Documente Academic
Documente Profesional
Documente Cultură
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.
4
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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].
6
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
Fig. 5. Schematic overview of the experimental setup. Squares marked with an N in the 6LoWPAN network are sensor
nodes (motes).
Fig. 7. A stack view of how internet connected systems is connected with proprietary communication protocol.
8
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
10
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
11
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
13
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
D 1 k
r , avg,13941 * H * (2)
2 2
14
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
16
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
Main contribution to energy saving is originated from Concept Pipe T T E E Net E Energ. Energ.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
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:
19
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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]
20
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
22
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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)
23
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
Temperature ( C)
T sr0= 29.9 C , Qrel= 25 %
speed in Fig. 2. The supply temperature and the mass 45
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
24
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
fan
Tsr,0=39 Tsr,Fan=34
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
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
2.2 W
Fig. 8 Increased cooling of secondary system.
25
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
The relative heat output from the radiator with and Panel radiator Pfan = 2.7 W
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
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
50 Tss,Fan
Fig. 9 Modified secondary temperature program. Tsr,Fan
45
New space heating temperature program - results
T [C]
26
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
By applying the first strategy, both Tpr and the mass flow 10
saving [C]
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
0
1 0.8 0.6 0.4 0.2 0
rel heatload
27
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
P = 3.0 W T unchanged
fan ps
P = 2.2 W T unchanged
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.
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
10
heating is needed, +17 C.
28
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
29
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
30
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
A CHP utilizing biomass might consist of the following system camber Control
System
production unit production unit
Chipping, packing, transport, local roads Figure 5 Segment/selection of part of the architecture
Intermediate storage, transport regional roads information model.
35
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
37
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
[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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
44
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
10
48
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
$2,300,000
$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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
$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 (-)
52
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
56
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
57
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
66
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
67
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
68
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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)
70
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
72
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
75
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
76
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
77
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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)
[4]
[5]
[6]
[7]
78
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
The governing equations then change to: Table 8 Pipe temperature predication comparison (supply
o
outlet temperature is controlled at 50 C)
[8]
Where
[9]
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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:
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)
83
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
84
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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]
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.
87
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
90
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
91
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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)
92
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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:
93
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
0,029
at the start can be an effect of not having steady-state 0,028
conditions before starting the temperature decline. 0,027
94
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
International Symposium on District Heating and
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
98
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
99
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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)
100
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
0,12
0,06 Down(mm)
0,04
Time (s)
mm
0,02
-0,04
-0,06
0,00 1,00 2,00 3,00 4,00
time (s)
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
14
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
104
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
105
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
12th floor
11th floor
10th floor
9th floor
T high
8th floor
6th floor
5th 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
11th 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
1st floor
HWS
DHWS
HWR
DHWR HWR
HWR
for only one case. HWS Heating water supply 10th floor
7th floor
HWR Hot water return
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
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
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
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
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
HW
R
7th floor
HWR Hot water return
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
109
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
110
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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%
111
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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).
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
114
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
115
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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]
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
Heat loss of buried piping [W/m]
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
120
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
35.0
PEX/PUR 162A110 22.3 0.049
30.0
PB/PE 63A32 15.2 0.038
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).
123
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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]).
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
128
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
131
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
132
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
134
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
136
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
4 44 11,6 4
5 29 11,6 -
137
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
Scenario 1 Scenario 3
IRR 5% 5%
Capital cost 70 % 90 %
Financial support 40 % 30 %
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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
[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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
141
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
142
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
(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)
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
150
payments, out-payments, initial outlays and terminal
value.
100
50
100
In-payments Out-payments
Comb.
Comb.
Comb.
0
Ref.
Ref.
Ref.
Ref.
-50
-150
Fig. 4. Resource efficiency of biomass quantified as -200
GHG reduction per used quantity of biomass.
-250
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%
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
infrastructure. Only relying on local demand for biogas Yearly Cash Flows (M)
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
155
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
158
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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 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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
(a) Case A
(b) Case B
Fig. 10. Comparison of energy load prediction: elec. Load
161
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
163
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
Fig. 18. Heating load and recovered waste heat for gas
turbine CHP
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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
168
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
concept is possible within the operational limits of the Heat Duration Curve - Multiperiod Model
CHP plant. 25
Integrated Case - Lower Loads
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
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
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.
169
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
170
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
171
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
Primary Energy Factor In this study PEFs as shown in table 2 have been used:
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
[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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
177
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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).
180
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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%)
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
193
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
Heat storage content (%)
Electricity price (/MWh)
PE_BP_IN
Heat production (MW)
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)
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
mean value of heat per hour of the same day divided 4,5
Systen A
System B
Ph Pd
3,5
h
Total daily variation, d
Pa
[h/day]
2,5
the three example systems. The figure shows that the 1,5
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
0,3
Momentary daily variation, h
0,2
0,1
Total annual daily variation is a variable that is
[h/h]
-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
200
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
CONCLUSIONS
Total annual daily variation
SYSTEM B
350
250 SYSTEM C
150
201
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
202
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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).
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
REFERENCES
205
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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).
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
Temperature [ C]
50
40 To,dam p
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
%
10
Flow [l/s]
0.5
20
210
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
Test done
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
213
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
215
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
218
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
222
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
Pulp/Paper
Oth. non-metallic
minerals
Cement
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
230
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
CO2_40
CO2_50
200
heating and hot water. In the long term, the cooling 150
CO2_60
CO2_70
commercial sectors.
Figure 3: CO2 prices of the different scenarios
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
1000.00
emission certificate prices (-22 TWh in 2020 between
CO2_110 and CO2_010), the demand is higher by 800.00
400.00
resulting in lower electricity demand in 2020, while by
2030 the switch to electricity based technologies to fulfil 200.00
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
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
232
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
the fossil fuel switch from petroleum products and coal 2000
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
SUM
1800 4180
25%
compared to CO 2 price of 27 /t
20%
District heat generation in [PJ]
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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 [%]
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
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
International Symposium on District Heating and
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
[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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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 )
Distribution and
Transmission (el)
Electricity Energy carrier (z)
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
the insulation material itself, the storage tank, the Hotel and restaurant
building
complete local piping system of the substation system 0,960 Educational 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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
[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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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:
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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
253
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
15
Heat loss [W/m]
10
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
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]
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
255
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
Temp erature inner service pip e [C]
Heat loss c omparison Thermaflex and test institute two
Competitive sample 1 15
Competitive sample 2
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
256
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
Thermaflex International Holding B.V.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
50C
Heat source capacity Heat sink capacity
24.3C 15.5MW 19MW
Flue-gas
Flue-gas- 34.2C
cleaning 59.2C
3. THE KATRI VALA PLANT IN HELSINKI:
COMBINED HEATING AND COOLING Steam
turbin
- Generator
e
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
260
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
261
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
263
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
264
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
272
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
costumer heat use and how the heat use will develop in 150
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
Heat use
kWh/m2
400
350
300
250
200
150
100
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
CONCLUSION
The main conclusions from the analysis were:
276
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
277
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
282
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
0
heating networks whose consumption of energy 0 100 200 300 400 500
84 %
PART 1.
INTRODUCTION
Pum p II, pow er: 131 kW
Capital
O&M
1. COSTS Energy
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
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]
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
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 %
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)
0.6 %
The used pumping energy in different companies was
analyzed by comparing the pumping energy to the 0.4 %
1.7 %
Specific pumping energy (electrical power
1. Heat supply
1.5 %
Heat supply is the same as sold heat + losses.
1.3 %
1.1 %
0.1 %
0 100 200 300 400 500
Length of the DH net, km
285
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
Specific pumping energy vs. length of the DH net Specific pumping energy vs. heat density
Length 0 - 70 km
1.7 %
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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 %
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
REFERENCES
[1] DH statistics 2007, Energiateollisuus ry, 2008
287
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
288
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
[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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
297
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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.
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)
140
25%
100
[W/m]
20%
80
60
15%
40
20
0 10%
0 200 400 600 800 1000
nominal diameter [DN]
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
3
Parameter Value Unit 3
17 /m
60,000 16 /m 10.0%
300
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
301
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
self-compacting material
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
renovation measure 0 - 0 ++ ++ 0 -- ++ + --
construction site in an
+ ++ -- ++ 0 + + + 0 ++
urban area
construction site in a
+ + ++ 0 ++ ++ ++ + + 0
rural area
303
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
304
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
305
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
(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
(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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
308
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
[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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
50
45
40
35
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
314
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
1200
1000
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.
1200
1000
Pipe Length [m]
800
600
400
200
315
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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]
317
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
(1)
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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):
(6)
322
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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.
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
The 12th International Symposium on District Heating and Cooling,
th th
September 5 to September 7 , 2010, Tallinn, Estonia
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