make the data more interpretive. For this calculation a 2 m test section and two 50 cm end guards is used.
the linearization method described in the European At the time of writing no further information on the test
standard [1] is used. In Graph 4 the data from Table 3 facility was available. This paragraph will be updated
is presented as a graph. when this information becomes available.
Table 3, Results test institute one for the Flexalen 600 Method of testing
products
Prior to testing, the sample is prepared by placing
Heat loss of the Flexalen 600 series in W/m tested at thermocouples in various locations on the inner service
test institute one pipe and outer casing. Subsequently the sample is
Product 40 C 50 C 60 C 70 C 80 C placed in the temperature controlled room and the
heating probe is inserted. By setting the power supply
50A25 5.6 8,8 11.9 15.1 18.3
to a calculated value for all three heating coils the
160A90 9.1 15.1 21.1 27.1 33.0 heating process of the sample is started. Because of
200A110 9.8 15.1 20.5 25.8 31.2 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
Heat loss results test institute one the guarded ends, the actual test is performed. The test
40 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
30
bandwidth.
Heat loss [W/m]
20 Test results test institute two

The test results of institute two are given in Table 4 and
10
Graph 5. As not all data was available during writing
there tables and graphs will be updated.
0
40 50 60 70 80 Table 4, Results test institute two for the competitive
products
Temp erature inner service pip e [C]
Flexalen 50A25
Heat loss of competitive products in W/m
Flexalen 160A90 tested at test institute two
Flexalen 200A110
Graph 4, Results test institute one for the Flexalen 600 Sample 40 C 50 C 60 C 70 C 80 C
products
TEST INSTITUTE TWO 90DN25 1.95 4.00 6.02 8.06 10.09
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.
255
th th
Heat loss results test institute tw o competitive products show consistency with test
15 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
Heat loss [W/m]
10
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.
5
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
0
40 50 60 70 80
heat loss equipment and will be updated as more data
comes available.
Heat loss c omparison Thermaflex and test institute two
Competitive sample 1 15
Competitive sample 2
Graph 5, Results test institute two for the competitive
Heat loss [W/m]

products 10
COMPARISON OF THE TEST RESULTS 5
Comparison of the Thermaflex flexalen 600 series:

Although both methods, guarded end and calculated 0
end, are approved and described in the European 40 50 60 70 80
standard [1], the difference between the results is Temp erature inner service pip e [C]
substantial as displayed in Graph 6. Moreover all Competitive samp le 1
Competitive samp le 1
results vary more as the temperature difference Competitive samp le 2
increases. This can be explained by the use of the Competitive samp le 2
calculated end caps that conduct more energy at
Graph 7, Comparison results of the heat loss equipment
higher temperature differences. As these end cap
and test institute two
losses increase, the corrected thermal conduction for
the sample also increases, resulting in a higher
CONCLUSION
calculated heat-loss.
During this research it has become clear that the
Heat loss c omparison Thermaflex and test institute one
35
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
30
25 guarded end cap method varies from the result of the

Heat loss [W/m]
20 calculated end cap method, however no assumptions

15
can be made based on only one comparing
measurement. Further study that is being conducted at
10
this moment will provide more comparison data. This
5 will be updated with this data as soon as becomes
0 available. This new data could point out that the van
40 50 60 70 80
Rinsum theory is not suitable for accurate heat loss
measurement of plastic piping systems.
Flexalen 50A25 Thermaflex result
Flexalen 50A25 Institute one result The comparison of the guarded end method results
Flexalen 160A90 Thermaflex result
Flexalen 160A90 Institute one result
from test institute two and the Thermaflex heat loss
Flexalen 160A90 Thermaflex result equipment conclude that the results of the heat loss
Flexalen 160A90 Institute one result equipment are correct and comply with the European
Graph 6, Comparison results of the heat loss equipment standard [1]. This validation makes the results of the
and test institute one Thermaflex heat loss equipment valid for not only in
house testing but also for publication as done in Heat
Comparison of competitive products: loss of flexible plastic pipe systems analysis and
Although the Thermaflex heat loss equipment was optimization (E. van der Ven et Al.) [4] and Perfor-
designed for Flexalen series, test results on mance of pre insulated pipes (I. Smits et Al.) [6].
256
th th
FURTHER INFORMATION for their devotion on all the heat loss measurements
they performed during this research in a short amount
Questions concerning the paper can be addressed to:
of time.
Veerweg 1 REFERENCES
5145NS Waalwijk
The Netherlands [1] NEN-EN 15632 and NEN-EN-ISO 8497
Liandon B.V. [2] P. Wallentn, steady-state heat loss from
Dijkgraaf 4 insulated pipes, Lund Institute of Technology,
6920AB Duiven Sweden, 1991
The Netherlands
[3] J. Korsman and G. Baars, Heat loss system
ACKNOWLEDGEMENT optimization, 12th ISDHC 2010
Acknowledgments go to both the test institutes for their [4] E. van der Ven and R. van Arendonk, Heat loss
open and honest explanation of their testing methods analysis and optimization, 12th ISDHC 2010
and facilities and for even showing the entire facility
[5] C. Engel and G. Baars, New economical
and methods.
connection solution for flexible piping systems,
Furthermore acknowledgements go to all involved 12th ISDHC 2010.
employees of Thermaflex Isolatie B.V. and Liandon
[6] I. Smits and E van der Ven, Performance of pre
B.V. who made this research possible. Special
insulated pipes, 12th ISDHC 2010.
acknowledgements go to P. Blom and P. van Rijswijk
257
th th
DISTRICT HEATING AND COOLING WITH LARGE CENTRIFUGAL CHILLER-HEAT

PUMPS
Ulrich Pietrucha
Friotherm AG, Switzerland
ABSTRACT Number of units 2

With prices for primary energy resources soaring, the Type UNITOP 50 FY/34FY
recovery of "waste energy" was getting into the focus of
attention within the last years. Also the global climate Refrigerant R134a
change reminded us to limit the use of primary energy Heat source medium Raw waste water
resources to a minimum, thus exploiting "waste energy"
potentials wherever feasible. The process of upgrading Raw sewage water inlet 10.0 C ... ~ 15 C
low grade waste heat is especially interesting where Raw sewage flow water flow 3800 m3/h
large amounts of such energy are available at one
point, e.g. next to sewage water treatment plants, Heating water temp. in/out 60 / 90 C
alongside main sewers, in power plants or close to Heating water flow 824 m3/h
ground water sources.
Power at terminal 9750 kW
Even if the "waste energy" potential is abundant and
easily exploitable, the aspect of overall thermal Heat capacity 27600 kW
efficiency is considered crucial for the final decision to
Coefficient of performance 2.83 up to >3.0
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 One of the Skoyen heat pumps
raw sewage water heat pump/chiller installation. This
plant is operated successful since 1989.
2. THE SANDVIKA PLANT IN OSLO: COMBINED
Application 3: Combined heating and cooling: a
HEATING AND COOLING FROM A RAW SEWAGE
combination of cooling with simultaneously heat
WATER HEAT PUMP
production in summer and heat recovery from cleaned
sewage water in winter. This is the oldest combined chiller/ heat pump
installation in the world, producing simultaneously
Application 4: Heat recovery from wet flue gas
cooling, taking out heat from raw sewage water and
cleaning process
producing heating capacity for the district heating
Application 5: Combined heating and cooling in system.
Stockholm
The heat pumps are in successful operation since 1989
and each one has an additional heat exchanger, which
1. SKOYEN VEST PLANT IN OSLO: HOT WATER is used either as raw sewage evaporator or as raw
PRODUCTION AT 90 C sewage water condenser.
This is the world's largest heat pump plant using raw Each heat pump has an overall operating time of about
waste water as heat source. It is installed in a cavern 160'000 hours, means the heat pumps were operated
alongside one of the main waste water channels in since 1989 each year for more than 8'400 hours.
Oslo. With 2 heat pumps a heating capacity of 27'600
A 3rd larger heat pump was taken into operation in
kW is generated by recovering heat from raw waste
2008.
water.
258
th th
points with lower district heating temperatures where a

COP of up to 6.5 can be achieved.
2 Friotherm heat pumps From district-

Type UNITOP 28CX-71210U heating-network
50C
Heat source capacity Heat sink capacity
24.3C 15.5MW 19MW
Flue-gas
Flue- gas- Power consumption

condensing 3.5MW
Combined chiller / heat pump at Sandvika plant
Flue-gas- 34.2C
cleaning 59.2C
3. THE KATRI VALA PLANT IN HELSINKI:
COMBINED HEATING AND COOLING Steam
turbin
- Generator
e
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 Boiler
achieved (150MW / 30MW). Waste-to-Energy plant SYSAV Malm Sweden
During Winter season the required cooling is done by

sea water, while heat is produced by using cleaned 5. NIMROD STOCKHOLM: COMBINED HEATING
waste water as heat source. 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
Typical Unitop 50FY heat pump (Qheat 15 to 23MW) required as there is sufficient capacity available from
the existing heat pumps.
The same units are producing during Spring, Autumn
4. SYSAV MALM: HEAT RECOVERY FROM WET
and Winter a cooling capacity 24MW with a full heat
FLUE GAS CLEANING PROCESS
recovery of 35.6MW at a temperature level of 78 C.
SYSAV Malm in Sweden has built a new waste-to- For heat recovery operation mode the centrifugal
energy plant. An important part in this plant was the compressors are switched in series.
installation of a 19MW heat pump using the flue gas
Each chiller / heat pump consists of two centrifugal
condensation as heat source. The heat pump is
compressors Type Uniturbo 33CX and 28CX and is
supplying hot water with a temperature of up to 70 C
able to operate at the following modes, described
to the district heating system of the community of
below:
Malm.
The two heat pumps are connected in series on the 5.1 Cooling only:
heat source side and on the heat sink side; this During Summer with high cooling demand, the waste
improves considerably the COP. There are operating heat from the condenser is removed with sea water of
259
th th
max. 22 C, therefore the condenser and sub cooler Nimrod single stage operation
are equipped with Titanium tubes.
2 compressors in parallel
The two compressors Uniturbo 33CX and 28CX are
then working in parallel, in a single stage mode, with a
Sea water
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 2
8
both compressors can be put in operation. 8
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 Cooling capacity 12 MW
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
Nimrod two stage operation
capacity of 35.6MW. 2 compressors in series
The two compressors Uniturbo 33CX and 28CX are
Heating water
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 28
volume flow is working as 1st stage and the Type
Uniturbo 28CX with the smaller volume flow as 2nd
stage compressor. 33
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: Heating capacity 9 MW
Cooling capacity 6 MW
CONCLUSION
Today developments are the extension of the heat
Reliability of technology, future developments and
pump operation range in temperature and capacity to
challenges
exploit new heat sources and to extent the field of
About more than 140 heat pumps, producing hot water applications.
with temperatures above 70 C, are installed world
wide since 1980. The heat pump plants described in The adaptation of the centrifugal heat pumps to new
this article are only showing a small part of the refrigerants with GWP close to zero is already on the
nowadays available applications. way.
Almost all of the installed heat pumps plants starting

REFERENCES
from the early 1980's are today still in operation, which
is showing the high reliability of this technology. Text and pictures from Friotherm AG / Switzerland.
260
th th
NEW ECONOMICAL CONNECTION SOLUTION FOR FLEXIBLE PIPING SYSTEMS

Christian Engel, Gerrit-Jan Baars
ABSTRACT Together with the University of Leoben, the long term

durability of two types of plastic medium pipes, made of
Most Energy Companies are facing the same problem:
PB and PE-X were investigated. The research made by
Connection costs per house shall be cheaper and Dipl.Ing. E.Kramer and Univ.Prof.Dr.J. Koppelmann [2]
faster to install to reach more customers. At the same was based on OIT (oxygen induction time), tear
time high level of durability and a system free of strength, elongation at break and internal pressure
maintenance must be guaranteed. tests to determine the lifetime of plastic pipes at 80, 95
and 110 C. The final results were in favour for pipes
This paper shall give an insight of the practical
made of PB. The calculation of the lifetime for PB pipes
experience with new solutions showing the economic
was based on a typical temperature profile used in
and ecological advantages in projects with several
secondary district heating networks of STEWEAG. The
Energy companies.
lifetime expectancy was stated with 36 years.
INTRODUCTION The decision of STEWEAG was made for PB pipes due

to their more homogenous structure, superior flexibility
District Heating & Cooling networks are a major cost and allowance for welded joints. More than 250 km of
factor for Energy Providers and subject to permanent this system have been installed since 1981 in
search for cost improvements. secondary networks operated by STEWEAG. See also
Flexible plastic pipe systems have been a major step Univ. Prof. Dr. E. Hnninger, STEWEAG [6].
for cost reduction in low temperature networks. With
the new EN 15632 [1] the necessary basis for WHAT CAN BE SOLVED WITH FLEXIBLE PLASTIC
certification of these systems has been laid. This is a SYSTEMS
milestone in terms of acknowledgement for flexible
Apart from the high and long term investment costs, the
systems as a proven part of future network
following main problems had to be solved as well:
developments.
Corrosion problems in conventional Systems made
As flow temperatures and pressures are reduced, the
of Steel/PUR/PE or Cu/PUR/PE or Cu/Mineral
field of application for flexible plastic systems is
wool/PE
increasing. Until recently only a small percentage of
Heat loss due to wet and aged insulation
District Heating companies have started to use plastic
System shut downs for maintenance and repair
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.: 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
Photo no.:1 first Flexalen installation 1981 costs for new district heating projects. The new system
261
th th
developed is called FLEXALEN 600, the improved up to d63 there trenches can be compared to cable
version of the system used by STEWEAG. The system ducts. Another fact is that connections are most of the
consists of a PB (Polybutene) medium pipe and PO time only necessary at branches, for sticks trenches
(Polyolefin) insulation foam welded to a HDPE (high need to be suitable for execution of the welding and the
density Polyethylene) outer casing. insulation process.
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.: 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).
Photo no.:3 Flexalen 600 longitudinal cut
FLEXALEN has been the first system to pass a BRANCH SOLUTIONS

certification and 3rd party control by KIWA, which is
Until now 2 types of branch solutions have been used.
similar to the new EN 15632.
On site welded solution with Half-shells plus insulation
to cover (see Photo no.:4). This technology has been
DECREASING INSTALLATION COSTS used for smaller networks. Due to homogenous
The most obvious advantage is the chance to reduce welding techniques, either with polyfusion or with
the installation time with flexible systems supplied in electrofuction fittings, the branches are corrosion free
coil lengths of 100m and more. Compared to rigid and offer the same inner diameter as the pipes.
systems the following relative costs have been realized
in actual projects
Table 1
Pre-insulated
FLEXALEN
steel pipes
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
Photo no.: 5 Polyfusion welded Flexalen branch
pipe sizes and lengths according to the advantages of
PB pipes and connection systems, as described later Pre-insulated Tees are another way to secure more
on, the material costs can be reduced for Flexalen. reliable network quality due the reduced number of
Installation costs are proven in practical experiences joints to be made on site (see Photo no.: 6). The
since almost 30 years. Flexalen systems can be straight connections are insulated with special kits with
installed 5 times faster than rigid systems. a robust slide over HDPE tube, which is sealed to the
outer casing with heat shrinks.
Lower costs in trenching is related to the fact that
Flexalen systems are supplied in double line systems
262
th th
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.: 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.
Photo no.: 8 Flexalen T-Link
PRE-FABRICATED NETWORKS
This solution combines the following advantages:
Compared to the branch solutions described before,
Factory made welding and branch insulation all
these new solutions take full advantage of flexible
watertight and pressure tested.
welded systems, in order to further reduce the number
Connections are made under clean manufacturing
of joints on sites.
circumstances. No weather influences, no failure
A new type of pre-fabricated network has been costs.
developed. High flexibility and a minimum of Customer made connections according to the real
connections was the goal. situation.
Light weight and flexible for easy sliding into the
The first application was the district heating network
trench or under the house
Capelle a/d Ijssel in the surroundings of Rotterdam for
Fast Installing time (first project experience 10
renovation in difficult circumstances under houses
houses/day)
(high ground water level) to replace corroded heating
Reducing connection costs in new building
and sanitary distribution systems.
projects.
The space under the houses is so small that neither Less system parts on the building area.
welding nor mechanical connections can be carried out
This development covered all the wishes and
in a safe way. Steel welding is even forbidden under
requirements of the Dutch Energy Provider ENECO.
these conditions.
They have already ordered this system for 800 house
connections in 2010.
263
th th
Photo no.: 9 Installation of Flexalen T-Link
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
Photo no.: 10 Installation of Flexalen T-Link Purmerend) is very critical. Due to higher ground water
levels than expected the current metal pipe systems
have corroded and need to be replaced.
These 2 photos (no 8 and 9) demonstrate how easy it
was to slide the connection into the duct under the As the network has been installed under the basement
house. of the attached housing schemes, the space for the
installer is very tight and it is not allowed to use any
Also further connections between the pre-fabricated steel welding process in these circumstances. Flexalen
sections were made before sliding the entire system T-Link has been identified to be the ideal solution.
under the house. Only the last connection had to be
made in the duct. See also photos no.:11 & 12. 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
Photo no.: 11 pipes under the building connection costs.
Both Dutch Energy Provider, Eneco and Nuon are
investigating this new solution for new building projects.
264
th th
Fig. 1 network scheme suitable for T-Link
Especially for networks with short distances between PB systems can be operated with much higher flow
the branches and the connection to the houses, a high speed; hence smaller dimensions can be used for the
degree of pre-fabrication can be offered. same load requirement. See also J. Korsman, I.M.
Smits, E. van der Ven [4].
One solution is the T-Link as described before. Another
one can be a main line up to 100 m with factory welded With relation to the topic of this paper, the following
and insulated Tees. This reduces the work on site to additional savings can be made during the network
just 1 welding for the house connection line. This design:
solution is interesting for longer distance house
Looking for a new building area, mostly streets with
connections.
block of houses there are two possibilities:
Installing under the floor or Installing in the streets.
THE IMPORTANCE OF THE NETWORK DESIGN For every house connection under the floor only
Flexible PB piping systems offer important advantages two welds and two insulation sets are necessary.
compared to other plastic and steel systems in terms of Reduction of the installing time/costs by 5060%.
layout and design. For every house connection in the street, a
pre-fabrication e.g. for 810 house-connections
Compared to steel pipes flexible PB systems can be built in into one 100m coil can safe installing
laid more direct as the system is flexible and fully self- time/costs totally including excavating the trenches
compensating. Expansion loops and elbows can be of 70%
saved. The saving in pipe length can be calculated with Taking all these possibilities for savings and
710%. optimization into account, the next most important topic
PB systems offer low friction loss and show no for Energy Provider, the efficiency of the network in
calcification or incrustation during the lifetime. The operation, can be tackled as well. Due to the possible
polyfusion welded fittings have at least the same inner reduction in network length and pipe diameter, the
diameter as the pipe and offer the same high abrasion overall heat loss can be reduced as well. See also
resistance. Taking this into account, some extra results from the work of I.M. Smits, J. Korsman, J.T.
security factors used in pipe dimensioning can be van Wijnkoop and E.J.H.M. van der Ven [5] and J.T.
eliminated. van Wijnkoop, E.J.H.M. van der Ven [3].
265
th th
CONCLUSION REFERENCES
New connection solutions, which meet the [1] EN 15632 District heating pipes, Pre-insulated
requirements of Energy Companies in terms of lower flexible pipe systems, Requirements and test
investment costs, faster installation time and durability, methods
have been co-created with leading Energy Suppliers in
Austria and The Netherlands. [2] Dipl. Ing. E. Kramer, Univ.Prof. Dr. J. Koppelmann,
Untersuchung zur Dimensionierung einer flexiblen
These solutions are based on flexible and weld-able Fernwrmeleitung aus Kunststoff, University
plastic systems and have been used successfully up to Leoben, Austria; 1984.
29 years in secondary networks with maximum
operation temperatures of 95 C (peak temperature) [3] J.T. van Wijnkoop, E. van der Ven, Verification of
and maximum pressure of 8 bars. heat loss measurements, 12th ISDHC 2010.
The latest development is going into the direction of a [4] J. Korsman, I.M. Smits, E.J.H.M. van der Ven
higher degree of pre-fabrication, by including the entire Heat loss analysis and optimization of a flexible
connection line to the houses as well as parts of the piping system, 12th ISDHC 2010.
main line into one piece, made up and fully pressure [5] I.M. Smits, J. Korsman, J.T. van Wijnkoop and
tested in the factory. E.J.H.M. van der Ven, Comparison of competitive
The experiences in recent projects are showing (semi)flexible piping systems by means of heat
installation times 510 times faster compared to loss measurement, 12th ISDHC 2010.
conventional pre-insulated steel. The number of [6] Univ. Prof. Dr. E. Hnninger, Sekundrnetze
connections to be made on site is significantly reduced. frdern die Fernwrmeanwendung, STEWEAG,
Successful projects with Energy Suppliers in The Fernwrme International 14/85.
Netherlands are confirming the advantages of this new [7] C. Engel, Polybutene The alternative material for
connection solution. heating and domestic hot & cold water systems,
PLASTIC PIPES IX, Edinburgh 1995.
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.
266
th th
COMPETITIVENESS OF COMBINED HEAT AND POWER PLANT TECHNOLOGIES

IN ESTONIAN CONDITIONS
1 1
E. Latov and A. Siirde
1
Department of Thermal Engineering
Tallinn University of Technology, Tallinn, ESTONIA
ABSTRACT energy consumption effective methods, where CHP

production from the renewable fuels is preferable. [1]
The goal of this paper is to evaluate competitiveness of
market ready combined heat and power (hereafter During the last 2 years a few CHP plants working on
called as CHP) technologies for CHP expansion woodchips and peat were build in Estonia. A few of
potential locations in Estonian. The main criteria to biomass CHP plants are under active development. All
indicate preference of CHP technology is a heat price of them are planed or constructed in major Estonian
by which the internal rate of return is equal to investors cities and are based on backpressure steam turbine
expectations. technology. At the same time feed-in tariffs as well as
possibilities to get grants for expanding of CHP and
Calculation results shows, that in spite of the
usage of renewable fuels makes CHP expansion more
advantages of gas engines (relatively low investment
attractive for locations with a lower heat demands.
costs and high electrical efficiency) the calculated heat
prices are the highest. Heat price for expected 7% IRR Steam turbine technology is a classic for CHP plants.
is 5361 EURO/MWhheat depending on heat demand. But in relatively small-scale boilers and district heating
It is mainly because of relatively high natural gas price. systems use of steam turbines is connected to
Under 5 MWel ORC is competitive to steam economically less efficient operation (commonly higher
turbine/engine technology. Heat prices are lower for specific investment costs, O&M costs and lower
14 EURO/MWhfuel, depending on heat demands. electrical efficiency) where use of other alternative CHP
technologies could be preferable.
Heat prices for places with annual heat demand under
20 000 MWh are mainly above 45 EURO/MWhfuel The goal of this paper is to evaluate competitiveness of
(average heat prices for biomass boiler houses in market ready CHP technologies for CHP expansion
Estonia is between 4045 EURO/MWh). Developing of potential locations in Estonian. The main criteria to
CHP plants in such areas is feasible in the case of indicate preference of CHP technology is a heat price
grant payments for investments. CHP plant by which the internal rate of return (hereafter called as
development based on wood chips or peat could be IRR) is equal to investors expectations.
feasible without grant payments in the places where
The paper is structured as follows. After an overview of
heat demand exceed 30 00040 000 MWh annual.
places where construction of CHP plants can be
Carefully selected CHP technology and capacity can
reasonable the paper provides principles for evaluation
afford higher IRR when keeping competitive heat
of CHP technologies competitiveness. Next sections
prices.
provide an overview of the CHP technologies which
The most feasible places for CHP expansion in Estonia can be used in CHP plants and descriptions of main
are Maardu, Viljandi, Rakvere, Valga, Haapsalu, Vru, fuel sources for energy production in Estonia. The last
Paide and Plva. section provides heat price calculation examples based
on proposed principles for evaluation of CHP
INTRODUCTION technologies competitiveness.
This paper draws on ongoing project Analysis on the
LOCATIONS OF POTENTIAL BIOFUELED CHP
technical and economic consequences of renewable
PLANTS IN ESTONIA
energy based CHP systems in new areas with the
lowered useful heat demand or after implementation of Fig. 1 shows major Estonian cities and municipalities
energy conservation measures in the areas with older which are distributed by the annual heat demands.
buildings within the project Primary Energy Efficiency
Places where CHP plants are already constructed or
partly financed by NER, which contributes to the effort
under construction, as well as in a state of active
of enhancing the primary energy efficiency (PEE) and
development are marked separately.
reducing CO2 emissions in the energy sector.
Fig. 1 reflects well known principles, where the
Present-day world energy policy is based on two main
consumers with higher annual heat consumptions are
directions: energy efficiency and environmental
more preferable.
protection. Efficient CHP production is one of the
267
th th
Fig. 1 Distribution of Estonian cities and municipalities by the annual heat demands
PRINCIPLES FOR EVALUATION OF CHP Income from the heat sells depends on amount of sold
TECHNOLOGIES COMPETITIVENESS heat (computable value) and heat price.
The revenues of a CHP company are generated from Knowing investment costs (specified in section CHP
the heat and electricity sales. Theoretically they must technologies), and other above mentioned costs and
cover the operation and maintenance (hereafter called incomes the power plant operation annual net cash flows
as O&M) costs of the CHP plant completely and can be calculated and IRR defined.
provide an expected IRR. Main CHP plant related costs
The principle for evaluation of CHP technologies
and incomes are shown in Fig. 2.
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),
Fig. 2 CHP plant incomes and costs steam engine (hereafter called as SE), ORC
technology (hereafter called as ORC) and gas engine
Fuel costs, pollution charges and ash handling costs are (hereafter called as GE). Hereafter SE and ST are
mainly depend on used fuel properties and are considered jointly, where capacities less than 1 MWel
estimated in section Fuel sources for energy production. correspond to SE by default.
CHP technology related fixed operation and For CHP plant economical calculations it is important to
maintenance costs depend on selected CHP technology know such CHP plant parameters as efficiencies, price
and are defined in % from the investment costs annual. and O&M costs.
They are estimated in section CHP technologies.
Above mentioned parameters are obtained and
Electricity sells depends on amount of produced systemized on the basis of information regarding CHP
electricity (computable value) and fuel prices. Fuel prices plants collected from different information sources such
are estimated by taken into account feed-in tariffs as [3, 4, 5, 6, 7, 8].
described in Electricity Market Act [2].
268
th th
Fig. 3 Main prime mover CHP technologies
Values for investments as well as nominal electrical heat load is 35% for steam engine/turbine, 80% for gas
capacities used to calculate fuel prices depend on CHP engine and 85% for ORC from the nominal electrical
nominal electrical capacities and CHP technology, as efficiency.
shown in Table 1.
It is assumed, that CHP technology related fixed O&M
Table 1. Values for investments and nominal electrical costs for SE/ST, ORC and GE are relatively 2.5%, 2%
capacities for selected CHP technologies and 3.5% from the investment costs annual.
Specific Electrical
Capacity investment nominal FUEL SOURCES FOR ENERGY PRODUCTION
Technology costs efficiency
MW el MEURO/MW el %
Main fuel sources for under 10 MWel CHP plants in
Estonia are natural gas, peat and wood chips.
0,1 ST/SE 10,3 10
0,1 GE 1,6 32
Fuel prices
1 ST/SE 5,1 15
The fuel prices taken as basis for heat price calculations
1 ORC 5,8 15
are as follows:
1 GE 1,0 40
5 ST/SE 3,2 22 Peat price 11.7 EUR/MWhfuel. Proposed price is
5 ORC 4,5 16 based on average peat price levels obtained from
5 GE 0,8 41 Tootsi Turvas AS, the biggest peat milling and
exporting enterprise in Estonia.
10 ST/SE 2,9 22
10 ORC 4,2 16 Wood chips price 12.8 EUR/MWhfuel. Proposed
10 GE 0,8 42 price is based on latest data, published by Estonian
Institute of Economic research in their web based
In this paper investment means all costs before CHP price information system [9].
plant commissioning.
Natural gas 35 EUR/MWhfuel. Proposed price is
For the evaluation of CHP competitiveness the efficiency an average price for the latest data published by
drop working at partial load is taken into account. It is Statistics Department of Estoni [10].
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
269
th th
Ash handling costs Taking into account above mentioned information the
Ash handling costs calculations are based on ash handling costs per MWh of fuel energy content for
assumptions, that: the peat and wood chips are ~0.19 and 0.72 EURO
respectively.
Peat ash content is 5%. Average calorific value is
3.3 MWh/t [11]; Pollution charges
Wood chips ash content is 1%, calorific value Pollution charges and levels are calculated base on the
2.4 MWh/t; Environmental Charges Act [12], Regulation No 99/2004
[13] and No 94/2004 of Estonian Minister of Environment
Natural gas combustion does not emit any ash; [14].
Regarding to information obtained from different The method described in [13] takes into account different
landfill owners, an average for year 2012 expected combustion technologies, flue gas cleaning
ash removal costs (ash transportation to landfill, and technologies, control devices as well as capacities to
storing) are 45 EUR/t. define emission factors of pollutants.
The combustion plant is equipped with dry ash
removing system.
Table 2. Summarised results of the heat price calculations for different CHP expansion scenarios
District heating area CHP plant capacity Heat price, EURO/MWh
Heat Maximum Investment

Heat Electrical IRR 7% IRR 12%
demand heat capacity Technology Fuel
MWh MW MW h MW e MEURO Without With Without With

grant grant grant 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 ---
To avoid complexity of the analysis to be issued from Combustion plant is equipped with the most
different combinations of capacities, combustion effective flue gas treatment technology mentioned
technologies, fuel gas cleaning and control equipment in [13].
it is assumed that:
Calculated levels for pollution charges for year 2013
Thermal capacity of combustion plants is below are:
50MW;
~0.07 EUR/MWhfuel for wood chips;
Selected combustion technology provides lowest
~0.95 EUR/MWhfuel for peat;
emission level than the others in [13] mentioned
combustion technologies; ~0.43 EUR/MWhfuel for natural gas.
Combustion plant is equipped with the most
effective control systems mentioned in [13];
270
th th
HEAT PRICE CALCULATIONS Heat prices for places with annual heat demand under
20 000 MWh are mainly above 45 EURO/MWhfuel
Calculations of heat prices are provided in
where an average heat prices for biomass boiler
correspondence with principles described in section
houses are between 4045 EURO/MWh [16].
Principles for evaluation of CHP technologies
Developing of CHP plants in such heat demand areas
competitiveness.
is feasible in the case of receiving of grant payments
Heat prices are evaluated for different scenarios. for investments.
Scenarios include described heat demands, considered
CHP plant development based on wood chips or peat
fuels and technologies.
could be feasible without grant payments in the places
Heat prices are calculated for 7% and 12% IRR. where heat demand exceed 3000-40000 MWh annual.
Carefully selected CHP technology and capacity can
Heat price for CHP plant developing scenarios which
afford higher IRR when keeping competitive heat
satisfy the requirements described in regulation [X],
prices.
which define conditions for grant payments to expand
renewable energy production and construction of CHP The most feasible places for CHP expansion in Estonia
plants in Estonia, are calculated separately. are Maardu, Viljandi, Rakvere, Valga, Haapsalu, Vru,
Paide and Plva.
For calculating heat prices in addition to information
from previous paper sections, some other figures have Calculation results are valid for assumed cases only.
to be specified: Other particular cases should be calculated
individually.
Cash flows are calculated for 20 years;
CHP starts energy production in the beginning of
2013; REFERENCES
Expected rate of inflation is 1.5%; [1] C. Dtsch and A. Jentsch, District heating (DH) in
Heat loses in district heating network are 15%; areas with low heat demand density (HDD):
Heat load profile is estimated based on heat load A chance for the integration of renewable energy
model described in [15] taking as a basis the heat sources (RES), 10th International Symposium on
load duration curve shape of Tallinn. District Heating and Cooling, 35, September
2006, p. 2
The results matrix of heat price calculations is shown in
www: http://www.lsta.lt/files/events/20_doetsch.pdf
Table 2.
[20.01.2010]
[2] Electricity Market Act www:
CONCLUSION
https://www.riigiteataja.ee/ert/act.jsp?id=13279771
The technologies for smaller CHP applications are [14.05.2010]
more expensive (specific price) and less efficient than
[3] Schwaiger, H., Jungmeier, G, (2007) Overview of
those for larger CHP plants.
CHP plants in Europe and Life Cycle Assessment
At present peat is considered as a good alternative for (LCA) of GHG emissions for Biomass and Fossil
wood chips. Lower fuel price (11.7 EUR/MWh) smooth Fuel CHP Systems CIBE Conference
over higher than for wood chips ash handling costs and Cognration biomasse dans l'industrie et sur les
pollution charges. At the same time wood chips are rseaux de chaleur opportunits retours
more preferable because of higher feed-in tariffs for d'exprience-perspectives
produced electricity.
[4] Obernberger, I., Thek, G, Techno-economic
The advantages of gas engine CHP plants are evoluation of selected decentralised CHP
relatively low investment costs and high electrical appications based on biomass combustion in IEA
efficiency. But because of high natural gas price partner countries Graz (2010)
(MWhfuel price is 2.53 times higher than for wood
[5] Bryson, T., Major, W., Darrow, Ken. Assessment of
chips and peat) and relatively high fixed O&M costs the
On-Site Power. Opportunities in the Industrial
calculated heat prices are the highest. Heat price for
Sector, Carlsbad (2001) www:
expected 7% IRR is between 53 and
http://www.uschpa.org/files/public/Assessment%20
61 EURO/MWhheat depending on heat demand.
of%20Onsite%20Power%2001.pdf
Under 5 MWel ORC is competitive to SE/ST [14.05.2009]
technology. Calculated heat prices are lower for
[6] Kirjavainen, M., Sipil, K., Savola, T. Small-scale
14 EURO/MWhfuel, where higher fuel price difference
biomass CHP technologies. Situation in Finland,
corresponds to places with lower heat demands.
Denmark and Sweden, VTT Processes (2004)
271
th th
www: [12] Environmental Charges Act, [14.05.2010] www:

http://www.opet- http://www.riigiteataja.ee/ert/act.jsp?id=13316043
chp.net/download/wp2/small_scale_biomass_chp_t [14.05.2010]
echnologies.pdf [14.05.2010]
[13] Procedure and Methods for Determining Emissions
[7] Institute for Thermal Turbomachinery and Machine of Pollutants from Combustion Plants into Ambient
Dynamics, Cogeneration (CHP) Technology Air www:
Portrait, Vienna (2002) www: http://www.riigiteataja.ee/ert/act.jsp?id=789462
http://www.energytech.at/pdf/techportrait_kwk_en.pdf [14.05.2010]
[14.05.2010]
[14] Vlishu eralduva ssinikdioksiidi heitkoguse
[8] U. S. Environmental Protection Agency Combined mramismeetod www:
Heat and Power Partnership, Biomass Combined http://www.riigiteataja.ee/ert/act.jsp?id=127572 15
Heat and Power Catalog of Technologies, (2007) [14.05.2010]
www:
http://www.epa.gov/chp/documents/biomass_chp_ [15] Latov, E., Siirde, A. (2010). Heat load model for
catalog.pdf [14.05.2010] small-scale CHP planning. In: Proceedings of
International Conference on Renewable Energies
[9] Estonian Institute of Economic research www: and Power Quality: International Conference on
http://www.ki.ee [14.05.2010] Renewable Energies and Power Quality
[10] Statistics Estonia www: www.stat.ee [14.05.2010] (ICREPQ10), Granada (Spain), 23-25th March,
2010., 2010.
[11] Paappanen, T., Leinonen,A. Fuel peat industry in
EU, 2005, p. 134 www: [16] Estonian Competition Authorities approved district
http://turbaliit.ee/index.php?picfile=21 [14.05.2010] heat maximum prices (without VAT) to end-users
www:
http://www.konkurentsiamet.ee/file.php?15416
[14.05.2010]
272
th th
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 corresponding to 310 million square metres in multi-

family buildings and service sector premises. The
The current heat use refers normally to the average
survey sample thus constituted a sizable portion of the
heat use in a country or a sector during the course of a
entire building stock.
year. But it is also important to be aware of the
distribution of high to low use when estimating the This energy statistical data, published in the annual
potential for reducing total heat use. reports from Statistics Sweden, have been
supplemented with a deeper analysis of the distribution
Energy statistical data published in the annual report
of the heat use and the systematic causes regarding
from Statistics Sweden have been supplemented by a
high heat use. Independent variables for explanation of
deeper analysis of distribution of heat use and
variations were number of degree-days, construction
systematic causes regarding high heat use.
year, ventilation system, energy efficiency measure,
The aim of this paper is to explain the variation in heat and co-use of heat supply. High and low users were
use with respect to construction year, degree days and also analysed by location, construction year, heat
energy efficiency measures. supply method, ownership, and building size [3]. In this
short paper, the specific heat use will be presented by
In the Swedish energy efficiency debate, many voices
its distribution, construction year, degree days and
refer to systematic causes for high heat use. However,
energy efficiency measures.
the results from this study do not support this opinion,
since the use distribution mostly comes from individual
1. Distribution of heat use
causes. The most important implication of the study
results is that systematic policy measures will have a The total distribution of specific heat use as a function
low impact on the total national energy efficiency. of the percentage of the building area of all multi-family
buildings and service sector premises in Sweden is
shown in Fig. 1.
INTRODUCTION
Multi-family residential buildings and service sector Heat use
kWh/m2
premises constitute 80% of the customer stock in the
Swedish district heating systems. The level of future
400
Multi-family buildings
heat use in these buildings will thus have a strong 350
Premises
influence on the future district heating economy and the 300
corresponding investment demand. It is therefore of 250
interest to collect information and make analyses of the 200
costumer heat use and how the heat use will develop in 150
the future. 100
50
Specific heat use in multi-family buildings and service 0
sector premises has decreased considerably since the 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
1970s. In 2006, the specific heat use in multi-family Share of all national building space
buildings has decreased by 38% compared to the heat Fig. 1 Heat use distribution during 2006 as a function of
the share of all national building space. The diagram is an
use in 1972. The lower heat use is due to increasing
estimation for all multi-family and service buildings in
energy prices and more energy efficient buildings. Sweden.
An extensive study of the current heat use for buildings
The area under each curve is the total heat used in
in Sweden has been performed. The input information
multi-family buildings and service sector premises
for this study was constituted by the anonymous
during 2006. The figure shows that 13% of the area in
responses to the annual survey of energy use in multi-
multi-family buildings had a specific heat use of more
family buildings and service sector premises performed
than 200 kWh/m2, and 12% of the area in service
for 2006 by Statistics Sweden, [1] & [2]. The responses
sector premises had a specific heat use of more than
provided input data from 11253 buildings having a total
200 kWh/m2. This result shows that there are no major
area of 77.6 million square metres. By using scaling
differences between the percentages of the building
factors, estimates could be made for the entire country,
273
th th
area with high heat use in multi-family buildings and The relationship between construction year and high
service sector premises. heat use in buildings has been analysed. The definition
of high heat use is 200 kWh/m2 or more. Fig. 4 shows
The results in Fig. 1 also show that 11% of the building
the results for multi-family buildings. There were a total
areas in multi-family buildings, and 31% of the building
of 179.3 millions square metres in multi-family buildings
area in service sector premises have a specific heat
in 2006 and 13% of the heated area had heat use of at
use lower than 100 kWh/m2.
least 200 kWh/m2.
Buildings with heat use 200 kWh/m2 or more have
Of special interest are buildings built during the period
been further analysed and the results are presented in
196574, when a large part of the existing buildings in
the following section considering construction year.
Sweden were built. During this period there were no
requirements for low energy use in buildings.
2. Construction year
Fig. 2 shows specific heat use in multi-family buildings In multi-family buildings built during the period
as a function of construction year. The figure also 196574, 30% of the total area had heat use of at least
includes the average value each year, together giving 200 kWh/m2 and for buildings built in the period
the total average specific heat use of 152 kWh/m2. 194160. 42% of the total building area had heat use
There are no major differences in heat use in buildings of 200 kWh/m2 or more.
constructed before 1980. After 1980, the heat use was
approximated 15% lower than the average heat use for
all buildings in Sweden.
Heat use
kWh/m2
400
350
300
250
200
150
100
Fig. 4 Total square metres where heat use is higher or
equal to 200 kWh/m2 in multi-family buildings categorised
50
0 by construction year.
1930 1940 1950 1960 1970 1980 1990 2000 2010
Construction year
Fig. 2 Specific heat use as a function of construction year

for 4285 multi-family buildings
Heat use
kWh/m2
400
350
300
250
200
150
100
50 Fig. 5 Total square metres where heat use is higher or

0
equal to 200 kWh/m2 in service sector premises
1930 1940 1950 1960 1970 1980 1990 2000 2010 categorised by construction year.
Construction year
Fig. 3 Specific heat use as a function of construction year The relationship between high heat use in service
for 4061 service sector buildings. sector premises and construction year is shown in
Fig. 5. During 2006, 15.6 million square metres had
The heat use in service sector premises is shown in
heat use of at least 200 kWh/m2. Service sector
Fig. 3. Also in these buildings, the average heat use
premises built between 1965 and 1974 had high heat
after 1980 is lower (about 10%) than the average heat
use in 3.5 million square metres.
use in all service sector premises in Sweden.
274
th th
The results show that the period 19651974 did not in the theoretical analysis of the optimal wall insulation
have a dramatically higher heat use in the construction as a function of degree-days.
year analysis.
The results show that the average difference between
Northern and Southern Sweden was small, implying a
3. Degree days
small climatic impact on heat use. The main conclusion
The climate in Sweden varies with a much colder from this analysis is that the individual variation in each
climate in the northern part compared to the southern climate area is much higher than the local impact of
part. Since the statistical data consist of buildings from climate. This astonishing conclusion can have several
different parts of Sweden, the influence of the local different explanations:
climate on the heat use in buildings can be analysed.
This has been done by analysing the correlation Higher awareness and consequences of low
between the number of degree days for the location of building heat resistances in Northern Sweden
a building and the corresponding specific heat use. Lower regional GDP in Northern Sweden giving
higher incentive to reduce heat costs
The number of degree days, according to the Swedish
definition, varies from approximately 3000 in the south More frequent snow cover in Northern Sweden
up to 7000 in the north of Sweden. Each building in the giving extra heat resistance during the winter.
analysis was connected to one of 14 climate areas.
4. Energy efficiency measures.
Heat use The statistical data shows the energy efficiency
kWh/m2
400
measures during the period 19952005. The energy-
350
efficiency measures were:
300 a. Supplementary insulation
b. More energy efficient windows
250
200
150
y = 15,63x0,28
c. Balancing heating- and ventilation systems
100 d. Electrical efficiency measures
50
e. Heat recovery in ventilation systems
0
2000 2500 3000 3500 4000 4500 5000 In multi-family buildings, one or several energy
Degree-days
efficiency measures were implemented for an
Fig. 6 Specific heat use for 5111 multi-family buildings as estimated floor area of 57.6 million square metres
a function of the number degree days in each climate
during the period 19952005. No energy efficiency
area.
measures had been performed for an estimated floor
area of 92.2 million square metres during the same
Heat use period.
kWh/m2
400 In service sector premises, with an estimated floor area
350 of 37.2 million square metres, one or several measures
300 had been taken during the period 1995-2005. During
250 the same period, no measures had been taken for an
200 estimated floor area of 70.3 million square metres.
150
y = 10,37x0,30 The most common measures in multi-family buildings
100
and service sector premises were balancing of heating-
50
and ventilation systems.
0
In many buildings, a combination of two or several

2000 2500 3000 3500 4000 4500 5000
Number of degree-days
Fig. 7 Specific heat use for 6041 service buildings as a energy efficiency measures had been taken in the
function of degree days in each climate area. same building. In some buildings, up to five measures
have been taken in the same building.
Fig. 6 and Fig. 7 show the specific heat use as a The average heat use in multi-family buildings and
function of degree days for multi-family buildings and service sector premises in relation to measures taken
service sector premises. The figures also show the is shown by bars in Fig. 8 and Fig. 9. The horizontal
average curve and its equation for specific heat use as lines show the average heat use in buildings, in which
a function of degree-days. You should also note that no energy efficiency measure was performed.
the exponent in the fitted equations has only the
magnitude of 0.3 instead of the 0.5 exponent obtained
275
th th
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

Fig. 8 Average heat use in multi-family buildings in for lower heat use in the Swedish multi-family
relation to the measures performed. The measure figures and service sector buildings.
correspond to the measures defined in the text.
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.
Fig. 9 Average heat use in service sector premises in
relation to the measures performed. The measure figures REFERENCES
correspond to the measures defined in the text.
[1] Statistics Sweden, Energistatistik fr
flerbostadshus 2006 (Energy statistics for multi-
As shown in the figures 8 and 9, there were no family houses during 2006). Statistiska
substantial differences in heat use between buildings Meddelanden EN16SM0702.
where energy-saving measures had been taken and
those where they had not. The conclusion from this [2] Statistics Sweden, Energistatistik fr lokaler 2006
analysis is that the measures taken during these 10 (Energy statistics for premises during 2006).
years were taken by late-comers rather than by early Statistiska Meddelanden EN16SM0703.
adopters, since heat use after measures were taken [3] Andreasson M, Borgstrm M, Werner
generally corresponds to the average level for all S, Vrmeanvndning i flerbostadshus och lokaler
buildings. (Heat use in multi-family buildings and premises
2006) Fjrrsyn report 2009:4, Stockholm 2009.
Available at www.svenskfjarrvarme.se
276
th th
DAMAGES OF THE TALLINN DISTRICT HEATING NETWORKS AND INDICATIVE

PARAMETERS FOR AN ESTIMATION OF THE NETWORKS GENERAL CONDITION
1 1 2 1 1
Aleksandr Hlebnikov , Anna Volkova , Olga Duba , Arvi Poobus , lo Kask
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 a different heat supply alternative. Often the

decentralized heating is not an effective solution for
District heating networks in Estonia are mostly old and
regional heat supply strategy and it decreases potential
in bad condition. The state of the district heating
of combined heat and power production.[3].
networks of Tallinn is typical for the rest of Estonian DH
networks. The paper includes analysis of the Tallinn Nowadays DH systems operate both in big cities and in
district heating networks. Valid data about damages in small towns, which means, that there is enough heat
district heating systems received for the last 12 years load for the installation of new cogeneration equipment.
were used for an analysis of the networks damages.
But before new energy sources installation it is
Different types of network damages are analysed: important to define and analyse the situation with DH
external corrosion, internal corrosion, defect of networks.
installation, factory defects, defect of construction and
The purpose of this paper is to define the valid
other reasons. The number of damages for the different
condition of typical old networks in Estonia, to define
elements of networks is compared in the paper:
the reasons of damage occurrence on the basis of
armature, compensator, construction and pipes. Main
operational data and to make forecasts for operation of
factors, which influence damages in district heating
a DH network for the next 20 years. The paper includes
networks, are the age of networks, the quality of
analysis of Tallinn district heating networks. The valid
construction works and the network operation
data about damages in district heating systems
conditions.
collected during past 12 years was used for analysis of
The damage quantity dependence on the age of networks damages.
networks is also defined and analysed in the paper.
The number of damages can be diminished by THE PRESENT CONDITION OF TALLINN DISTRICT
reducing the average age of networks. This is possible HEATING SYSTEM
by replacing old pipelines and other network system
elements. Pipes average age changes for 20 years District heating networks in Estonia are mostly old and
period are simulated according different intensities of in bad condition. The state of the district heating
renovation works. networks of Tallinn is typical for the rest of Estonian DH
networks. In Tallinn the heat is transmitted to the
consumers through a 406-kilometres long heating
INTRODUCTION
network including the 93 km of pre-insulated pipes
District heating (DH) allows centralized heat production (23%). District heating systems of Tallinn were
for an area and hot water transportation to the buildings constructed mostly during the 1960-1980 period and
through a network of pipes. District heating systems their average age is 22 years.
offer the potential to use energy-efficient and
The AS Tallinna Kte enterprise makes operation of
renewable heat generation technologies, such as
the bigger part of district heating networks and boiler-
cogeneration technologies which implement both fossil
houses of Tallinn.
fuels, as long as biomass and waste [1]. District
heating system is traditional in Estonia. It has formed District heating systems of Tallinn consist of five
approximately 70 per cent of all heating in the country. districts of the central heat supply: Kesklinna district
The share of heat produced by combined heat and (total length ~92 km, length on the balance of
power production stations is approximately one third. At AS Tallinna Kte ~76 km), Lne district (total length
the same time, the technical situation of the district ~162 km, length on the balance of AS Tallinna Kte
heating networks (and production equipment) is poor. ~141 km), Lne district local networks (total length
[2] Unsatisfactory condition of DH networks and ~12 km, length on the balance of AS Tallinna Kte
unreliable heat supply can doubt on future of district ~11 km), Lasname district (total length ~114 km,
heating and the consumers can make a choice towards length on the balance of AS Tallinna Kte ~106 km),
277
th th
Maardu district (total length ~25 km, length on the (natural gas, 232 MWth); the CHP Vo (wood chips, 25
balance of AS Tallinna Kte ~14 km). [4] MWel, 65 MWth); the boiler house Mustame (natural
gas, 390 Wth); the boiler house Kadaka (natural gas,
District heating systems of the areas Kesklinna and
290 MWth).
Lasname are connected through the pump station
Laagna. The total length of heating networks is 406 km Besides the abovementioned there are some small-
from which on the balance of AS Tallinna Kte there scale boiler houses. In Fig. 1 is displayed the basic
are 348 km, or 85,7%. scheme of Tallinn heat supply.
The following CHP stations and boiler-houses supply District heating systems of Tallinn were constructed
heat to the districts of Tallinn: the CHP Iru (natural gas, mostly during the 19601980 period and their average
190 MWel, 748 MWth), the boiler house lemiste age is 22 years.
Mustame
boiler-house Kadaka Iru CHP
390 MW boiler-house lemiste 748 MW (190 MW)
290 MW boiler-house
390 MW 232 MW
(in reserve)
Laagna
pump station
200 MW
Mustame network Kesklinna Lasname
Maardu
325 MW network network
network
180 MW 268 MW
Vo CHP
65 MW
Fig. 1 The basic scheme of Tallinn district heating system

The state of DH networks varies for the different (districts Mustame and isme). Initially there had
districts of Tallinn. been two separate networks which were merged later
on as a result of growth. In the area Lne the
In Lasname the construction of district heating
construction of district heating systems began in 1960.
systems began in 1970, and the network length is
The length of the Lne area network is ~141 km. The
~106 km at present time. Assuming the actual load the
diameters of the main pipelines are less than those in
heating systems of Lasname district are the most
the Lasname area.
overloaded in town.
The length of the main pipelines with diameter
The length of main pipelines DN10001200 is ~19 km,
DN400900 is ~27,8 km. The heat losses of the
the length of pipes DN400-800 is ~4,4 km. The share of
network in 2008 were 16% from the total produced
the main networks is quite big and it is ~22% of total
heat.
network length in Tallinn. Thermal isolation is made of
glass wool according to old soviet building norms and it The speciality about the heating system of the area
is the reason of big heat losses in the network. The Lne is that in past there was an open system of hot
heat losses in Lasname network in 2008 were 21% water supply. The water added to the system had no
from the total produced heat. time to purify sufficiently and oxygen and water
hardness led to an intensive internal corrosion of pipes.
The interconnected district heating systems of boiler-
houses Mustame, Kadaka and Karjamaa (not in In Kesklinn area the network construction began in
operation at present time) are related to the Lne area 1959. Initially the heat supply was carried out by the
278
th th
Tallinna Soojuselektrijaam heat and power station and 30 years and by today they are already worn out. The
later on by the boiler-house lemiste. The district probability of failures sharply increases. By today the
heating system of Kesklinn area is the oldest in Tallinn. 84% of all compensators should be replaced. Some
The average age of the Kesklinn area network is parts of the old locking armature also have to be
25 years, the total length is ~76 km. replaced. The service life of armature has exceeded
25 years. Armature and compensators are partly
The length of the main pipelines with diameter DN400-
renovated; however some pieces of it are old and also
900 is ~13,8 km. The share of main pipelines in
require replacement. [5]
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 300
are less. The reasons for this are: the bigger network
250
loading, the not oversized pipes and the significant
share of preinsulated pipes.[5] 200 armature
compensator
THE ASSESSMENT OF DAMAGES 150
construction
The analysis of networks damage statistics for Tallinn 100 pipes
is made on the basis of valid data collected during the
past 20 years. 50
The distribution of damages of Tallinn district heating 0

network is shown in Fig. 2 according the periods of 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
construction. It is obvious that the most critical situation
is with the sites constructed during the 19801985 Fig. 3 Places of damage in Tallinn district heating
networks elements
period. It can be explained by the poor quality of both
construction works and materials used in construction. In Fig. 4 the nature of damages is summarized. There
During that period the networks were being constructed are no data about the character of damage for all the
in a hurry and with lack of proper supervision. areas of Tallinn within past 10 years, thats why the
damage allocation by character of damage is shown for
damages during
1998-2009
a five year period.
In Tallinn network the significant part of damages is
400
350 caused by external corrosion of pipes. Main reasons of

300
external corrosion are the bad waterproofing of
underground channels and chambers and the
collapsed drainage. Amongst other reasons are the
250
200 defects of pipe supports and the destruction of

150
concrete channels.
100
90
50
80
0
- 1965 - 1970 - 1975 - 1980 - 1985 - 1990 - 1995 - 2000 - 2005 - 70 external corrosion
1965 1970 1975 1980 1985 1990 1995 2000 2005 2008
years of construction 60 internal corrosion
Fig. 2 Damages of Tallinn district heating networks 50 deffect of construction
according the periods of construction deffect of installation
40
In Fig. 3 the places of damage in the network elements 30 wrong service

are shown: armature, compensators of thermal 20 other reasons
lengthening, construction and pipes. The major part of 10
all damages was the pipes.
0
During the 19972003 period there were many 2004 2005 2006 2007 2008
problems with armature and compensators; after 2003
the quantity of damages to these elements had Fig. 4 Nature of damages in Tallinn district heating
networks
considerably decreased. The oldest thermal
lengthening compensators work since 1959. The
The second main cause of damages is the internal
resource of axial compensators is no more than
corrosion. In 2004 many pipes damaged by internal
279
th th
corrosion were revealed. Internal corrosion is the most D 0,0096 A2 1.8985 A 1.0496 (1),
serious problem in Lne network where an open
system of hot water supply earlier has been used. where
Besides the damages caused by defects of installation, D Number of damages/100 km per year
defects of construction, factory defects and improper
A Age of networks
maintenance, the other reasons have also been
registered. Before using this regression for further calculations, we
should check if this equation is appropriate. One of the
The main factors, which have an affect on the damages main parameters for estimation of regression equation
in district heating networks, are the age of networks, is the correlation coefficient. It is considered, that the
the quality of construction works and the network correlation is good in case when R>0.8. In the case of
operation conditions. The two latter can be regulated damage dependence on pipes age, R is 0.802.
by control authorities and proper legislation, however, R2=0.643, which means that the equation
the influence of these factors has been reduced in characterizes the 64,3% of damage number changes,
comparison with the 19701990 period. Then quality of but the 35,7% of changes are characterized by another
construction works was very low, drainage systems factors. There is still an influence of other factors,
were installed incorrectly or were not installed at all and which can not be changed, such as construction and
isolation materials were not qualitative. As regards installation problems in the past.
district heating operation conditions, the
aforementioned open vented hot water supply system Data about damages allocation by the group and
used in some networks has led to intensive internal approximation of these data is shown in Fig. 5.
corrosion of pipes. The regression equation can be used for the damage
One important reason for damages reduction is that in forecasts in future.
recent years the networks have significantly reduced damages/100 km
pressure. The network works in a stable temperature per year
mode, the reliability of heat sources is improved and 70
the quantity of equipment emergency stops forced by 60
sharp fluctuations of the heat-carrier temperature has
50
decreased.
40
Other operation condition factor which influenced
number of district heating system damages was higher 30
water temperatures in networks (up to 130 t C) than 20
nowadays (up to 110 t C). Finally we can conclude
10
that such factors as quality of construction works and
quality of network operation are close to their optimum 0
at present time in comparison with previous years. 0 5 10 15 20
Age of networks
25 30 35
Damage quantity also depends on the age of networks. Fig. 5 Damage number dependence on the age of pipes in
The number of damages can be reduced by reducing district heating systems for the 20052007 period
the average age of the networks. This is possible by
As it has been mentioned before, the age of networks
replacing the old pipelines and other networks systems
depends on the intensity of renovation works.
elements.
In Fig. 6 the length of all repaired sites is shown split by
Reconstruction and replacement works are made in
years.
Tallinn, but the intensity of replacement is rather low
and not enough for a stable system operation. It is Since 1980 the serial repair of Tallinn district heating
important to define, how intensive the network system is being carried out.
reconstruction should be.
Basically the investments have been directed towards
Data for the three past years were used for defining the the increase of reliability and the reduction of quantity
damage dependence (number of damages/km/year) on and duration of faults in heat supply. It has been
the age of networks. Data about damages were invested a lot in the locking armature.
collected for 7 age groups (05 years, 510 years,
For the past 10 years ~35 km of district heating
1015 years, 1520 years, 2025 years, 2530 years,
pipelines have been replaced, which is 10% of total
3035 years).
length of the district heating systems in Tallinn area.
Using least squares analysis, a regression equation for
this dependence was defined.
280
th th
The annual replacement of pipes is in average about The average age of pipes for each year was calculated
3,06 km per year, which is less than 1 percent from the according equation (2)
length of Tallinn DH system pipelines. j b b
Length, km
li ( j i) li ( j i) ( j c) (l j li )
8 Aav j i a i a i a
(2)
lj
7
b c
6 i=b, if li l j ; i=c, if
ia
l
i a
i lj
5
4 where
3 Aav is average age of pipes in j year
2 li is length of pipes, constructed in i year
1 I year of construction;
0 J current year;
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
a year of construction of the oldest pipes, operating in
Fig. 6 Length of replaced pipelines by years in Tallinn
the current year.
district heating network As a result of simulations, seven forecasts for pipes
average age were calculated according different
THE FORECASTS FOR DISTRICT HEATING
intensity of renovation works: for current intensity of
SYSTEM AGE
renovation (3,06 km/year) and for intensities when 1%,
One of the tasks was to assess, how big the renovation 1,5%, 2%, 2,5%, 3% and 4% of total DH system length
works should be in order to stop increasing the average would be annually renovated. The forecasts were
age of pipes. A simulation model, which uses both real simulated for the 20 year long period.
data and also some assumptions, was created for such
The results of simulation are shown in Fig. 8.
estimation.
Assuming that the length of pipes (360,67 km) will not
age, years
40
change during the forecast period and that the annual
35
scope of renovation works will remain the same during
whole of the period means that the length of renovated 30
pipes also will not change. Besides its was assumed

25
that every year just the oldest pipes would be 1%
2%
renovated; however in reality the renovation works are 20 3%
4%
based on the pipes actual state estimation. 15
1,50%
2,50%
current
Allocation of pipes ages for starting point (2008) is 10
shown on Fig. 7 [5].
5
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
length, km
25
Fig. 8 Pipe age forecasts for different intensity of network
renovation works
20
As it can be seen from Fig. 8 in case the renovation
stays on the same level, the average age of pipes will
15
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.
10
When the 2% of DH system length is annually

5
renovated there will be the minimal changes in age
during first 5 years, after that the age will start rising
0
and only after 15 years it will begin to decrease.
49
46
43
40
37
34
31
28
25
22
19
16
13
10
7
4
1
age, years
If renovation intensity is 2,5% of the length or higher,
Fig. 7 Length of DH networks by pipes age (in 2008) 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
281
th th
sites should be at least 9 km/year. This way the Seven forecasts for pipes average age according
process of ageing will slow down and also the average different intensity of renovation works were simulated:
age will stabilize on a certain mark. One of the possible for current intensity of renovation (3,06 km/year) and
solutions is to replace the pipes with higher intensity of for intensities when 1%, 1,5%, 2%, 2,5%, 3% and 4%
34% until reaching the 1720 years average age and of total DH system length would be annually renovated.
then reduce the length of renovated pipes per year to It was concluded, that for maintaining the networks
the 22,5% of the whole length of DH network. average age at least at former level, the rate of old
pipelines replacement should exceed the 2,5% of the
CONCLUSIONS whole length of DH system.
District heating networks in Estonia are mostly old and

AKNOWLEDGMENT
in bad condition. The state of the district heating
networks of Tallinn is typical for the rest of Estonian DH This work has been partly supported by the European
networks. Thats why the result of damage analysis Social Fund within the researcher mobility programme
made for the DH network of Tallinn can be used for the MOBILITAS (20082015), 01140B/2009
other networks in Estonia.
The AS Tallinna Kte enterprise makes operation of REFERENCES
85% from the length of district heating networks in
[1] Cogeneration and district energy sustainable
Tallinn. Tallinna Kte data about the damages were
energy technologies for todayand tomorrow,
used for assessment.
International Energy Agency, 2009.
Places of damages in the DH system are following:
[2] Long-term Public Fuel and Energy Sector
armature, compensator, pipes and construction. Most
Development Plan until 2015, Riigi. Teataja, RT I,
of the damages happened in the pipes.
23.12.2004, 88, 601
As regards the character of damages, the typical
[3] Hlebnikov, A.; Siirde, A. The major characteristic
damages are caused by external corrosion, internal
parameters of the estonian district heating
corrosion, defect of construction, defect of installation
networks, their problems and development. // The
and wrong service. The major part of damages is
11th International Symposium on District Heating
caused by external corrosion of pipes.
and Cooling: University of Iceland, 2008, 141148.
The age of networks, the quality of construction works
[4] Tallinna kte webpage, www.soojus.ee
and the network operation conditions are the most
important factors, which influence the damages in [5] A. Hlebnikov "The analysis of efficiency and
district heating networks. The number of damages can optimization of district heating networks in
be reduced by reducing the average age of the Estonia", Doctoral Thesis, Tallinn University of
networks. This is possible by replacing the old pipelines Technologies, 2010.
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.
282
th th
EFFICIENCY OF DISTRICT HEATING WATER PUMPING IN FINLAND

1 1 1
Antti Hakulinen , Jarkko Lampinen and Janne Lavanti
1
Pyry Finland Oy
ABSTRACT The investment costs of a pump
The objective of this study was to determine the 90 000
savings potential in district heating pumping in Finland. 80 000
A measurement method was also developed to quickly 70 000
60 000
estimate the efficiency of district heating pumping. 50 000
Total
Pump
The work was based on the data gathered from district 40 000 Frequency converter
Motor
30 000
heating statistics. The work is divided into two parts. 20 000
The district heating statistics reveal a number of district 10 000
0
heating networks whose consumption of energy 0 100 200 300 400 500
needed for pumping is exceptionally high. These pow er kW
companies should clarify the reasons for that.

Fig. 1. The investment costs of a pump.
In addition, companies with an exceptionally low
1.2 Total costs of pumping
consumption of pumping energy should check their
measurements and data gathering routines. Total costs of pumping include capital, maintenance
and energy used in pumping. The pump lifetime costs
On average the electricity needed for district heating are mainly energy costs as we can see from figure 2 on
pumping should not be over 0.5 per cent of the total the next page. The lifetime costs are calculated with
energy supply (=sold+losses). If the density the following assumptions: energy price 60/MW/h,
(supply/length of the network) of the district heating operating lifetime 15 years, utilization period of
network is less than 3 GWh/km, the energy needed for maximum load 5000 h/a, interest rate 5 per cent and
pumping may rise. In any case the proportional the O&M 1.2 per cent of the investment.
pumping energy should be lower than 1 per cent of
total energy supply. Pum p I, pow er: 16 kW
The Finnish potential for saving in district heating 14 %
pumping is estimated to be 20 per cent of the current 2%
pumping energy i.e. 30 GWh/a. This is equivalent to a

Capital
O&M
yearly saving of approximately 2 million. Energy
84 %
PART 1.
INTRODUCTION
Pum p II, pow er: 131 kW
In the Finnish district heating systems no typical

pumping arrangements have been used at heat 7%
1%
production plants or at booster pump stations. The Capital

ways of dimensioning and connecting pumps have O&M
varied a lot. This has led to incorrect dimensioning and

Energy
connections of pumps, which in turn has caused higher 92 %
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 Pum p III, pow er: 283 kW
Finnish district heating system is based on the variable 5%

flow operation (consumer driven scheme)
1%
Capital
O&M
1. COSTS Energy
1.1 The investment costs of a pump (including 94 %
motor and controls)

The calculated investment costs of a pump including
motor, control, and pump are shown Fig. 1. Fig. 2. An example of lifetime costs of three different sized
pumps.
283
th th
As we can see the pumps efficiency plays a huge role cent or 2*70 per cent in parallel connection with
because the lifetime costs mainly consist of the individual rotating speed controls. In that way the
operational energy (8494 per cent). For that reason a pumping of maximum heat load can be managed and
lot of attention should be paid to the efficiency when there is a room for possible expansion of the district
making the investment. Pumps with a low efficiency heating network. The other pump will act as a summer
may eat into the savings of the investment many fold. pump so that the efficiency of pumping remains high
also when the heat load is low.
1.3 Booster pump station and costs
By dividing the pumping capacity between many
A booster pump station should be considered when the pumps it is possible to save pumping energy even if
primary pumps of an energy station do not have pumping is handled from one point or from the heat
enough capacity to ensure the pressure difference at production plant and the booster pump station. The
the last customer. Typical reasons for the building of a possibilities to divide the pumping must be examined
booster pump station can be: long transmission lines, case by case by taking into account every single thing
expansion of network, optimization of pumping energy that might have an effect on the costs.
and controlling of pressure level.
The investments of a booster pump station including 3. MAXIMUM WATER FLOW
pump, motor, frequency converter, building, automation
The actual cooling of the district heating system in
systems, etc. are shown in the following figures 3
operational conditions of pumping should be taken into
and 4.
account when determining the calculated maximum
The investment costs of a booster pump station water flow. It is worthwhile to specify the water flow
according to slightly worse cooling than the actual
500 000
conditions require so that there is some design margin
480 000
460 000
for unusual conditions.
440 000
420 000 4. OPERATION POINT

400 000
To change the rotating speed of a pump with a
380 000
frequency converter is a good way regarding energy
360 000
340 000
efficiency because the pumps efficiency often remains
320 000
on high level within the whole adjusting area but the
300 000
need for power reduces strongly when the rotating
0 50 100 150 200 250 300 350 400 450 500
speed goes down.
pow er [kW]
Fig. 3. The investment costs of a booster pump station

(only 1 pump).
The investment cost of a booster pump station (flow + return)
760 000
710 000
660 000
610 000
560 000
510 000
460 000
410 000
360 000
0 50 100 150 200 250 300 350 400 450 500
pow er [kW]
Fig. 5. An example functional diagram of a pump.
Fig. 4. The investment costs of a booster pump station An example functional diagram is shown in Fig. 5.
(with 2 pumps). When the rotating speed changes, the efficiency
remains good regardless of the changing rotating
2. PUMPING ARRANGEMENTS
speed. The pumping of a district heating network
At the primary station it is usually sensible to divide the follows this theoretical situation very well. However
pumping between a few pumps, for example 2*60 per when choosing a district heating pump it is important to
284
th th
pay attention to its rotating speed which should be at Specific pumping energy vs. heat supply
Supply 0 - 2 500 GWh/a
the minimum from 50 to 60 per cent of the nominal 1.6 %
rotating speed.
Specific pumping energy (electrical power /

1.4 %
1.2 %
PART 2.
1.0 %
INTRODUCTION
heat supply)
0.8 %
The goal of the second part was to motivate the district
heating companies to analyse their pumping methods 0.6 %
and, hopefully, to lower their pumping costs. 0.4 %
Total savings potential in district heating pumping in 0.2 %
Finland was also estimated. 0.0 %

0 500 1000 1500 2000 2500
This part is based on the Finnish district heating Heat supply, GWh/a
statistics of the year 2007 [1]. The statistics cover the Fig. 6a. Example Electricity used for pumping in relation to
data of nearly 200 district heating companies but only the size of a district heating company, heat supply
about 60 of them have reported the electric power used 02 500 GWh/a.
in district heating pumping. Specific pumping energy vs. heat supply
Supply 0 - 300 GWh/a
The biggest companies have reported the pumping 1.6 %
energy, thus, the pumping figure is available to Specific pumping energy (electrical power /
1.4 %
companies which supply almost 70 per cent of all
district heat in Finland.
1.2 %
1.0 %
heat supply)
STATISTICAL FINDINGS 0.8 %
0.6 %
The used pumping energy in different companies was
analyzed by comparing the pumping energy to the 0.4 %
following parameters: 0.2 %
Heat supply (sold heat + losses) 0.0 %

0 50 100 150 200 250 300
Length of the district heating network Heat supply, GWh/a
Heat density (supplied heat energy divided by the

Fig. 6b. Electricity used for pumping in relation to the size
length of the DH network)
of a district heating company, heat supply 0300 GWh/a.
The following parameters were also examined but no
clear correlation was to be seen, and the results are The average pumping energy is 0.6 per cent of the
therefore not reported in this paper: heat supply. The bigger the company, the smaller the
proportional pumping energy.
Heating output density (daily maximum heating
output divided by the length of the DH network) Companies with exceptionally high pumping energy are
CHP production marked with a circle.
Share of small (< 30 kW) consumers
2. Length of the district heating network
Peak load utilization time
Specific pumping energy vs. length of the DH net
Losses of DH network Length 0 - 500 km
1.7 %
Specific pumping energy (electrical power
1. Heat supply
1.5 %
Heat supply is the same as sold heat + losses.
1.3 %
The assumption used was that the bigger the company

/ heat supply)
1.1 %
the smaller the proportional pumping energy. 0.9 %
The situation is presented in figures 6a and 6b, where 0.7 %
only the companies which supply less than 0.5 %
2500 GWh/a (Helsinki not included) are shown. 0.3 %
0.1 %
0 100 200 300 400 500
Length of the DH net, km
Fig. 7a. Electricity used for pumping in relation to the

length of the district heating network, 01 300 km
285
th th
Specific pumping energy vs. length of the DH net Specific pumping energy vs. heat density
Length 0 - 70 km
1.7 %

1.7 %
Specific pumping energy (electrical power
1.5 %
1.5 %
1.3 %
1.3 %
1.1 %
heat supply)
/ heat supply)
1.1 %
0.9 %
0.9 %
0.7 % 0.7 %
0.5 % 0.5 %
0.3 % 0.3 %
0.1 % 0.1 %
0 10 20 30 40 50 60 70 0 1 2 3 4 5 6 7
Length of the DH net, km Heat density, GWh/km
Fig. 8. Electricity used for pumping in relation to the heat

Fig. 7b. Electricity used for pumping in relation to the
density.
length of the district heating network, 0500 km.
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.
Specific pumping energy vs. length of the DH net
1.6 % FURTHER INFORMATION:

1.4 %
Pyry Finland Oy
1.2 % PL 93 (Tekniikantie 4 A)
1.0 % FI-02151 Espoo
heat supply)
Finland
0.8 %
antti.hakulinen@poyry.com
0.6 %
0.4 %
CONCLUSION
0.2 %
District heating networks enlarge and change
0.0 %
0 200 400 600 800 1000 1200 1400
continuously and therefore the conditions of pumping
Length of the DH net, km will also change. For that reason, it is important to
check every now and then if the actual operating point
Fig. 7c. Electricity used for pumping in relation to the of the pump is as designed and what the efficiency of
length of the district heating network, 070 km. the present operating point is. The pumping could still
work technically well but the pumps could be operating
For big companies the proportional pumping energy is with low efficiency.
almost constant 0.5 per cent of heat supply.
The most important issues in designing and operating
The longer the DH network, the smaller the proportional of district heating pumping are:
pumping energy. The result is partly the same as in the
previous chapter: the bigger companies have smaller A sufficient but not too big pressure difference
proportional pumping energies. must be guaranteed for customers.
There must be enough pressure in all parts of the
If a company seems to have a high proportional network at all circumstances and at the same time
pumping energy in figures 7a7c it may be due to poor the maximum pressure level must not be
heating density (lots of pipes in areas with not so much exceeded.
consumers). When designing pumping it is important to study
all possible pumping cases.
3. Heat density
Good operating point should be verified when
Heat density is the heat supply divided by the length of designing and operating pumps.
the district heating net.
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.
286
th th
The figure below is the same as the Fig. 8 added with a The following figure illustrates an example case in
red line to help the reader estimate the pumping energy which the heat density is over 2.5 GWh/km. The figure
of his own plant. If the pumping energy is above the red can be utilized when estimating the losses in real
line some measures ought to be taken. money if the proportional pumping energy is over the
average of 0.5 percent.
Specific pumping energy vs. heat density
Value of "excess" pumping energy
Heat density > 2.5 GWh/km, Value of power 60 EUR/MWh
1.7 % 1.7 %
400
Specific pumping energy (electrical power / heat
1.5 % 1.5 %
350 Proportional share of pumping energy 0.8 %
Proportional share of pumping energy 0.7 %
Value of excess pumping energy, 1 000 EUR/a

1.3 % 1.3 %
300
Proportional share of pumping energy 0.6 %
1.1 % 1.1 % 250

supply)
0.9 % 0.9 % 200
0.7 % 0.7 % 150
0.5 % 0.5 % 100
50
0.3 % 0.3 %
0
0.1 % 0.1 %
100 600 1100 1600 2100
0 1 2 3 4 5 6 7 8 Heat supply, GWh/a
Heat density, GWh/km
For example, if the heat supply of the company is 1.1

Fig. 9. Electricity used for pumping in relation to the heat TWh/a and the proportional pumping energy is
density of the district heating network + trend line 0.7 percent, the losses of unnecessarily high pumping
energy is 130 000 per year.
Figure 9 shows that on average the electricity needed
for district heating pumping should not be over 0.5 per Some of the pumping energy is converted to heat. This
cent of the total energy supply (=sold+losses). If the decreases the value of the losses.
density (supply/length of the network) of the district
In total, the potential savings in all Finnish district
heating network is less than 3 GWh/km, the energy
heating companies are approximately 20 percent of the
needed for pumping may rise. In any case the
current pumping energy, i.e. 30 GWh/a. This is
proportional pumping energy should be lower than
equivalent to a yearly saving of approximately
1 percent.
2 million.
REFERENCES
[1] DH statistics 2007, Energiateollisuus ry, 2008
287
th th
MODELLING DISTRICT HEATING COOPERATIONS IN STOCKHOLM AN

INTERDISCIPLINARY STUDY OF A REGIONAL ENERGY SYSTEM
1 2
D. Magnusson , D. Djuric Ilic
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 Karlstad in 1948 and during the following decades

the largest cities built their own systems, as was the
In this paper, a combination of methods from social
case in Stockholm [2]. Because of the large amount of
science (interviews) and technical science (modelling)
energy in the systems, the fuel used in the plants has a
have been used to analyse the potential for
major impact on greenhouse gas (GHG) emissions,
cooperation in the present and future district heating
and there is also a large potential for using combined
system in Stockholm. The aim of the paper is to explore
heat and power (CHP) technology in the systems. CHP
barriers and driving forces for energy cooperation in the
technology is becoming more important as a part of
Stockholm district heating system and to analyse the
creating sustainable energy systems, which for
potential for combined heat and power generation in
example can be seen in the EU directive for promotion
the system. In the study it was found that with better
of cogeneration [3]. In Sweden, as well as in
connectivity in existing systems, the annual system
Stockholm, large investments are made in building new
cost would decrease by approximately 10 million , and
CHP plants, in large part thanks to the electricity
with new CHP plants a similar potential exists. There is
certificate system [1]. Another important potential with
also a large potential for decreasing the local and
CHP generation is through the Electricity Directive of
global emissions of CO2 with CHP plants. The results
1996, in which the EU prescribed common rules for
from the interviews showed that the existing
creation of an open and competitive electricity market
cooperation has a long history and is working well
[4]. With a fully integrated electricity market, the
today. The advantages are higher supply security and
Swedish prices of electricity can be expected to
economic benefits, while disadvantages are a need for
increase. However, as long as they are lower than
more administration and control because of a more
Europes there is a large potential for exporting
complex system. That the barriers to cooperation are
electricity. From a marginal power production
seldom technical is another conclusion. With the
perspective, which will be discussed further in the
combination of methods, we have gained a better
paper, there is a potential for decreasing global
understanding of the actual potential for the
emissions of CO2, if the exported electricity comes from
development of the system.
non-fossil fuels.
NOMENCLATURE A large enough system is an important prerequisite for

investment in CHP plants, in order to take advantage of
CO2 carbon dioxide; the economy of scale of district heating and CHP
LECO2 local emissions of CO2; generation. In Stockholm, the largest urban region in
Sweden, there are already well-developed district
GECO2 global emissions of CO2; heating systems. The systems started as smaller units
CHP combined heat and power; that gradually have been interconnected and today
consist of three large networks. However, since there
BCHP CHP plants fuelled by solid biomass; are eight different energy companies in the city region,
NGCHP CHP plants fuelled by natural gas; a working cooperation between the energy companies
is important. With this in mind we will analyze how the
TGC tradable green certificates;
actors perceive existing and future cooperation. The
GHG greenhousegas. study is conducted with an interdisciplinary approach
where interviews have been combined with modelling
1. INTRODUCTION the systems' performance with present and possible
future interconnections, present plants and future CHP
Swedish district heating has a long history and is today plants, and finally with a hypothetical introduction of
one of the dominant heating forms with approximately natural gas. The aim of the paper is to explore barriers
55% of market share, and an annual energy production and driving forces for energy cooperation in the
of approximately 55 TWh[1]. The first system was built
288
th th
Stockholm district heating system and to analyse the them, although with some differences depending on the
potential for CHP generation in the system. company. Because a semi-structured interview is a
qualitative method, the possibility of using open
2. CASE STUDY questions is an advantage, and since we are interested
in a specific situation, the interviewees have the
There are three large district heating networks in chance to give their opinion. It also gives the
Stockholm that deliver more than 12 TWh of heat opportunity to analyze the answers in different ways, to
annually, produced in some 70 heating plants [5]. understand the opinions expressed [7].
Table I shows the heat production, types of base
production and installed heat and electricity capacity in 3.2 Modelling Stockholms district heating system
those networks. Six of the plants in the system are
Based on the data from Open district heating network
CHP plants with total installed electricity capacity of
in greater Stockholm [5] a model of Stockholms
about 600 MW, which gives a possibility for production
district heating system has been constructed.
of over 2 TWh of electricity annually [5].
Purchases and sales prices of electricity, taxes and
Table I. Major district heating networks in Stockholm. [5] tradable green certificates (TGC) are included in the
model (Table II) as well as the operating and
South- North- South-
maintenance costs for all plants and fuel prices.
central west east
However, due to agreements with the contact persons
Heat production in
from the district heating companies, the prices for fuel
the year 2005 [TWh] 9.4 2.2 0.53 are not presented in the paper.
Installed heat
Table II The average annual purchases and sales prices
capacity [MW] 4000 700 300
of electricity, including all taxes and TGC. [8], [9], [1]
Installed electricity
capacity [MW] 493 105 20 Current price of electricity [/MWh]
Base production CHP BCHP NGCHP Purchase Sale Sale with TGC
waste, included
CHP 70.10 35.46 67.56
coal
European price of electricity [/MWh]
Purchase Sale Sale with TGC
3. METHODS included
A combination of methods from social science and 83.30 48.65 80.77

technical science has been used; modelling with
MODEST and semi-structured interviews with Carbon dioxide emissions used in this paper are shown
representatives from the largest energy companies. in Table III [10]. However, since the greenhouse effect
MODEST is an energy-system optimisation model with is a global problem, carbon dioxide (CO2) emissions
time-dependent components that was developed at are not simply analysed from a local perspective but
Linkping University in Sweden. MODEST uses linear also in regard to a global perspective. The global
programming to calculate the most profitable emissions of CO2 (GECO2) of the system are calculated
combination of existing and potential new facilities and with the assumption that electricity produced in the
shows which investment options are financially plants is going to replace marginal power production in
viable [5]. the integrated European electricity market. Since coal-
fired condensing power plants have the highest
3.1 The structure of the interviews variable cost compared with other sources of electricity
in the EU, they work as the marginal power
The interviews were conducted during the spring of
production [11]. When assuming that the coal-fired
2009, with representatives from the five largest energy
condensing power plants have an electricity efficiency
companies producing and/or distributing district heating
of 33%, each megawatt-hour of electricity generated in
in the Stockholm region. These are Fortum, Norrenergi,
such a plant releases approximately one tonne of CO2.
Sderenergi, E.ON and Vattenfall. The representatives
According to that, any increase in electricity production
were chosen by the companies themselves, since they
in Stockholms district heating system can lead to
could better decide who would be most appropriate to
reduced production in the marginal coal condensing
answer questions regarding interconnections,
power plants, and consequently to a reduction of
cooperation and future strategies. We decided to let the
one tonne of CO2 emissions. However, it is necessary
respondents remain anonymous, since one of the
to mention that considering the EU Emissions Trading
interviewees wanted this. The interviews were semi-
Scheme (EU ETS), the decrease of CO2 emissions in
structured, as we had similar questions for most of
289
th th
electricity production sector does not necessarily to Since electricity generation will probably be the primary
lead to reduction of GECO2 [12]. But the marginal production in all district heating companies in the
electricity concept still has significance for future future, when the Swedish electricity price becomes as
measurement of and planning for future limitations of high as the typical European price, in scenarios 6-9 our
CO2 emissions and the future trading system. research focuses on the cogeneration potential in
Stockholm's district heating system. Scenario 1 has
Table III. Net emissions of CO2 [10].
been used as a reference scenario for scenario 6.
Fuel Emissions kg/MWhfuel Scenario 3, where the influences of a higher electricity
Oil 280 price on the system with the existing plants have been
Coal 330 analysed, has been used as a reference scenario for
scenarios 79. Scenarios 6-9 are analysed as possible
Waste 100
future cases that may exist more than 10 years from
Biomass 0
today. Because of that, all plants in the scenarios are
Electricity 950
new so the investment costs for all plants are
Natural gas 230 considered. While in the scenarios 6, 7 and 8 the
system consists of 31 CHP plants fuelled by solid
3.3 Description of chosen scenarios biomass (BCHP), there are a total of 46 CHP plants
fuelled by natural gas (NGCHP) in scenario 9. In
Nine different scenarios have been analysed
scenario 9 it is assumed that the natural gas network
considering the possible future cases (Table IV), with
exists along the Swedish east cost.
special attention to economic and environmental
aspects. The characteristics of the CHP plants that have been
integrated in the model of the district heating system in
The existing district heating system (scenario 1) and
scenarios 49 are presented in Table V [13].
the system with three new CHP plants that are planned
to bee built according to the interviews and documents
Table V. The characteristics of the new integrated CHP
(scenario 4) have been analysed. Since the base
plants in scenarios 49 [13].
productions in the networks differ, the differences
between the productions costs in different parts of the Technical characteristics
system are notable. Because of that, in both cases
Fuel Electrical Fuel *
(scenarios 1 and 4) the influences of a better output efficiency
connectivity between networks have been studied Sc.
(scenarios 2 and 5). MWe %
biomass 30 110 0.45
Table IV. List of the chosen scenarios.
45 waste 20 91 0.32
Plants biomass 80 110 0.46
in the
Sc. district Connectivity Electricity TGC 68 biomass 80 113 0.51
heating price 9 natural gas 150 89 1.41
system
Economic characteristics
1 existing existing Nordpool exist
2 existing one Nordpool exist Process Operating and maintenance
1
network plant cost
3 existing existing EU exist /KWe % of PPC /MWh fuel
4 + new existing Nordpool exist
2 745 1.5 2.45
CHPP
45 5 440 3 9.31
5 + new one network Nordpool exist
CHPP 2 110 1.5 2.45
6 BCHP one network Nordpool exist 68 2 110 1.5 2.45
7 BCHP one network EU exist 9 715 2.5 0.9
8 BCHP one network EU do not
exist * electrical/thermal output
9 NGCHP one network EU -
3.4 Previous studies
1) Interconnections between the south-central and the Two studies regarding Stockholms district heating
north-west networks have been introduced as well as system were done in the years 2005 [10] and 2006
interconnections between south-central and south-east [14], and the results showed that benefits for better
networks. Capacities for existing pipes have been connectivity between some parts of the system existed.
increased.
290
th th
It was also shown that if all plants in the system are 4. RESULTS OF THE SCENARIOS
replaced with BCHP plants, with an electricity-to-heat
The results from the scenarios are presented in
output ratio 0.46, up to 10TWh electricity can be
Table VI and Table VII.
produced and the potential for decrease of GECO2 of
the system would be 3 tons CO2 annually. If all plants According to the optimisation results, if better
in the system are replaced with NGCHP plants, with an connectivity is introduced, some economic benefits
electricity-to-heat output ratio 1.2, the electricity exist. In both cases the case with only existing plants in
generation in the system can increase to 11TWh and the system and the case where the new plants are
the potential for decrease of the GECO2 of the system introduced in the model (scenarios 2 and 5) the
would be about 5 tons CO2 annually. However, since decrease in system costs would be about 10 million
these two studies were done, a new connectivity annually. The potential for decrease of the
between networks has been built and the total installed environmental impact of the system is more notable. If
electricity capacity in the system has increased by 20% better connectivity were introduced in the system
[5]. Furthermore, new CHP technologies are constantly today, the biomass share in total fuel use would be 8%
being developed, which enable greater electricity higher and consequently both the local emissions of
efficiency and consequently greater benefits from CO2 (LECO2) and GECO2 of the system would be
economic, energy and environmental viewpoints. about 0.25 million tons lower annually. The potential for
decrease of GECO2 of the system if better connectivity
Regarding interconnection and cooperation of DH
is introduced after the building of new CHP plants
systems, some studies have been conducted in a
(scenarios 4 and 5) is 0.4 million tons annually.
Swedish context. However, none of them have focused
on cooperation between energy companies. They have Table VI. Results for the scenarios economic aspects.
instead focused on cooperation between energy
companies and industry.
Thollander et al. [15] found that technical aspects are
Annual CHP heat Electricity
seldom barriers to cooperation. The barriers are rather system
risk, different aspects of information during Sc. costs production The
negotiations, and other social factors such as inertia share income
among personnel. Driving forces have been economic Annual from
factors such as an aim for lower costs and means of production electricity
control, as well as environmental values. In a study with
a similar aim, Fors [16] found the same results, that million % TWh Million
technical aspects are seldom barriers. Information
1 258 47 2.30 122
during negotiations, stable contracts and the
importance of involving the personnel at the plants in 2 245 47 2.31 125
the process are important factors. It is also important
that the cooperation benefits both parties. Grnkvist 3 243 48 2.35 150
et al.[17] reached a similar conclusion in a study that
4 204 58 2.96 164
emphasises the importance of the willingness of people
on both sides to cooperate. The main advantages of 5 192 62 3.15 176
the cooperation are lower costs and benefits for the
environment, while the main disadvantages are less 6 403 424
flexibility as both parties work under contracts.
7 344 100 6.39 482
Historically, interconnection of technical systems has
been seen in the theory of Large Technical Systems as 8 546 281
one way for systems to grow. Systems start in a local
9 504 100 17.66 777
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
As the electricity price increases, the system would
other. Interconnection of systems can also be
earn extra income from the electricity sold, and thus the
explained through the fact that larger systems have a
heat production cost would decrease (scenarios 1, 3).
higher load factor and better economic mix [18], [19],
This gives an even bigger advantage to CHP
[20].
generation compared with pure heat production.
291
th th
Table VII. Results for the scenarios environmental also notable that in scenarios 6, 7 and 9 the annual
aspects. income from electricity is higher than the annual
system costs. However, since all plants in those
scenarios are new, the total investments are high.
Biomass LECO2 GECO2 of Because of that, if the analysed time period is just
share in the system 10 years, the annual system costs are much higher
Sc.
the then today.
system
The lowest GECO2 of the system are in the scenarios
% [million [million
where all plants in the system are BCHP (scenarios 6-
tons/year] tons/year]
8) and NGCHP (scenario 9) plants. In those two cases
1 48 2.50 0.32 GECO2 in Sweden, which is about 60 million tons
annually [21] would be reduced by approximately 9%
2 52 2.25 0.06
and 15% respectively, with the assumption that the
3 49 2.46 0.23 electricity produced would replace the marginal
4 52 2.12 0.69 electricity. LECO2 in the system is highest in the
5 55 1.91 1.08 scenario where all plants are NGCHP but at the same
time GECO2 of the system is lower because of the high
6
electricity production.
7 100 0 6.07
8 5. RESULTS FROM THE INTERVIEWS
9 0 7.80 8.98
In the following section the results from the interviews
will be presented. The interconnections between the
The income from the electricity sold in scenario 3 is systems make it possible to cooperate regarding heat
about 30 million higher then the income in scenario 1, production and distribution.
and because of that the system cost is 6% lower. The
difference between the electricity production in 5.1 The system today
scenarios 1 and 3 is not significant, but in spite of that,
The interviews show that the interconnections have a
the decrease of GECO2 of the system in scenario 3 is
historical background. Most of them were made during
almost 100%. The reason is higher biomass share in
a period when a regional energy company called
the total fuel used in the system in scenario 3, and
STOSEB (Greater Stockholm Energy Company)
consequently lower LECO2 in the system.
existed, where the municipalities, which to a large
The introduction of three new plants in the system extent owned the systems then, were represented. The
(scenario 4) would lead to a significant reduction of the main reason for the interconnections then was supply
heat production cost compared with the system today. security. When the systems were interconnected, the
The income from the electricity sold would be 35% companies could help each other during stops, and this
higher and, as a result, the annual system costs would is still the case. All representatives say this, and the
be 20% lower. This confirms that heat production in representative from Sderenergi expresses it this way:
CHP plants has a major influence on the economic
At the same time it is a common good. It is good that
efficiency of the district heating system. With the
the systems are interconnected. It is an extra security if
assumption that the electricity produced would replace
one plant should stop for some reason [22].
the marginal electricity in the European electricity
market, reduction of GECO2 of the system would be The advantages historically and foremost today are
almost 1 million tons annually. also economic. The emissions trading makes it
advantageous, since the companies can use the
If all plants in the system are BCHP (scenarios 68) or
production better by making capacity trades and even
NGCHP (scenario 9) plants, the annual electricity
out the production cost between the companies:
production would be as high as 4.5% and 12% of the
total electricity production in Sweden, which was about We see that we can use existing production more
145 TWh in the year 2008 [1]. The annual income from effectively. Most of the trades are a trade to mid-price
the electricity sold in those scenarios is much higher so to speak. You can say that we split the profit.
then the income from the electricity sold in the other Capacity trading (effektkp) is also common. Like we
scenarios. In the scenarios with typical European have here with Sderenergi, we have partly
electricity price, (scenarios 79), the income from the a production cooperation and partly we buy capacity.
electricity sold is 220%, 90% and even 420% higher They have more capacity than they need today [23].
then in scenario 3, where the system with the existing
plants is analysed with the higher electricity price. It is
292
th th
The main advantage with capacity trading is to avoid modelling, the systems are also already well
peak load, which often is oil-based, which is costly both interconnected:
for the fuel price but also because of the emissions.
Yes, the principal structure is already established. ()
Another point is that, as the Fortum representative
It is this connection, between the central and the
said, it is possible to even out effect between systems.
northwest system, it is the only one. That is not solved
One system may have cheaper base load than the
yet [23].
other, and for tax reasons it may be cheaper to buy
from the other than to use peak load. This particular connection would interconnect the two
main systems, and has been discussed in some
One factor that is pointed out for a successful
investigations [26], [27]. However, it is yet to be done.
cooperation is that both parties can benefit from it. As
This connection is most important for Fortum, as for
in all business, it is important that the cooperation be
example E.ON thought that it made little difference to
correct from a business standpoint and that both parts
them.
are satisfied [24].
The other main connection still missing is connection
The extent of cooperation varies between the
between the south system and Vattenfall's system in
companies. Some have more extensive cooperation
the southeast. Vattenfall thinks that the question has
with daily trades, like Fortum and Sderenergi, while
been raised on occasion, although never realized. They
others, for example Fortum and E.ON, do not trade
give no specific reason for this; they state that all
every day. In the latter case, they normally do not trade
cooperation is important and that different
as much during winter, although sometimes when peak
investigations have shown the advantages, although it
load is needed it is decided quickly [25]. Another
is difficult to quantify what it means practically [28].
advantage with the interconnections is that the
Stockholms Energi (now Fortum) previously owned one
companies can cooperate regarding revisions of the
of the plants, and there were plans to interconnect the
plants. While one company has revision during
systems then. Fortum gives no explanation for why the
summer, the other can produce for the other company.
interconnection has not been done earlier or now.
The factors that are seen as barriers are seldom
Although no direct comments regarding the lack of
technical. The companies think that the technical
interconnection were made, one of the interviewees
problems often can be solved while making the
who previously worked at Vattenfall said that there was
interconnection and at that point there is a need to
an opinion at Vattenfall that they prefer to keep to
negotiate certain aspects. For example, who provides
themselves, without interconnections, and should not
the electrical energy for the pumps and takes
work towards cooperation. Comments without a
responsibility for the regulation of the water pressure in
specific direction were also expressed in interviews that
the culverts and repairing the system in a joint part of
there was a lack of will to cooperate from some
the system? However, this is often solved:
companies. There is also a history of rivalry between
Yes, the other things we can handle while building the Vattenfall and the former Stockholms Energi [29]. It is
technical parts. At that point we hopefully have possible that this rivalry stills exists. Fortum also
identified all technical barriers so that they can be expressed opinions about the fact that other companies
taken into account. They should not appear during are building their own CHP plants instead of trying to
production. Settlement of account and such things, find regional solutions.
they are not a big problem although complicated.
However, it is nothing that makes you pass on a 5.3 Building CHP in the system
profitable cooperation [23]. As seen in the scenarios, in the near future in Sweden
In the above quote, we see one of the disadvantages many CHP plants are planned and will start to be built.
with today's cooperation, on which all the companies In Stockholm most of the companies have plans for
agree, and that is the settlement of accounts. It is CHP, and two of them have already built in the last
complicated to control the systems and the trades, and years, for example Igelsta (Sderenergi) and Jordbro
it requires staff to do so. (Vattenfall). Other companies are making plans, such
as Norrenergi, EON and Fortum. The reasons for
5.2 Barriers towards more co-operations building CHP are varied, but the most clear is that they
see economic advantages in selling electricity, and our
In the interviews the companies expressed satisfaction
stagnating heat load ahead. By selling electricity there
with the present cooperation. Few actual barriers as
is a possibility to keep profits high, even with a
such were expressed, except the ones that todays
stagnating heat load. The system is also relatively old
situation creates. For example, it is almost
and well established; the potential for further
geographically impossible to expand the systems to
connections are getting smaller as saturation in the
smaller systems nearby. As could be seen in the
293
th th
heating market for district heating makes it more actors say that they can optimise the system's
difficult to expand: performance, and our scenarios have shown that more
cooperation could benefit them even more
A rough rule of thumb has been that the expansion with
economically. Even though the gain is not extremely
new customers that have been, (...), has been eaten up
high, since the lower system cost would be
by the efficiency we could achieve together with the
approximately 5%, there is potential. However, since
customers in their buildings. So basically, the heat load
there seems to be reluctance to cooperate between
has been static in our area for quite some time. ()
some actors, it is difficult to fulfil the potential.
[The reason for building CHP] is the electricity. We,
as the producing company, have the problem that we Advantages with the cooperation are said to be a
can not expand. We have our two customers and possibility to even out the production in the system and
district heating is not a new thing in the municipalities thus avoid peak load. The disadvantages with the
so the chance of getting new customers is limited [22]. cooperation are the need for more administrative work
to control the system and the trades; the control of the
The other representatives are of a similar opinion, that
system becomes more complex. This study also
a stagnating load can be expected, and CHP is a way
confirms previous studies that have pointed out that
to keep profits high. The Swedish certificate system
technical aspects are seldom barriers to cooperation.
also makes it advantageous to build new bio-fuelled
Most things are solved while the systems are being
CHP-plants. Another reason, arguably of a more
interconnected, and the will of the persons involved to
rhetorical character, is that building CHP is more
cooperate is important.
economically and environmentally correct since the fuel
efficiency is higher with CHPs. As scenarios 48 show, There is a large potential in building new CHP plants,
there is major potential for reducing local and global both from an economic and an environmental
CO2 emissions. perspective. If all the plants in the system were
replaced by BCHP or NGCHP, the electricity produced
In the interviews we also asked questions about the
could make up to 4.5 or 12% of total Swedish electricity
possibility of an introduction of natural gas in the
production, based on the fact that total production in
region. Investigations have been made earlier by the
Sweden in 2008 was 146 TWh [1]. The reason that the
above mentioned STOSEB; however, the plans never
difference between the electricity productions in those
came to reality. Generally the representatives did not
two cases is so large is a big difference between the
think that an introduction would happen. Since most of
electrical/thermal outputs (see Table V). The intro-
them also have strategies to be climate neutral, natural
duction of NGCHP is a less likely future since it can be
gas probably is not an option. The large investments in
considered only with the assumption that the natural
infrastructure are another barrier:
gas network already exists along the Swedish east
These are such large infrastructure investments and cost. On the other hand, introducing more BCHP in the
natural gas is not especially cheap either. It is difficult district heating system would increase the systems
to come in with natural gas in this energy system. It is dependence on biomass availability and the heat
rather stable [23]. production cost would become highly sensitive to the
solid biomass cost. The actors are highly aware of the
What the representative here points at is also the
potential for CHPs. Since they are expecting a
inertia in the system. In LTS terms it is called
stagnating heat load, the sale of electricity is a way to
momentum: as the system is stable, it is difficult to
keep profits high. However, none of them think that
change the structure [18], [19].
natural gas will become a reality in the near future, and
even if it did, the introduction is expected to be
6. CONCLUDING DISCUSSION somewhat problematic, since the fuel can be
The study has shown the advantages of an considered fossil fuel and substantial infrastructure is
interdisciplinary approach. Advantages with needed.
interconnections and CHP have been shown in the The study has shown a potential for decreased LECO2
modelling; however, as there are many different actors and GECO2. The largest potential from a local
involved, there is a need for a will to cooperate. The perspective is from BCHP; so, since the LECO2 would
interconnections have a historical background, with an be low and with high electricity production, the potential
aim for higher supply security, and today most of them for lower GECO2 would exist. The high electricity-to-
continue to cooperate, despite the fact that the heat output ratio in NGCHP has a high potential for
structure and ownership of the companies in some decreasing GECO2 of the system. If all plants in the
cases have changed since the deregulation of the system would be replaced with NGCHP the GECO2 of
electricity market in 1996. As previous studies have the system would be -9 million tons annually. However,
shown, the main advantages with cooperation have in that case LECO2 would be much higher than today.
been economic, as is also the case in this system. The
294
th th
Modelling of a system gives one side of the truth, as [9] M. Melkerson and S-O. Sderberg, Dynamiska
does interviewing the actors involved. When combining elpriser elprissttning p en integrerad europeisk
the methods there is a possibility of getting a better and elmarknad (Dynamic electricity prices pricing in
deeper understanding of the actual potential for an integrated European electricity market in
cooperation. The historical and social aspects cannot Swedish), Sweden: Institute of Technology, Dept of
be neglected; they can in many cases explain why Mech Eng, Linkping University, Linkping (2004)
potentially beneficial cooperation is or is not done,
[10] N. Levinson and R. Freiman, Optimal kraftvrme
while modelling can show the actual potential.
och ntinvestering I Stockholms fjrrvrmesystem
(Optimal CHP and district heating network
7. ACKNOWLEDGEMENTS investments in Stockholm), Linkping University,
This article has been carried out in two PhD projects in Linkping (2005)
the Energy System Program, financed by the Swedish [11] Statens energimyndighet (Swedish Energy
Energy Agency. The authors would also like to thank Agency). Marginal elproduktion och CO2-utslpp i
Jenny Palm (Linkping University) and Louise Trygg Sverige (Marginal electricity production and CO2-
(Linkping University) for valuable comments on the emissions in Sweden, in Swedish), Swedish
paper. National Energy Administration, ER 14:2002,
Eskilstuna, Sweden. (2002)
8. REFERENCES
[12] E. Dotzauer, Greenhouse gas emissions from
[1] Statens energimyndighet (Swedish Energy power generation and consumption in a nordic
Agency), Energy in Sweden 2009. Swedish perspective, Energy Policy 2010, Vol. 38(2), pp.
National Energy Administration, ET 2009:30, 701-704.
Eskilstuna, Sweden (2009) [13] H. Hansson, S-E. Larsson, O. Nystrm, F. Olsson
[2] S. Werner, Fjrrvrmens utbredning och utveckling and B. Ridell, El frn nya anlggningar (Electric
(The development and expansion of district Power from New Plants), Elforsk, Stockholm
heating), Vrmefreningen, Stockholm (1989) (2007)
[3] European Union, Directive 2004/8/EC, Directive on [14] M. Danestig, A. Gebremehdin and B. Karlsson,
the promotion of cogeneration based on a useful Stockholms CHP potential An opportunity for
heat demand in the internal energy market, CO2 reductions? Energy Policy 2007, Vol. 35(9),
(2010), homepage: pp. 4650-4660.
http://www.managenergy.net/products/R81.htm, [15] P. Thollander and I-L. Svensson, Vgen till
2010-04-28 framgngsrika vrmesamarbeten en fallstudie
[4] European Union, , Directive 96/92/EC Second (Road to succesful heating co-operations a case
report to the Council and the European Parliament study), In: L. Trygg, L. et al (2009) optimala
on harmonisation requirements concerning fjrrvrmesystemi symbios med industri och
common rules for the internal market in electricity, samhlle, Rapport 2009:13, Svensk fjrrvrme
(2010), homepage: (Swedish district heating association) (2009)
http://eur-lex.europa.eu/LexUriServ/LexUriServ.do? [16] J. Fors, Spillvrme frn industri till fjrrvrment
uri=CELEX:32003L0054:EN:NOT , 2010-04-28 sammanfattning av intervjuer p 5 orter, (Excess
[5] B. Dahlroth, ppnade fjrrvrment i heat from industry to district heating systems)
Storstockholm (Open district heating network in Rapport 2004:5, Svensk fjrrvrme (Swedish
greater Stockholm), Stockholm, Sweden: district heating association) (2004)
Fastighetsgarna Stockholm, (2009) [17] S. Grnkvist and S. Sandberg, Driving forces and
[6] D. Henning, Optimisation of Local and National obstacles with regard to co-operation
Energy Systems. Development and use of the betweenmunicipal energy companies and process
MODEST model. Dissertation No. 559. Linkpings industries in Sweden, Energy Policy 2006, Vol. 34,
universitet, Linkping (1999) pp. 15081519.
[7] S. Kvale and S. Brinkmann, Den kvalitativa [18] T.P. Hughes, Networks of Power: Electrification in
forskningsintervjun (The qualitative research Western Society 1880 1930, John Hopkins
interview), Studentlitteratur, Lund (2009) University Press, Baltimore (1983)
[8] Nordpool. Nordpool electricity spot market, (2009), [19] B. Joerges, Large Technical Systems: Concepts
homepage: http://www. nordpool.com. 2009-12-18. and issues, In: R. Mayntz and T.P. Hughes, The
295
th th
Development of Large Technical Systems, [26] STOSEB, STOSEB 92 Energiframtider fr

Campus Verlag, Frankfurt (1988) Stockholms ln (STOSEB 92 Energy futures for
[20] J. Summerton, Changing Large Technical Stockholms County), STOSEB, Stockholm (1992)
Systems, Westview Press, Boulder, CO (1994) [27] Fortum & Stadsbyggnadskontoret, Mjlighets-
[21] SCB (Statistics Sweden). Utslpp av vxhusgaser, studie: Ntintegration Storstockholm, (Possibility
(2010), homepage: study: Netintegration in greater Stockholm) Fortum
http://www. scb.se. 2010-02-17. & Stadsbyggnadskontoret, Stockholm (2005)
[22] Sderenergi, Production manager, 090318 [28] Vattenfall, Head of business development and
Senior advisor, 090320
[23] Fortum, Site manager and Senior advisor, 090325
[29] STOSEB, 25 Energiska r Om Stor-Stockholms
[24] Norrenergi, Production manager, 030304 Energi AB (25 Energic years about greater
[25] E.ON., Group manager production, 090323 Stockholms Energy AB), STOSEB: Stockholm,
(2003)
296
th th
CUTTING COSTS OF DISTRICT HEATING SYSTEMS

BY USING OPTIMIZED LAYING TECHNIQUES
1 1
Alexander Goebel , Dr. Stefan Holler
1
MVV Energie AG, Mannheim, Germany
ABSTRACT MATERIALS AND METHODS

The soil covered plastic jacket pipe is the common The cost saving potentials of the alternatives, con-
state of the art laying technique in the district heating cerning the digging of the trench and the backfill, are
sector: A preferable shallow trench is dug out and mainly evaluated by outlining the results of research
backfilled with cable sand after the installation of the reports. Calculations are used in order to estimate the
two pipes. Alternative possibilities concerning the insulation properties of special backfill material. Also
digging of the trench, the backfill and the piping itself the cost saving potentials of pipes with low friction
are evaluated in this paper. Results show, that an losses are evaluated with simple equations.
optimized laying technique can save construction or
running costs under the right boundary conditions: RESULTS
Backfill materials with insulation properties can reduce
Reuse of the excavated material [1], [4]
the heat losses by about 25 %. Using glass-reinforced
plastic pipes (GRP) instead of steel pipes leads to Earlier research activities have proven, that plastic
pump energy savings of about 40 %. 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
INTRODUCTION
but also for the pipes. Field tests have shown, that
In the first place, excavation costs could be cut by there are promising money saving potentials because
digging smaller and shallower trenches. However, this of the significant reduction of transport and disposal
is only possible if the location of the construction site is costs. A consideration of reusing the excavated soil is
appropriate. In an urban area the situation is com- also reasonable from an environmental point of view.
pletely different from a rural area concerning space and
Following points are important, when it comes to an
regulations. The paper describes the boundary con-
evaluation of this possibility at an individual
ditions and compares different methods from the tech-
construction site:
nical as well as the economical perspective using the
example of the district heating system in Mannheim, the grading of the excavated material
Germany. sandy or cohesive ground
compacting properties
The second approach which will be presented in the
the friction between the ground and the jacket pipe
paper is the potential to reuse the excavated material
protection measures for muffles and the pipe
and to use self-compacting material when refilling the
underground construction regulations
trench. Furthermore, it is also possible to use new
a place for the storage of the excavated material
materials with better insulation properties in order to cut
(beside the trench, container or any place near the
down heat losses. In the paper the different properties
construction site)
of the new materials will be compared and evaluated.
formation of dust (especially in the summer)
A third possibility to reduce costs is the use of contamination
specialized piping systems wherever possible. pH-value
Nowadays a wide range of products is available on the an improvement of the excavated material with
market. In many cases a specialized system fits some lime (especially with cohesive ground)
applications better than a standard system does. Not a removal of the coarse material
only insulation properties but also compensation, a separation of the material, if the ground is split up
ductility and friction losses are important characteristics in different layers
of modern piping systems. In the paper it will be shown,
how costs could be reduced by using less or no The use of self-compacting material [1], [4], [5]
compensation measures (cold laying, flexible pipes, It is important to distinguish between the following two
fibre pipes), by avoiding welding measures (flexible types of self-compacting material:
pipes for house connections, fibre pipes) or by reducing
friction losses (fibre pipes). stabilised sand mix
297
th th
excavated material mixed with water and special An economical justification is only achievable, if the
additives in order to get a self-compacting transport costs are low and the heat price is high. From
behaviour a technical point of view, the compaction behaviour has
The use of self-compacting materials offers a wide to meet the requirements and regulations. The jacket
range of advantages and applications: temperature must not exceed the maximum of 50 C
and the friction between pipe and the material should
it is possible to dig out a narrower trench, because
be in the common range.
no machines are needed for the critical compaction
around the pipes A calculation method for heat losses of plastic jacket
the backfill process is significant faster after the pipes is described in EN 13941 ANNEX D [2]. Figure 1
trench is filled up, it takes normally only one day shows the influence of the thermal conductivity of soil
until the material is hard enough to walk on s on the heat losses. Normally the value of s lies in
self-compaction is more reliable within difficult between 1,0 and 2,0 W/m*K [2]. The curve becomes
conditions (many crossing pipes etc) very non-linear below a value of 1,0 W/m*K. This
without the use of compaction machines, buildings indicates that it is necessary to customise the
nearby the construction site are stressed less (no calculation method in order to get realistic results. The
vibrations) heat losses are cut down by 30%, if the s is reduced
there is less inconvenience for residents living from 1,5 to 0,35 W/m*K.
nearby the construction site, because of the noise 75
reduction
in combination with the pipeline laying technique, heat losses of the flow and return pipe
70
the sheeting can be omitted, because nobody 65

needs to work in the trench
f + r [W/m]
60
A common problem is the local availability of the
technology. The price is also an issue, if the reason of 55
the application is the approach to save money.
50
Another problem concerning the dimensioning of the
45
compensation measures is the bad predictability of the
friction between the jacket pipe and the self- 40
compacting material. Depending on whether the pipes 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
are taken into service during or after the hardening Thermal conductivity of the soil S [W/m*K]
time, which is about a month long, a more or less Fig. 1 Heat losses of a district heating line as a function of
crucial tunnel effect is observed [4]. s (DN 250, 120/50 C, Z = 0,6 m, C = 0,55 m,
The reuse of the excavated soil as base material is i = 0,03 W/m*K)
more elegant, than the stabilised sand mix, because of The insulation material should solely be integrated in
the recycling aspect. Research projects have even the calculation as an additional thermal resistivity
shown that sharp particles are less problematic, (R,embedment), since the soil around the pipes is not
because they are enclosed in the self-compacting made completely out of it. The heat dependency of the
mass. An advantage of the stabilised sand mix is the insulation foams thermal resistivity should also be
easier application. taken into account. Figure 2 illustrates, what is meant
If the district heating line does not run under a street, with additional insulation layer.
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. Fig. 2 The different layers of the heat conductivity problem
298
th th
The modified calculation procedure is made up of the

following equations: 2 (Z R ) 2

ln1
1
Rh 0 s
1 4 embedment DC s embedment
R (7)
Di (1)
1
Uflow
1 D Rs Ri ,flow Rh R R ,steel R , jacket R ,embedment
R ,steel ln o
2 steel Di
(2)
(8)
D 1
1 Ureturn
Ri ln insulation Rs Ri ,return Rh R R ,steel R , jacket R ,embedment
2 i Do (3) (9)
1 DC
R , jacket ln
2 C Dinsulation (4) i i ,50C 0,0001 Ti ,average 50K (10)
1 D 2 sembedment
R ,embedment ln C
Tflow Treturn
2 embedment DC (5) f r U flow U return Tsoil (11)
2
1 4 (Z R0 s )
Rs ln
2 s DC 2 sembedment Ti ,average Tfluid Q Ri R R,steel R, jacket (12)

(6)
180 without insulation material 30%

with insulation material
160
reduction in %
heat losses of the district heating line
140
25%
120 heat loss reduction
100
[W/m]
20%
80
60
15%
40
20
0 10%
0 200 400 600 800 1000
nominal diameter [DN]
Fig. 3 Reduction of the heat losses for DN 15 to DN 1000
The average temperature of the insulation was temperature was at 120 C and the return temperature
calculated with following equation and put back into was at 50 C.
(10). A VBA script was used to iterate five times.
Fig. 3 shows, that savings are significant lower with
Fig. 3 shows the results of the calculation. The small diameters. Also the specific thickness of the PUR
insulation material was taken into account with a value insulation, which differs because of standardised jacket
of 0,33 W/m*K (embedment). Around and in between the pipe diameters, has an impact.
flow and the return pipe a space of 0,2 m for each pipe
The heat losses of a DN 250 pipe are reduced from 67
size was chosen (sembedment). The depth of cover had a
to 51 W/m (24%). This means, that the heat loss
value of 1 m (Z). Like in the previous example, the flow
reduction is 6% less compared to Fig.1.
299
th th
Since the use of an insulating backfill is more efficient A common value of 6% was chosen for the required
with huge diameters, a DN 700 pipe was chosen for an rate of return (i).
example scenario. An annual average for the flow and
The net present value C0 was calculated with the
return temperature was taken into account. For the
following equation:
calculation of the required backfill volume in the
embedment, 0.2 m space in every direction of the pipes T
was estimated. It is important for the calculation to take

C0 I (Rt ) (1 i ) t
t 1 (13)
only the additional costs into account. That means the
price difference between cable sand and the insulation The internal rate of return shown in Fig. 4 was
material including the transportation costs. calculated with the IRR- function in Excel.
Table 1 Scenario for insulation material
3
Parameter Value Unit 3
17 /m
60,000 16 /m 10.0%
net present value (20 years) []

Length of the district
5000 m
heating line: 50,000
internal rate of return

8.0%
Nominal diameter: DN 700 40,000

6.0%
Average flow 16 /m
3
95 C 30,000
temperature:
NPV 4.0%
20,000
Average return
50 C
temperature: 2.0%
10,000 internal
3 rate of
Annual hours of 17 /m
8760 h return
operation: 0 0.0%
0.0% 1.0% 2.0% 3.0%
Heat price (at the time of growth rate of the heat price
15 /MWhth
the invest):
Fig. 4 NPV and IRR of the scenario defined in Table 1
Required volume of 3 depending on the additional specific costs of the insulation
1.85 m
insulation material: material.
3
Additional specific costs: 16 - 17 /m The results in Fig. 4 show, that the additional specific
costs should be below 17 /m3 in order to get a positive
Required volume for the 3 value spread, assumed that the required rate of return
9250 m
whole line: is 6 %. A reduction of specific costs of 5% (16 /m3)
results in an increase of the value spread by 1%. The
Heat loss with use of the
69.6 W/m net present value after 20 years rises about 10,000 .
material:
The growth rate of the heat price is difficult to predict,
Heat loss without use of
90.7 W/m but has an important influence within the given period
the material:
of 20 years. Presumably the heat price is mainly
influenced by emission trading, governmental subsidies
Energy savings: 23.3 %
and the development of the fossil fuel price.
Annual savings of the Other scenarios may estimate higher growth rates, but
924 MWhth
whole line: in order to get realistic results, the rate was varied from
0% to 3,5%.
Additional investment (I): 148,000 157,000
Table 2 gives an example of materials with low heat
Required rate of return: 6.0 % conductivity that could be interesting to use as backfill.
It is obvious to look for natural products, because of the
Time of cash flow: 20 a price and environmental regulations.
Annual growth rate of
0 3.5 %
the heat price:
300
th th
Table 2 Heat conductivity of different materials Flexible systems with a corrugated service pipe have
significant higher friction losses, which has to be taken
W/m*K into consideration (dimensioning).
hard plaster [9] When it comes to money saving potentials, the most
important properties of flexible pipes are the following:
3
600 kg/m 0.18
less welding measures
3 self-compensating
900 kg/m 0.30
less insulation work
1200 kg/m
3
0.43 less work concerning the monitoring system
less head access holes, because of the reduced
1500 kg/m
3
0.56 welding measures
less risk of leaks, because of less weld joins
light sediment natural stone [9] 0.85 faster laying of the pipes
porous rock, e.g. lava [9] 0.55 Pipes with low friction losses
Service pipes made of glass-reinforced plastic (GRP)
natural pumice [9] 0.12
have significant less friction losses than steel service
pipes. Because of their chemical resistance, GRP
bitumen [10]
pipes are used mainly in the chemical industry. It is
2100 kg/m
3
0.70 important do distinguish between filament-wound pipes
and centrifugally cast pipes. Because of the Poissons
as matter, 1050 kg/m
3
0.17 effect almost no compensations measures are needed,
if filament-wound pipes are used. Centrifugally cast
membrane, 1100 kg/m
3
0.23 pipes need to be compensated, but have an even
smoother inner surface, which means the lowest
expanded volcanic rock (perlite) [11] possible friction losses. Also the temperature
resistance is a little bit higher. The greatest problem of
3
loose perlite, 50 - 130 kg/m 0.07 GRP pipes is the fact that the service life is cut down
by high temperatures in combination with high
perlite compressed with filaments, pressures (derating factor).
3
0.06
170 - 200 kg/m Fig. 5 shows the possible savings, if a GRP pipe with a
surface roughness of k = 0,01 mm is compared to a
Thermosand [8] 0.33
steel pipe with a roughness of k = 0,2 mm.
The following equations were used:
Flexible pipes
The calculations of the Reynolds number:
Flexible pipe systems, which are defined in EN 15632
[3], are mainly distinguished by the material of the w d
service pipe: Re
(14)
plastic (e.g. PE-Xa, Polybuten)
The value of the kinematic viscosity () was taken with
copper
2.941*10-7 m2/s, the density () with 958.77 kg/m3
mild steel
(water with 100 C and a pressure of 10 bar) [7]
corrugated stainless steel
The pipe friction factor was calculated with the
Flexible pipes have a significant higher operating
following equation [6]:
pressure (16 or 25 bar) [3], if the service pipe is made
of metal. Also the maximum and continuous operating k 68
0,25
temperatures differ much. Because of this fact, 0,11

systems with plastic service pipes could normally not
d Re (15)
be used within huge district heating networks. In a The pressure loss was calculated with the following
smaller network with lower flow temperatures, which equation:
e.g. was built to distribute the heat of a small block
heating station, a system with a plastic service pipe L w 2 (16)
p
might have an application. d 2
301
th th
The equation for the pump power: CONCLUSION

A consideration of an alternative laying technique is
V p
Ppump usually worth the work, because the money saving
(17) 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
Equation (17) shows, that the correlation between of technologies/materials etc), whether a different
pump power and pressure loss is linear. The savings technique makes sense from an economical point of
are expressed as a percentage. They do not depend view.
on the diameter or length of the pipe, because only the
friction factor differs. Containing a bunch of alternatives in district laying
techniques, Table 3 gives a rough overview with a
47.5% 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
reduction of pressure loss
45.0%
and will be evaluated in future studies. When the word
42.5% alternative is used, it means every technical aspect,
which differs from the standard laying technique
40.0% defined in the abstract.
37.5% Table 3 Overview of alternative laying techniques
35.0%
++ highly recommended to take into
consideration from an economical point of
view
32.5%
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6
+ a closer look seems promising
flow speed [m/s]
0 an economical benefit can be achieved, if
Fig. 5 Pump energy savings of a GRP pipe special boundary conditions are given
Another important aspect of GRP pipes are the joints: - because of technological or economical
the pipe ends are glued together with a two component reasons, the use of the technology is not
adhesive, which is heated up for the curing process. recommended
This can be an advantage, because welding measures -- the technology is not available or can not
on a construction site are often problematic (lack of be applied reasonable under the given
space, wind). Statistics show that in most cases leaks boundary conditions
are caused by bad weld seams [12].
302
th th
pipeline laying technique
self-compacting material
reuse of excavated soil
insulation as a casting
combined laying with
other supply pipes
trench-less laying
insulating backfill
stacked laying
flexible pipe
GRP pipes
techniques
compound
nominal diameter
+ ++ 0 ++ + 0 + ++ - +
< DN 150
nominal diameter
++ -- ++ + ++ 0 ++ 0 ++ ++
> DN 150
new district heating line ++ + + ++ ++ ++ ++ - + ++
renovation measure 0 - 0 ++ ++ 0 -- ++ + --
construction site in an
+ ++ -- ++ 0 + + + 0 ++
urban area
construction site in a
+ + ++ 0 ++ ++ ++ + + 0
rural area
yet to be built housing

+ + ++ + + ++ ++ - + 0
estate
existing housing estate + ++ -- ++ + + + - 0 ++
NOMENCLATURE R,jacket insulance of the jacket pipe

s the coefficient of thermal conductivity for R insulance of the convective heat transfer
the soil inside the pipe
i the coefficient of thermal conductivity for R0 surface transition insulance
the PUR insulation
R,embedment insulance of the insulating material used
i,50 C the coefficient of thermal conductivity for in the embedment
the PUR insulation at 50 C
Di inner diameter of the service pipe
C the coefficient of thermal conductivity for
D0 outer diameter of the service pipe
the jacket pipe
Dinsulation outer diameter of the PUR insulation
steel the coefficient of thermal conductivity for
the service pipe DC outer diameter of the jacket pipe
embedmen tthecoefficient of thermal conductivity for Tflow flow temperature

the insulating material used in the
Treturn return temperature
embedment
Tsoil temperature of the soil
heat transfer coefficient
Ti,average average temperature of the PUR
Rh insulance of the heat exchange between
insulation
flow and return pipe
Tfluid flow or return temperature
Rs insulance of the soil
Uflow heat loss coefficient for the flow pipe
Ri insulance of the insulation material
Ureturn heat loss coefficient for the return pipe
R,steel insulance of the service pipe
303
th th
Z distance from surface to the middle of the REFERENCES

pipe
[1] Alexander Goebel, Alternative Fernwrme-
C distance between the centre lines of the Verlegesysteme, Mannheim (2010)
two pipes
[2] European Committee for Standardization,
sembedment space between the pipe and the trench EN 13941 ANNEX D, Brussels (2009)
wall
[3] European Committee for Standardization,
f + r heat loss per pipe pair EN 15632, Brussels (2009)
[4] Dipl.-Ing. Heinz-Werner Hoffmann, Dipl.-Ing.
C0 net present value Torsten Ghler and Dr.-Ing. Manfred Klpsch,
Fernwrmeleitungsbau mit Recyclingmaterial, MVV
I investment Forschungsbericht, Mannheim (2006)
Rt net cash flow (annual savings) [5] Dipl.-Ing. Heinz-Werner Hoffmann and Zoltan
i required rate of return Dioszeghy-Gnter, Neuartige Verlegetechniken fr
das Kunststoff-Verbundmantelrohr-System Band 1,
T given period MVV Forschungsbericht, Mannheim (1995)
t the time of the cash flow [6] Fratzscher et. al., Energiewirtschaft fr
Verfahrenstechniker, VEB Deutscher Verlag fr
Grundstoffindustrie, Leipzig (1982)
k surface roughness of the service pipe
[7] VDI-Wrmeatlas, Verein Deutscher Ingenieure,
Re Reynolds number Heidelberg (2006), Dba 5, Dba 13
w flow speed [8] KE KELIT Kunststoffwerk Gesellschaft m.b.H.,
d inner diameter of the service pipe Thermosand (Broschre), Linz (2006), p. 6
the kinematic viscosity [9] VDI-Wrmeatlas, Verein Deutscher Ingenieure,

Heidelberg (2006), Ded 12
the pipe friction factor
[10] VDI-Wrmeatlas, Verein Deutscher Ingenieure,
p pressure loss Heidelberg (2006), Ded 10
L length of the pipe [11] Heinz Schmid, Excel mit VBA in der
the density of the heating water Wrmetechnik, C. F. Mller Verlag, Heidelberg
(2008), p. 26
Ppump pump power
[12] Dipl.-Ing. (FH) Frank Espig, Schadensstatistik KMR

V 2007 des AGFW, article published in the
flow rate
EuroHeat&Power (2008), issue 10
pump efficiency
304
th th
ANALYSIS OF HEAT TRANSFER IN HEAT EXCHANGERS BY USING

THE NTU METHOD AND EMPIRICAL RELATIONS
1
O. Gudmundsson, O. P. Palsson and H. Palsson
Faculty of Industrial Engineering, Mechanical Engineering and Computer
Science Hjardarhagi 2-6, IS-107 Reykjavik, Iceland
ABSTRACT a steady state condition it has proven to be relatively

simple, since analytical and empirical relations can be
Heat exchangers are widely used in domestic and
derived for different heat exchanger types and used for
industrial applications involving transfer of energy from
all necessary calculation regarding time invariant
one fluid to another, for example in district heating
conditions, see e.g. [1]. If a dynamic operation exists it
systems. The wide usage underlines the importance to
becomes more complex to monitor the condition of the
have a good technique to detect if the effectiveness of
heat exchanger and more complex models are used,
an heat exchanger is diminishing. There are number of
see e.g. [2] and [3].
things that can cause diminishing effectiveness of an
heat exchanger, for example fouling, changes in fluid In this study, a mathematical model is used that has
properties as well as corrosion. In many cases the been developed to simulate accurately the temperature
fouling is a particular problem, for example when and flow transients in a cross flow heat exchanger. The
geothermal water is used. Geothermal water is very model is based on the finite volume method (FVM)
mineral rich which can cause serious fouling problems. where a mathematical representation of a general
The method presented in this paper is simple and easy cross flow heat exchanger is solved numerically. One
to use and can be used to detect a diminishing heat possible application of such a model is to generate
transfer coefficient in many types of heat exchangers, data that can be used to compare and tune more
in this paper the method is used on cross flow heat simple dynamic models based on either black box
exchanger. The method uses measurements of the methods or state space modelling. An important
inlet and outlet temperatures as well as the mass flows, application in this context involves methods to detect
these measurements are usually easy to gather under fouling in heat exchangers under dynamic operation.
normal operation. The method uses the well known Description of the model can be seen in [4].
Number of Transfer Units (NTU) method as well as
Fouling in heat exchanger can be categorized in the
empirical relations to estimate the overall heat transfer
following categories, precipitation fouling, chemical
coefficient, which is then statistically analyzed. The
reaction fouling, corrosion fouling, particulate fouling,
data used in this study was gathered from a simulated
biological fouling and freezing fouling. Usually fouling is
cross-flow heat exchanger where the overall heat
a combination of the categories. The fouling process in
transfer coefficient was gradually decreased to
heat exchanger can be described as a process where
simulate diminishing effectiveness of the heat
the separating metal inside the heat exchanger
exchanger. The conclusion of this study shows that the
accumulates deposits from the fluids. This is very
derived detection method can detect fouling based on
common and poses problems and results in reduced
the data from a simulated cross-flow heat exchanger,
efficiency of the heat exchangers. There are numerous
with a good accuracy and consistency. Further analysis
methods available to address the effect of fouling, see
on real data is scheduled.
[58]. Finally, decrease in the thermal efficiency of a
heat exchanger due to property changes in a working
INTRODUCTION fluid will have similar effect on the heat exchanger as
Heat exchangers are widely used in domestic and fouling.
industrial applications involving transfer of energy from There are number of ways to detect fouling but
one fluid to another. General classification of heat according to [9], classical methods involve
exchangers are parallel flow, counter flow and cross a) examination of the heat transfer coefficient,
flow. Their size and complexity can also vary greatly. b) simultaneous observations of pressure drops and
Their operating conditions can be classified into two mass flow rates, c) temperature measurements,
main classes, steady operation where mass flow and d) ultrasonic or electrical measurements and
temperatures are relatively constant and dynamic e) weighing of the heat exchanger plates. Methods
operation where mass flow and temperatures can vary ac) require the heat exchanger to be operating in
greatly with time. steady state condition, d) can only monitor local fouling
During operation it is important to have some and e) requires the process to be stopped. These
knowledge of the condition of the heat exchanger. For restrictions can be too strict or costly. Another
305
th th
approach is to model the heat exchanger and look for allowed to progress to a maximum of Rf=0.00033,
discrepancy between model predictions and what is which corresponds to 25% decrease in the overall heat
actually measured, see [10] and [4]. The method used transfer coefficient.
in this study falls into category a). To make the method
valid for dynamic operating conditions, empirical THE DETECTION METHOD
relations for the mass flow rates are furthermore used.
The fouling detection is done by estimating the overall
Although district heating systems usually operate in heat transfer coefficient, U, by using NTU relations and
relatively steady state it can be argued that methods monitor the means of U for shift that can be related to
that work well to detect diminishing efficiency under diminishing efficiency either because of accumulation
dynamic operation should work very well under steady of fouling or property changes of the working fluid.
state condition. NTU method is commonly known and a description of it
can be seen in [1].
DATA USED
It is known that effectiveness of a heat exchanger can
The data used in this study was the same data as was be calculated by
used in [4]. The data was generated by a simulator
representing an unmixed cross flow heat exchanger.
(1)
The advantage of using simulated data is that it is
possible to control when and how much fouling will The minimum fluid is the fluid that has the minimum
occur in addition to controlling the inlet temperatures value of the production of mass flow and specific heat,
and the mass flows. The data used had temperatures . Effectiveness for a unmixed cross flow heat
for the hot side in the interval [53, 67] C and the cold exchanger can also be calculated by the following
side [12, 27] C, the mass flow rates for the hot and relations of the effectiveness to NTU.
cold side were in the interval [0.30, 1.45] kg/s.
Description of the simulator can be found in [4].
Fouling
(2)
During design a heat exchanger is commonly designed
to operate under mild fouling by assuming a fouling In normal use, the overall heat transfer is usually
factor in the interval 0.0001 to 0.0007. According to [11] unknown and it is therefore not possible to calculate
and [12] there is usually an induction time before a NTU directly. It is therefore necessary to estimate NTU
noticeable amount of fouling has accumulated. In [13] it from the relation between NTU and the effectiveness.
is shown that the fouling will grow with increased rate The estimation is done by minimizing a score function
during the fouling period. Figure 1 shows the evolution with respect to NTU. The minimization
of the fouling factor from the time the heat exchanger was done by using the minimization routine fmincon in
starts to accumulate fouling until the simulation is Matlab, see [14].
stopped. A dimensionless time is used to make easy
The parameter NTU is defined by
comparison between different lengths of data series.
(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
Figure 1. Evolution of the fouling factor from the time where big variations can occur during operation, it is
hard to see shift in the overall heat transfer coefficient
The simulated data sets used in this study include 200 that can be related to diminishing efficiency in the heat
sets without fouling and 200 sets with fouling, the data exchanger. In [15] it is proposed to use empirical
sets are further divided equally between slow and fast relations of U to make a heat exchanger model valid
fouling. In the fouled cases the data set was without over a wide range of operating conditions. The heat
fouling for the first 25% and then the fouling factor was transfer coefficient can be written as
306
th th
The effect of the empirical relations can be seen in

Figure 2. It can clearly be seen that including the
(5)
empirical relations really helps to reduce the variations
in the overall heat transfer coefficient.
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) Figure 2. The figure shows the evolution of the number of

where y is the exponent of the Reynolds number. In [1] transfer units and the overall heat transfer coefficient with
and without the empirical relations.
it is recommended to use y=0.8 for turbulent flow,
which is expected in a heat exchanger. To detect fouling a CuSum chart is used, see [16]. The
It can be practical to normalize U with a reference CuSum chart was chosen since it is known to be
mass flow. effective to detect shift in mean values. When using
CuSum charts it is necessary to define two CuSum
The overall heat transfer coefficient according to the parameters, a decision limit to prevent false detection
reference mass flow and is similarly 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
(8) consistent in detecting diminishing efficiency. Figure 4
shows the detection if no empirical relations are used.
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 Figure 3. The CuSum chart quickly detects the shift in the
measurements more realistic. Measurement errors of overall heat transfer coefficient.
0.2 C were assumed on the temperatures and 12%
measurement errors to the mass flows.
307
th th
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
Figure 4. The CuSum chart quickly detects the shift in the
overall heat transfer coefficient. conventional methods, this method can detect fouling
in heat exchangers that are not operated in steady
Comparison of Figures 3 and 4 shows that it is possible state conditions. The fouling detection is performed
to detect fouling in heat exchangers operating in within the designed fouling factor interval.
dynamic condition with quite good accuracy by using
Further work will include application of the method on
the NTU method and empirical relations.
data from a real heat exchanger.
In Table 1 a comparison between the method in [4] and
the method presented in this paper is shown. From the ACKNOWLEDGEMENT
table it is apparent that the method presented in this
paper gives better results. The fouling detection interval This work has been supported by the Environmental
for the drift corresponds to fouling factors on the and Energy Research Fund of Orkuveita Reykjavkur,
intervals [0.00002, 0.00004] and [0.00001, 0.00003] National Energy Fund and Energy Research Fund of
respectively for the fast and slow fouling. The method Landsvirkjun.
is therefore giving considerable better results than the
method described in [4]. REFERENCES
Table 1: Comparison of detection time between the two [1] J. P. Holman Heat Transfer. Ninth edition, McGraw
methods, where method 1 is from [4]
Hill, 2002.
Method 1 Method 2 [2] M. Mishra, P. K. Das and S. Sarangi. "Effect of
Percentiles Fast temperature and flow non-uniformity on transient
behaviour of crossflow heat exchanger".
2.5% 0.59 0.26 International Journal of Heat and Mass Transfer,
50% 0.83 0.35 2008, p. 2583-2592.
97.5% 0.98 0.40 [3] H. Kou and P. Yuan. "Thermal performance of

crossflow heat exchanger with nonuniform inlet
Percentiles Slow temperatures". International Communications in
2.5% 0.63 0.23 Heat and Mass Transfer, 1997; 51(9-10):357-370.
50% 0.81 0.30 [4] O. Gudmundsson, H. Palsson and O. P. Palsson.

"Simulation of fouling in cross-flow heat exchanger
97.5% 0.93 0.35 and a fouling detection based on physical
modeling". In: Proceeding of The 50th Conference
on Simulation and Modelling, Fredericia, Denmark,
Typical fouling factors are, as stated above, on the 7-8th of October, 2009.
interval [0.0001, 0.0007]. The results therefore indicate
that the method can be used to detect fouling in cross [5] W. L. Pope, H. S. Pines, R. L. Fulton and P. A.
flow heat exchangers that are operating in non-steady Doyle. "Heat exchanger design "why guess a
state condition prior to the time a typical fouling factor fouling factor when it can be optimized?". Energy
heat exchangers are designed for is reached. The Technology Conference and Exhibition. Huston,
Texas, 1978.
308
th th
[6] A. Nejim, C. Jeynes, Q. Zhao and H. Mller- [11] B. Bansal and X. D. Chen. "Fouling of heat
Steinhagen. "Ion implantation of stainless steel exchangers by dairy fluids a review". In:
heater alloys for anti-fouling applications". In: Proceeding of Heat Exchanger Fouling and
Proceedings of the International Conference on Ion Cleaning Challenges and Opportunities, Kloster
Implantation Technology, 1999;2:869-872. Irsee, Germany, June 5-10, 2005.
[7] P. K. Nema and A. K. Datta. "A computer based [12] F. Fahiminia, A. P. Watkinson and N. Epstein.
solution to check the drop in milk outlet "Calcium sulfate scaling delay times under sensible
temperature due to fouling in a tubular heat heating conditions". In: Proceeding of Heat
exchanger". Journal of Food Engineering. Exchanger Fouling and Cleaning Challenges and
2005;71:133-142. Opportunities, Kloster Irsee, Germany, June 5-10,
[8] S. Sanaye and B. Niroomand. "Simulation of heat 2005.
exchanger network (HEN) and planning the [13] M. W. Bohnet. "Crystallization fouling on heat
optimum cleaning schedule". Energy Conversion transfer surfaces 25 Years research in
and Management. 2007; 48:1450-1461. Braunschweig". In: Proceeding of Heat Exchanger
[9] G. R. Jonsson, S. Lalot, O. P. Palsson and B. Fouling and Cleaning - Challenges and
Desmet. "Use of extended Kalman filtering in Opportunities, Kloster Irsee, Germany, 5-10th of
detecting fouling in heat exchangers". International June, 2005.
Journal of Heat and Mass Transfer, July, [14] MathWorks http://www.mathworks.com/. 20th of
2007;50(13-14):2643-2655. April 2010.
[10] O. Gudmundsson, O. P. Palsson, H. Palsson and [15] G. R. Jonsson and O. P. Palsson. "Use of
S. Lalot. "Fouling detection in a cross flow heat empirical relations in the parameters of heat-
exchanger based on physical modeling". In: exchanger models". Industrial and Engineering
Proceeding of Heat Exchanger Fouling and Chemistry Research, June, 1991;30(6):1193-1199.
Cleaning, Schladming, Austria, 14-19th of June,
2009. [16] NIST/SEMATECH e-Handbook of Statistical
Methods, April 30, 2009,
http://www.itl.nist.gov/div898/handbook/.
309
th th
HEAT LOSS ANALYSIS AND OPTIMIZATION OF A FLEXIBLE PIPING SYSTEM

1 1 2
J. Korsman , I.M. Smits and E.J.H.M. van der Ven
1
Liandon B.V.
2
ABSTRACT (PP) or Polybutylene (PB) improves flexibility. From

these, PE does not have adequate strength at higher
The object of this paper is to evaluate heat losses of a
temperatures and PP is rather stiff. This leaves PE-X
flexible PB-PE-PE piping system in the field, compared
and PB, of which the latter can be welded without
to a conventional rigid Steel-PUR-PE piping system.
difficulty. This is therefore the material of choice for this
The flexible system is optimized in both insulation
study. In accordance with the temperature duration
quantity (thickness) and quality.
profile mentioned in the BRL5609/EN15632, PB is
The heat loss for pairs of pipes in the field, with 70 C suitable up to a maximum temperature of 95 C.
supply and 40 C return temperature, is based on heat
As with other plastics, PB is prone to some diffusion of
loss measurements in the laboratory and has been
oxygen and water vapor. These effects have been
evaluated using the multipole method.
investigated by Korsman et al. 2008 [7]. To prevent
Since the hydraulic properties of Polybutylene and oxygen diffusion, an EVOH oxygen diffusion barrier
steel medium pipes differ, hydraulic calculations of a may be used. Unless fully submerged for years, the
demonstration distribution network, fitted with either diffusion of water vapor will not be much of a problem.
system, are made. When kept completely under water, it will take at least
30 years for all cells to fill with condensate.
Total system heat losses for this demonstration
network are calculated by summing the product of the It is not all that easy to compare the heat loss of a
heat loss per pair of pipe and the amount of pipe used. conventional SteelPURPE piping system to a flexible
PBPEPE piping system. Internal diameters differ, as
INTRODUCTION does the friction coefficient, because PB is smoother
than steel. A pipe for pipe comparison yields skewed
Flexible piping systems for district heating and cooling results. One way around this problem is to compare
have several advantages when compared to rigid complete distribution systems, as was done in
piping, mainly during installation, some even in use. Korsman et al. 2008 [2]. A demonstration (or reference)
Flexible pipe can be utilized much like cable, arrives on network is used to design and compare similar
large reels, requires less engineering and fewer has networks. For reference purposes, the same network is
joints. However, flexibility comes at a price. It seems used in this study.
harder to reach comparable levels of insulation, see
Smits et al. 2010 [1]. The differing properties that complicate comparison
between piping systems, can also be used to minimize
The reason for this lies in the specific properties of the distribution system heat loss. The object of this study is
material most commonly used for insulation: to reach comparable heat loss for the flexible system,
Polyurethane foam. PUR foam has a crystalline by exploiting specific properties, whilst transporting the
structure and tends to be quite rigid. It is not very same amount of heat with comparable pressure losses.
suitable for flexible applications. Bending may lead to a
breakdown of the crystalline structure and may also
1. HEAT LOSS IN THE GROUND
compromise the bonding between foam and medium
pipe, thus creating a channel. This channel may Heat losses have been measured on test rigs as
accelerate the exchange of foaming agent and air with described by van Wijnkoop et al. 2010 [3] and have
the environment, thereby speeding up the ageing been evaluated by van der Ven et al. 2010 [4].
process. Flexible variants of PUR are available, but do
With the results of these tests, the in-ground heat
not seem quite as good. Insulation foam made of
losses are calculated using the multipole method by
polyolefins show ample flexibility and quite good
Johan Cleasson and Camilla Persson in 2005 [5]. Note
insulation properties for small diameters. Furthermore,
that the mentioned heat losses are calculated for a pair
aging typically is a faster process than in rigid systems.
of pipes, run at 70 C supply and 40 C return
As for the medium pipe, metals may be flexible enough temperature.
for the smaller diameters, but are too rigid for the
For rigid piping, some room between the pipes is
bigger pipes. Again, using polyolefins like Polyethylene
required for welding, see Fig. 1a.
(PE), cross linked Polyethylene (PE-X), Polypropylene
310
th th
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.
PB std insulation thickness PB close
50
45
40
35
Heat Loss [W/m]

30
Fig. 1a, Supply and return pipe 25
For flexible systems, there is no space requirement 20

between the pipes for installation purposes, and pipes 15
are best installed right next to each other, see Fig. 1b. 10
5
0
16 20 25 32 40 50 63 75 90 110
Nominal diameter [mm]
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
Fig. 1b, Supply and return closely together are different between Steel-PUR-PE and PB-PE-PE
When supply and return are installed closely together and therefore will lead to a different selection of
the temperature profile in the ground is altered in diameters in the engineering process, an approximate
benefit of the in-ground heat loss. comparison can be made, see Fig. 4.
Dependant on the refill, it may be difficult to achieve

St std PB close
defined compaction when supply and return are
installed too closely together. However, similar 50
temperature effects can be reached by installing 45
likewise in vertical orientation. 40
As mentioned before, the in-ground heat losses are 35
Heat Loss [W/m]
calculated using the multipole method by Johan 30

Cleasson and Camilla Persson in 2005 [5].
25
See Fig. 2 for a calculated temperature profile. 20
15
10
5
0
16 20 25 32 40 50 63 75 90 110
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
Fig. 2, Temperature profile supply and return closely smaller diameters does not create a heat loss
together in the ground reduction.
311
th th
4. SELECTION OF INSULATION THICKNESS The graph in Fig. 6 shows that the modified range (in
blue) is much closer to the reference (Steel-PUR-PE, in
The reason for the relatively high heat losses for the
purple). Heat loss is (almost) proportional to the
two smallest diameters is explained by Fig. 5.
diameter, which seems about right. However, the
modified range of insulation thickness in Fig. 5 (purple)
PB std insulation thickness PB extended insulation thickness
still shows a somewhat erratic distribution, which
45 suggests that the current range of customary
40 dimensions for the outer casing does not lead to an
optimal distribution of insulation thickness. It may prove
Thickness insulation [mm]
35
worthwhile to develop a new range of outer
30 dimensions, adapted to the diameter of the medium
25 pipe.
20
5. IMPROVEMENT OF INSULATION QUALITY
15
Due to the new testing facilities described by van
10
Wijnkoop et al. in [3], the process of product
5 improvement has been speeded up considerably.
0
During the course of the investigations, resulting in this
16 20 25 32 40 50 63 75 90 110 paper, it is becoming clear that further improvement of
Nominal diameter [mm] the insulation quality is feasible. The measurement
principle used for the determination of heat losses does
Fig. 5, Insulation thickness, standard and increased not allow for direct measurement of the insulation
properties of the foam; however, some sort of
In red, this graph contains the insulation thickness of
equivalent lambda can be derived from the data by
the current PB-PE-PE product range. The somewhat
calculation. As explained by van der Ven et al. in [4],
erratic distribution of insulation thickness over the
insulation quality differs for different diameters. For
range is caused by the use of customary dimensions
production reasons, it is not expected that insulation
for the outer casing. It can be seen that for the two
quality will reach the same level over the entire product
smallest diameters the insulation is rather thin, which
range. Typically, the higher values will be reached in
explains the relatively high heat losses.
the smaller dimensions. Still, an educated guess can
In purple, the graph in Fig. 5 shows a modified range, be made as to which levels are feasible from a
with increased insulation thickness for some of the technical viewpoint, see Table 1.
smaller diameters, as in general it is easier to achieve a
Table 1, Improved insulation quality, equivalent or
good insulation quality for the smaller dimensions. The
synthetic lambda at 50 C mean temperature
heat losses of the modified range were calculated and
are presented in Fig. 6.
Type Area Lambda fresh Lambda Degassed
50A25 1074 0.0283 0.0326
St std
63A32 1701 0.0287 0.0330
PB std insulation thickness
PB extended insulation thickness 75A40 2364 0.0291 0.0335
50 90A50 2993 0.0295 0.0340
45 90A40 3670 0.0345
0.0300
40
90A32 3886 0.0301 0.0346
Heat Loss [W/m]
35
30
125A63 6204 0.0316 0.0363
25 160A90 10790 0.0345 0.0397
20 160A75 12611 0.0357 0.0411
15 200A110 16879 0.0385 0.0442
10
5 Please note: The lambda values in Table 1 are not the
0 measured lambdas of samples of the insulation foam,
16 20 25 32 40 50 63 75 90 110 and may not be interpreted as actual physical
Nominal diameter [mm] properties of the insulation material. The values were
Fig. 6, Heat loss per pair, including increased insulation calculated on the basis of heat loss measurements of
thickness sections of pipe according to EN15632, and therefore
are some sort of synthetic system lambdas.
312
th th
The values presented do of course largely depend on The red graph in Fig. 7 represents the predicted heat
the actual physical lambdas (in W/m.K) of the insulation loss values for the combined effect of both increased
material, but the underlying measurement data suggest insulation thickness and insulation quality
that other factors come into play as well, such as the improvement. For most diameters, these are on par
geometry of foam in combination with the temperature with or slightly better than the reference in Steel-PUR-
dependence of the physical lambda of the foam. PE. These data are valid only for the recently produced
Therefore, the values in Table 1 are valid only for or fresh product. As there is no experimental data
calculation / prediction purposes, in exactly the same available on the rate of degassing and therefore the
calculation model from which they were derived (also rate of ageing, it is difficult to predict heat loss over the
according to EN15632). The values in Table 1 are life time of the product.
supported by experimental data on four samples at the
However, it is possible to speed up the process of
time of writing this paper.
ageing artificially, until all the foaming agent has been
When all parameters are known, equation 1 can be replaced by air. The predicted values for this condition
used to calculate heat loss: are also presented in Table 1, as lambda degassed.
These are synthetic as well, and suitable for

2 Tprobe Tcasing
i (1) calculation purposes only. Calculated heat loss results
with these values are presented in Fig. 8.
1 d2 1 d3 1 d4
ln ln ln Pump I, power: 16 kW
s
d1 i d2 c d3 St std PB impr fresh PB impr degassed
Where: 45
14 %
Tprobe, Tcasing represent probe (medium) and casing 40
temperature 2%
35
d1 to d4 represent inner/outer diameters of service
Heat Loss [W/m]
30
pipe and casing
s, i, c = heat coefficient of service pipe, 25
insulation and casing 20
In this case, s and c are known: s = 0.19 W/m.K 15

and c= 0.40 W/m.K. On a test rig, T probe, Tcasing and 10
84 %
heat loss are measured, so for specific test samples,
5
eq. 1 can be used backwards to calculate synthetic
values for i in Table 1. 0
16 20 25 32 40 50 63 75 90 110
For the Steel-PUR-PE reference, see [1], values can Nominal diameter [mm]
be determined in a similar fashion. Synthetic i Fig. 8, Heat loss per pair, including improved insulation
values for PUR foam, determined from quality, fully degassed
measurement of samples, were typically in the range In Fig. 8, the purple graph represents the reference,
of 0.030 to 0.032 W/m.K, with s for steel 50 W/m.K Steel-PUR-PE as measured, see Smits et al. 2010 [1].
The red graph represents the prediction of improved,
St std fresh PB-PE-PE and green the prediction of fully
PB std insulation thickness
PB extended insulation thickness degassed PB-PE-PE. The values vary a bit, but are
PB impr. fresh
generally in the same range. During the lifetime of the
50
product, heat loss is expected to increase from the red
45
40
values to the green values.
Heat Loss [W/m]
35 Of course, ageing is also applicable to the reference

30 product, but not included here for two reasons. First,
25
the ageing process for rigid systems is expected to be
20
significantly slower than for flexible systems, and
15
second, the reference samples were not fresh, as could
10
be judged by the gas content. Therefore, it is not likely
5
that the values presented for the reference system will
0
16 20 25 32 40 50 63 75 90 110 deteriorate much further during lifetime.
Nominal diameter [mm] Ageing can be slowed down considerably if measures
Fig. 7, Heat loss per pair, including improved insulation are taken to prevent the exchange of blowing agent
quality with the environment. If successful, these measures
313
th th
would result in the red heat loss values during The graph in Fig. 10 represents the pressure in the
lifetime. Moreover, a new generation of blowing agents supply network (in m water column), as a function of
is under development. These new agents aim at lower the distance from the source. For standard symmetrical
conductivity values for the gas and larger molecules. networks, the return network is similar, but mirrored
This may result in lower conductivity values for the over a horizontal axis.
product as well as a slower ageing process.
Using the flexible and smooth PB pipes allows for
smaller diameters, mainly because PB is less prone to
6. HYDRAULIC CALCULATIONS the transmission of hydraulic noises. This is due to the
The pipe per pipe comparison between Steel-PUR-PE low modulus of elasticity of PB when compared to
and PB-PE-PE as demonstrated in Fig. 8, gives an steel. In contrast, a steel pipe filled with water is quite a
indication of field results, but is not conclusive. Internal good conductor of sound. To prevent noise caused by
diameters differ, as do friction coefficients. Therefore, high flow velocities, these are limited in the design for
for the comparison between distribution systems fitted steel networks to 1 m/s.
with either pipe, hydraulic calculations are needed. To A network, specifically designed for PB, is shown in
this end, a reference network is introduced in Korsman Fig. 11. Smaller diameters in the periphery of the
et al. 2008 [2]. The same network is used here. It is network as a result of a higher permitted fluid velocity
installed in a housing estate near Arnhem, the causes higher pressure drops. This has to be
Netherlands, and has been designed using Pipelab, compensated by bigger pipes closer to the source to
developed by Prof. Dr. Pll Valdimarsson in 1995 [6]. reach the same overall pressure drop.
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. 11, Design pressure drop PB network
In the design of district heating networks, the maximum

Fig. 10 shows an output graph of Pipelab.
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
Fig. 10, Design pressure drop steel network calculation is shown in Fig. 12, which can be compared
to Fig. 10.
314
th th
(1) ST.PUR.PE Ref (2) PB as measured std eng.
1200
1000
Pipe Length [m]

800
600
400
200
0
16 20 25 32 40 50 63 75 90 110
Fig. 12, Pressure drop steel network with increased Nominal diameter [mm]
service pipe flow
Fig. 14, Pipe length histogram Steel and PB
There is not a lot of difference between both graphs in
Fig. 10 and Fig. 12. The reason for this is that the As a result of the use of smaller diameters with PB, the
design maximum fluid velocity is rather low for steel. distribution of pipe lengths generally shifts to the left in
the pipe histogram. As heat loss increases with
This may prove different for the PB network, which is diameter (see Fig. 8) this should have a positive effect
designed with smaller diameters in the periphery. See on the total distribution system heat loss.
Fig. 13, which can be compared to Fig. 11.
This shift to the left may be taken one step further,
The graph in Fig. 13 indeed shows an increased since the wall thickness of the smallest PB medium
pressure drop in the service pipes connecting the pipes currently is chosen a bit larger than the strength
houses, when the flow in those pipes is artificially class (SDR11) requires. This is done for ease of
increased to 100% of the installed power. However, the installation. If the thickness of these pipes is chosen no
total pressure drop stays within the same limits as does larger than SDR11, there is a slight additional shift to
the steel network under similar conditions (Fig. 12). the left, see Fig. 15.
(1) ST.PUR.PE Ref (2) PB as measured std eng.

(3) PB impr. Fresh all SDR11
1200
1000
Pipe Length [m]
800
600
400
200
Fig. 13, Pressure drop PB network with increased service 0

pipe flow 16 20 25 32 40 50 63 75 90 110
The result of both design calculations is plotted in
Fig. 14, steel in red and PB in green. Fig. 15, Pipe length histogram steel, PB and PB SDR11
networks
315
th th
6. TOTAL SYSTEM HEAT LOSS There is currently no experimental data on the rate of
ageing of PB-PE-PE as a result of the exchange of
To calculate the total system heat loss, pipe lengths as
blowing agent with air. However, it is possible to
shown in the pipe length histograms are to be
calculate a worst case situation (see fig. 18), using the
multiplied by the respective heat losses per pipe pair,
predicted values plotted in fig. 8.
as shown in the heat loss value histograms.
Fig. 15 multiplied by Fig. 6 leads to the total system
heat losses in Fig. 16. ST.PUR.PE Ref
PB impr. insulation thickness/quality
PB impr. Insulation thickness/quality degassed
ST.PUR.PE Ref
70.0
PB as measured std eng.
PB increased insulation thickness
60.0
PB increased insulation thickness all SDR11
80.0 50.0
Heat Loss [kW]

70.0 40.0
60.0
30.0
Heat Loss [kW]
50.0
40.0 20.0
30.0 10.0
20.0
0.0
10.0
system [-]
0.0 Fig. 18, Total system heat loss, including worst case
system [-]
Fig. 16, Total system heat loss, current insulation quality In practice and over time, the predicted total system
heat loss will slowly shift from the fresh value in purple
The graphs in Fig. 16 show the reference heat loss for to the worst case value in blue. Average heat loss
steel-PUR-PE in red and the currently measured heat during lifetime will be somewhere in-between.
loss for PB-PE-PE with non-optimized insulation
thickness in orange. In green, total system heat loss is
FUTURE RESEARCH
shown for current insulation quality but with optimized
insulation thickness. The exclusive use of SDR11 Hydraulic calculations in combination with insulation
(blue) has a rather small effect. thickness form an interesting optimization problem:
what diameter to select and which insulation thickness
Improving insulation quality, as described in paragraph
to choose?
5 and shown in Fig. 7, leads to slightly lower total
system heat loss for freshly produced PB-PE-PE when Current design strategies for hydraulic networks,
compared to the reference, see Fig. 17. aiming at linear pressure drop with distance, seem too
adventitious to be optimal. In addition, heat loss
ST.PUR.PE Ref PB impr. insulation thickness/quality calculations using standard casing dimensions show
60.0 rapidly diminishing yields with each step up in
insulation thickness, suggesting the optimum is
50.0 somewhere in-between.
First attempts have been made to use Pipelab [6] in a
40.0
double optimization routine, trying to find optimal
Heat Loss [kW]
hydraulic performance in combination with optimal

30.0
insulation thickness distribution over the network.
20.0 Given the specific hydraulic properties of PB (high fluid
velocities permitted) and the specific insulation
10.0 properties of PE foam (better at small size), this may
lead to rather different design strategies when
0.0 compared to conventional rigid piping systems for
system [-] district heating and cooling.
Fig. 17, Total system heat loss, reference and prediction
for improved insulation quality
316
th th
FURTHER INFORMATION REFERENCES

Questions concerning the paper can be addressed to: [1] I.M. Smits, J.T. van Wijnkoop, E.J.H.M. van der
Liandon B.V. Ven, Comparison of competitive (semi) flexible
www.liandon.com piping systems by means of heat loss
www.pipelab.nl measurement, in Proc. of the 12th International
Dijkgraaf 4 Symposium on District Heating and Cooling,
6920 AB Duiven Tallinn, Estonia (2010).
The Netherlands [2] J. Korsman, I.M. Smits, S. de Boer, System
Thermaflex International Holding B.V. optimization of a new plastic piping system, in
www.thermaflex.com Proc. of the 11th International Symposium on
Veerweg 1 District Heating and Cooling, Reykjavik, Iceland
5145 NS Waalwijk (2008).
The Netherlands [3] J.T. van Wijnkoop, E.J.H.M. van der Ven,
Verification of heat loss measurements, in Proc.
CONCLUSION of the 12th International Symposium on District
Heating and Cooling, Tallinn, Estonia (2010).
Flexible Polybutylene piping, insulated with
Polyethylene foam is a recent development, leaving [4] E.J.H.M. van der Ven, R.J. van Arendonk, Heat
ample room for product improvement. Experimental loss of flexible plastic pipe systems, analysis and
data shows rapid improvement in heat loss optimization, in Proc. of the 12th International
performance. Symposium on District Heating and Cooling,
Tallinn, Estonia (2010).
By optimizing both quantity (thickness) and quality of
flexible Polyethylene foam, and by using the specific [5] J. Claesson, C. Persson, Steady-state thermal
hydraulic properties of Polybutylene piping, the heat problem of insulated pipes solved with the
loss performance of conventional rigid steel piping multipole method, Chalmers University of
systems insulated with polyurethane foam, is within Technology, Report 2005:3. (2005)
reach.
[6] P. Valdimarsson, "Graph-theoretical calculation
If the current rate of improvement of the PB-PE-PE model for simulation of water and energy flow in
flexible piping is maintained, total distribution system district heating systems", in Proc. of the 5th
heat losses will be comparable to conventional rigid International Symposium on Automation of District
Steel-PUR-PE piping. The evident benefits of flexibility Heating Systems, Helsinki, Finland. (1995).
would become available without the current heat loss
[7] J. Korsman, S. de Boer, I.M. Smits, Cost benefits
penalty.
and long term behaviour of a new all plastic piping
system, IEA DHC|CHP Annex VIII research report
ACKNOWLEDGEMENT (2008)
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.
317
th th
FREE OPTIMIZATION TOOLS FOR DISTRICT HEATING SYSTEMS

1 1
Stefan Gnchtel , Sebastian Gro
1
Institute of Power Engineering, Technische Universitt Dresden, 01062 Dresden, Germany
ABSTRACT modify the system configuration to check how the

system reacts under new conditions and how the
At the Technische Universitt Dresden, Institute of
Power Engineering, Chair of Power Systems operating costs will change. In all cases the tool gives
Engineering as part of the project LowEx Fernwrme valuable information.
Multilevel District Heating [1] supporting by the
So FreeOpt provides help for any municipality that
Federal Ministry of economy and technology (FKZ
need a first guess on the feasibility and operating costs
0327400B), two public available and cost free software
of a new district heating network or who need to
tools have been developed, which enable the user to
improve the operation of an existing one. And of course
find
it does not matter if the network is supposed to be
A) the optimal unit commitment of district heating extended or build from scratch or an existing system
generators: FreeOpt just to be analysed.
B) the optimal pipeline route with the optimal pipe

diameter of district heating networks: STEFaN How does the software tool work?
First the simulation time is divided into time steps. For
at a minimum of costs. Both software tools are very every time step several variables (i.e. generated heat
easy and intuitive to handle. Lead time required to and electric power, amount of fuel, energy
learn how to operate the programs is short. Both tools consumption) exist inside given boundaries as well as
intend to support general design decision pro district individual costs and proceeds are stated (i.e. for fuel,
heating systems. In this paper an overview on their CHP refund). Different kinds of generators can be
advantages and fields of application, as well as chosen like heat plants, combined heat and power
example calculations are presented. plants, solar thermal plants or heat pumps. Hot water
storage tanks and electricity transferred in or out the
PROGRAM FREEOPT grid via contracts improve the flexibility of the whole
system (Fig. 1 ).
FreeOpt is an optimization tool to find the optimal unit
commitment of district heating generators for a given System boundary
time domain. Optimal decisions are found to minimize

total costs related to thermal and electric loads. Block heat
and power
SQfuel
Local district heat networks are becoming more plant
Power
common, so it is important to know how to operate demand
Combined Pdem
even small systems in terms of minimal costs and heat and
highest efficiency. When should which generator be power
plant
switched on or off? How to handle the storage? Which Grid
connection
influences have contracts for electric power? Heat
Qdem
plant/
boiler
There already is a lot of existing software for unit
commitment. As commercial and generalised software Heat
demand
Solar
for large systems it is mostly very expensive. So thermal
FreeOpt has been developed for any cases of district system
heating networks for most efficient operation of all heat

and power generators. Mainly operators of smaller Heat Heat
supply areas purchasing an expensive software pump storage
solution would be uneconomical can reach monetary

savings.
A stable version of FreeOpt is already finished. The Fig. 1 System boundary and interaction plan
programs power is demonstrated with a simple
The modular design of FreeOpt allows to form easily
example determining cost optimal operation of a
any generator system. All one has to know are the
specific district heating network. Furthermore with the
respective figure lines and parameters for all available
help of parameters and figure lines it is very easy to
generators and the network. After fixing the demand for
318
th th
heat and electric power as well as the parameters the The most important ones are the two balance
optimization problem is defined automatically. FreeOpt equations for power (2) and heat (3).
calculates the minimal costs to satisfy the demand. For
testing purpose there is no limitation of the values.
(2)
The mathematical model belongs to the mixed integer
stands for the generated electric power, for
problems. So the objective function and all constraints
are linear, all variables are continuous or discrete. It is power transferred in or out the grid via contracts,
written in the mathematical modelling language GMPL for the electric power demand and for the own
[2] and solved with the COIN-OR brunch-and-cut solver
consumption of electric power.
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 (3)
boundaries, cost coefficients, starting values, demand
values or figure lines in a very easy way. The internal stands for the generated thermal power, for
data flow between user interface, optimization model the heat demand, the heat disposed in the auxiliary
and solver is realised with help of txt-files (Fig. 2).
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. .
Electricity network
Fig. 2 FreeOpt user interface
Running the generators cause costs. Therefore the Heat

main aim is to minimize the total cost Heat
boiler
plant CHP
CHP1 CHP 2 CHP 3
storage
storage
(1)
in which are the following operating costs (and

proceeds):
costs for fuel Fig. 3 Flow scheme of a local heat network
costs or proceeds for transferred
electricity The heat demand of the customer is provided by
costs for start-up procedures 3 CHPs (Block Heat & Power Plants) base
costs for network access load
Heat plant (natural gas) peak load
costs for maintenance
Heat storage used for optimization
costs for CO2-cerfiticates
and is given for every hour of one year.
CHP-refund
EEG-refund The electricity demand is not directly considered
because of intern clearings inside the energy supply
proceeds for avoiding network access company. The whole generated power is transferred in
costs for electricity tax the grid and refunded as well as used to satisfied the
own consumption. It is also possible to transfer
penal costs for balance violation
electrical power out the grid when all CHPs are
(virtual costs)
switched off.
As already noted several variables exist for every time
step limited by some boundaries and connected by
parameters in lots of equations and inequations.
319
th th
Tab.1 gives an overview on the design parameters of all Analysis operation mode
heat and power generation units as well as on the cost First the real operation mode of whole year 2008 is
assumptions for fuel and the refund of the energy tax. analysed retrospectively (case I) and the optimal
Tab. 1 Design parameters of generation units operation mode is determined with the help of FreeOpt
(case II). Following the annual operation costs are
Generation CHP 1 CHP 2 CHP 3 Heat calculated for both cases (Tab. 3).
unit Plant
The calculations in Tab.3 show that the costs for fuel
/ kW 911 774 911 - increase but so the proceeds through electricity sale
and CHP-refund increase too.
/ kW 911 774 911 -
Tab. 3 Comparison annual operating costs and proceeds
(case I and case II)
/ kW 1200 1020 1200 200
Annual operation Real operation Optimal
/ kW 1200 1020 1200 2000
costs and mode operation
3
0.35 0.35 0.35 0.87 proceeds / 10 (case I) mode
/%
(case II)
( ) ( ) ( ) ( )
Costs fuel 1326.12 1331.13
0.76 0.76 0.76 -
Proceeds
Own electricity 468.48 473.65
0.5 %
consumption 2.0 % 2.0 % 2.0 % contracts
(of
(of ) (of ) (of ) CHP-refund 58.17 58.67
)
Start up costs 10.01 6.62
Costs fuel / 0.045
0.04444 0.04444 0.04444 Costs network
/kWh 74 2.83 2.47
access
Refund
0.002 Total costs 812.30 807.90
energy tax / 0.0055 0.0055 0.0055
2
/kWh
Start up Furthermore costs for start ups and network access

cost / 12 12 12 6 can be reduced. But altogether the total annual
/start up operating costs for case I and case II are nearly equal.
In this particular example savings are under 1% of the
One specific characteristic of the energy system is that operation costs. So the energy system is operated
all three CHP cannot be operated in part load, so the nearly in an optimal way.
value of the maximum and the minimum power have to
be equal. System configuration changes
The energy supply company considers to replace the
The base own consumption of the whole energy
CHP 2 through a new one. Of course the new CHP has
system is 50 kW. In the 100 m3 heat storage it is
new parameters (Tab. 4).
possible to store 4000 kWh. Tab.2 gives an overview of
all intern electricity contracts. Tab. 4 System configuration changes CHP 2
Generation unit Old CHP New CHP

Tab. 2 Electricity contracts
/ kW 774 404
Electricity purchase Electricity
/ kW 774 404
sale
6am 22pm 22pm 6 am
/ kW 1020 535
Costs /
0.08150 0.05700 0.04735
/kWh 1020 535
/ kW
The CHP-refund is 0.0056 /kWh and costs for network /% 0.35 0.36
access are 0.0386 /kWh.
Start up cost /
12 9.5
/start up
320
th th
All other parameters are the same like the old CHP 2. curve helps to determine the recommended operation
Main difference is the smaller range of performance. mode.
The remaining energy system left unchanged except
The red line marks the given heat demand which is
for the CHP-refund expiring the next year.
satisfied at all time steps. The small-sized CHP 2
The optimal operation mode considering the old operates continuous. When the heat demand is higher
(case III) and the new (case IV) parameters of the than the output of the CHP 2 the heat plant is switched
CHP 2 are determined. Case III serves as a reference on or the heat storage is discharged mostly. Charging
case. The calculated operation costs per week are the heat storage takes place in low demand times and
shown in Fig. 4 for both cases. by switching on one of the others CHPs (CHP 1 or
CHP 3) for up to four hours. Using the CHP is more
Operating costs per week
35
economic than using the heat plant but start up costs
and the size of the heat storage restrict the operation of
30
a second CHP.
Total cost / 103
25
20
Heatbalance
15 1800
1600 charge of
10
1400 storage
heat flow / kW
5 by CHP
1200 heat
plant
0
1000 heat demand
J F M A M J A S O N D
Month 800
Old CHP 2 (case III) New CHP 2 (case IV) 600
400
Fig. 4 Operation costs per week (case III and case IV) 200
CHP production discharge of storage
0
In Fig. 4 it can be seen that the operation costs for 0 12 24 36 48 60 72 84 96 108 120 132 144 156 168
time / h
case IV are below the operation costs for case III in CHP production storage load heat plant storage unload heat demand
every week, especial in the summer months because
the smaller size of the new CHP 2 suits better to the Fig. 5 heat balance curve on one summer week
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 Finally it can be summarized that first experiences and
savings which settle easily the decreasing proceeds calculations show, how the FreeOpt allows in an easy
through electricity sales. By given investment cost it is and quick way to check beforehand if certain system
very simple to check if the renewal of the old CHP 2 is configurations are useful or contra productive.
economic reasonable.
PROGRAM STEFAN
Tab. 5 Comparison annual operating costs and proceeds Application field
(case III and case IV)
The network optimization is a special case of the
Annual opera- Old CHP 2 New CHP 2 research-main focus optimization of the technical
tion costs and structure of district heating systems of the 5th energy
(case III) (case IV)
3
proceeds / 10 research program of the German Federal Government.
Due to the relatively high net costs of district heating
Costs fuel 1323.87 1208.79
systems it is necessary (beside the application of
Proceeds actual piping systems) to optimize the nets concerning
electricity 462.87 367.12 their design parameters, in particular the pipe diameter
contracts and the pipe routing.
Start up costs 7.72 1.20 Therefore the software tool STEFaN has been
Costs network created for the combined pipeline routes and diameter
1.92 0.00 optimization. This Windows program for the support of
access
the application of the district heating has interfaces to
Total costs 870.64 842.87 geographical information systems (GIS) and is
complementary with these. Its application is possible in
As example the heat balance curve on one summer 3 planning phases:
week (168 hourly time steps) is shown in Fig. 5. Such a
321
th th
1. In the conceptual design planning phase for the If the edge is not used for the site development ( ),
cost estimate and the principle decision for the then holds.
district heating (yes or no).
As variables are required beside the diameter further
2. In the detailed planning phase the localization of variable than auxiliary variables to the formulation of
the pipeline route occurs for the approval the constraints:
planning.
vector of the mass flows of the edges
3. In the execution planning phase the final vector of the pressures of the vertices
determination of the dimension occurs, but no
binary variable to the capture of the jump at
mechanical calculation (stress-strain analysis) is
done by the program. Still the required proofs Thus the constraints can be formulated. These are the
according to e.g. EN 13941 have to be done. equation (6), local and technical limitations as well as
In addition, the program can be used for the hydraulic equations. (8) and (9).
calculation of existing district heating networks. First Kirchhoffs law: Point rule. The sum of all mass
flows in a vertex is equal zero. ( - vertex matrix)
Model
The hydraulic calculations establish the technical basis (8)
which performs constraints of the optimization model. Second Kirchhoffs law: Mesh rule. The sum of the
pressure losses along a mesh is equal zero ( - mesh
In district heating systems a distinctive turbulent flow
can be presumed. In this case a good approximation matrix):
with the surface roughness of the pipe and the (9)

coefficient of friction is applied (4). This mathematical model is simple to describe, but
difficult to solve (already for medium-sized graphs). If
(4) the diameters are eliminated by the equation (6) as a
variable, the variables and whose impact on the
In the mentioned planning phases the coefficients of
objective function is discussed in detail in [5] remain.
drag are included blanket into the pressure loss
according to (5): The principal dependency of the objective function on
the vector of the mass flows is displayed in Fig 6
(5) schematically.
1 6
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) K konkav
in the model.
K Baum
(7)
0
The bracket of the first summand contains the

0
Fig. 6 Schematic dependence of the objective function on

investment costs of the route per meter as a total lump-
the mass flow
sum price and must be multiplied according to by the
length of pipeline. In the second summand On the abscissa the circulatory mass flow of a mesh
"obstacles" can be included as direct costs dependent
is displayed. The ordinate shows the non-convex
from the diameter. The exponent is set =1 in the
objective function which shows jumps by the binary
present program version for linear dependence. The
parameters and are input data. variables with the rhombuses (the filled rhombus is the
function value).
322
th th
This shape leads to the fact that as a solution under the Project processing
requirements mentioned above a graph free of The help file and the user's manual of this program
meshes, thus a tree, arises. contain detailed instructions to its operation and for the
This complex course of the objective function requires project processing which occurs typically in six steps in
especially suitable methods. the change of STEFaN (steps 2, 3, 5 and 6 a) and of a
GIS (steps 1, 4 and 6 b).
Development of mathematical procedures These necessary steps 1 to 6 are demonstrated at a
Different mathematical methods are used: fictive example. As a GIS system the recommended
and provided program ShapeUp (www.nilione.com) is
a) The classical non-linear optimization which is
used. It is freeware too.
applied for the diameter optimization for fixed
development ways.
Example
b) Topological optimizations to the determination of Step 1: Gathering of Geographical information
the shortest ways (shortest path problem) and the With the GIS layers (themes) with geo-referenced
shortest networks (spanning tree problem) which information (vertices for the source and for the
are combined under use of the procedure from a) customer as well as edges for consisting and possible
to a special iteration process. routes) invested and in a special standardized format
c) Stochastic methods for the improvement of the (MIF MapInfo Interchange format) exports:
optimization results of the algorithm of b): A special a) The geo-referenced background image is imported
implementing of the Monte Carlo Method and a ( OpenStreetMap, pale colors in Fig. 7). The
special implementing of the Evolutionary Algorithm. figures of the buildings (darkly) and the courses of
the streets (white) allow a good orientation.
Fig. 7 step 1
323
th th
b) Directly input (or import) of the heat source (red Step 3: Generating the network topography
pentagon on the top left in the Fig. 7, through violet
After the import of the files invested by the GIS the
circle marked) and the sinks (customer yellow
network topography of the possible routes is created by
flags in the Fig. 7, as house service connection the program. Gaps between the inputted routes (step
stations in the house lines arranged) with the given 1c) are complemented by the program to a graph with
attribute for the heat demand. entire and varied development.
c) The input of possible routes (thick green lines in Step 4: Verification of the generated network
Fig. 7) by using the mouse and the assignment of topography
the attributes (table in the right section of Fig. 7),
for quite available pipes with attribute With this step the files generated by the program can
be imported in the GIS: The generated edges (thin blue
(branch pipe from the source in lines in Fig. 8) and vertices (blue dots in fig. 8)
DN 150 input value: ) and complete the entered network topography and can be
for the laying procedure ( for street and checked.
ground, for cellar corridor and available Step 5: Determination of the optimal development
channel). The route optimization is carried out by the program.
Step 2: Providing the non-geographical data Step 6: Evaluation
The files with the non-geographical data (general a) Output of a result report and export of the
entries to the network as for example media optimization results to the GIS.
temperatures as well as economic data) are entered on
forms.
Fig. 8 step 4
324
th th
b) The local representation and if necessary related data is displayed: Length of pipeline in m,
treatment of the results by the GIS: The result can
be visualized in the GIS (Fig. 9). The different the diameter in mm, the mass flow in kg/s
colored lines in the left section show the and the pressure difference in bar.
ascertained route planning, and in the right section
of Fig. 9 a part of the data base with the site-
Fig 9 step 6
CONCLUSION REFERENCES
Stable versions of FreeOpt and of STEFaN are already [1] www.bmwi.de: LowEx Fernwrme Multilevel
finished. The programs power was demonstrated with District Heating, Frdergeber: Bundesministerium
simple examples. They can be downloaded from [6].
fr Wirtschaft und Technologie, FKZ 0327400B.
FreeOpt calculates the optimal operating solution of
[2] www.gnu.org/software/glpk:
district heating networks at a minimum of costs to estimate
GNU Linear .Pro-gramming Kit
saving potentials. With the help of parameters and figure
lines it is very easy to modify the system configuration to [3] www.coin-or.org: Computational Infrastructure for
check how the system reacts under new conditions and Operations Research.
how the operating costs change. In all cases the tool gives
valuable information. [4] www.gnu.org/licenses/gpl.html: GNU General
Public License
Unfortunately, the user guide, the help file and the
manual are only available in German at the moment for [5] S. Gnchtel, Ein Beitrag zur Strukturoptimierung
both programs. 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
325

12th International Symposium On District Heating PDF

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

12th International Symposium On District Heating PDF

Загружено:

Авторское право:

Доступные форматы

12thInternationalSymposiumonDistrictHeating

Aadu Paist, professor

INTEGRATION OF AN IP BASED LOW-POWER SENSOR NETWORK

ABSTRACT most equipment currently used in district heating

Individual measurements NETWORK TECHNOLOGY

Visualized energy efficiency

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

Network setup The physical connection between the Ethernet network

All of the substation devices used in this study were

Fig. 10. A Mulle sensor platform integrated with a

REFERENCES [6] F. Wernstedt, P. Davidsson, and C. Johansson,

ON THE RADIAL CONTACT PRESSURE

For the case of parallel buried pipes former analytical

Fig. 4. Ratio B2/B dependent on relative overburden

A calculation method for parallel buried pipes in

Fig. 3. Method to determine the concentration factors for

The calculations by Rizkallah and Achmus showed only

Sand in a medium dense to dense state was assumed

Definition and Unit Size

In order to keep the model as simple as possible the

From the DIN EN 13941 regulation the average radial

In Table II the results of the numerical investigation are

Regarding the average radial contact pressure the

EN 13941 regulation for single pipes and the calculated REFERENCES

ANALYSIS ON FLAT STATION CONCEPT.

THE PARAMETERS ADRESSED

Secondly, a situation is analysed where heat loss is not

installed hot distribution pipes. To begin with, it is

Delay until reaching 45C

Fig. 6. Basic application for flat station, including boundary

8 effect is similar for concepts compared.

IMPROVED TEMPERATURE PERFORMANCE OF RADIATOR HEATING SYSTEM

ABSTRACT level. In order to reduce the temperature demand in

For many years, return temperatures in DH networks radiator

THEORY Qrad C Tm3 (7)

Radiation Nu Pr 1/3 (0.037 Re 0.8 871) 5 105 Re 10 7 (13)

Impact of air speed on space heating temperature T ss0= 60 C , Qrel= 100 %

output for a radiator designed for the temperature 50

program 60/45 C is illustrated as a function of the air T ss0= 33.6 C , Qrel= 25 %

flow through the radiator are kept constant. Two cases 40

Fig. 2 Calculated heat output improvements at Tss=60 C, 20

Field study object

Panel radiator, see Fig. 5 Data acquisition

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

Fig. 7 Log file from field study.

MODIFYING SPACE HEATING TEMPERATURE

constant and not dependent on the relative heat output.

Panel 2.7 W 360 W* 25

without fan operation is calculated from equation 1. 60

In this paper, the ms has been considered constant, 25

0 0.2 0.4 0.6 0.8 1 1.2 1.4

30 Fig. 11 Modified temperature program for panel radiator,

The procedure described above has been applied to all 40

collected data. Results are shown in Fig. 10 and Fig. 11 35

As seen in the figures, the temperature program is now 25

significantly reduced for both the panel and the column 20

Column radiator Pfan = 2.2 W DH primary supply temperature

Pfan = 2.7 W mp unchanged

in the DH network is reduced. The second strategy

building with heat and domestic hot water (DHW) and

heating load a flow-weighted mean value for Tps and Tpr

load from both DHW and DHW circulation. The

Fig. 16 Resulting Tps reduction with column radiator.