12th International Symposium on District Heating and Cooling by Viivi Jokk

12th International Symposium on District Heating and Cooling September 5th–September 7th, 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 Pärnu. 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-Järve 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 Pärnu. Plant will start the operation in the end of 2010. CHP power plants can also be an efficient source to supply district cooling. This is very promising concept for conditions where both low winter temperatures and high summer temperatures prevail. Even in Estonia, with just a few weeks of hot weather in normal summer (not, as this year`s summer), large shopping centres and office buildings seem to be a good application area. The research in the district heating and cooling field is very important and at this symposium we will hear forty five technical presentations, divided into ten sessions: conceptions and studies in district heating and cooling, efficiency issues of district heating and cooling, district heating in areas for low density, district heating systems-pipes properties, renewable district heating and cooling, the aspects of district heating-CHP, district heating-case studies and optimisation and stimulation in district heating. Two plenary speakers cover the activities on district heating and cooling in the European Union: the prospects for district heating and cooling seen from the EU commission point of view and the prospect of district heating and district heating research in Germany. This symposium is organized at the Tallinn University of Technology in cooperation with the Scientific Committee consisting of members of a Nordic Research group called Primary Energy Efficiency (PEE), one member from Tallinn University of Technology and Riga Technical University. On behalf of the organizers we want to express our sincere thanks to the members of the Scientific Committee, the Local Organizing Committee and the members of the Advisory Committee who gave us valuable support. We also thank Nordic Energy Research and Tallinn University of Technology for financial support. We would like to thank the individual authors for their submitted papers and the reviewers for their time and help. We hope that you will experience an enjoyable stay in our sweet old Tallinn and also that this conference will improve further cooperation in the field of district heating and cooling research and development.

Aadu Paist, professor Andres Siirde, professor

TABLE OF CONTENTS PREFACE ...................................................................................................................................................................... 1 INTEGRATION OF AN IP BASED LOW-POWER SENSOR NETWORK IN DISTRICT HEATING SUBSTATIONS ... 4 J. Gustafsson, H. Mäkitaavola, J. Delsing and J. van Deventer ON THE RADIAL CONTACT PRESSURE OF PARALLEL BURIED PIPES FOR DISTRICT HEATING ................... 12 I. Weidlich, M. Achmus ANALYSIS ON FLAT STATION CONCEPT. PREPARING DHW DECENTRALISED IN FLATS .............................. 16 Thorsen, Jan Eric IMPROVED TEMPERATURE PERFORMANCE OF RADIATOR HEATING SYSTEM CONNECTED TO DISTRICT HEATING BY USING ADD-ON-FAN BLOWERS ........................................................................................................ 22 Per-Olof Johansson, Janusz Wollerstrand PRIMARY ENERGY EFFICIENCY AND SYSTEMS ENGINEERING ......................................................................... 31 M.Berner, R. Ulseth, J.Stang ENHANCED DISTRICT HEATING AND COOLING SYSTEMS – REALISATION OF THE LOW-EX CONCEPT ..... 39 Stefan Bargel, Clemens Pollerberg, Armin Knels, Li Huang, Dirk Müller and Christian Dötsch APPLICATION OF EXERGOECONOMICS TO THE OPTIMIZATION OF BUILDING HEATING SYSTEMS CONNECTED TO DISTRICT HEATING NETWORKS ................................................................................................ 45 C. W. Snoek and S. C. Kluiters SLIMNET: AN INNOVATIVE INTEGRAL APPROACH FOR IMPROVING EFFICIENSIES OF DISTRICT HEATING NETWORKS ................................................................................................................................................................ 53 M. W. P. van Lier A DIRECT HEAT EXCHANGER UNIT USED FOR DOMESTIC HOT WATER SUPPLY IN A SINGLE-FAMILY HOUSE SUPPLIED BY LOW ENERGY DISTRICT HEATING .................................................................................... 60 Marek Brand, Jan Eric Thorsen, Svend Svendsen and Christian Holm Christiansen CHALLENGES ON LOW HEAT DENSITY DISTRICT HEATING NETWORK DESIGN ............................................. 69 M. Rämä and K. Sipilä DESIGN OF LOW TEMPERATURE DISTRICT HEATING NETWORK WITH SUPPLY WATER RECIRCULATION 73 Hongwei Li, Alessandro Dalla Rosa, Svend Svendsen STEADY STATE HEAT LOSSES IN PRE-INSULATED PIPES FOR LOW-ENERGY DISTRICT HEATING ............ 81 A. Dalla Rosa, H. Li, S. Svendsen TRANSIENT THERMAL CONDUCTIVITY OF FLEXIBLE DISTRICT HEATING TWIN PIPES ................................. 90 C. Reidhav and J. Claesson DISTRICT HEATING PIPES 200 MM BELOW SURFACE IN A STREET WITH HEAVY TRAFFIC .......................... 96 Anders Fransson and Sven-Erik Sällberg STUDY ON THE HEAT LOSS REDUCTION METHOD FROM THE SECONDARY PIPELINES IN THE APARTMENT COMPLEX .................................................................................................................................................................. 105 Byung-Sik Park, Yong-Eun Kim, Sung-Hwan Park, Yong-Hoon Im, Hyouck-Ju Kim, Dae-Hun Chung, Mo Chung HEAT LOSS OF FLEXIBLE PLASTIC PIPE SYSTEMS, ANALYSIS AND OPTIMIZATION ..................................... 112 EJ.H.M. van der Ven, R.J. van Arendonk COMPARISON OF COMPETITIVE (SEMI) FLEXIBLE PIPING SYSTEMS BY MEANS OF HEAT LOSS MEASUREMENT ....................................................................................................................................................... 119 I.M. Smits, J. Korsman, J.T. van Wijnkoop and E.J.H.M. van der Ven EFFECTIVE WIDTH – THE RELATIVE DEMAND FOR DISTRICT HEATING PIPE LENGTHS IN CITY AREAS .. 128 Urban Persson, Sven Werner INTEGRATING RENEWABLE ENERGY INTO LARGE-SCALE DISTRICT HEATING SYSTEMS .......................... 132 Peter Begerow, Dr. Stefan Holler SOLAR DISTRICT HEATING (SDH): TECHNOLOGIES USED IN LARGE SCALE SDH PLANTS IN GRAZ – OPERATIONAL EXPERIENCES AND FURTHER DEVELOPMENTS ...................................................................... 140 M. Schubert, C. Holter and R. Soell BIOENERGY COMBINES IN DISTRICT HEATING SYSTEMS: PROSPECTS FOR A FUTURE GROWTH INDUSTRY? ............................................................................................................................................................... 143 E. Axelsson, A. Sandoff, C. Overland SEA WATER DISTRTICT COOLING FEASIBILITY ANALYSIS FOR TALLINN ....................................................... 153 A. Hani, I. Britikovski, H. Voll and T.-A. Kõiv ANALYSIS FOR THE OPERATION BEHAVIOR AND OPTIMIZATION OF CHP SYSTEM IN DISTRICT HEATING AND COOLING NETWORK....................................................................................................................................... 157 Yong Hoon Im, Hwa-Choon Park, Byung-Sik Park and Mo Chung

IMPROVED PRIMARY ENERGY EFFICIENCY OF DISTRICT HEATING NETWORKS BY INTEGRATION OF COMMUNAL BIOMASS-FIRED COMBINED HEAT AND POWER PLANTS WITH BIOMASS PYROLYSIS ........... 168 T. Kohl, N.A. Pambudi, T. Laukkanen and C.-J. Fogelholm CHP OR POWER STATION? – QUESTION FOR LATVIA ....................................................................................... 177 D. Blumberga, G. Kuplais, F. Romagnoli and E. Vigants LCA OF COMBINED HEAT AND POWER PRODUCTION AT HELLISHEIÐI GEOTHERMAL POWER PLANT WITH FOCUS ON PRIMARY ENERGY EFFICIENCY ........................................................................................................ 184 Marta Ros Karlsdottir, Olafur Petur Palsson, Halldor Palsson FLEXIBILITY FROM DISTRICT HEATING TO DECREASE WIND POWER INTEGRATION COSTS .................... 193 J. Kiviluoma and P. Meibom DAILY HEAT LOAD VARIATION IN SWEDISH DISTRICT HEATING SYSTEMS .................................................... 199 H. Gadd and S. Werner DISTRICT HEATING AS PART OF THE ENERGY SYSTEM: AN ENVIRONMENTAL PERSPECTIVE ON ‗PASSIVE HOUSES‘ AND HEAT REPLACING ELECTRICITY USE ....................................................................... 202 Morgan Fröling and Ingrid Nyström ADAPTIVE CONTROL OF RADIATOR SYSTEMS FOR A LOWEST POSSIBLE RETURN TEMPERATURE ........ 206 P. Lauenburg and J. Wollerstrand POLICIES AND BARRIERS FOR DISTRICT HEATING AND COOLING OUTSIDE EU COUNTRIES ................... 215 A. Nuorkivi and B. Kalkum BARRIERS TO DISTRICT HEATING DEVELOPMENT IN SOME EUROPEAN COUNTRIES ............................... 223 Dag Henning and Olle Mårdsjö IMPACT OF THE PRICE OF CO2 CERTIFICATES ON CHP AND DISTRICT HEAT IN THE EU27 ...................... 229 Markus Blesl CONSIDERATIONS AND CALCULATIONS ON SYSTEM EFFICIENCIES OF HEATING SYSTEMS IN BUILDINGS CONNECTED TO DISTRICT HEATING .................................................................................................................... 238 Maria Justo Alonso, Rolf Ulseth and Jacob Stang HEAT LOAD REDUCTIONS AND THEIR EFFECT ON ENERGY CONSUMPTION ................................................ 244 Christian Johansson and Fredrik Wernstedt VERIFICATION OF HEAT LOSS MEASUREMENTS ............................................................................................... 250 J.T. van Wijnkoop, E. van der Ven DISTRICT HEATING AND COOLING WITH LARGE CENTRIFUGAL CHILLER-HEAT PUMPS ............................. 258 Ulrich Pietrucha NEW ECONOMICAL CONNECTION SOLUTION FOR FLEXIBLE PIPING SYSTEMS ........................................... 261 Christian Engel, Gerrit-Jan Baars COMPETITIVENESS OF COMBINED HEAT AND POWER PLANT TECHNOLOGIES IN ESTONIAN CONDITIONS..................................................................................................................................... 267 E. Latõšov and A. Siirde DISTRIBUTION OF HEAT USE IN SWEDEN ........................................................................................................... 273 Margaretha Borgström, Sven Werner DAMAGES OF THE TALLINN DISTRICT HEATING NETWORKS AND INDICATIVE PARAMETERS FOR AN ESTIMATION OF THE NETWORKS GENERAL CONDITION .................................................................................. 277 Aleksandr Hlebnikov, Anna Volkova, Olga Džuba, Arvi Poobus, Ülo Kask EFFICIENCY OF DISTRICT HEATING WATER PUMPING IN FINLAND ................................................................ 283 Antti Hakulinen, Jarkko Lampinen and Janne Lavanti MODELLING DISTRICT HEATING COOPERATIONS IN STOCKHOLM – AN INTERDISCIPLINARY STUDY OF A REGIONAL ENERGY SYSTEM ................................................................................................................................. 288 D. Magnusson, D. Djuric Ilic CUTTING COSTS OF DISTRICT HEATING SYSTEMS BY USING OPTIMIZED LAYING TECHNIQUES ............. 297 Alexander Goebel, Dr. Stefan Holler ANALYSIS OF HEAT TRANSFER IN HEAT EXCHANGERS BY USING THE NTU METHOD AND EMPIRICAL RELATIONS ............................................................................................................................................................... 305 O. Gudmundsson, O. P. Palsson and H. Palsson HEAT LOSS ANALYSIS AND OPTIMIZATION OF A FLEXIBLE PIPING SYSTEM ................................................. 310 J. Korsman, I.M. Smits and E.J.H.M. van der Ven FREE OPTIMIZATION TOOLS FOR DISTRICT HEATING SYSTEMS .................................................................... 318 Stefan Gnüchtel, Sebastian Groß

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

INTEGRATION OF AN IP BASED LOW-POWER SENSOR NETWORK IN DISTRICT HEATING SUBSTATIONS J. Gustafsson, H. M채kitaavola, J. Delsing and J. van Deventer Div. of EISLAB, Dept. of Computer Science and Electrical Engineering Lule책 University of Technology, 971 87 Lule책, SWEDEN most equipment currently used in district heating substations is antiquated.

ABSTRACT In this study, the implementation of a wireless, lowpower, sensor network with IP capabilities in a district heating substation was evaluated. The aim of the study was to show that an open standard solution is technically feasible. Low-power wireless communication was established using IPv6/6LoWPAN on an IEEE 802.15.4 wireless network. An experimental district heating substation was equipped with sensor platforms in vital devices located within or near a district heating substation. As a result, all connected devices could obtain a direct internet connection.

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

A system with open standards facilitates the introduction of new energy services such as individual measurements and improved space heating control. In this study, we found that resource-limited batterypowered devices possess a life expectancy of over 10 years, using small batteries while participating in IPv6 compatible communication.

If heat meters, control systems, and other non-district heating equipment could communicate, new services that have impact on both economy and the environment could be developed. The infrastructure required to achieve wireless device communication may be attained with low-power wireless technology. Small sensor platforms with direct internet access through standardized wireless technology can provide a solid platform for new services.

INTRODUCTION Embedding low-power wireless devices in district heating substations and surrounding equipment such as temperature sensors could provide useful services to consumers and producers. Currently, many different substation control systems on the market can connect to the internet and have various wireless sensor reading systems. However, these systems tend to be specialized and are only compatible with equipment from the same manufacturer. Moreover, internet-compatible control systems are often also relatively expensive, and provide bad scalability.

A lack of standardized communication protocols is commonly encountered when connecting electronic devices from different vendors. In general, devices manufactured by different companies use different communication protocols, which limits the functionality of the substation.

In general, commercially available heat meters cannot communicate through the current infrastructure; thus, specialized communication methods such as mbus, pulse, and infrared readings must be employed. Therefore, poor communication standards limit the current usage of heat meters and other equipment in the substation. However, by sharing information with other devices in the substation, the heat meter could provide useful feedback and sensing information, which can be used to improve the substation control functionality.

District heating substations can be divided into sections based on metering, space heat control, and tap water control. For a visual overview of a common parallel connected district heating substation, see Fig. 2, this is also the substation type used in the study. Typically, information is not shared between these sections; thus, each system can only be locally optimized. To achieve complete substation optimization, information must be shared between sections. To this end, wireless sensor-platforms were installed in temperature-sensors, heat-meters, circulation pumps, and control valves, and new control methods and services were tested. This empowers us

Fig. 1 provides an overview of the development of sensor networks over the last 20 years. Unfortunately, 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 new services to heat suppliers, building owners and end users.

temperature of the returning distribution medium should be minimized. Unfortunately, there are many challenges in maintaining the efficiency of a district heating network. Problems related to the equipment that controls the temperature of radiator water and hot tap water are often encountered. These devices tend to be calibrated to satisfy the desires of the customer only; thus, the effects on the energy efficiency of the entire district heating system are often ignored. One key factor in obtaining a high ΔT across a district heating substation is the radiator circuit supply temperature. The radiator circuit supply temperature does not only affect the indoor comfort, but also the primary return temperature as the returning radiator circuit media cools the primary media through the heat exchanging unit. Specifically, water returned from the radiator circuit cools the primary supply through the heat exchange unit. Currently, the radiator circuit supply temperature is based on the local outdoor temperature, which produces a stable indoor temperature. However, the primary supply temperature also affects the ideal radiator supply temperature and the radiator circuit flow. The relationship between outdoor temperature and primary supply is often assumed to be linear (colder outdoor air leads to a warmer primary supply). However, significant deviations from the ideal curve are common. More information on the effect of primary supply temperature and radiator control on the indoor air temperature and ΔT of the system can be found in [2].

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

SERVICES To control or reduce their energy bill, district heating customers require specific information to determine the appropriate action. Currently, the only information available to the customer is the information provided in the bill or on the heat-meter display.

Adaptive radiator control is another intelligent way of controlling the radiator circuit and obtaining a high ΔT. More information on this method can be found in previous studies by Lauenburg [3].

If information on all devices was available online, customer could easily monitor their usage and interact with the substation. Examples of services that could be provided by the substation are explained in the following sections.

Fault detection Control valves in the district heating substation often possess inappropriate dimensions, resulting in intermittent control, pressure shocks, and high return temperatures. Due to the high thermal time constant of a building, the indoor temperature is not directly affected. Therefore, an error in the control valve may go unnoticed for a considerable amount of time.

Improved substation control Combined heat and power plants are becoming more common; thus, the importance of the distribution system ΔT is increasing. In a combined heat and power plant with a flue-gas condensation system, a high ΔT is even more important to obtain satisfactory fuel efficiency.

Error identification can be achieved by evaluating high frequent meter readings, which to some extent are done today.

To maintain high energy efficiency, the hot water produced by the plant must be delivered to customers with a minimal heat loss. Once the hot water is transported to the customer, a maximum amount of energy per volume of water should be extracted and used for heating purposes, such as hot tap water and space heating. To achieve a maximum ΔT, energy transfer between the distribution medium to the point of consumption should be maximized, while the

A fouling valve that is stuck or does not move in accordance with the control signal may also be difficult to detect. A direct comparison of the valve control signal with the heat meter, which measures the primary flow through the district heating substation, can be used to identify a broken fouling valve [4]. 5

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Individual measurements

NETWORK TECHNOLOGY

Individual measurements are common in some countries and are gaining interest in others. To obtain measurements of each apartment, tap point, or radiator, new metering devices must be installed. The most straight forward method is to install flow meters at each tap point and/or radiator. In general, high resolution flow meters are quite expensive; thus, installing one on every tap point/radiator can be costprohibitive.

A common method of visualizing a network communication protocol is in the form of stack. A stack consists of layers that are separated by function; thus, a communication stack contains different layers of tasks related to data transportation. The layers can be divided and visualized in many ways. For example, the five-layer internet model has been used extensively in previous studies and is displayed in Fig. 4 [7]. In this paper, only the layers that are significant to the results of this research will be discussed. Thus, the network, link, and physical layers are considered in more detail.

An alternative method has been evaluated by Yliniemi [5]. In this method, temperature sensors were installed at each tap point, and one central flow meter was used to measure the flow through a section, which contained up to 40 tap points. The flow recorded by the meter and the temperature measured at the tap points were synchronized, and the integrity of each tapping point was verified by installing inexpensive temperature sensors at each site and a limited number of central flow meters throughout the building. Load balancing Dynamic load balancing is a method used to remove heat load peaks and divide power consumption between buildings. Dynamic load balancing is based on the presence of a large thermal time constant of each building. For instance, in a building with a high thermal time constant, the heating system can be turned off when the price of heat is high or during peak energy hours. An online automatic and independent auction system is used to decide which buildings will be shut down or provided a limited amount of thermal power. In this system, all connected buildings are involved in the bidding process. Specific details on dynamic load balancing are provided in the literature [6].

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

Visualized energy efficiency If a large number of district heating substations were connected to the internet, the performance of different substations could be compared. For instance, the supply/return temperature, Î&#x201D;T, energy usage, etc. of all substations could be plotted in a graph, table or map. Fig. 3 displays a map of the return temperature of a substation, which allows the consumer to compare the performance of their house to others in the area. Moreover, the map provides the utility company with an overview of the network and improves the detection of leaks and short circuits. Moreover, the utility company can identify deteriorating substations or individual installations that perform poorly.

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

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

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

IPv6 prefix from the router to the local link layer address, creating a complete IPv6 address. To ensure that another device does not possess the same IP address, the device broadcasts a neighbor solicitation message to search for a duplicate address. If another device has the same IP number, the new device shuts down.

IPv4 Currently, internet protocol version 4 is the most widely used IP, and almost all computers connected to the internet use this version. An IPv4 address is 32 bits long and is typically written in 4 sections divided by dots (e.g., 192.168.100.123). The theoretical number of IPv4 addresses is 232 (approximately 4.2 billion); however, a fraction of addresses is reserved and cannot be used for online purposes. The total number of usable IPv4 addresses is approximately 3.7 billion. As the number of devices connected to the internet increases, IPv4 addresses are beginning to run out.

6LOWPAN 6LoWPAN is an adaptation layer that separates the network and data link layer of the protocol stack. The purpose of the layer is to compress IPv6 headers and minimize unnecessary data transmission while maintaining IPv6 compatibility. According to the literature, [8] the 6LoWPAN header uses less than 10% of the total energy used during packet transmission.

Technology such as network address translation (NAT) and port address translation (PAT) have postponed the depletion of IPv4 addresses; however, the number of available IPv4 addresses decreases every day.

IEEE 802.15.4 physical and data link layers are often used in combination with 6LoWPAN; however, other standards can also be applied.

IPV6 IPv6 was developed to compensate for the limited number of IPv4 addresses. IPv6 uses a longer address than IPv4 and has several convenient features. IPv6 uses a 128 bit address, which means that there are 2128 possible addresses. Thus, the number of address per square millimeter of the earth‘s surface is 6.7·1017. Hopefully, the addresses obtained through the implementation of IPv6 will last for a long time.

802.15.4 (Link and Physical Layer) The most common data link and physical layer used with 6LoWPAN networks is IEEE 802.15.4; however, 6LoWPAN is also compatible with other layers. Moreover, IEEE 802.15.4 is also the basis for ZigBee, Wireless HART, and MiWi. The IEEE 802.15.4 standard specifies operation at low frequency bands such as 868 MHz (EU), 915 MHz (US), and 950 MHz (JP), and high frequency bands including 2.4 GHz (World Wide) [9]. The main practical differences between low and high frequency bands are the bandwidth and communication range. The 2.4 GHz band supports a higher bandwidth but the range is limited, especially in armored concrete buildings. The low frequency bands have a moderate bandwidth and a considerably larger range. In a district heating substation, bandwidth usage can be minimized because rapid changes are uncommon (compared to many other control/measurement situations) and low frequency groups are preferred. However, only 2.4 GHz sensor platforms were available at the beginning of this study; thus, these platforms were used in most of the tests.

With the additional address space, it is possible to give every small device its own unique IP number without implementing NAT. Thus, direct communication over the internet can be achieved without any special gateways. However, the new address space increases the overhead of data packages, which negatively impacts small, low-power devices because more battery energy is wasted on header data in every wireless data transmission. However, a new adaptation layer (6LoWPAN) was developed to limit the amount of lost energy. More information on 6LoWPAN can be found in the next section and in [1]. In addition to a wider address space, IPv6 also includes stateless autoconfiguration, which is a function that can be used to automatically configure newly connected devices without any special servers. To obtain stateless autoconfiguration, newly connected devices broadcast a router solicitation (RS) message to every listening device. When a router receives the message, it responds with a router advertisement (RA) message. The device adds the

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. 6. A stack view of the experimental setup.

Fig. 7. A stack view of how internet connected systems is connected with proprietary communication protocol.

The physical connection between the Ethernet network and the IEEE 802.15.4 wireless network is done at the edge router (in Fig. 5). Here is where the 6LoWPAN header compression & decompression is performed to the passing IPv6 packages. This can also be viewed in Fig 6, where a stack view of the experimental setup is depicted. As a comparison to Fig. 6, Fig. 7 illustrates an example of how vendor specific products currently connect to the internet, using proprietary protocols, which makes them incompatible with devices from other manufacturer. The scalability is hence limited to devices from the same manufacturer, which tends to be a short tem solution.

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

Sensor platforms The sensor platform used in the study, is the Mulle v5.2 [10]. Other similar available sensor platforms are among others Micaz [11], AVR Raven [12] and Sensinode [13]. We choose to work with the Mulle platform since it has a very good performance to energy ratio. 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, the lightweight operating systems Contiki [14] and TinyOS [15] have been successfully ported to the Mulle platform. Both operating systems were specifically designed to be compatible with resource limited embedded systems such as Mulle. Moreover, Contiki and TinyOS both support IPv6 and 6LoWPAN. However, TinyOS was selected for this study because stability issues due to edge-routing problems with Contiki.

a part of the 6LoWPAN standard. Additionally, TinyOS uses short addressing, while Contiki employs long addressing. The type of addressing and header compression used by the OS can be changed, but in this particular test, default settings were used. For payload sizes greater than 60/90 bytes, the IP packet had to be divided into two separate 802.15.4 frames because the maximum frame size of IEEE 805.15.4 is 127 bytes. The separation of IP packets increased energy usage and decreased the expected lifetime of the sensor. Thus, software developers should consider the maximum frame size if absolute maximization of sensor lifetime targeted. However increased payload sizes can of course be compensated with a larger battery.

Sensor platform energy usage Obtaining an acceptable life expectancy is one of the biggest challenges to battery powered, wireless devices. In Sweden, heat meters are inspected every 5 to 10 years, depending on the size of the meter. The life expectancy of wireless devices should be equivalent to the inspection period to avoid frequent and expensive battery replacements. All sensor nodes do however not need to be battery powered. In the case of available electric power in close proximity, e.g. for platforms mounted in pumps or valves there is no explicit need for batteries since there are electricity available. At other sensor platforms, battery power is the only feasible solution, for instance outdoor temperature sensors.

As shown in Fig. 8, the fixed transmission interval was set to 15 minutes, and the effect of transmission interval on the expected lifetime of the sensor was analyzed. Additionally, sensor lifetime was evaluated at various transmission frequencies and a fixed payload of 80 bytes, as shown in Fig. 9. In accordance to theory the results indicated that a low transmission frequency has a positive effect on sensor lifetime. In the case of context aware sensors, which only transmit data when required e.g. when a measured temperature exceeds a set threshold, sensor life expectancy will in most cases be increased. However, the impact of the sleep/standby energy usage will make up a larger percentage of the total energy usage, which hence will mean that the importance of keeping the sleep current low will be even bigger.

To determine the amount of energy used by a wireless sensing device, the current at the sensor platform associated with IPv6/6LoWPAN communication was measured. To measure the current used by the device, a 1 ohm high precision resistor was connected in series to the Mulle power connector. The voltage drop generated across the resistor was amplified 100 times with a MAX4372H amplifier circuit. Using an analog acquisition card, the amplified signal was measured and stored in an ordinary PC. Due to poor precision at very low current, complementary measurements were performed with a high precision ampere-meter to determine the current usage of the Mulle, when it was in deep sleep mode. To evaluate the energy cost of transmitting data packets with UDP on IPv6/6LoWPAN, packets with payload sizes between 1 and 100 bytes were transmitted, and the expected lifetime of the sensor was calculated. Fig. 8 displays the expected lifetime of a sensor with a 500 mA battery and a 15 minute transmission interval. Out of curiosity, both TinyOS and Contiki were programmed to transmit UDP packets of different sizes at consecutive time intervals to observe any differences in energy usage between the two. The results indicated that the energy usage of 50 to 80-byte payloads in Contiki and Tiny OS were significantly different. The observed difference between operating systems is most likely related to the method of header compression. Specifically, Contiki uses HC1, while TinyOS is based on HC01. However, both methods are

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

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

All of the substation devices used in this study were module-based, which allows manufacturers to produce 6LoWPAN module for large scale deployment. RESULTS Wireless devices in a district heating substation were successful integrated to support a IPv6/6LoWPAN network. Due to the range limitations of 2.4 GHz modules, deployment of several platforms was restricted. However, new 868 MHz platforms are now available and show excellent preliminary results. 2.4 GHz platforms will be replaced with 868 MHz platforms during the spring/summer of 2010.

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

A lifetime of 10+ years can be achieved with 500 mAh battery and an average transmission interval of 15 minutes using IPv6 compatible communication; thus, the life expectancy of battery powered sensors did not have a negative effect on integration.

The predictive life expectancy calculations did not take into account the fact that batteries loose energy over time, even if they are not in use. Depending on battery type, this can significantly reduce the expected lifetime of a sensor.

CONCLUSION Integrating an IPv6/6LoWPAN wireless network in a district heating substation can significantly increase the functionality and scalability of the substation and supply new services to both producers and consumers.

SENSOR INTEGRATION To provide wireless accessibility to devices in the district heating substation, some simple interface electronics were developed to integrate Mulle with device hardware. As shown in Fig. 10, a heat meter was integrated with a Mulle in the bottom module location.

Using an open, well documented, and tested protocol increases the possibility of interoperability between products of different manufacturers. This study revealed that available technology can be used to achieve IP-based wireless communication. However, a considerable amount of work on smart application layers must be conducted before wireless sensor networks in district heating substations can be deployed and used to its full potential.

When digital communication interfaces were available (heat meter and circulation pump), the corresponding application protocols were kindly provided by the vendors (Kamstrup and Grundfos). The control valve (Siemens SQS-65) was not equipped with any digital communication interface; however, an analog 0â&#x20AC;&#x201C;10 V input used to control the position of the valve and a 0â&#x20AC;&#x201C;10 V output used to read the position of the valve were available.

FUTURE WORK To achieve complete device compatibility, the application layer(s) of the integrated network must further developed. One interesting approach is to adapt the service oriented architecture in web-based services to low-power sensors. Available service oriented architectures (SOA) such as DPWS1 are developed primarily for large enterprises and are not intended to be used with a resource limited device that possesses a low-bandwidth link. However, the functionality of this architecture would support a convenient solution for direct sensor integration in enterprise systems. The integration of sensors and SOA such as DPWS is a challenging but intriguing task.

Fig. 10. A Mulle sensor platform integrated with a Kamstrup Multical 601 heat meter.

Mulle is marked by a blue square, and the interface card is indicated by a purple square.

Device Profile for Web Services

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

[6] F. Wernstedt, P. Davidsson, and C. Johansson, ―Demand side management in district heating systems,‖ in AAMAS ‘07: Proceedings of the 6th international joint conference on Autonomous agents and multiagent systems. New York, NY, USA: ACM, 2007, pp. 1–7.

REFERENCES [1] Z. Shelby and C. Bormann, 6LoWPAN: The Wireless Embedded Internet, November 2009. [2] J. Gustafsson, J. Delsing, and J. van Deventer, ―Improved district heating substation efficiency with a new control strategy,‖ Applied Energy, vol. 87, no. 6, pp. 1996–2004, 2010. [Online]. Available: http://www.sciencedirect.com/science/article/B6V1T-4Y648K9-1/2/14e2e71a60c1335c8def21f6328bb9a0

[7] J. Kurose and K. Ross, Computer Networking a Top-Down Approach featuring the Internet, 2nd ed. Pearson Education International, 2003. [8] G. Mulligan and 6lowPAN Working Group, ―The 6lowpan architecture,‖ in Proceedings of the 4th workshop on Embedded networked sensors, June 2007.

[3] P. Lauenburg, ―Improved supply of district heat to hydronic space heating systems,‖ Ph.D. dissertation, Dept. och Energy Sciences, Lund University, P.O Box 118, SE-22100, Lund, December 2009.

[9] ―IEEE 802.15.4-2006 standard‖, http://standards.ieee.org/getieee802/802.15.html, April 2010.

[4] K. Yliniemi, Fault detection in district heating substations. Licentiate thesis, Div. of EISLAB, Dep. of Computer Science and Electrical Engineering, Luleå University of Technology, 971 87 Luleå, Sweden: Luleå University of Technology, 2005.

[10] ―Embedded internet system technology botnia AB,‖ http://www.eistec.se/, March 2010. [11] ―Crossbow technology,‖ March 2010.

[5] K. Yliniemi, ―Individuell mätning av varmvattenförbrukning,‖ http://www.svenskfjarrvarme.se/download/4774/Ki mmo Yliniemi.pdf, 2007.

http://www.xbow.com,

[12] ―AVR raven,‖ http://www.atmel.com, April 2010. [13] Sensinode,‖ http://www.sensinode.com, April 2010. [14] ―Contiki,‖ http://www.sics.se/contiki/, March 2010. [15] ―Tinyos,‖ http://www.tinyos.net, March 2010.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

ON THE RADIAL CONTACT PRESSURE OF PARALLEL BURIED PIPES FOR DISTRICT HEATING 1

I. Weidlich , M. Achmus

AGFW, German Heat and Power Association, Research & Development, Stresemannallee 28, 60596 Frankfurt am Main, i.weidlich@agfw.de 2 Institute of Soil Mechanics, Foundation Engineering and Waterpower Engineering, Leibniz University of Hannover, Appelstr. 9A, 30167 Hannover, achmus@igbe.uni-hannover.de ABSTRACT For the design and calculation of buried district heating pipe systems the magnitude of radial contact pressures acting on the pipes is of importance, since these pressures affect the friction forces which may be mobilized. For parallel buried pipes, the stress distribution is generally expected to be different from the case of a single pipe. The present investigation compares radial stresses according to current design directives for buried single pipes with numerically calculated stresses for parallel buried pipes. The calculations show a deviation of the radial stress distributions in particular for the springline area. The results are compared with former theoretical investigations, which predicted a reduction of radial contact pressures between the two pipes. This is verified for small-diameter pipes. With larger pipe diameters a stress increase was identified between the pipes. However, with regard to the average radial pressure only slight differences between single pipes and parallel buried pipes were found.

Fig. 1. Typical trench condition for DH-pipes after FLOSS [2]

The distance between the two pipes depends on the requirements of the laying technique and procedure. For small distances between the two pipes an interaction between the two pipes is to be expected.

INTRODUCTION

PREVIOUS WORK

As a part of the underground infrastructure of modern settlements, district heating pipe networks are an important medium of economic heat transportation. Hot water is pumped in a flow pipe from the supply station to the consumer at a high temperature and under high pressure, and the used water is pumped back to the supply station in a return pipe.

Previous theoretical investigations were based on the calculation method developed by Leonhardt, taking into consideration the deformation behaviour of pipe and soil and their influence on each other [3]. Leonhardt introduced the ―shear resistant beam on elastic bedding‖ theory, in which the backfill above the pipe is considered to be a shear resistant beam, which is able to transfer shear loads, but no bending moments. Using this model it is possible to determine the shear forces activated by the deformation of the ―shear resistant beam‖ caused by different stiffnesses of the pipe and the surrounding soil, which leads to a redistribution of stresses in the soil with corresponding concentration factors .

For buried district heating pipes the earth pressure on the pipe, respectively the radial contact pressure, is an important value for the design, since it affects the friction forces which may be mobilized. The friction forces determine the axial deflections of the pipe and the distribution of normal stresses, which are induced by the temperature loading of the pipe. According to the European Standard EN 13941, the normal stress on the pipe coating is calculated for single pipe trench conditions dependent on the overburden weight of the soil, the diameter, the pipe weight and an earth pressure coefficient [1]. However, in practice flow and return district heating supply pipes are buried side by side in the same trench. Fig. 1 shows a typical situation for buried district heating pipes according to Floss [2].

For practical application in Germany regulation ATV A 127 was published employing Leonhardt‘s theory for buried pipes [4]. This regulation can be applied analogously to all kinds of buried pipes. The special application of regulation ATV A 127 for buried preinsulated district heating pipes was first investigated by Beilke [5]. 12

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

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

A calculation method for parallel buried pipes in stepped trenches was proposed by Hornung and Kittel [7]. With this calculation method the total loading on one pipe is derived from the sum of the partial loadings, which correspond to the trench shape to the right and left of the pipe. The typical trench condition for district heating pipes provides a non stepped trench with the flow and return pipes installed on the same underground level. The presented study was therefore carried out without employing the Hornung and Kittel calculation method.

Fig. 2. Shear resistant beam for parallel buried pipes

The concentration factors B. B for the area beside and between the pipes were found with the following assumptions: 



The influenced area for the determination of the concentration factor B beside the pipe (B1) is defined by a line with an inclination of 60° shown in Fig. 3. This angle coincides with the theoretical slope inclination of a noncohesive soil with an internal angle of friction of ' = 30°.

NUMERICAL INVESTIGATIONS Numerical calculations were carried out with the two dimensional finite element program PLAXIS, version 8.6. Two standard situations with different outer pipe diameters D (DN65, D=140 mm; DN250, D= 400 mm) of two parallel buried district heating pipes were investigated. The distance between the pipes was chosen to be A=10 cm (see Fig. 1). The overburden height of the backfill material of the trench was H/D=3.0. The finite element mesh used for the DN65 pipe is shown in Fig. 5 as an example.

Between the pipes (B2) the full interspace is taken as the area of influence.

Fig. 3. Method to determine the concentration factors for parallel buried pipes

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

Fig. 5. Finite element mesh for the case DN65, H/D=3

The installation process was simulated by a ―staged construction‖ process, considering a retained trench and the backfilling procedure with several layers. The compaction process was accounted for by applying a static distributed load of p=10 kN/m² on each of the layers. Ground water was not considered in this investigation.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Sand in a medium dense to dense state was assumed as backfill material. The mechanical behaviour of the soil was modelled with the Mohr-Coulomb constitutive law, which is a linear elastic / ideal plastic material model. The parameters used for the model are shown in Table I. Table I. – Soil parameters used for sand in Mohr-Coulomb material model Definition and Unit

Size

Unit weight  [kN/m³]

Oedometric Elasticity Modulus Eoed [MPa]

70.5

Poisson‘s ratio 

0.3

Internal angle of friction ‘ [°]

Angle of dilatancy  [°]

Interface friction Rinter [1]

0.536

Fig. 7. Horizontal effective stresses h (DN250 pipe, H/D=3)

In Fig. 8 the stress concentration is shown by the distribution of the radial contact pressure for the lefthand pipe. In the springline area a maximum value of r=21.44 kN/m² for the radial pressure was obtained. Compared with the calculated average radial pressure of

Between pipe and soil, the Coulomb friction law with a

according to Eq. (1).

tan  i  Rinter * tan  '

r,avg,calc= 18.81 kN/m² the deviation is about 12.2%. (1)

In order to keep the model as simple as possible the pipes were assumed to be rigid. In the numerical calculations the initial soil stress state due to the soil unit weight was established first. The installation procedure was then simulated and the results were evaluated. In the first model of pipes with an outer diameter of D=140mm (DN65), no significant stress concentration between the pipes was observed. The radial contact pressure obtained for both pipes is shown in Fig. 6.

Fig. 8. Contact pressure on the left-hand DN250 pipe, H/D=3

From the DIN EN 13941 regulation the average radial pressure on a single buried pipe can be derived for the investigated trench condition according to Eq. (2).

 

 r , avg,13941   *  H 

D  1 k  *  2  2 

(2)

In Table II the results of the numerical investigation are compared to the expected radial pressure from the DIN EN 13941 regulation.

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

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

Table II. – Average contact pressure r,avg for H/D=3.0 DN

Single pipe according to Parallel buried pipe DIN EN 13941 according to numerical results

6.15 kN/m²

7.25 kN/m²

250

17.58 kN/m²

18.81 kN/m²

Regarding the average radial contact pressure the difference between the expected values from the DIN 14

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

EN 13941 regulation for single pipes and the calculated values from the numerical simulations for parallel buried pipes is rather small. The values for the parallel buried pipes are slightly larger.

REFERENCES [1] DIN EN 13941, Berechnung und Verlegung von werkmäßig gedämmten Verbundmantelrohren für Fernwärme, Deutsches Institut für Normung e.V., Beuth Verlag Berlin, 2003.

CONCLUSION

[2] R Floss, „Handbuch ZTVE-StB 94/1997―, Kommentar mit Kompendium Erd- und Felsbau, 3. Auflage, Kirschbaum-Verlag, 2006.

The earth pressure on district heating pipes is an important design value and should be determined as exactly as possible. In the presented work the earth pressure on parallel buried pipes was investigated.

[3] ]G. Leonhardt, „Belastung von starren Rohrleitungen unter Dämmen―, PhD Thesis, Institute of Soil Mechanics, Foundation Engineering and Waterpower Engineering, University of Hannover, 1973.

The evaluation of the radial stresses on the pipe in numerical calculations showed a stress concentration between two pipes buried in the same trench for short pipe distances and large diameters. However, former theoretical investigations led to a reduction of radial contact pressure between the two pipes, which was observed in the numerical simulations for small diameters and small overburden heights.

[4] ATV A 127, Richtlinie für die statische Berechnung von Entwässerungskanälen und -leitungen, Arbeitsblatt A 127 der Abwassertechnischen Vereinigung e.V., 2000. [5] O. Beilke, „Interaktionsverhalten des Bauwerks „Fernwärmeleitung–Baugrund―, Institute of Soil Mechanics, Foundation Engineering and Waterpower Engineering, University of Hannover, 1993.

Because typical trench conditions with two parallel buried pipes are not considered in current design directives for district heating pipes the numerical results were compared with the values derived from the current design regulations. For the observed systems only small deviations regarding the average normal pressure between single pipe and parallel buried pipes were found. Thus, as long as the exact distribution of stresses along the pipe perimeter is not of particular relevance, current calculation directives are also suitable for parallel buried pipes. Only for conditions with large pipe diameters and small distances between the pipes and also relatively large overburden heights is a significant deviation to be expected.

[6] V. Rizkallah, M. Achmus, ―Zur Größe der Reibungskräfte an erdverlegten Fernwärmeleitungen‖, Forschungsvorhaben Wechselwirkungen Fernwärmeleitung – Bettungsmaterial, Institute of Soil Mechanics, Foundation Engineering and Waterpower Engineering, University of Hannover, 1993. [7] K. Hornung, D. Kittel, „Statik erdüberdeckter Rohre―, Bauverlag GmbH Wiesbaden und Berlin, ISBN 3-7625-2039-9, 1983.

Furthermore, inhomogeneous backfill compaction, which is probable for small pipe distances under in situ conditions, affects the contact pressure. In order to take into account the real compaction process within the trench, only direct measurements seem to lead to correct results. Further research work is necessary at this point. In order to avoid large deviations in the contact pressures, good and consistent backfill compaction and a certain minimum distance between the flows and return pipes is recommended.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

ANALYSIS ON FLAT STATION CONCEPT. PREPARING DHW DECENTRALISED IN FLATS Thorsen, Jan Eric Senior Project Manager, M.Sc., Danfoss District Energy, DK-6440 Nordborg jet@danfoss.com ABSTRACT

Investments

In some countries the flat station concept is becoming a common way of realising heating and domestic hot water (dhw) installation in blocks of flats. Anyhow, in other countries it is at the very beginning. Experience from those countries reveal a number of questions when understanding and evaluating the flat station concept. A number of parameters can be addressed to and be evaluated to disclose qualities and performance of the flat station concept in relation to traditional concepts for heating and dhw installations.

Reference for comparing the flat system concept with a conventional concept is based on modern way of making block pipe distribution systems [1]. In both cases it is a horizontal pipe layout in flats with a vertical pipe tunnel for distribution. Pipe distribution systems are shown in fig. 1. Main differences are to be seen in the number of pipes installed. Since dhw is prepared decentralized in flats, dhw pipe and dhw circulation pipe are eliminated. Centrally located dhw station in the basement is replaced by decentralized flat stations. Balancing valves for heating as well as for dhw distribution is saved for the flat station concept. Regarding metering then the dhw meter is eliminated, since the primary supply to the flat station covers flat heating and dhw as well. According to measurements of more than 2500 dwellings in Denmark, including detached houses as well as multi storey buildings, individual metering, say individual billing, resulted in savings of 15–30%, [2]. Therefore, this analysis assumes metering of all thermal energy deliveries to flats.

This paper aims at analysing main parameters regarding quality (comfort) and performance of the flat station concept, covering block distribution system, flat station itself and flat installation. Parameters in focus are: riser system, instantaneous dhw principles, heat losses, comfort of dhw, investments and energy savings, metering and hygienic issues for dhw. INTRODUCTION Areas of district heating distribution systems, building heating installations and domestic hot water (dhw) installations show a high degree of conservatism and traditions, which are reasonable due to their lifetime. But this also implies a number of questions when new concepts like the flat station concept are to be introduced. Not only questions addressed to the flat station concept but also to existing systems, where detailed knowledge is faded out due to the maturity of concepts. This paper aims at analysing main parameters regarding quality and performance for the flat station concept, covering block distribution system, flat station itself and flat installation. THE PARAMETERS ADRESSED Investments: –Distribution system –Basement sub station versus flat stations –Energy meters Energy Savings: – eat loss in primary distribution system –dhw circulation pump consumption Comfort: –dhw temperature stability and variation –dhw recovery time after idle period Hygienic issues – considerations on Legionella related to the system‘s physical layout.

Fig 1. Pipe distribution systems in blocks of flats. C: Modern reference principle. F: Flat station principle.

A recent investment example comparing flat station concept (F) with traditional system (C) is included in fig. 2. Data are based on a Danish case from Århus area where a block, built in 4 levels and a basement level consisting of 24 flats, will generally be modernised. Investments compared are based on concepts presented in fig. 1. Main conclusion is that investment level is approx. break-even for the two systems, for this 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 components/costs levels, level could change. In general, the experience is that flat stations are on break-even cost level or slightly higher. This is valid for new buildings as well as for renovation projects.

demand for a 1970 Danish block building (not including energy for dhw). Table 1. Energy losses for traditional system C, and flat system F based on the Århus case. Concept

Pipe length [m]

Saved eneregy meter for dhw Saved balancing vavles for dhw circulation Saved balancing valves for heating distribution Saved dhw pipes, incl. circulation Saved dhw preparation centrally located Invested in Flat Station

T T E E pipe amb. loss loss/y [W/mK] [°C]

Net E loss/y

Energ. Energ. price costs [ /kWh] [ /year]

[°C]

[W]

[kWh]

Trad. con. C Trad. con. C

Sum. flow Sum. return

120 120

0.20 0.20

40 25

20 20

480 120

4205 1051

2102 526

0.05 0.05

105 26

Flat st. con.F Flat st. con.F

Sum. flow Sum. return

120 120

0.20 0.20

55 30

20 20

840 240

7358 2102

3679 1051

0.05 0.05

184 53

Trad. con. C Flat st. con.F

Unit heat loss 1 pcs. Unit heat loss 24 pcs.

300 W/unit 25 W/unit

2628 3816

1314 1908

0.05 0.05

66 95

dhw circ. C dhw circ. elec.

Summer Sum. + win.

0.20 -

1584 13876 30 260

6938 -

0.05 0.25

347 65

Trad. c. C Flat st. c. F

Total Total

Diff.

10000

5000

-5000

-10000

-15000

-20000

Item

Investments comparison Flatstation concept Århus Case, block of 24 flats

240 -

53 -

20 -

10880 6638

544 332

4242

212

C-F Total (ex. electrical consumption)

Secondly, a situation is analysed where heat loss is not utilised in the building distribution system at all. Winter energy losses for the flat station is assumed to be usable and no floor heating is active during summer.

Euro

Fig. 2. Investment balance for traditional system C and flat system F. Block of 24 flats.

Table 2. Energy losses for traditional system C, and flat system F based on the Århus case.

Energy savings Main contribution to energy saving is originated from installed hot distribution pipes. To begin with, it is assumed that half the yearly distribution energy loss is net loss (summer time), meaning not contributing to heating up the building. Wintertime temperatures are assumed to be identical for the two concepts, because for this period the heating system defines temperature levels. To quantify losses a room temperature of 20 °C is assumed. Danish Technical Insulation Standard [3] requires minimum allowable heat loss constants (W/m), depending on temperatures, annual operation time and pipe diameter. These constants turn out to be quite similar to all pipes in question. To simplify preconditions a heat loss coefficient of 0.20 W/mK has been chosen for all hot pipes. Table 1 shows a comparison of pipe temperatures, heat loss and electrical dhw circulation pump.

Concept

Pipe length [m]

T T E E Net E Energ. Energ. pipe amb. loss loss/y loss/y price costs

[W/mK] [°C] [°C]

[W] [kWh] [kWh] [€/kWh] [€/year]

C C

Sum. flow Sum. return

120 120

0.20 0.20

20 20

0 0

0.05 0.05

0 0

Flat st. con. F Flat st. con. F

Sum. flow Sum. return

120 120

0.20 0.20

55 30

20 20

840 240

7358 2102

3679 1051

0.05 0.05

184 53

Trad. con. Trad. con.

C C

Winter flow Winter return

120 120

0.20 0.20

70 30

20 20

1200 10512 240 2102

5256 1051

0.05 0.05

263 53

Flat st. con. F Flat st. con. F

Winter flow Winter return

120 120

0.20 0.20

70 30

20 20

1200 10512 240 2102

5256 1051

0.05 0.05

263 53

Trad. con. C Flat st. con. F

Unit heat loss 1 pcs. Unit heat loss 24 pcs.

300 W/unit 25 W/unit

2628 3816

2628 1908

0.05 0.05

131 95

dhw circ. C dhw circ. elec.

Sum. + win. Sum. + win.

0.20 -

1584 13876 13876 30 260 -

0.05 0.25

694 65

Trad. c. Flat st. c.

C F

Total Total

Diff.

C-F Total (ex. electrical consumption)

Trad. con. Trad. con.

240 -

53 -

20 -

22811 12946

1141 647

9865

493

Comparing the two systems regarding heat loss, then favour is again towards the flat station concept. For the Århus case it means approximately 9900 kWh/year savings corresponding to 490 Euro/year (ex. pump. costs). This means a saving of approx. 4 kWh/m2/year. This represents a saving of approx. 4% of the yearly heat demand for a 1970 Danish block building. Additionally, as for the flat station concept there is no need for dhw circulation pump, thus no need for the electric energy of 260 kwh/year. A part of this saving is anyhow spent for the flat station concept due to additional circulation of primary water. It is assumed that this is approx. half the electric energy for dhw circulation pump of 130 kwh/year.

Flats in this first case are provided with floor heating in bathrooms; therefore, heating is active all year. Due to floor heating, temperatures for the traditional concept are lower during summer season compared to the flat station concept, since floor heating typically operates at lower temperatures. For the flat station concept a dhw temperature at 45 °C is assumed, demanding a primary temperature of 55 °C. Comparing the two systems regarding heat loss, then favour is towards the flat station concept. For the Århus case it means approximately 4200 kWh/year savings corresponding to 210 Euro/year (ex. pump. costs). This means a saving of approx. 2 kWh/m2/year. This represents a saving of approx. 2% of the yearly heat 17

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

dynamics are heavily influenced by idle bypass thermostat setting. Also pump dynamics are influencing, meaning how fast is the primary circulation pump reacting on rapid changes of hydraulic conditions, say opening of primary valve.

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

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

Fig. 4 shows a flat system with cold heat exchanger. Bypass temperature setting corresponds to primary supply temperature (Tf.dh) of 40 °C and primary return temperature (Tr.dh) of 30 °C. This setting is in the very ―low‖ end, but in the ―high‖ end regarding energy saving. Available differential pressure is 1 bar, but drops to 0.25 bar at the beginning of the tapping. In this case temperature in circulation (Tsupply) is approx. 67 °C. Primary branch pipe from supply to the flat station is 4m, ø 20 mm. Measurements show that primary supply has a delay of approx. 7 sec. to reach a level of 55 °C. Additional delay is then caused by heating up the heat exchanger and dhw water, this delay is additional approx. 3 sec. to reach a minimum demanded level of 45 °C. After 5 meter of pex pipe of ø 22 mm additional delay is approx. 7 sec. By this the total delay from tapping the start to reach 45° at the tap is approx. 17 sec. In this example a very long idle branch pipe length is used, more realistic would be 0–2 m, resulting in a ―primary side‖ delay of not more than a few seconds. Also diameter of secondary dhw pipe is rather big and represents a typical shared pipe dimension, representing one pipe for several taps.

Fig. 3. Dynamic control performance (step test) for thermostatic and pressure controlled heat exchanger for dhw production [6]

Fig. 3 shows that stability, temperature peaks at load change and total dhw temperature (T22) variation is limited to 3–4 °C. Regarding oscillations at low tapping flow it should be noted that T22 is measured at heat exchanger outlet. As example a 5 m ø 22 mm pex pipe reduces peaks and amplitudes additionally, dependent on frequencies, but typically 50%. This example is for very high primary supply conditions. Oscillations appear at tap flow of 100 l/h or below. This level shall be seen in relation to the fact that a typical tapping flow for one tap is 200–400 l/h.

Anyhow, this delay is only relevant for the first tapping, since thermal capacities combined with efficient insulation is maintaining temperature, typically with time constants of 1–2 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 thermostat setting and/or a ―hot‖ heat exchanger during idle. Fig. 5 shows an example of flat station with ―hot‖ heat exchanger and thermostatic controlled heat exchanger [7]. Idle temperature is approx. 50 °C corresponding to dhw tapping temperature.

For simulations a branch pipe flow (Q1) of 800 l/h is assumed. This represents a situation where the thermostat is fully open until the desired set temperature is reached. Further a step vice flow change from zero to Q1 or zero to Q2 is assumed. Tapping flow is assumed to be on a high level flow for one tap, which is typically applied when opening the dhw. Q2=400 l/h for all simulations. Delay until reaching 45°C L2=5m & 10m (internal ø10mm) - Heat Exchanger hot & cold at idle 16

hot - dt at T2 - L2=5m hot - dt at T4 - L2=5m cold - dt at T3 - L2=5m cold - dt at T4 - L2=10m

dT [sec]

hot - dt at T3 - L2=5m hot - dt at T4 - L2=10m cold - dt at T4 - L2=5m

10 8 6 4 2

Fig. 5. Dynamic control performance (idle recovery) for thermostatic controlled heat exchanger for dhw production. Heat exchanger is warm during idle. [7]

0 0

Fig. 5 shows a flat system with ―hot‖ heat exchanger at idle. Bypass temperature setting corresponds to a primary supply temperature (T11) of 58 °C and primary return temperature (T12) of 44 °C. This setting is the high end, meaning in ―high‖ end regarding comfort. For this system there are no primary delays, and dhw tapping temperature at the flat station is available after approx. 2 sec. Additional delay due to dhw piping towards tap would be similar to previous example.

3 4 L1 [m] (internal ø20mm)

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

Heat exchanger simulated is Danfoss XB06H-40 [6]. It can be seen from figure 7, that influence on hot or cold heat exchanger is in the range of 2 sec. delay. Branch pipe length (L1) has minor impact on time delay. This is due to the fact that temperature is maintained with a temperature gradient along pipe during idle, reflecting T1 to T2. Basically water in branch pipe is heated to a certain level already before tapping. Anyhow, due to energy loss and return temperature, idle bypass temperature is lower than dhw tapping temperature in this case.

In many practical matters a compromise between the two examples regarding idle temperature setting fulfils demands for good comfort with reasonable energy consumption. In the following a general trade off is included between branch pipe length, dhw pipe length, idle condition for heat exchanger and temperature delay on dhw, based on dynamic simulations. Pipes are simplified by simple delay models with no heat loss. Heat exchanger is based on a lumped capacity model described in [5].

Main influence on time delay is dhw pipe diameter and length (L2). Connection in flats shall be of ―star coupling‖ principle where every tap has its own supply pipe with a small inner diameter. Temperature in dhw pipe water is assumed to be room temperature prior to tapping. In general, additional delays of typically 3 to 6 seconds shall be expected due to thermal interaction with thermal capacities along the way to tap and hydraulic dynamics on branch pipe side and hydraulic dynamics on dhw side. Simulated waiting time for a dhw temperature of 40 °C is included in figure below:

Fig. 6. Basic application for flat station, including boundary conditions for dynamic simulations. 19

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

rather Legionella bacteria concentration in the dhw. Facts influencing on potential for Legionella concentration growth are dhw temperature, exchange rate of hot water in distribution pipes, and volume of dhw water in the entire hot system. Also other factors are influencing, e.g. systematic cleaning of shower outlets, but this will be not addressed to here, since the effect is similar for concepts compared.

Delay until reaching 40°C L2=5m & 10m (internal ø10mm) - Heat Exchanger hot & cold at idle 16

hot - dt at T4 - l"=5m cold - dt at T4 - l2=5m cold - dt at T4 - L2=10m

dT [sec]

hot - dt at T4 - L2=10m cold - dt at T3 - L2=5m

10 8 6

Comparing volumes of dhw in pipes for concepts, the flats station solution has significantly lower volume compared to the traditional system. Furthermore dhw pipes should be ―star‖ connected, meaning one small (diameter) pipe from the flat station to each individual hot tap. This eliminates problematic dead end or low flow areas.

4 2 0 0

3 4 L1 [m] (internal ø20mm)

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

Typically volume of heat exchanger is 0.25 to 0.50 litre. Typical dhw pipe volume is 0.10 l/m, equal to 1.0 litre for 10 m pipe. In total this is a volume of 1.5 to 2 litre pr. flat. The comparable centrally placed dhw system with dhw distribution will have a volume of 5–7 litre pr. flat. By installing a dhw storage tank this will increase significantly. The German DVGW regulations states that heating dhw up to 60 °C, due to e.g. Legionella, is not required if volumes of heat exchanger or volume of dhw pipes is less than 3 litres [8]. Based on those physical concept differences Legionella bacteria risk is reduced for the flat station concept.

First of all it can be seen that time delay for reaching 40 °C at tap is only a bit shorter than reaching 45 °C. This is due to the fact that the T4 temperature profile has an almost step vice nature, i.e. if temperature goes up after the dhw pipe is flushed through, it goes almost like a step. Different dhw controllers have different performance regarding time delay. In case of pure proportional control for dhw system, time delay is longer at part load. This is because primary flow is proportional to secondary flow, and the lower the flow the longer the waiting time. Looking at the example for Q1=800l/h, Q2=400 l/h, L1=4 m, L2=5 m then time delay (dt) at T4 is 6.9 sec to reach 45 °C dhw temperature. In case of proportional controller with parameters Q1=400 l/h, Q2=400 l/h, L1=4m, L2=5 m then dt=11.0 sec to reach 45 °C. This has considerable effect on time delay as L1 gets longer. In case of a thermostatically controlled dhw system or a combination of a thermostatically and proportionally controlled dhw system, time delay is shorter because no matter how small tapping is, as long as the desired set point temperature is not reached, the primary valve will be fully or almost fully open resulting in high primary flow. Regarding delay to reach a dhw temperature of 40 °C this is only related to dhw pipe dimension since 40 °C is the bypass temperature if heat exchanger is hot during idle. In case the heat exchanger is cold during idle, then this introduces an additional time delay as described above. In all cases, time delay is dependent on dhw flow, resulting in delay in the dhw pipe.

Future energy supply/demand perspective One important challenge for DH is to convert to 4th generation DH systems. Intention is to realise efficient DH systems for urban areas where heat demands will decrease due to modernisation and new building energy saving codes. In this context one way to go is to reduce temperatures in DH networks [9]/[10]. This allows for cost effective geothermal sources as well as other renewable low temperature sources. For dhw, normal temperature level is 45 to 60°, where higher temperatures typically are based on considerations towards Legionella. A way to reduce temperature levels in DH networks is to set dhw temperature at 45 °C. By this a primary temperature at sub station of 50 to 55 °C will be sufficient. A precondition for this is to use heat exchangers for dhw production, like the flat station concept. CONCLUSION The two pipe flat station concepts, consisting of decentralised instantaneous dhw production, open the possibility of reducing general DH net work temperature, which for the future will be even more relevant due to decreasing building heat demand and increased availability of renewable energy. For building owners, the investigated case shows that the flat station concept is on brake-even investment level compared to

Hygienic considerations Legionella is a well-known bacterial risk for dhw systems. Normally it is not the question whether Legionella is present in the dhw system or not, but 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 energy saving due to less installed hot pipes. Energy savings are in the range of 2 to 4 kWh/m^2/y for the investigated cases. Comfort level has been investigated, revealing well acceptable dynamic control performance. Dhw temperature recovery after an idle period for the instantaneous preparation of dhw is, however, a trade-off between comfort and energy saving. Related to Legionella, then risk can be reduced when installing flat stations as presented in this paper.

[5] Thorsen, J. E. Control Concepts for DH Compact Stations Investigated by Simulations, The 9th International Symposium on District Heating and Cooling 2004. [6] http://www.danfoss.com/Products/Categories/List/H E/Temperature-Controllers/Temperaturecontrollers/IHPT-and-XB-06/b1c8a73c-59f1-4fef8b52-f49c97b6019b.html [7] http://www.danfoss.com/Products/Categories/Group /HE/District-Heating-Substations/SubstationsDirect-Heating/Flat-Stations/8f81605b-bab9-4644961b-51a3f0503f05.html

REFERENCES [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 Behaviour‖, the magazine ―News from DBDH‖, #3/2006.

[9] Olsen, P.K., Lambertsen, H., Hummelshøj, R., Bøhm, B., Christiansen, C.H., Svendsen, S., Larsen, C.T., Worm, J. A new Low-Temperature District Heating System for Low-Energy Buildings, The 11th International Symposium on District Heating and Cooling 2008.

[3] DS 452, Code of practise for thermal insulation of technical service and supply systems in buildings, 2. Revision, Dansk Standard, 1999

[10] Paulsen, O., Jianhua, F., Furbo, S., Thorsen, J. E. Consumer Unit for Low Energy District Heating Net Works. The 11th International Symposium on District Heating and Cooling 2008.

[4] Thorsen, J.E. Cost considerations on Storage Tank versus Heat exchanger for htw preparation, The 10th International Symposium on District Heating and Cooling 2006.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

IMPROVED TEMPERATURE PERFORMANCE OF RADIATOR HEATING SYSTEM CONNECTED TO DISTRICT HEATING BY USING ADD-ON-FAN BLOWERS 1

Per-Olof Johansson , Janusz Wollerstrand 1&2

Lund University, Department of Energy Sciences, Division for Efficient Energy Systems Corresponding author: per-olof.johansson@energy.lth.se, Energy Sciences, Lund University, P.O. Box 118, 221 00 Lund, Sweden, Phone: + 46 46 222 40 43, Fax: + 46 46 222 47 17 level. In order to reduce the temperature demand in existing buildings the idea of using small add-on-fan blowers placed under the radiator to increase the heat output due to an increased share of forced convection came up.

ABSTRACT District heating (DH), which is the most common heat source in multifamily houses and commercial buildings in Sweden, can be produced in several different type of production units. In order to gain thermal efficiency in a DH system it is important that DH supply and return temperatures are kept low. The temperature demand in the DH system is, during the heating season, dependent on the temperature level in the heating system of the DH connected buildings. Many production units benefit from a lowered DH return temperature, while others are more affected by a reduced supply temperature. In a CHPstation the heat to power ratio will increase when the DH supply temperature is decreasing. In order to reduce the temperature demand, low temperature heating systems are of interest, as well as systems resulting in a low DH return temperature.

Objective The field study presented in this paper investigates the possibility to reduce the space heating temperature program and estimates the impact on the DH supply and return temperature. Possible reduction of the DHflow rate is also calculated. This paper is focusing on buildings indirectly connected to the DH network through a substation with heat exchangers (HEX). DESCRIPTION OF ADD-ON-FAN BLOWER The add-on-fan blower that is tested in this study consists of several regular DC motor driven fans, originally used for cooling, mounted under a radiator, see Fig. 1. In the study, two different kinds of radiators were tested, a panel and a column radiator.

To increase the heat output in an existing radiator heating system, the radiators can be complemented with small electric fans resulting in an increased share of forced convection in the heating system. Field studies have shown that the heat output, with constant supply temperature and mass flow through the radiator, can increase with more than 50%.

Tss

Space heating radiator

For many years, return temperatures in DH networks have been an important issue for DH research. A low DH return temperature is in many cases favorable for the DH production units. However, if also the supply temperature could be kept at a low level the share of electricity produced in a CHP station could increase. This would lead the way towards an increased share of electricity produced by non fossil fuels. In Sweden more than 30 % of the DH is produced in CHP stations [4].

Outer wall

INTRODUCTION

Tsr Add-on-fan Increased air flow Floor

In many reports the gain from a reduced temperature level in the DH network has been discussed and quantified in economic terms, see e.g. [12], [13].

Fig. 1 The add-on-fan blower mounted on a panel radiator.

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

The DH supply temperature level in the DH network is, during heating season, dependent of the temperature demand in the DH-connected buildings heating system. In modern buildings low temperature heating systems are common, which may allow reduced DH temperature 22

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

(7)

Qrad  C  Tm3  

THEORY In this section a theoretic analysis of the impact of forced airflow on heat output from radiator with a length of (L) 1 m and height of (h) 0.59 m is described. The radiator is in this study approximated by a flat vertical plate. The indoor temperature is assumed to be constant at 21°C and equal to Tinf.

Convection The convection that arises due to the temperature difference between the radiator surface and the surrounding air is a function of the Nusselt number (Nu), see equation 8. (8)

 conv  Nu   / h

Heat output The heat output from the radiator to the room arises from convection and radiation. The heat transfer process from heating water to the room through a radiator is summarized in equation 1 [7], [8].

The heat output due to convection is divided into three sections, natural, mixed and forced convection. For natural convection, the Nu number is dependent on the Rayleigh number (Ra), which is a product of the Prandtl number (Pr) and the Grashof number (Gr). For air, Pr can be considered constant, Pr=0.71, wile

(1)

 s  c p (Tss  Tsr )  (k  A) Qm

k is the heat transfer (convection) from the water to the surrounding metal, conduction through the metal and convection from the outer surface of the radiator to the room according to equation 2.

 1 1 1   metal  k  watermetal metal  conv   rad

Gr  g     

(9)

2

  1 / Tinf  1 / Ti

(2)

where g is the gravity force, ν is kinematic viscosity and β is the coefficient of expansion.

The dominating parameters in this equation are the convection and radiation between the radiator and the room (αconv and αrad), while the other terms, in this case, can be neglected. This results in a new equation for energy output, see equation 3.

Several empirical relations describing Nu are available. In this study a relation described by Churchill has been used [9], see equation 10 and 11.

Q  Qrad  Qconv   conv  Aconv     rad  Arad   The temperature Δθ is the logarithmic temperature difference according to equation 4.

Nu  0.68 

(3) Nu

mean

0.5

 0.825 

0.387  Ra 1 / 6 [1  (0.492 / Pr) 9 / 16 ]8 / 27

Ra  109

(10)

Ra  10 9

(11)

For forced convection the Nu number is calculated by equations described by Holman [10], see equations 12 and 13.

(4)

T  Tsr   ss T  Ti ln ss Tsr  Ti

0.67  Ra 0.25 [1  (0.492 / Pr)9 / 16 ]4 / 9

(12)

Nu  0.664  Re 0.5  Pr1/3

Re  5  105

Radiation

Nu  Pr 1/3  (0.037  Re 0.8  871)

5  105  Re  10 7 (13)

According to Trüschel [8] the heat output from radiation can be estimated according to equation 5.

and the Reynolds number, Re, is described as:

Qrad   rad  Arad   

Re 

 4

 rad    Tm3  Arad   Arad  rad   (1   rad ) Aradiator

(5)

(Tsf  Tsr ) / 2  Ti 2

(14)



The product of Gr/Re2 describes the dominating type of convection. If Gr/Re2>10, natural convection is dominating, if Gr/Re2≈1, both natural and forced convection is of importance and if Gr/Re2<<1, forced convection is dominating. When a mix of forced and natural convection occurs, the Nusselt number is calculated according to equation 15 [11].

Where the temperature, Tm, is the mean temperature of the radiator surface and the surfaces in the rooms, see equation 6. For a panel radiator the Arad/Aradiation=1 [8].

Tm  Tradiator  Troom,surface 

uL

(6)

Nu  ( Nu forced  Nu natural)1 / 3

Since the Arad and emissivity, εrad, are constant for a specific radiator, the relation can be simplified to equation 7. 23

(15)

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Impact of air speed on space heating temperature

T sr0= 45 C , Qrel= 100 %

Temperature (  C)

% additional Q, rad&conv

 C and %

T sr0= 29.9 C , Qrel= 25 %

T ss0= 33.6 C , Qrel= 25 %

40 35

15 20 air velocity (m/s)

Fig. 3 Possible Tss and Tsr to for three heat load situations at different air speeds. Qrad=65% at DOT.

New temperature programs have been derived for some moderate air speeds, see Fig. 4. As seen the impact of an increased air flow, expressed in °C, is larger at high relative heat load.

150

100

NEW TEMPERATURE PROGRAM

65 60

8 10 air speed (m/s)

Temperature (  C)

T sr0= 35.6 C , Qrel= 50 %

20 0

% additional Q, conv T sr,rad&conv ( C)

 (W/m K)

T sr,conv ( C)

20 10 0 0

T ss0= 43.1 C , Qrel= 50 %

T sf ( C)

0 0

T ss0= 60 C , Qrel= 100 %

Results from the theoretical analysis, using the equations above, are shown in Fig. 2 to Fig. 4. The heat output for a radiator designed for the temperature program 60/45 °C is illustrated as a function of the air speed in Fig. 2. The supply temperature and the mass flow through the radiator are kept constant. Two cases have been derived, one with heat output only from convection, and one with heat output from both radiation and convection. With ε=0.9, the share of heat output from radiation will be 65% at DOT.

8 10 air speed (m/s)

T sf0 U= 0.0m/s T sr0 U= 0.0m/s U= 0.5m/s U= 1.0m/s U= 2.0m/s U= 3.0m/s

40 35 30 25

Fig. 2 Calculated heat output improvements at Tss=60 °C, Tsr0=45 °C with increasing airspeed. ms is kept constant.

20 15 0

As seen, the additional heat output from the radiator is increasing rapidly when the air flow is increased. With radiation taken into account, the increase is somewhat lower since the mean temperature, Tm, is decreased, see equation 7.

40 60 relative heatoutput (%)

100

Fig. 4 New space heating temperature programs at different air speeds. Red lines: Tss, Blue lines: Tsr. Qrad=65% at DOT.

In the calculations performed, the radiator is assumed to have the same heat output from both sides of the radiator. The air flow is assumed to be uniformly distributed through the length and height of the panel radiator. This is not the case in the real add-on-fan blower applications, however, one can expect results following the same pattern.

In Fig. 2 the heat output is increasing. In Fig. 3 and Fig. 4 the supply temperature to the radiator is reduced instead to keep the heat output constant. New Tss and Tsr can now be calculated under the assumption that the total heat output and the mass flow (ms) through the radiator are constant. The impact of the air flow is described for three different heat loads (Qrel=100%, 50%, 25%) with standard 60/45 °C temperature program as a reference. See Fig. 3.

EXPERIMENTAL STUDY To investigate the performance of the add-on-fan blower, two radiators of different type were supplied with such device during the heating season 2009/2010. The power supply to the fans was scheduled to switch on and off while the mass-flow (ms) through the radiator was kept at a constant level. Field study object The radiators are situated in two offices at Lund University. The original temperature program for the 24

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 60/45ºC at DOT (represented by narrow black lines in Fig. 8 trough Fig. 13).

* Calculated for new radiators of the same dimensions manufactured by Lenhovda radiator factory [3]

The radiators are located in traditional environment in a building built in 1960.

The radiator types tested were: 

Panel radiator, see Fig. 5



Column radiator, see Fig. 6.

office

Data acquisition Measured parameters in the test were: secondary supply and return temperature (Tss and Tsr), indoor temperature (Ti) and outdoor temperature (Tout). The impact on return temperature for a given supply temperature was then calculated. In Fig. 7, a screenshot from the logger software is shown. The return temperature is decreased by 5°C when the fan is switched on. This results in an increased heat output by more than 60 %. fan

Tsr,0=39 Tsr,Fan=34

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

Tss (ºC) Tsr (ºC) Ti (ºC) Tout (ºC) Ufan (V)

Fig. 7 Log file from field study.

New reduced temperature program will be derived in next section. MODIFYING SPACE HEATING TEMPERATURE PROGRAM Method When the add-on-fan blower is switched on, the Tpr is decreasing, causing an additional heat output since ms is kept constant. See Fig. 8.

Fig. 6 Add-on-fan blower mounted on a column radiator. For each radiator the fans have been run at two different rotation speeds. The net electric power consumption (Pfan) has been measured. See Table 1 for Pfan and the design heat energy output at DOT. Note that the electric power to the add-on-fan blower is constant and not dependent on the relative heat output.

Temperature [C]

Pfan (el)

Q @ DOT, 60/45ºC (Heat)

Panel

2.7 W 1.9 W

360 W*

Column

3.0 W 2.2 W

430 W*

Tsr,0 Tsr,Fan

Table 1 Radiator and add-on-fan blower design. Radiator type

Tss

Original cooling in radiator

Additional cooling with add-on-fan-operation

35 30 25 20

0.2

0.4 0.6 Relative heatload

0.8

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 without fan operation is calculated from equation 1.

Qrel ,0  m s  C p (Tss  Tsr 0 ) QDOT Qrel ,Fan  m s  C p (Tss  Tsr ,Fan ) QDOT

  

Panel radiator Pfan = 2.7 W 60

(1)

Tsr,0 Tss,Fan

(T  Tsr ,Fan )  ss Qrel ,0 (Tss  Tsr 0 )

Tsr,Fan

T [C]

Since the temperature drop in the radiator is increasing with the fan in operation, the radiator now could be considered oversized. Then, with the same type of reasoning as in e.g. [2], the Tss program or ms needs to be adjusted in order to avoid overheating of the building. In this paper, the ms has been considered constant, allowing us to compute the new relative space heating load for a given Tss according to equation 2. Qrel,0 is computed using the original space heating temperature program.

Qrel ,Fan

Tss,0

40 35 30 25 20

0.2

0.4

0.6

0.8

(2)

Panel radiator Pfan = 1.9 W 60 Tss,0

Tsr,0 Tss,Fan

Tsr,Fan

T [C]

Tss,Fan

45 s

ss,0

60 55

40 35

sr,0

sr,Fan

0.2

0.4

0.6

0.8

1.2

1.4

Qrel

Fig. 11 Modified temperature program for panel radiator, Pfan=1.9 W.

30 25 20

1.4

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

Knowing Qrel,Fan, a new temperature curve, which will result in correct heat output from the radiator with the fan in operation, can be calculated. The curve appears to the right in the diagram, see Fig. 9.

1.2

Qrel

Column radiator Pfan = 3.0 W 60 0

0.2

0.4

0.6

0.8

1.2

Tss,0

1.4

Qrel

New space heating temperature program - results

T [C]

Fig. 9 Modified secondary temperature program.

The procedure described above has been applied to all collected data. Results are shown in Fig. 10 and Fig. 11 for the panel radiator, and Fig. 12 and Fig. 13 for the column radiator.

Tsr,0 Tss,Fan Tsr,Fan

40 35 30

As seen in the figures, the temperature program is now significantly reduced for both the panel and the column radiator. The new temperature program shows a similar pattern for both types of radiators. For the panel radiator, the measured return temperature is more concentrated, especially at the higher fan speed. This could be explained by a more favorable air flow pattern due to the physics of the radiator.

25 20

0.2

0.4

0.6

0.8

1.2

1.4

Qrel

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

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia Column radiator Pfan = 2.2 W

DH primary supply temperature 110

60 Tss,0

Tsr,0 Tsr,Fan

Tps [C]

T [C]

100

Tss,Fan

35 30

25 60

0.2

0.4

0.6

0.8

Qrel

0.6

0.4 0.2 rel heatload

Results

INFLUENCE ON DH NETWORK

The first control strategy is in Fig. 15 – Fig. 20 noted as „Tps unchanged‟, and the second strategy is noted as „mp unchanged‟.

Knowing the reduced temperature level on the secondary side of the HEX, the impact on the DH network can be estimated. The impact is calculated based on two different strategies:

0.8

Fig. 14 DH primary supply temperature.

Fig. 13 Modified temperature program for column radiator, Pfan=2.2 W.

In Fig. 15 and Fig. 16 the possible reduction of DH supply temperature is shown.

Primary supply temperature (Tps) is kept at the same level as before

Panel radiator T

saving

15 Pfan = 2.7 W mp unchanged

The primary flow (mp) through the HEX is kept constant

= 1.9 W m unchanged p

saving [C]

By applying the first strategy, both Tpr and the mass flow in the DH network is reduced. The second strategy results in a reduced Tps and Tpr without changing the flow rate in the DH network.

fan

Results so far will now be applied to a DH substation dimensioned as recommended by the Swedish district heating association [1]. The calculations are made with a parallel connected DH substation serving a building with 20 apartments. The substation is providing the building with heat and domestic hot water (DHW) and DHW circulation. The assumed DHW usage is 125 l/apartment&day, space heating load at DOT is 3 kW/apartment. The heat loss from DHW circulation is assumed to be 0.1 kW/apartment. For each space heating load a flow-weighted mean value for Tps and Tpr is calculated for a time period of 24 h, including heat load from both DHW and DHW circulation. The reference DH supply temperature, dependent on the space heating load, is assumed as illustrated in Fig. 14.

0.8

0.6

0.4 rel heatload

0.2

Fig. 15 Resulting Tps reduction with panel radiator. Column radiator T

saving

15 Pfan = 3.0 W mp unchanged fan

= 2.2 W m unchanged p

saving [C]

0.8

0.6

0.4 rel heatload

0.2

Fig. 16 Resulting Tps reduction with column radiator.

In Fig. 17 and Fig. 18 the reduction of Tpr is shown. As seen the reduction of Tpr is of the same magnitude independently of which control strategy is used. 27

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

Column radiator m saving (% ) p

15 pr

saving

m saving [%]

Panel radiator T 15

Pfan = 2.7 W Tps unchanged fan fan

unchanged

= 2.7 W m unchanged p

Pfan = 1.9 W mp unchanged

P P 0

saving [C]

= 1.9 W T

0.8

0.6

fan fan

= 3.0 W T = 2.2 W T

0.4 rel heatload

unchanged unchanged

0.2

Fig. 20 Resulting mp reduction with column radiator.

Annual gain in Tps and Tpr during heating season 0

0.8

0.6

0.4 rel heatload

0.2

In order to evaluate the annual impact on the primary temperature level, the outdoor temperature has to be considered. In this case measured values for the outdoor temperature in Malmö have been used, see Fig. 21.

Fig. 17 Resulting Tpr reduction with panel radiator. Column radiator T

saving

Pfan = 3.0 W Tps unchanged unchanged

= 3.0 W m unchanged

Pfan = 2.2 W mp unchanged

ps p

T out ( C)

fan

= 2.2 W T

saving [C]

fan

20 15 10

5 5 0 -5

0.8

0.6

0.4 rel heatload

0.2

-10 May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May

Fig. 21 Outdoor temperature in Mamö st th 1 May 2006 – 30 April 2007.

Fig. 18 Resulting Tpr reduction with column radiator. Panel radiator m saving (% ) p

When calculating the annual gain for a DH-network a comparison of flow-compensated mean temperature during the heating season has been made. For the calculations we assume a maximum heat output (QDOT) at -15 °C and the balance temperature, when no space heating is needed, +17 °C.

Table 2 Reduction in annual primary temperature level during heating season

m saving [%]

P 0

0.8

0.6

fan fan

= 2.7 W T = 1.9 W T

0.4 rel heatload

0.2

Tps unchanged

unchanged unchanged

ΔTpr

mp unchanged

T ps  T pr 2

Panel Column radiator radiator

Fig. 19 Resulting mp reduction with panel radiator.

Pfan= 2.2 W

-2.2 °C

-2.4 °C

Pfan= 3.0 W

-5.7 °C

-6.6 °C

Pfan= 1.9 W

-0.8 °C

-0.9 °C

Pfan= 2.7 W

-2.5 °C

-2.7 °C

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Note that the annual flow-compensated mean temperature in Table 2 is based on results from the field study where measured values for low relative heat load are missing, which makes the values in the table somewhat underestimated.

ACKNOWLEDGEMENT

CONCLUSION AND DISCUSSION

Abbreviations

This work is part of the Primary Energy Efficiency project of Nordic Energy Research. NOMECLATURE

By installing the add-on-fan blower application on existing radiators the temperature level in the heating system can be substantially reduced. This will also have impacts on the DH network and DH production units. The impact on the DH network can be applied based on two principles: 1) DH supply temperature kept at the same level as without the add-on-fan blowers. This will result in reduced primary flow rate.

CHP

Combined heat and power station

District heating

DHW

Domestic hot water

DOT

Design outdoor temperature

HEX

Heat exchanger (DH substation)

Variables α Heat transfer coefficient (W/m2.K)

Gr Grashof number (-)

β coefficient of expansion (K-1)

h Height (m)

The first strategy could be applied immediately, since the primary supply temperature is kept as the same level as before. This means that not all heating systems connected to the DH network need to be modified in order to apply this method. The lowered secondary temperature level results not only in reduced DH-return temperature, but also in a reduction of the DH flow rate. The reduced flow rate could be used to increase the number of buildings connected to the DH network, or to avoid bottlenecks in the DH network. The magnitude of the reduction of the DH supply temperature is between 9 and 12 °C at DOT and at the same time the flow rate is decreased with more than 10 %. On annual basis the possible reduction of temperature level in the DH network is in the magnitude of several degrees Celsius.

δ Thickness (m)

k Heat transfer coefficient (W/m2.K)

2) Reduced DH supply temperature while primary flow rate is kept constant.

Conductivity (W/m.K)

L Length (m)

Emissisivity (-)

m

mass flow (kg/s)

ν Kinematic viscosity (m2/s)

Nu Nusselt number (-)

ζ Stephan-Boltzman constant

P Electric power (W or kW)

Δθ Logarithmic mean temperature difference (K)

Pr Prandtl number (-)

A Area (m2)

Q Heat output (W or kW)

cp Heat capacity (J/kgK)

Ra Rayleigh number (-)

In order to apply the second strategy the demand for a high temperature level in the DH network needs to be reduced for all the connected buildings. Otherwise the DH flow rate will increase. Calculations based on the results from the field study in this paper shows that the DH supply temperature can be reduced with about 10 °C at DOT without affecting the DH flow rate. At the same time the DH return temperature will be reduced with as much as 10 °C at DOT.

C Constant

Re Reynolds number (-)

The performance of the tested add-on-fan blowers corresponds to the pattern of theoretical calculations. However, the results are not comparable since the air flow in the pilot project has not been measured.

g Gravity force (N/s )

T Temperature (ºC or K)

Subscripts 0

Design condition (without fan)

Primary (side)

Fan Add-on-fan blower in operation

Return

indoor

rad radiation

Mean

rel

Relative

Secondary (side) or Supply

out outdoor

The results presented here are an important part in the evaluation of effects of improvements in consumer heating systems on primary energy efficiency in DH systems including production plants, especially CHP.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

REFERENCES

[7] J. A. Myhren, S. Holmberg, ‖Design consideration with ventilation-radiators: Comparisons to traditional two-panel radiators‖, Energy and buildings 41, p. 92-100, 2009

[1] Svensk Fjärrvärme, ―District heating substations design and installation, Technical regulations F:101‖, The Swedish District Heating Association, 2008

[8] A. Trüschell, ―Värmesystem med luftvärmare och radiatorer, En analys av funktion och prestanda‖, Licentiate Thesis, Chalmers, Göteborg, 1999

[2] P. Ljunggren, P-O. Johansson, J. Wollerstrand, ―Optimized space heating system operation with the aim of lowering the primary return temperature‖, in Proc. of the 11th International Symposium on District Heating and Cooling, 2008, Reykjavik

[9] S. W. Churchill, ―Correlating equations for laminar and turbulent free convection from a vertical plate‖, Int. J. Heat Mass Transfer, Vol. 18, p. 1323-1329, 1975

[3] http://www.lenhovdaradiatorfabrik.se/display_sub.a sp?apid=20, 2010-04-16, Downloaded spread sheet for calculating heat output.

[10] J. P. Holman, ―Heat transfer‖, 9th edition, 2002 [11] Discussion with professor B. Sundén, April 2010

[4] Swedish Energy Agency, 2008, ―Energy Indicators 2008, Theme: Renewable energy‖, 2008

[12] P. Selinder, H. Walletun, ‖Modell för ändrade förutsättningar i fjärrvärmenät‖, Rapport 2009:50, Svensk Fjärrvärme, 2009

[5] http://a-energi.jetshop.se/, 20010-04-20 [6] EN 15316-4-5:2007, ―Heating systems in buildings. Method for calculation of system energy requirements and system efficiencies‖, CEN, Brussels, 2007

[13] S. Werner, FVB Sverige AB, ‖Nytta med svensk fjärrvärmeforskning‖, FoU – orientering 2004:9, Svensk Fjärrvärme, 2004

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

PRIMARY ENERGY EFFICIENCY AND SYSTEMS ENGINEERING 1

M.Berner , R. Ulseth , J.Stang 1

Distribution

Transmission

Generation

Transport

Storage

The revised Energy Performance of Building Directive (EPBD) [1] emphasizes that the energy performance of a building shall be calculated by use of Primary Energy Factors (PEF). Calculation of CO2 emission will not be mandatory so far. Thus EPBD will reduce the use of non-renewable energy, incite the use of energy from combined heat and power generation (CHP) and reduce the energy consumption in the building sector.

Processing

Extraction

ABSTRACT

Transformation

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

Figure 1 A Typical energy chain

Methodology An energy chain might consist of several elements or processes from extraction, through processes such as drying, storage, transport, power/heat/cool generation, and distribution to the end user. In order to ensure that there is a correct PEF, all elements that influence the energy flow have to be accounted for. The energy balance or calculation of the energy efficiency of a process focuses primary on the energy input in the form of fuel and the output in kWh, and lacks information on the energy used to build infrastructures such as the power plant, distribution net, transportation and the extraction.

A simplified method that enables comparison of the PEF from different energy chains is required. However, calculation of all the parameters affecting the PEF values like energy used for extraction, transportation, power and heat generation etc. is time-consuming. The method described in EN 15603 [2] is rather general l and provides PEF values for 13 energy carriers and chains. This is based on average European values. Life Cycle Assessment methods include several of the relevant steps, but a complete LCA often imply collection of more than 6000 parameters.

Life Cycle Assessment (LCA) might contribute to provide such information in a generic method. However, the number of input parameters, often more than 6000 in an ordinary LCA analysis demonstrates the need for an easily accessible method.

The systems engineering method used here have demonstrated the feasibility of developing a generic method that provides credible data for calculating primary energy efficiency. It applies the generic method on energy chains in the Nordic region which is relevant to CHP plants utilising bio based fuel.

Systems engineering is a method that has been developed gradually with increasing complexity of projects and systems. Systems engineering is often considered to have started at Bell Laboratories in the 1940s, later applied in organizations such as NASA and formalized as a separate engineering field with the formation of INCOSE [9] in 1990. The benefits of systems engineering is the possibility to treat complex systems with several subsystems. Therefore, as a first step in the development of a method a systems engineering approach has been chosen. The main objective is to develop systems and methods that enable a sufficiently reliable calculation to be made of the primary energy factor (PEF) in general and for different energy chains with required level of details.

INTRODUCTION Background The terms Primary Energy, Primary Energy Efficiency and Primary Energy Factors (PEF) are introduced [3] [8] in order to compare different energy sources and chains based on losses and a calculated environmental impact. Primary energy is energy that has not been subject to any conversion or transformation process. The use of primary energy factors takes into account the energy that are used from the extraction of the energy carrier and all of the losses until energy is delivered to the end use in the desired form such as heat, cooling or electricity .

At present systems engineering approaches have not been found to have been previously applied on the development of generic PEF methods for different energy chains.

The primary energy factor (PEF) expresses how much primary energy is needed to deliver 1 unit of power, heat or cooling to the end user. The term primary energy efficiency (PEE) therefore is used to describe the total use of energy from extraction to the end user.

Objective 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 different energy chains is developed. The paper will show an overall approach and will not describe all the necessary iterations in detail.

Measures of effectiveness (MOE)

The definition of MOE are: ‖A small subset of the requirements that are so important that the system will fail if they are not met and will be a huge success if they are met‖ [11].

SYSTEMS ENGINEERING

The system engineering process

Development of information models

The different information models describe the observed system in relation to legislation, physical architecture and a system interface model. Four separate models are developed

A systematic approach such as systems engineering is essential to be able to develop a generic model describing a complex system with several subsystems. The intention with the systems engineering process is to analyse and describe complex systems. Often the method is used in the design process, to make sure that the subsystems are connected properly, that the process is optimized and that the different components are described, implemented and integrated precisely. A common feature of all systems engineering processes is an indefinite number of iterations at all different steps.



Requirement traceability model



System architecture model



Behaviour model



System interface model

Trade-offs

The trade-off phase is essential in the development of a method. Each of the steps is carried out in iterative loops gradually increasing detailing level. After satisfactory trade-offs have been performed and consistent information models obtained, a theoretical method is developed. Real data are collected and trade -off between the model and the gathered data are performed.

Systems engineering principles are often applied when a new system or products are developed. The methodology alters slightly between development and re-engineering. Re-engineering methods are applied when an existing system is described. The energy chains considered are already designed and built, and a re-engineering technique is selected in order to develop a method that calculates the PEF for different kind of energy chains.

Documentation

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 following six different tasks according [2]. Some of them might seem unnecessary, but they all contribute to the decomposing of a system and development of a method. 1.

Iterate to find feasible solution

Create requirement traceability model

Feasible solution

4 1 Establish problem statement

Establish problem statement;

This comprises the definition of the problem approach, which includes development of a problem statement describing the problem/challenge, its importance and a state of the art. To be able to establish the problem statement; four questions must be answered:

2 Asssess available information

3 Define effectiveness measures

Create system architect. model

Tradeoffs

and 4 Create behaviour model 4 Create context model

What is the problem?

Figure 2 The system re-engineering process described as a functional block diagram (FFBD), ref. [10]

Why is it important What have others done?

ESTABLISH PROBLEM STATEMENT

What must be done? 2.

No feasible solution

What is the problem?

Assess available information assessment

Use of Primary Energy Factor (PEF) will provide information on the energy losses and consequently the environmental impact of different kind of energy sources, power production processes and energy

Provide available information including an overview of possible stakeholders.

6 Document current system design

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 accessible calculation method.

and process lines (chains) primary in Norway and the Nordic countries. Detailed data must be provided such as efficiency and loss from the different systems and mix of systems, or at least provide the necessary parameters. Since the systems engineering approach is chosen, the problem approach must be defined, a theoretical method developed and data collect. This includes performing of a trade-off between the theoretical model and available information. The method shall be tested by selected case studies and finally adjusted.

Different countries have different energy chains and energy supply systems. Analysis of even the most actual processes and process lines does not exist neither for Norway or Europe [12] – [13]. In order to compare and choose different energy chain there is a need for standardized methods. The lack of objective and reliable data of the different elements in the energy chain might prevent an efficient use of energy, and contributes to wrong choices and unnecessary CO2 emissions. [14] – [15].

Main hypothesis

The method is principally described in EN 15603 [2] and provides only single PEF values for 9 energy carriers and 4 energy chains, and is based on average European values. Without an easy accessible method or methods is it not possible to compare PEF values and calculate the actual environmental impact of different energy chains. Some studies [16] -[21] have described parts of this topic, but they lack a holistic view of the energy chains from cradle to grave, often the chosen system boundaries are different, time scale varies, detailing level different and the, approach/ method varies. Results from different studies therefore are not comparable.

As a part of the systems engineering process, one or several (systems engineering) hypothesis is developed. The success of a system engineering process is related to the fulfilment of the hypothesis. In this project the system engineering method must prove two main hypotheses;

Why is it important

It is possible to develop a generic method that provides credible data for calculating primary energy use by use of PEF values.

It is possible to apply the generic method on energy systems in the Nordic region for CHP plants utilising bio based fuel.

Stakeholder analysis

PEF is a key indicator to be able to evaluate energy use (for different purposes) especially with regards to the goals of the EPBD [1]. PEF is an over all energy efficiency indicator which makes it possible to compare and collocate different energy sources and energy carriers by a single number. The same method can be used to calculate the CO2 emission.

A stakeholder is a party having a right, share or claim in the system [16]. The intention with the stakeholder analysis is to reveal the different kinds of stakeholders since they might have requirements influencing a possible method in a legal way. Stakeholders with mutual interest are aggregated in groups; some of them might not be in incompliance with each other.

What have others done? Different CEN standards describe, and partly discuss, the theory. In the EC-mandated CEN standards related to EPBD mainly one single reference are referred [14] whilst the PEF values have been gradually changed over time. An extended literature survey has showed discontinuity between some of the studies performed and lack of details in the calculations. Methods developed to provide PEF values for heating systems in buildings might be useful, but they will not totally comply with a whole energy chain approach. Life Cycle Assessment (LCA) might also contribute to a generic method, but the vast number of input parameters, often more than 6000 in a traditionally LCA demonstrates the need for a more easy accessible method. What must be done? In order to develop a method a systems engineering approach will be used. The most important task in this context is the identification of relevant energy systems 33



Energy producer, distributors, energy companies; Business profitability is the main issue by optimizing production from different energy carriers according to cost-benefit



Investors (energy and building); The electricity markets are opening gradually throughout Europe, e.g. Nord Pool Financial Marked and investments in Power production and the introduction of so called Green Electricity Certificates might be a new or extended business area. Investors in the building marked might be interested in the actual PEF values and primary energy use when choosing between different investment objects.



Building owners, end user; Correct calculation of PEF values and primary energy use is supposed to have significant importance for the choice of energy supply system, building services, insulation level, especially for new buildings and major rehabilitation projects. Future operating cost might

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia



depend of PEF since taxes might be dependant of the primary energy use and/or the CO2 emission.

INFORMATION MODELS

Developers& Building and construction industry; Technical equipment in the building and design strategies depends on the actual use of the specific energy carrier. The use of PEF values in the primary energy calculations might change the value of traditionally installed equipment due to overall energy costs and also create a demand for more energy flexible solutions.

In systems engineering shall the requirement traceability information model “aim to show the break down of requirements from source documents to final allocation functions to stakeholders ― [2]

The requirement traceability information model

This model is an important tool to keep track of the different requirements, source documents and eventually what the system accomplishes and who or what are in charge. Usually an Entity-RelationshipAttribute method is used [23], where the entities (objects) represent the legislation, requirements, etc. whilst the relationship shows the association between the system/process.

Politicians, government, Regulators, Community planning; Most European countries have affiliated the Kyoto Protocol, and a possible method to increase the use of renewable energy policy tools and subsidy schemes might be based on the use of PEF values for the different solutions, besides possible tax on systems with high primary energy use. National regulators mandatory monitor and report emission levels and this influences national legislation, local and urban planning

EPBD Energy Performance.

Landfill Directive

Source

Building regulation TEK

EN 15316-44:2007)

ËN 15603:2007

Waste regulations

Source

Documents

Incorporates

96/62/EC Ambient Air Quality Source Documents 1999/30/EC Limit values NOx… Requirements Incorporates Pollution Control Act

Building Guide REN Source

Research groups, Universities; Different research communities might be interested in development of other PEF calculation methods or adjustments of methods and development of new solutions and systems

Specifie NS3031 s function Allocated to Building Permit stakeholder

Requirements Specifie s Disharge permit

Specifies Discharge permit

Requirements Allocated to

Requirements Specifies

NOx emissions

Internal control system

Boiler

function Allocated to NOx emission

As earlier stated the measures of effectiveness (MOE) should be independent of any solutions and not concerned with internal details [22]. It might also be fruitful to develop MOE for the different kind of stakeholders since they often might have a different opinion regarding MOE.

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

1.2.2.1 Stakeholder

Most EC directives are enforced and implemented in laws, directions, regulations and guidelines, both in the EU and associated EEC countries, hence the order of entities in Figure 3. Several directives influence the national laws and regulations. Since the directives usually are enforced through national laws, the law includes requirements from more than one directive like the Norwegian Planning and Building Act [26], which includes requirements from EPBD [27], Directive on the promotion of the use of energy from renewable sources [30], Directive on the landfill of waste [29], The Pollution Control Act [30] amongst others.

In this context MOE are primarily described for the ongoing Nordic PhD-project Primary Energy Efficiency (PEE). A further detailing level, by including the stakeholders, might provide valuable information, but that is considered to lie outside the scope of this work. The methods (tools) developed during the project should be suitable for different kind of energy chains.



Working Environm. Act

Documents

DESCRIPTION OF MEASURES OF EFFECTIVENESS



Planning and Building Act

The results should be utilised by the different kinds of stakeholders e.g. the building owner, the architect/designers of the building, the energy supplier and producer and finally politicians and governments.

The requirements traceability model provides important information about constraints regarding an energy chain. Some of the elements such as the the Working Environment Act [31] might seem irrelevant, but the regulations set limits for the pollutant inside the working area, introducing need for e.g. conveyor belts.

The methods will enable the different stakeholders to choose between different energy systems and furthermore be able to reduce primary energy use and CO2 emissions from stationary energy purposes.

Each function consists of several entities for instance the discharge permit from regulators like The Norwegian Climate and Pollution Agency will set restrictions on the authorised discharge levels of different gasses not only NOx as illustrated in the Figure 3.

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 instance the Planning and Building Act will affect both end users, construction industry and energy distributors.

An energy chain consists of several sub systems as described in Figure 4. A more detailed architecture information model is also developed. Each of those sub elements can be spilt up into sub elements as shown in Figure 5. The final or basis element can be described as Figure 6.

Architecture information model The architecture information model shows the physical components of a system with subsystems.

Energy Transformatio n system

In order to describe the possible physical systems a generic model is developed [11], detailed description of some of the most relevant energy chains are carried out in the actual PhD thesis, Figure 1 shows a principal descriptions of a single energy chain. An end user will typically be supplied with energy from a various number of energy chains, and each element might represent parallel processes.

Consist of Energy Feeding system

Fertilizing, cultivation, logging, logging track, loop of twigs, trimming, transport



Chipping, packing, transport, local roads



Intermediate storage, transport regional roads



Transport central and regional roads



Building, operation demolition of power plant, technology, efficiency, part-load, size, Lifetime, waste treatment, gas cleaning supply of additives, internal transport



Transformation to central net, building, operation, demolition of infrastructure, heat/power loss



Transmission to local net, building, operation, demolition of infrastructure (pipes, high-tension lines heat/power loss (insulation, temperature levels (supply, return, ground), twin/single pipes, length)



Elec prod Other (chem etc) Fuel in

(1) Fuel

Dismantling Hea t

Heat prod Dismantling

Purification system Component

Combustion camber

Energy Transformatio n system

Component

Component Consist of

Internal Control System

Component

Stakeholder

Electricity production unit

Heat production unit

Component

Consist of

Built of Filter

Heat Transport system

Component

Heat Storage system

Component

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

Additional PEF

PEFin

Infrastructure, buildings, machinery etc.

Operation and Demolition maintenance

PEF out

Loss

Figure 6 Architecture information model, basis element

Since the Primary Energy Efficiency (PEE) of an energy chain consist of all of the elements from extraction to delivery the PEF for a chain can be calculated by E Chain  E Fuel  E Extraction  E Processing  E Storage  E Transport   E Generation

(1)

 E Transformation   E Transmission   E Distribution Construction Dist. net

Cold Heat

Dismantling Waste handlin g

Component Consist of

Combustion process

Consist of

Construction

Component

Consist of

Dismantling

Cooling Col prod d

Waste handling system

Consist of

Distribution to end user, building, operation, demolition of infrastructure (pipes, lines, substations) heat/power loss

Construction

Consist of

Energy Production system

Component

A CHP utilizing biomass might consist of the following elements: 

Component

Consist of

Construction

Sub station

Where E is the primary energy input to the system

Heat Cool

The Power Bonus Method

Dismantlin g

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

Electricit

E P   ( Edel,i  f P,del,i )   ( Eexp,i  f P,exp,i )

Figure 4 Architecturey information model for a part of the energy chain from generation including distribution, based on [25]

where EP – Primary energy input to the system Edel,I– Delivered energy, energy carrier i 35

(2)

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Figure 4 and Figure 5. This is an iterative process and the detailing level is the first steps gradually increasing, until the analysis (trade-off) of the different factors influencing the PEF value enables a removal of factors with an impact of 1% or less.

fP,del,i – Primary energy factor, delivered energy carrier i Eexp, – Exported energy, energy carrier i fP,expl,i – Primary energy factor, exported energy carrier i Power exported from the CHP plant multiplied with the PEF value for the replaced power shall be subtracted from the delivered primary energy to the buildings when calculating the PEF-value for the for the district heating system [7]. The power bonus method is enforced in order to promote CHP, and the subtraction of power produced and delivered outside the system boundary significantly reduces the PEF value for the energy chain. This implies that the PEF value for a CHP will be dependent on the power to heat ratio.

System interface information model The system interface model also denoted the context information model shows the systems interface with its surroundings and the environment. The model provides information on the core system and other interconnecting systems; this means a description on how things relate to each other. The context is according to [10] ―the interrelated conditions in which something exists or occurs‖.

Behavioural model The behavioural model is another information model the ―what it does“[10], but also described as ―the way in which an organism, organ, or substance acts, especially in response to a stimulus” [23].

Energy prod. Raw material

Energy

Energy source carrier Requirements

A behavioural model consists of functions, inputs and outputs and control operators. This implies that it is supposed to provide information on what is happening, in which order and what kind of iterations are performed.

Energy

Energy productio n

Requirements

Energy

Energy transport system

Requirements Requirements

The system boundary is drawn with a dashed line, and the system assessed lies within. Since this is a simplified model the relation towards investors, national regulators, constructors etc. are not shown. In this system energy source/carrier is closely connected to Extraction, Energy Source consists of storage and transport, Energy production corresponds with Generation and Energy transport system to Transformation, transmission and distribution in Figure 1.

Assess available information Define MOE And Energy production

End user

Figure 8 A simplified context information model.

Establish problem statement

Energy source

Energy

Energy transport system

The main issue has been to show the connection between the energy chains from production to end use. Politicians and national regulators might have specific requirements on each level. The building industry, constructors may likewise have interest on several of the levels, but a final listing is not possible to provide within this paper.

And Trade-off Develop generic model Collect data case study

Another important question is the definition of the system boundary. Precise definitions of the system boundaries are essential in order to be able to compare different studies. The system boundaries must distinguish between what is included and what lies outside of the task of the LCA, since the method must also rely on data collected by other parties and the use of different constraints might influence the quality of the method.

Test, evaluate method Simplify method Publish model

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

A more detailed partition of the energy chains have been applied in the development of the method. The energy chain is divided in subsystems as shown in 36

In order to provide information of the whole energy chain, all major elements have to be included i.e. the extraction phase is an integrated part of the chain.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

often significant lower than the actual exchange rate e.g. pipelines might have a twice times higher - more than 60 years. The use of yearly average efficiency and appurtenant power-to-heat ratio will often increase the PEF value for the whole system due to the impact of the power bonus method.

Trade-offs In the ongoing PhD-project of the first author a cut-off rule is set to 1%. That implies that factors with less than 1% impact on the final result can be removed. The trade-off considerations are still an ongoing process, and it will be presented and documented in a later paper. According to the main hypothesis this is not meant be developed as an optimization tool, since the intention of the ongoing PhD-project is to develop a method describing different energy chains.

The reliability of the method will be influenced by possible lack of detailed data, but based of average data a reliable comparison of different energy chains might be performed.

A complete trade-off could preferably [32] be performed by use of computerizes programme like CORE. [http://www.vitechcorp.com/solutions/]. The complexity of the different kind of systems shows the utility value of more than manual tools, which has been applied.

More standardized values for some the different parameters needs to be developed, like lifetime, heat load curves and extraction of biomass. Some adjustment will be necessary for instance for extraction where the transport distances are an important parameter. The resulting model can form a basis for future optimization tools, since only elements with major influence on the PEF-values are included.

Document current system design The results of the iterative process are described in the figures mentioned above. Only selected parts of the chosen design are illustrated in this document due to limitation in size.

ACKNOWLEDGEMENT This paper is developed as a part of the PhD-project Primary Energy Efficiency (PEE) and the work title for thee PhD-Theses is "System, methods and credible data for calculating primary energy efficiency in general and for energy systems in the Nordic region with special focus on energy systems applying CHP-technology with bio based fuel in particular".

The final system design is carried out according to Figure 6 for each element. CONCLUSION By performing a system engineering process describing different energy chains an outline of the model have been developed. The method has proven to be efficient in structuring the thoughts and will hopefully reduce mistakes in the future development of the model. The decomposition process in different subsystems is valuable, and the generic model will be able to treat different kind of energy systems and chains.

The project is financed by Nordic Energy Research with financial support from the industry and includes six PhD-studies carried out in the respective countries; Estonia, Finland, Sweden, Iceland and Norway. The projects objective is to contribute to the effort of enhancing the primary energy efficiency (PEE) and reducing CO2-emissions in the energy sector.

The systems engineering process have demonstrated that;

FURTHER INFORMATION PhD.student Monica Berner, Norwegian University of Science and Technology (NTNU).

It is possible to develop a generic method that provides credible data for calculating PEF-values and the primary energy efficiency. 2. It is e.g. possible to apply the generic method on energy systems in the Nordic region with CHP plants utilising bio based fuel The system engineering process provides a new approach to the design and development of a generic model describing PEF-values for energy systems with different kind of energy carriers. The method might be used for more than systems using CHP-technology since the model development are generic and thereby utilizes different kind of energy carriers.

Address: Monica.Berner@ntnu.no REFERENCES [1] Proposal for a Directive of the European Parliament and of the Council on the Energy Performance of Buildings Recast SEC (2008) 2820, SEC (2008) 2821) [2] EN 15603 – Energy performance of buildings – Overall energy use and definition of energy ratings [3] EN 15603:2007 Energy performance of buildings – overall energy use and definition of energy ratings

The method can provide detailed data (e.g. efficiency, loss etc) from the different energy chains and mix of chains. A major challenge is the data collection, some of the parameters lack standardization. The life time of different equipment varies, the economical lifetime is

[4] EN 15316-1: 2007 Heating systems in buildings – Method for calculation of system energy requirements and system efficiencies – Part 1: General 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 – method for calculation of system energy requirements and system efficiencies part 2-1 space heating emission systems

Systems, ITEV-Report 1999:06, Dr.ing Thesis 1999:117, NTNU [18] Sarigiannis D.A., Triacchini G., Meso-scale lifecycle impact assessment of novel technology policies: The case of renewable energy, Journal of Hazardous Materials 78, 2000 p. 145-171

[6] EN 15316-2-3:2007 Heating systems in buildings – Method for calculation of system energy requirements and system efficiencies – Part 2-3 Space heating distribution systems:

[19] Alanne K., Salo A., Saari A., Gustafsson S., Multicriteria evaluation of residential energy supply systems, Energy and buildings 39, 2007 p 12181226.

[7] EN 15316-4-4:2007 Heating systems in buildings – Method for calculation of system energy requirements and system efficiencies – Part4-4 Heat generation systems, building-integrated cogeneration systems

[20] Eriksson O, Finnveden G, Ekvall T, Bjorklund A, Life cycle assessment of fuels for district heating: A comparison of waste incineration, biomass- and natural gas combustion, energy Policy 35, 2007 p.1346-1362.

[8] EN 15316-4-5:2007 Heating systems in buildings – Method for calculation of system energy requirements and system efficiencies – Part 4-5 Space heating generation systems, the performance and quality of district heating and large volume systems

[21] Mûnster M., Lund H., Use of waste for heat, electricity and transport – Challenges when performing energy system analysis. Energy 34, 2009 p. 636-644

[9] INCOSE, International Council on Systems Engineering, A Consensus of the INCOSE Fellows, www.incose.org

[22] Lenzen M., Life cycle energy and greenhouse gas emissions of nuclear energy: A review, energy Conversion &Management 49, 2008 p.2178-2199

[10] Dahl H J, Information modelling and systems reengineering – an efficient approach to assessing complex current Norwegian natural gas transport operations, Proceedings of the Tenth Annual International Symposium of the International Council on Systems Engineering (INCOCE), July 2000

[23] Sproles N, Coming to Grips with Measures of Effectiveness, John Wiley & Sons, Inc Syst Eng. 3:50-58, 2000 [24] Olivier, Merrian Webster 1981 [25] Berner M, Primary Energy Concept and Life Cycle Assessment (LCA), Report no: 2009/001, June 2010, The Norwegian University of Science and Technology

[11] Olivier DW, Kelliher TP, Keegan JG, Engineering complex systems with models and objects, ISBN 048188-1, McGraw-Hill, 1997

[26] Act of 14 June 1985 No. 77 the Planning and Building Act, The Ministry of the Environment and the Ministry of Local Government and Regional Development

[12] Joelsson. A. Primary Energy efficiency and CO2 mitigation in Residential buildings, Doctoral Thesis 58/2008, Mid Sweden University (Dissertation 3.October 2008)

[27] Directive 2002/91/EC of the European Parliament and of the Council of 16 December 2002 on the energy performance of buildings.

[13] Berner M., Ulseth R., The Primary Energy Concept, The 11th International Symposium on District Heating and Cooling, August 31 to September 2, 2008, Reykjavik, ICELAND

[28] Directive 2009/28/EC on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC

[14] Frischknecht, R, Jungbluth et al, 2007, Őkoinventare für energiesysteme –Grundlagen für den ökologishen Vergleich von Energiensystemen und den Einbezug von Energiesystemen in Őkobilanzen für die Schweiz , ETH, Zürich 1996

[29] Council Directive 1999/31/EC of 26 April 1999 on the landfill of waste

[15] CEN/ CLC BT JWG, Energy Management, 2005)

[30] Act of 13 March 1981 No.6 Concerning Protection Against Pollution and Concerning Waste, [The Pollution Control Act]

[16] Nørstebø V., Application of systems engineering and information models to optimize operation of gas export systems, Systems Engineering archive, Volume 11 , Issue 4 (November 2008), p: 329342, 2008, ISSN:1098-1241

[31] Act of 17 June 2005 No. 62 relating to working environment, working hours and employment protection, etc. as subsequently amended, last by Act of 23 February 2007 No. 10, (The Working Environment Act)

[17] Sæther S, Thermal Heat and Power Production with models for local and Regional energy

[32] Purves B, Information Models as a Prerequisite to 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

ENHANCED DISTRICT HEATING AND COOLING SYSTEMS – REALISATION OF THE LOW-EX CONCEPT 1

Stefan Bargel , Clemens Pollerberg , Armin Knels , Li Huang , Dirk Müller and Christian Dötsch

Fraunhofer Institute for Environmental, Safety, and Energy Technology UMSICHT, Osterfelder Strasse 3, 46047 Oberhausen, Germany, Phone: +49 (0) 208-8598-1276, Fax: +49 (0) 208-8598-1423, stefan.bargel@umsicht.fraunhofer.de, clemens.pollerberg@umsicht.fraunhofer.de 2 RWTH Aachen University, E.ON Energy Research Center - EBC, Mathieustr. 6, 52074 Aachen, Germany, Phone: +49 (0) 241-8049-780, Fax: +49 (0) 241-8049-769 domestic gas boiler used to provide space heating wastes a huge amount of exergy, since the exergy efficiency of such a system reaches only approximately 5%! This result is valid for arbitrary heating systems in the supply target (room) itself. Therefore it is mandatory to use an integral evaluation approach to decide whether an energy system is efficient or not.

ABSTRACT Since heating and cooling represent low-exergy energy streams, high efficiencies can be obtained, if the energy demand is covered by appropriate – meaning also low-exergy (low-ex) – input energy flows. In order to be able to employ great potentials of lowexergy heat from many different sources, it is important to develop technologies for the supply and the use of energy that allow network temperatures close to ambient temperature in return as well as in supply pipes. Two possible technologies are phase change slurries (PCS) and capillary tube mats (CTM).

With respect to district heating and cooling networks as energy supply systems two findings are important. First, it can be shown that the network subsystem itself as depicted in figure 2 reaches optimal exergetic efficiency at quite low temperatures since the heat losses dominate the pumping electricity effort. Secondly the overall energy supply system efficiency can be greatly enhanced by utilising low-exergy input energy flows such as industrial waste heat.

PCS are discussed as heat transfer fluid, which has an increased heat capacity compared to water. The use of PCS in energy supply networks instead of water leads to an improved energy transport capacity, which results in a reduction of the necessary temperature difference of the transfer fluid. To ensure the transfer of energy from the supply network into the building while the temperature difference between network and building is low, large heat transfer areas are required, which can be achieved by the use of CTM.

In order to be able to employ great potentials of lowtemperature waste heat from many different sources, it is important to develop technologies for the supply and the use of energy that allow network temperatures close to ambient temperature in return as well as in supply pipes.

This paper discusses opportunities for the realisation of cold supply networks and low-ex systems and presents exemplary technologies for their realisation. INTRODUCTION Temperature levels in district heating and cooling networks have long been discussed. During the last years a tendency towards low temperature networks can be observed. From a scientific point of view answers to the question for the optimal temperature levels can be given using exergy efficiencies as for example discussed in [1]. The main advantage of this evaluation parameter is the thermodynamically correct distinction of thermal (low-exergy) and non-thermal (high-exergy) energy flows. Since heating and cooling represent low-exergy flows, it is of uttermost importance to cover these demands by appropriate – meaning also low-exergy – input energy flows. For example a heating system based on a 39

Today, district heating and cooling networks use water as heat transfer fluid. The heat is transported as sensible heat and the transport capacity of the networks is determined by the heat capacity of water and the temperature difference between forward and backward flow. In cold supply networks as well as in low temperature heating networks, high volumetric flow rates are necessary to provide the required transport capacity due to the comparably small temperature difference between forward and backward flow. To overcome these restrictions, a new heat transfer fluid with an increased heat capacity is under development as an alternative to water, phase change slurries. PCS are mixtures of dispersed phase change material and a continuous liquid phase, which can be used as heat transfer fluid in district heating and cooling networks. PCS possess an increased heat capacity due to additional latent heat of fusion occurring during the phase transition of the phase change material. The use of such a dispersion in energy supply networks leads to an improved energy transport capacity, which in turn

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

mind that the annual average outdoor temperature for the heating period e.g. in Germany is about 3.5 °C, it becomes apparent that the exergy to energy ratio of the target energy flows - passing the building envelope at 20 °C – is very small (approx. 7%). On the other hand, exergy to energy ratios of conventional input energy flows are usually 100% as combustible fuels or electricity is used.

results in a reduced temperature difference or volumetric flow rate of the transfer fluid needed to transfer a given amount of heat. The application of PCS for thermal energy transportation is investigated and discussed for example in [2]. An improved transport capacity is one important point for the realisation of the low-ex concept; another important point is the use of the energy on the consumer side. To ensure the transfer of the energy from the supply network into the building while the temperature difference between network and building is low, large heat transfer areas are necessary. These heat transfer areas can be realised by using capillary tube mats integrated into the walls, the floors and the ceilings of buildings.

The low-ex concept acknowledges the fact that demand flows are ‗low-ex‘ - meaning that they possess small exergy to energy ratios. Hence the concept demands to supply energy on appropriate ‗exergy levels‘, instead of wasting exergy by transforming high exergy flows into low exergy ones. In doing so, this approach opens up a totally new dimension of enhancement potential since it deals with the quality aspect of the energy flows under consideration. Therefore, within the low-ex concept energy is no longer one-dimensional. In addition to decreasing the amount of energy demanded by the consumers – leading to insulation efforts – a kind of exergetic suitability has to be taken into account and the task at hand becomes a two-dimensional problem (cf. fig. 1). Consequently, the concept aims at maximizing the exergy efficiency of an energy supply system, which allows to utilize potentials in both dimensions, quantity AND quality.

The E.ON Energy Research Center of the RWTH Aachen and Fraunhofer UMSICHT investigated the possibilities to realise district heating and cooling networks as low-ex systems. These investigations include system modelling and analysing as well as the development and testing of technologies. 1.

Exergy as evaluation parameter

1.1. The low-ex concept Exergy can be understood as the theoretical maximum of mechanical work that can be utilised by equilibrating an energy flow whilst considering its ambient conditions.

The exergy efficiency can be defined as:

Consequently this property distinguishes between types of energy that can theoretically be transformed into each other without any losses - like mechanical work, electrical energy or combustible fuels - and thermal energy. The possibility to transform the latter into any other type of energy is limited by the second law of thermodynamics and therefore inevitably connected to losses.

 ex 

exerget ic suit abilit y enhancem ent exerget ic qualit y

The ultimate goal of heating and cooling is to keep a target (room) at a constant temperature of e.g. 20 °C. As the outdoor temperature varies additional heat has to be supplied or excess heat has to be disposed of to fulfil this task.

low -ex concept

insulat ion

Theoretically the supplied energy flow could be transferred to the room using infinitesimal small temperature differences between supply flow and target2. The real temperature differences occur due to heating and cooling techniques applied which are mainly limited by finite heat transfer areas. Keeping in This statement is analogously true for cooling applications.

(1)

In applying this efficiency the demand flows and particularly the supply flows have to be defined carefully (cf. chapter 1.2.).

This distinction is of importance if one analyses a system where both types of energy flows (thermal and non-thermal) occur and have to be related to each other – as is the case with heating and cooling applications.

 exergy (demand )  exergy (supply)

energy dem and (quant it y)

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

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

conversion losses

prim ary energy

net w ork heat losses

building heat losses

heat f low dist rict heat ing net w ork

heat generat or

heat f low

T = 20°C heat ing surf ace

heat f low

pum ping elect ricit y

pow er plant conversion losses

prim ary energy

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

Instead thermal input flows as industrial waste heat3 or geothermal energy should be applied. On the other hand, if combustible fuels are used to meet thermal demands, at least Combined Heat and

1.2. Integrated system evaluation When evaluating a system it is important to specify the evaluation boundaries (cf. fig. 2). It has to be pointed out that an integrated system evaluation is mandatory since otherwise results are ambiguous and misleading.

Power generation (CHP) with a maximum electrical degree of efficiency should be utilized. This allows transforming part of the high-exergy fuel into highexergy electric current. Heat is produced as ‗waste product‘ of this conversion.

This can be demonstrated by assuming e.g. evaluation of the building subsystem only. If two systems are compared, one consisting of a target room equipped with space heating and the other one with a target room equipped with conventional heating, one could arrive at the conclusion, that the system utilizing space heating is more efficient. However, assuming both systems are also equipped with an identical condensing gas boiler providing the heat, an evaluation comprising the total system (consisting of heat generation and heat transfer to the target) would arrive at a totally different conclusion. In this case, both systems possess the same exergy efficiency, which is approximately 5% for the outlined case. This is because a potentially more efficient heating system is not put to use as the same input and supply flows occur in both cases.

Optimization potentials within the distribution subsystem are basically indirect. At first glance, the distribution system has no influence at all since the network acts as connection between heat generation and heat consumption. Consequently, no thermal flows exist that pass the overall system evaluation boundaries. However, two aspects remain and need to be accounted for. One is heat losses occurring throughout the network that have to be compensated by additional heat generation. The other is pumping to maintain the heat transfer medium circulation, which is met by an unalterable high-exergy input (electricity). The main problem is that concepts, which lead to decreasing heat losses cause increasing pumping efforts and vice versa. Nevertheless, heat losses are the exergetically dominant influence, therefore the focus should be to confine these losses. Heat losses depend on the driving temperature difference between medium and surrounding ground and on surface area. Minimization of the losses can most easily be achieved by reducing the network temperatures since pipe dimensions are affixed due to demands so that surface areas are not a modifiable parameter. This approach is even more rewarding since it allows employing low

1.3 Efficiency enhancement potentials The complete energy supply system can be divided into three subsystems – generation, distribution and building (representing the consumption). These subsystems possess different potentials to enhance overall system efficiency. Currently heating demands are met by burning highexergy fuels, great enhancement potentials are available within the generation subsystem. Firstly, fuels should not be used to directly satisfy thermal demands at all since this embodies pure exergy destruction.

Industrial waste heat in this sense is heat that can no more be put to any use within the industrial production process.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

temperature thermal input flows and therefore represents the prerequisite for an efficient generation subsystem. The last subsystem possessing enhancement potential is the consumer. Since the target temperature determines the exergetic quality of the thermal demand, therein lays no significant optimization potential. However, as decreasing the amount of energy that has to be supplied is also part of the low-ex concept insulation can help to improve the system. On the other hand, benefits similar to those already discussed for the distribution subsystem can be identified for the consumer system as well. By choosing appropriate heating and cooling technologies, as e.g. investigated in [3], the exergy destruction during heat transfer to the room air can be minimized. This is achieved by applying low-temperature heating and high-temperature cooling devices. Inlet and outlet temperatures of the heating/cooling device simultaneously define constraints for the distribution network subsystem, which in turn set constraints for the generation. In the end supply temperatures close to the target temperature form the basis for a ‗low-ex ready‘ consumer. Without this step an exergetically optimal energy supply system would be greatly hindered. 2.

Figure3. Photograph of a paraffin/water dispersion

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

TCEF 

Applicable technologies for the realisation

PCS wh f , PCM  c p , PCM  T   1  wc p , w  T  w  c p , w  T

(2)

The TCEF is a function of the densities of the PCS ρPCS and water ρw, the mass concentration of the PCM w, the specific heat capacity of PCM cp,PCM and water cp,w the heat of fusion of the PCM Δhf,PCM and the temperature change ΔT of the fluids. The TCEF is calculated and plotted in the diagram figure 4 for temperature differences ΔT between the forward and backward flow of 10 and 15 K as function of the mass concentration w.

2.1. Phase Change Slurries The most used heat transfer fluid in district heating and cooling networks is water. In supply networks, the heat is transferred as sensible heat with a temperature difference between forward and backward flow. The heat transfer capacity of a network is determined by the temperature difference, the mass flow and the heat capacity of the heat transfer fluid. The temperature difference and the temperature level of the network are limited by technical restrictions and determine the necessary mass flow of the heat transfer fluid. To overcome these restrictions, fluids with higher heat capacities than the heat capacity of water are under development. An alternative to water could be PCS. PCS are mixtures of dispersed phase change material and a continuous liquid phase, which possess an increased heat capacity due to the additional latent heat of fusion occurring during the phase transition of the phase change material. The PCS remains pumpable even when the phase change material is frozen. Thus, the PCS can be used as heat transfer fluid in supply networks. A promising PCS for heat or cold supply networks is paraffin/water dispersion. Figure 3 is a photograph of a paraffin/water dispersion. Paraffin is the phase change material, which can be chosen according to the desired temperature of the phase transition, and water is the continuous phase of the dispersion. In [4] paraffin/water dispersions are investigated and their properties presented

TCEF [-] 3.5 3

delta T = 10 K delta T = 15 K

2.5 2 1.5 1 0

0.2

0.4

0.6

0.8

w [-]

Figure 4. TCEF – PCS compared to water for temperature differences 10 and 15 K, diagram calculated with the properties of water and RT-42 of the company Rubitherm [5]

Using PCS with a mass concentration w of 0.4 would increase the heat transport capacity of the supply network to 1.5 times of the value compared to water, if

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

the temperature difference of the supply network is 15 K, and even 2 times, if the temperature difference is only 10 K. Furthermore, the diagram shows that with rising temperature difference the gradient of the TCEF is lower, which means that the advantage of the PCS compared to water disappear at higher temperature differences. At the point where the gradient of the TCEF is 0, the water system and the PCS system have the same transport capacity. At that point, the mass concentration of paraffin w has no influence on the TCEF.

is related to the heat transfer coefficient U, the heat exchange area A and the temperature difference between the mean temperature of the heat release Tm as well as the room temperature Tr.

Q  U  A  Tm  Tr 

(4)

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 reduction of the necessary temperature difference or volumetric flow rate of the transfer fluid needed to transfer a given amount of heat.

Tm 

Tin  Tout T ln in Tout

(5)

Another technical issue of PCS systems is the increased pressure drop in the pipes due to the higher viscosity of the PCS. A calculation methods and measurement data can be found in [6, 7 and 8]. The viscosity of PCS is related to several influence quantities and can cause an incensement of the pressure drop up 100%. PCS are non-newtonian fluids.

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

2.2. Capillary Tube Mats

NTU 

The most often used heat exchanger type in heating systems is a convective radiator, which is installed in rooms close to the window. The size of a radiator should be small, so that also the heat exchange surface is small and the heating system must be operated on a high temperature level to ensure the heat transfer from the heating system into the room. An alternative to convective radiators are floor heating systems. Floor heating systems consist of a capillary tube mat, which is installed in the upper layer of the floor. Because of the bigger heat exchange surface compared to the convective radiator, the temperature level of the heating system is lower. A new approach to realise heating and cooling of buildings is via CTM, which are integrated in the floors of the building, as well as in the walls and ceilings. This system offers a big heat exchange area and allows the heating and the passive cooling of the building. Due to the increased heat exchanger area, a low temperature difference between the heating system and room is possible. For the further discussion, the following simple model is used to describe the heat release of the heating system in the building. The heating release system is evaluated by the number of transfer units (NTU). The heat capacity provided by the heating network Q is

(6)

The NTU values have been calculated for a convective radiator system and a CTM system. The assumed temperatures for the calculation and the results are given in table I. Table I. NTU for both heat release systems: conventional radiator and CTM parameter

convective radiator

CTM system

Tin [°C]

Tout [°C]

Tr [°C]

NTU [-]

0.4

0.43

The NTU value of the CTM system is 0.43 and as high as the NTU value of the convective radiator. This means that both systems have the same heat release capacity, although the inlet temperature Tin of the CTM system is lower and the temperature difference between inlet Tin and outlet Tout of the CTM system is smaller.

calculated by equation (3) with the inlet and outlet temperature Tin/out of the supply network, the mass flow m and heat capacity cp of the heat transfer fluid.

  c p  Tin  Tout  Q  m

U  A Tin  Tout   Tm  Tr  m  c p

CONCLUSION From the point of view of the low-ex concept the major task en route to an exergetically efficient energy supply system is the replacement of the combustible fuel boiler by utilization of low temperature thermal input

(3) 43

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

flows such as industrial waste heat or geothermal energy. To achieve this goal it is necessary to decrease the medium temperatures within the distribution networks first. A prerequisite is a low-ex ready consumer that allows meeting the thermal demands applying low temperatures.

ACKNOWLEDGEMENT This study was supported by the Project Management Juelich (PTJ) and the Federal Ministry of Economics and Technology (BMWi) under 0327471A. Comments of a highly constructive nature were received from Daniel Wolf, Jorrit Wronski and Astrid Pohlig.

A possible realisation employs CTM in the heating system of the building that allows applying inlet and outlet temperatures of approximately 37 °C and 31 °C, respectively. Within the district heating or cooling network, the utilization of PCS instead of pure water enables the application of small temperature differences between forward and backward flow while retaining the pipe dimensions. Since the backward flow temperature mainly depends on the outlet temperature of the consumer system, small temperature differences within the network automatically lead to low forward flow temperatures. Consequently, the exploitation of low temperature heat sources as input flows for the energy supply system is rendered possible.

REFERENCES [1] C. Kemal et al., Evaluation of energy and exergy losses in district heating network, Applied Thermal Engineering, 24 (2004), pp. 1009-1017. [2] H. Inaba, New challenge in advanced thermal energy transportation using functionally thermal fluids, International Journal of Thermal Sciences, 39 (2000), pp. 991-1003. [3] M. Ala-Juusela et al., LowExergy Systems for Heating and Cooling of Buildings, final report of the IEA ECBCS Annex 37.

Moreover, the decreasing temperatures in both forward and backward flows of the network reduce the transportation heat losses. This leads in the end to a reduction of energy input (quantitative aspect of the low-ex concept) into the supply system.

[4] L. Huang et al., Evaluation of paraffin/water emulsion as a phase change slurry for cooling applications, Energy, 34 (2009), pp. 1145-1155. [5] Rubitherm RT-42, datasheet 08/20/2009, http://www.rubitherm.de, Rubitherm Technologies GmbH, Berlin (2010).

The only drawback suffered occurs in terms of an increased pumping effort caused by a higher viscosity of the PCS in comparison with water. But, since heat losses are the predominant factor over circulation pump energy, an overall benefit should be accomplishable.

[6] Yinping Zhang, et al., Experimental research on laminar flow performance of phase change emulsion, Applied Thermal Engineering, 26 (2006), pp. 1238-1245.

Summarizing it should be pointed out that applying technologies such as CTM in the building heating or cooling system and PCS as alternate heat transfer medium for the distribution networks the low-ex concept can be realised, thus greatly enhancing the efficiency of energy supply systems.

[7] A., B. Metzner et al., Flow of Non-Newtonian Fluids – Correlation of the Laminar, Transition, and Turbulent-flow Regions, American Institute of Chemical Engineers Journal, Vol. 1, No. 4 (1955), pp. 434-440. [8] R. Rautenbach, Kennzeichnung nicht-Newtonscher Flüssigkeiten durch zwei Stoffkonstanten, ChemieIngenieur-Technik, 36 No. 3 (1964), pp. 277-282.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

APPLICATION OF EXERGOECONOMICS TO THE OPTIMIZATION OF BUILDING HEATING SYSTEMS CONNECTED TO DISTRICT HEATING NETWORKS C. W. Snoek and S. C. Kluiters Renewables and Integrated Energy Systems, CanmetENERGY, Natural Resources Canada, 1 Haanel Dr, Ottawa, K1A 1M1, Canada Often, omitted from consideration is the ―quality‖ of the energy that is needed to provide comfort to the occupants of a building. While the heating requirements of a building can be determined (in GJ or TJ), the nature or origin of this energy is not addressed in energy efficiency calculations. The total amount of Joules can be provided by oil, natural gas, electricity or low temperature ‗waste‘ heat. While the first three energy sources are considered high quality, and can be used to generate very high temperatures (over 1000 °C), run equipment such as computers, radio and TV transmitters and receivers, ‗waste heat‘ is of low quality and has no other use. Comfort heating does not require high temperatures and therefore using high quality fuel for low quality applications is considered wasteful.

ABSTRACT The concept of energy efficiency, defined as useful energy output as fraction of required energy input, has been used for years in technical systems assessments. In addition to energy efficiency, there are benefits to using exergy efficiency to assess system performance. Whether systems will be installed or not is ultimately determined by their economic performance. This performance is usually determined by comparing initial investment cost and operational cost with revenues throughout a system‘s lifetime in terms of payback time or net present value. This paper describes a novel methodology that uses the concept of exergy and the thermoeconomic factor, a ratio that compares investment-related cost and exergy destruction cost, for the economic optimization of a community energy system. It compares the cost of exergy and the required capital and operational costs including carbon taxes to accommodate this low quality energy. In doing so it enables a quick way to properly assess the value of a system‘s ability to use low exergy energy inputs. The method is compared to a more traditional economic analysis.

Energy quality is often expressed as ‗exergy‘. Exergy is defined as the maximum useful work possible during a process that brings the system into equilibrium with a heat reservoir. To illustrate the concept of exergy one can compare two different forms of the same amount of energy: 100 kJ of energy is equivalent to: – 12 V/2.3 Ah stored in a car battery, or – 1 kg of water at 43 °C in a room with an ambient temperature of 20 °C.

INTRODUCTION

Obviously, the energy contained in the battery is considered more useful and therefore has the higher quality or exergy.

In the last few years, we have become painfully aware of the effects of climate change. The burning of fossil fuels and the resulting emissions are thought to be a major contributor to the apparent increase of adverse weather events. While people need energy for comfort, in some cases there may be a choice in the source and nature of that energy. In addition to climate change, there is also a concern about the rapid depletion of the more valuable of fossil fuels, natural gas and oil. For these reasons it makes much sense to re-evaluate the sources of the energy we use and the effect of using them has on the environment.

The ratio of Exergy (E) to Energy (Q) can be expressed as:

T E  1  ambient Q Tsup ply

(1)

where T is given in K. Equation 1 shows that when the supply temperature of an energy source is high, the exergy converges to the energy value. Electricity and mechanical work are (nearly) perfectly convertible and the exergy content is therefore equal to the energy content. Conversely, when the supply temperature is closer to the environmental temperature, the value of the exergy becomes (much) smaller than that of the energy.

To lower energy requirements, energy efficiency has been practiced for many years. In terms of comfort heating in houses, most of the effort has gone into improving building insulation, better windows, building orientation with respect to the sun, shading from solar energy etc. In terms of energy conversion equipment, improving the efficiency often meets ‗natural‘ limits, such as those expressed by Carnot‘s 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 ethical society the value, worth and quality of different energy supplies should, as a minimum, be matched to the requirements of the different energy applications.

exergoeconomic factor is also found in other sources, such as Temir & Bilge [5]. It is beyond the scope of this paper to provide a comprehensive literature overview of thermoeconomic publications or even of the methods used in these publications. The aim of this paper is to apply one of these methods, using the above-mentioned exergoeconomic factor to optimize building heating systems connected to a district heating system. To the best of the authors‘ knowledge, so far this method has only been applied to optimize individual components.

Methods to design low exergy buildings are available today. For instance, Schmidt [2] developed a method and pre-design tool for low exergy buildings in which he compared different heating systems, such as boilers, condensing boilers, electric heating, GSHP and low temperature under-floor heating. However, this method does not directly address the effect of system heat transfer surface area on the overall economics.

This work ties in with research into advanced lowtemperature district energy systems currently carried out at the CanmetENERGY laboratories of Natural Resources Canada in Ottawa, Canada.

Also, there is an additional benefit realizing that a building that can accommodate low exergy streams is ready for future hook-up to other, perhaps renewable energy sources: GSHP, solar, waste heat from industry, energy from thermal storage to name a few. This is a distinct advantage when the move to a sustainable society gains momentum, and the concept of low-temperature heating should be incorporated in building codes.

The system considered consists of buildings with their heating system (radiators and cross-flow heat exchangers are considered), the energy centre with boilers and pumps and the pipeline to move the energy in the form of hot water to the community. The development of the methodology was the main object of the study, not the optimization itself.

This paper considers the cost of using the low quality part of the energy source and the (increased) capital cost and operating cost that are required to ‗accommodate‘ low quality energy. A methodology has been developed to determine the optimal cost of operation, based on the capital cost, operational cost and the cost of the exergy.

While the development of the optimization was related to economics, in other words, the least costly option, it should be noted that the concept of ‗exergy‘ opens up the notion of ―morals‖ and ―ethics‖. For new developments, the costs of resource depletion and environmental destruction should be considered as well. Just because a certain system is economic, it is not necessarily the best moral or ethical choice. Just because a certain system does not cause local problems, that does not mean that (environmental or other) problems caused by this system elsewhere can be ignored.

This type of analysis is considered part of the field of thermoeconomics, more in particular exergoeconomics. Wikipedia defines thermoeconomics in a very theoretical way as a school of economics that applies laws of thermodynamics to economy. Valero et al. [3] operationalize this definition by describing two aims of thermoeconomics, (1) optimization to minimize cost of a system or component, and (2) cost allocation of individual outputs of a plant producing a number of outputs.

Traditional Optimizations System optimization is often done by optimizing systems separately, and not by considering the overall efficiency of integrated systems. Often, an integrated approach leads to optimal solutions, as in electricity generation using a back pressure steam turbine. Accepting a lower efficiency of the turbine may lead to the residual energy in the condenser being useful in other applications, whereas in the separately optimized version this thermal energy would be useless. In the latter case, the turbine back pressure is kept as low as possible, to extract the maximum electrical power. This makes the condensate of too low a temperature to be useful in other applications. Optimizing integrated systems as a whole avoids this problem.

Valero and coworkers [3] date this research field back as far as 1932, when Keenan apportioned cost of heat and work taking into account irreversibility and thermodynamic efficiency instead of enthalpy only. However, they go on to say that Gaggioli, and Tribus and Evans in the early 1960s started off real development in thermoeconomics. Ever since, these fields have received tremendous attention. Valero and coworkers identify that an important problem in this body of research is the variety of methodologies used with accompanying nomenclature. Between them and Tsatsaronis [4] they already name a fair amount of methods. In doing so, Tsatsaronis introduces the exergoeconomic factor f, as a fraction that compares two sources contributing to cost increases, investmentrelated cost and exergy destruction cost. This

Exergoeconomic Optimization In an exergoeconomic optimization, the concept of exergy is used to determine the best and most economic solution to an energy conversion process or 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 remains the same, the exergy that delivers this energy may vary. Analyzing the required exergy with respect to the energy transfer equipment will result in an optimum economic solution allowing for integration of the system with other systems.

the design temperature for Ottawa. While this is an over-simplification of reality, it neither hinders the development of the methodology nor introduces serious errors of consequence.

In this exergoeconomic optimization, the system is treated as an integrated whole together with other systems. While the (comfort) energy supplied remains the same in any given scenario, the exergy required for this scenario varies and the cost implications of this variation are included in the analysis. Therefore, in this analysis the consumer of energy does not pay for the energy, but for the exergy, the real value of the energy supplied.

The pipe diameters were estimated using the RETScreen software tool. This means for diameters under 400 mm the pressure drop is kept below 200 Pa per meter of pipe and for larger diameters flow velocity is maximized at 3 m/s [6]. As RETScreen has a limit of 13 sections for district heating systems, the 1000 homes were assumed to be located along twelve 80home streets. The final 40 homes were located in a separate street.

Energy Transmission

The energy transfer fluid is water. The pipes are preinsulated steel or cross-linked polyethylene (PEX) pipes. The first iteration of the methodology accounted for heat losses from the pipes. Since it was found that this heat loss was a negligible fraction of the transmitted energy, it was omitted in subsequent versions. For a thorough analysis, it is recommended to include heat losses, especially if the piping system is extensive and the supply temperatures reach high levels.

SYSTEM DESCRIPTION The system considered here to develop the methodology is a district heating system supplying hot water for space heating to a 1000 home community in the Ottawa area in Canada. It was modelled using the RETScreen clean energy project analysis software tool [6] and in-house spreadsheet based models. The hot water is transported from an energy centre located centrally in the community to the 1000 detached homes. Inside the buildings, radiators or cross-flow heat exchangers (water-to-air fan coils) are employed to provide space heating.

A pressure drop analysis was used to determine the required pump energy. The electric motor driving the pump was estimated to have 90% efficiency while the pump was assigned an efficiency of 85%.

Energy Supply End-Use

The building temperature set point is kept constant at 20°C. The hot water supply temperature is determined by the outdoor temperature. If the outdoor temperature is above 5 °C, the supply temperature is 70 °C. When the outdoor temperature drops below -15 °C, the supply temperature equals 90 °C. Between 5 °C and -15 °C, the supply increases linearly from 70 °C to 90 °C. This is a common supply temperature profile used in many European district heating systems. It prevents excessive flows in the pipes at high loads and permits smaller heat transfer surfaces in the buildings due to the higher temperature difference between water and building air. When the heat transfer surface area was varied to reach an optimum solution the supply temperature was adjusted by a constant value over the entire load range. The water return temperature was set at 30 °C in all design calculations, but varied throughout the year according to the offdesign characteristics of the heating equipment used. The load of the buildings is related to the outdoor temperature. The annual heat consumption was set at 100 GJ per house, a typical value for detached homes in this area. The instantaneous load throughout the year is simply calculated as a linear relationship between zero and the maximum capacity, when the outdoor temperature varies between 20 °C and -28 °C,

The energy supplied to the pipeline was used to keep the building temperatures at set point. Therefore, regardless of the (size of) building heating system used, the same energy was used to keep the buildings warm. However, the exergy used was dependent of the system in place and of its size. The larger the size, the lower the required water temperature and hence less exergy was required to achieve the same end result. Two different technologies were used to model the transfer of energy into the building space: cross-flow heat exchangers, and radiators. Simulations were done for both technologies separately, and the technologies were never mixed. This was done to simplify the analysis. In reality, mixed systems will occur and should be analysed as such. While this will increase the level of modelling complexity, it is not difficult to do. DESCRIPTION OF MODELING Climate The local climate has a significant effect on the design of a building heating system. A maritime climate may have many degree days but not show the variability in demand that a building in a continental climate with an equal amount of degree days experiences. Even if both 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 demand load in the building with the continental climate may be far greater. Therefore, the climate plays an important role in the design of a heating system and should, therefore, be considered in this analysis.

temperature. The pumping energy required was included in the modeling. Since electrical energy is equivalent to exergy, the pumping energy calculated from the pressure drop calculations (including efficiencies), was numerically counted as exergy. Normally, during periods of no-load, the pumps keep operating to keep a supply of design temperature water close to the load. This is done with a thermostatically operated by-pass valve. Since this valve represents a constant effect which does not affect the optimization, it was not modelled for simplicity.

Supply Temperature To include the effect of the variability of the energy demand with time-of-day and the seasons, the statistical average hourly temperatures for the city of Ottawa were used. These temperature values are real values, with realistic variability (high and low temperatures), using time-periods from different years to provide for a correct average. In total, 8760 hourly values of temperature were used in the spreadsheet, as shown in Figure 1.

Design of Cross-Flow Heat Exchanger The design of the cross-flow (or fan-coil) heat exchanger was based on the assumption that the overall heat transfer coefficient ‗U‘ was 25 W/(m2K). The F-factor was set at 0.94. To meet the design load, heat exchangers with a combined area of 17,152 m2 were required.

40 30

Temperature (C)

Design of Radiator Heating System

The design of the radiators was done in a very simple manner. It is acknowledged that better methods exist, but the development of the methodology did not suffer because of this simplification. For any optimization, actual modelling of the equipment should take place.

0 0

1000

2000

3000

4000

5000

6000

7000

8000

9000

-10 -20 -30 -40

To determine the heat transfer from the panels, the general radiation equation

Hour (-)

Fig.1. Average hourly temperatures in Ottawa

4 4  Q   Tpanel  Troom

Solar Radiation

(2)

To simplify the spreadsheet calculations, the effects of solar radiation, plug loads and occupancy gains were neglected. When performing the optimization, these effects remain constant and so have little effect on the final outcome. When using this method for design, these contributors to the exergy balance should be considered.

was used with the ‗average‘ panel temperature. The surface emissivity ‗ε‘ was estimated at 0.9. Convection from the surfaces was not separately considered.

Determination of Instantaneous Load

Determination of Exergy Use

The maximum thermal load of the community for space heating was determined using the average hourly temperatures and the annual heat consumption per house of 100 GJ. It turned out to be 10,640 kW. When the outdoor temperature reaches -28 °C, the Ottawa design temperature, the community requires the maximum thermal load. At the ambient temperature of 20 °C, the load is nil. The modelling is set up so that between these ambient temperatures, the load varies linearly. For instance, at -4 °C, the load equals 5.32 MW.

As indicated in Section 1, the ratio of Energy to Exergy can be expressed as:

To meet the design load, 42,271 m2 of radiative surface was required to meet the design load.

T E  1  ambient Q Tsup ply

(1)

Where E is exergy, Q is energy and T is the temperature given in K. For this study, knowing the energy supplied, Equation (1) was used to calculate the supplied exergy for each hour interval. For each of the heating systems used and each of their variations is size, the amount of energy supplied to the heated space remained the same. However, due to the supply temperature

Pumping Power and Exergy To meet the load, the water had to be pumped from the supply source to the load. The amount of water pumped varied with the load and the supply 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 exergy used by each system was unique.

per ton CO2eq. Carbon intensity factors of 0.050 ton CO2eq/GJ were used for natural gas and 0.054 ton CO2eq/GJ for electricity (taken from RETScreen [6] as representative for Canada). This results in a $6.5/GJ energy charge for heat, a $38.9/GJ exergy charge for heat and an $18.6/GJ energy (or exergy) charge for electricity.

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

Thermoeconomic Factor The exergoeconomic or thermoeconomic factor ―f‖ compares two sources contributing to cost, investmentrelated cost and exergy destruction cost. It is defined here as the ratio of Capital Cost Rate (CCR, which includes O&M cost, but excludes heat and electricity cost) and the sum of Exergy Destruction Cost Rate (EDCR) and CCR. The CCR equals the cost per unit time for the installation, depreciation, maintenance, etc, while EDCR is the cost of exergy.

Table I. – Cost for heating technologies 2

Cross-flow heat exchanger

$250/m of heat transfer surface

Radiative system

$200/m of exposed panel

f 

Future cash flows were discounted at a rate of 8% and system lifetime was set at 40 years. Annual operating and maintenance (O&M) cost other than cost for heat and electricity were set at a fixed fraction of 1% of total investment cost.

CCR EDCR  CCR

(4)

Since CCR and EDCR have the dimensions of $/time, ―f‖ is dimensionless. A high value for ―f‖ indicates that the capital and maintenance costs are dominant. Also, a high f – value indicates good use of the exergy in the fuel. On the other hand, a low value for ―f‖ indicates an inefficient use of fuel resources. For each heating system variation, the average annual thermoeconomic factor was calculated.

To compare traditional optimization with exergoeconomic optimization three types of analyses were performed. The ‗classical analysis‘ applies the traditional optimization where heat is valued based on energy content, at a rate of $5/GJ, which is considered representative for heat from natural gas combustion. Electricity cost has been set at $17/GJ (just over $60/MWh).

MODELLING RESULTS Base case design

In the exergoeconomic analysis heat and electricity are priced based on the exergy content. The exergy charge was determined at $30/GJ for thermal energy, based on the above mentioned $5/GJ for heat, assuming a 1 to 6 ratio of exergy to energy content (applies to a temperature around 80 °C). The electrical energy to exergy ratio was taken as one, resulting in an exergy charge of $17/GJ for electricity. At first glace it may seem erroneous to charge more for exergy from the thermal source than that for the electricity for the pump, but it must be remembered that the (thermal) exergy is a fraction of the thermal energy.

Table II shows the main information for the base case designs for both the radiator and cross-flow heat exchanger systems. As expected, the distribution pipe diameters, required pump capacity, annual space heat consumption and annual heat cost are the same for both systems. As the water return temperatures throughout the year are generally lower for the radiator system, the required water flows and consequently the annual electricity consumption are lower for the radiator system. As both systems have a design supply temperature of 90 °C (and thus also the same off-design supply temperatures throughout the year), the annual exergy consumption is the same for both. The lower return temperatures for the radiator system also show in the higher fraction of energy provided under 60 °C. In terms of cost, the radiators are clearly more expensive resulting in higher annual investment and O&M cost, which is not offset by the somewhat lower electricity cost. Overall the more capital intensive radiator system has a higher f-factor than the cross-flow heat

The third type of analysis is a classical analysis corrected for the difference in value of low- and hightemperature heat, by assuming energy under 60°C is available free of charge (as waste heat from a nearby process). For energy over 60 °C the charge is still $5/GJ. To assess the influence of carbon taxes, two sets of results are presented. One assumes no carbon taxes are in place and the other assumes a carbon tax of $30 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 systems, the cross-flow heat exchanger system works out cheaper using all three types of analysis due to the large difference in investment cost.

Alternative designs – radiator system For both the radiator and the cross-flow heat exchanger system alternative designs with increased and decreased surface areas were costed. The district heat supply temperatures were modified accordingly, and as noted before, the required water flows and thus distribution pipe diameters and pumping power requirements were modified too. The effects of these variations on cost were taken into account.

Table II. – Main information base case designs without CO2 tax.

Radiator

Cross-flow heat exchanger

Surface area (m )

42,271

17,152

Distribution pipe diameters (mm)

DN80/DN65

Required pump capacity (kW)

32.2

Annual electricity consumption (GJ)

115.2

147.6

Annual exergy consumption (GJ)

16,810

Annual space heat consumption (GJ)

100,000

Fraction of energy < 60 °C 68.4%

64.4%

Installed cost heaters

$4,288,066

The results of all the modelling runs are shown in the figures below in the form of the relationship between the annual cost (the sum of capital investment, O&M cost and energy or exergy costs) and the f-factor. An increasing f-factor means increasing surface areas (and thus increasing capital and operating and maintenance cost) and decreasing heat supply temperatures (and thus decreasing exergy cost). Figure 2 shows results for the radiator based heating system with no carbon taxes in place. The slight jump in annual cost around an f-factor of 0.81-0.82 is caused by an increase in district heating piping diameter from DN65 to DN80 for the 80-house streets and from DN50 to DN65 for the 40-house. All lower f-factors shown have piping diameters of DN65 and DN50 and all higher f-factors shown have DN80 and DN65 respectively. $2,900,000

Investment cost district heating system

$11,556,386 $11,556,386

Annual O&M cost

$200,107

Annual cost ($)

$2,700,000

$8,454,298

$2,500,000 $2,300,000 $2,100,000 $1,900,000 $1,700,000

$158,445

$1,500,000 0.68

0.70

0.72

0.74

0.76

0.78

0.80

0.82

0.84

0.86

0.88

Annual charge investment and O&M cost $1,878,206

$1,487,163

Annual heat (energy) cost

$500,000

Fig. 2. Relation between f-factor and annual cost radiator system, no carbon tax.

Annual heat (exergy) cost

$504,193

Annual electricity cost

$1,958

$2,509

The classical analysis shows a continuous increase in annual cost with increasing f-factor.1 This makes sense because cost is not based on exergy but on energy. Therefore, an increasing surface area means increasing capital cost, but constant energy cost, so the lower exergy requirement does not offset the increase in capital cost. The classical analysis would tell us to optimize the system with minimum capital expenses. In reality there would be a limit as ever increasing temperatures will mean that we are dealing with more expensive materials and at a certain stage steam instead of hot water, requiring a more expensive district heating system. Also, heat losses to the environment

f-factor (-)

Total annual cost classical analysis $2,380,164

$1,989,672

Total annual cost exergoeconomic analysis

$2,384,357

$1,993,865

Total annual cost heat under 60 °C free analysis

$2,038,164

$1,667,672

f-factor

0.814

0.782

Classical analysis

Exergo-economic analysis

Heat under 60C free analysis

Although exergy is not explicitly costed in the classical analysis, we can still calculate an f-factor.

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, which is not modelled here.

Alternative designs – cross-flow heat exchanger system

The exergoeconomic analysis does take into account the increasing exergy requirements for systems with lower surface areas. Consequently, there is a minimum annual cost (for an f-factor in the range 0.71–0.73). Going to lower f-factors, exergy cost significantly increase, which results in increasing annual overall cost. Going to higher f-factors, the annual cost increase again because the decreasing exergy cost are more than offset by the increase in capital and O&M cost.

Figure 4 shows the results for the system with crossflow heat exchangers. Note again that the jump in annual cost at an f-factor around 0.78 is due to the increase in district heating pipe diameter. As for the radiator system, the classical analysis shows a steep slope with increasing f-factors as capital cost are dominant and lower exergy requirements do not translate into cost savings. Again the classical analysis would lead us to minimize the surface area (with the same limitations as applied to the radiator).

The analysis discounting heat under 60 °C does take the temperature level of energy supplied into account while determining costs, though there is no explicit price for exergy in the calculations. As a result, the line does not slope up as strongly with increasing f-factor as the line pertaining to the classical analysis. Going to higher f-factors, eventually all heat will be delivered under 60 °C, and all heat provided will be free. Going to lower f-factors, eventually all heat will be supplied at temperatures over 60 °C and the green line will coincide with the blue classical analysis line. The heat under 60 °C free analysis does not show an optimum and would suggest minimizing the f-factor. Like the classical analysis the economic analysis suggests that capital cost are dominant.

$2,200,000 $2,100,000

Annual cost ($)

$2,000,000

$1,400,000 0.65

Classical analysis

Annual cost ($)

0.75

0.77

0.79

0.81

0.83

Exergo-economic analysis

Heat under 60C free analysis

Figure 5 shows the effect of a carbon tax on the exergoeconomic analysis. As for the radiator system the carbon tax means higher annual cost and lower f-factors for the same system due to increased exergy cost. As there is not a clear optimum in either line, we can not conclude that the optimum f-factor is the same for both. However, it is clear that both level off in the higher f-factors range.

$2,500,000 $2,450,000 $2,400,000 $2,350,000 $2,300,000 $2,250,000 $2,200,000 0.85 0.87

f-factor (-) No carbon tax

0.73

The analysis with free heat under 60 °C shows a line gradually sloping up, though far less pronounced than the classical analysis line. Like the classical analysis line it would indicate that lower surface areas would optimize this system.

$2,550,000

0.81 0.83

0.71

The exergoeconomic analysis shows a downward sloping line. This is caused by the reduced capital cost and O&M cost compared to the radiator system and thus increased importance of exergy cost as fraction of the total cost. An increase in cost due to surface area is more than offset by a decrease in exergy cost. Contrary to the radiator system, though, the exergoeconomic analysis does not show a clear optimum, although it clearly levels off at higher f-factors. It is interesting to note here that the classical analysis and the exergoeconomic analysis lead to contradictory recommendations as to optimization.

$2,600,000

0.77 0.79

0.69

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

$2,650,000

0.73 0.75

0.67

f-factor (-)

$2,700,000

0.69 0.71

$1,700,000

$1,500,000

$2,750,000

0.65 0.67

$1,800,000

$1,600,000

Figure 3 shows the effect of introducing a carbon tax of $30/tonCO2eq on the exergoeconomic analysis. It is clear that the carbon tax leads to higher annual cost and lower f-factors for the same systems, both caused by the increased exergy cost. Both lines show an optimum for an f-factor in the range 0.71–0.73, but for the case without carbon tax the corresponding surface area is lower than for the case with carbon tax. This makes sense as increasing heat and exergy cost mean a shift to a system with higher surface areas and lower heat and exergy requirements. As figure 3 shows, the capital and O&M cost as a fraction of total cost (and consequently also the exergy cost as a fraction of total cost) remain in the same range.

0.63

$1,900,000

Carbon tax $30/tCO2

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

combining heating technologies and possibly including other technologies such as under-floor heating provide further opportunities to optimize system cost.

$2,500,000 $2,400,000

Annual cost ($)

$2,300,000 $2,200,000

In addition, the application of the methodology developed in this study should be applied to a heat pump, where the variations in COP with supply temperature would be included. This would result in the ability to match the heating equipment to the heat pump, resulting in an optimum operation.

$2,100,000 $2,000,000 $1,900,000 $1,800,000 $1,700,000 $1,600,000 0.60

0.62

0.64

0.66

0.68

0.70

0.72

0.74

0.76

0.78

0.80

0.82

f-factor (-) No carbon tax

Carbon tax $30/tCO2

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

ACKNOWLEDGEMENT During this work the authors have had very fruitful conversations with many colleagues: Mikhail Sorin, Evgueniy Entchev, Libing Yang, Ibrahim Dincer, Hajo Ribberink and Kirby Wittich. These discussions helped focus the work and stimulated further thinking in this interesting area of science. This is to thank all those who spent their valuable time listening and providing valuable comments.

From the foregoing, it is clear that useful comparisons can be made using this methodology. The results from exergoeconomic analyses can significantly deviate from those obtained with a classical analysis. Which of the two is the more relevant one will depend on the situation. For non-integrated systems, the classical analysis may be the one to follow, but for integrated energy systems, which are expected to become more and more important, the temperature level of heat becomes important, and the exergoeconomic analysis seems more appropriate. Using the f-factor will help in finding optimum solutions, especially for exergoeconomic analyses.

REFERENCES [1] G. Wall, ―Exergy and Morals‖, in Second law analysis of energy systems: towards the 21st century, E. Sciubba, M.J. Moran Eds, Circus, Roma (1995), ISBN 88-86662-0-9, pp. 21-29.

Variations in external factors, such as fuel costs or Government / utility incentives could change the shape of the curves to make the minimum more pronounced.

[2] D. Schmidt, ―Design of Low Exergy Buildings – Method and a Pre-Design Tool‖, in International Journal of Low Exergy and Sustainable Buildings, Vol. 3 (2003), pp. 120-126.

CONCLUSION AND SUGGESTIONS FOR FURTHER WORK

[3] A. Valero, L. Serra & J. Uche, ―Fundamentals of Exergy Cost Accounting and Thermoeconomics. Part I: Theory‖, in Journal of Energy Resources Technology, Vol. 128 (2006), pp. 1-8.

From the results of testing the methodology of exergoeconomic optimization using the f-factor, it is clear that it is a useful tool to determine the effects of different heating technologies and heat transfer surface sizes of these technologies on the annual overall operational costs. This is especially true if the heating system is integrated with other energy systems. It is also true if the temperature level of the heat is important for another reason. The methodology can be used to make informed choices regarding technologies to be used for heating homes or buildings and regarding the size of these technologies.

[4] G. Tsatsaronis, ―Application of Thermoeconomics to the Design and Synthesis of Energy Plants‖, in Exergy, Energy System Analysis, and Optimization, [ed. Christos A. Frongopoulos], in Encyclopedia of Life Support Systems (EOLSS), developed under auspices of the Unesco, Eolss Publishers, Oxford, UK (2007). [5] G. Temir, D. Bilge, ―Thermoeconomic analysis of a trigeneration system‖ in Applied Thermal Engineering, Vol. 24 (2004), pp. 2689-2699.

To continue this development work, it is recommended that more practical considerations will be incorporated into the models and analyses. Increasing temperatures do not just cost more in terms of exergy but also in more expensive materials, and steam based district heating systems are considerably more expensive than hot water based systems. Heat losses from the pipeline were small but may need to be considered in a followup study. Including passive heating of houses by solar radiation, plug loads and occupancy gains will also improve model predictions. Also mixed systems

[6] Clean Energy Project Analysis – RETScreen Engineering & Cases Textbook, 3rd edition, RETScreen International, Natural Resources Canada, Varennes (2005).

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

SLIMNET: AN INNOVATIVE INTEGRAL APPROACH FOR IMPROVING EFFICIENSIES OF DISTRICT HEATING NETWORKS M. W. P. van Lier Stadsverwarming Purmerend B.V., the Netherlands m.v.lier@svpbv.nl installations. Specific to the secondary network are the post-insulated steel distribution pipes and connections to customer installations hanging in narrow crawl spaces under blocks of buildings.

ABSTRACT This paper describes the innovative integral approach improving district heating network efficiency, SlimNet. SlimNet consists of five phases which lead to annual energy savings of about 227.000 GJ and almost 37.000 ton CO2 savings for the city of Purmerend in 2015.

In the distribution process no heat exchangers are used except from the production of hot tapping water in the houses. Hydraulics are controlled by decentralized pressurizing valves, differential pressure valves and pumps compensating for hydraulic deficiencies.

INTRODUCTION Company situation In 2007 the new company Stadsverwarming Purmerend B.V. (SVP) took over the responsibilities of the district heating network from the municipality in Purmerend, the Netherlands. With 25.000 customers the grid is the fourth largest grid of the Netherlands. District heating Purmerend started in 1980. The network expanded organically following the city expansions. While daily operations were outsourced to external and changing partners, the final responsibility stayed with the municipality.

The supply temperature from production is directly related to the ambient temperature (i.g. 95 C at Ta=-10 C and 75 C when Ta=15 C). The maximum supply pressure to the primary network is 6,8 bars and to the secondary network 4,5 bars. NETWORK CONDITION Part of the business analysis was an extensive technical research program covering all technical aspects of the grid and finally entire district heating chain. The main conclusions were:

A comprehensive business analysis performed by the new management in 2008 showed severe problems. In the present state the company would remain structurally loss giving, (future) heat delivery was not ensured, and sustainability and customer satisfaction were below benchmark standards. Fall 2009 a new business plan was presented that sets course for a future proof company, based on sustainable, costeffective and 80% renewable heat. On the technical side this is achieved by two major project programs, a. improving network efficiency, SlimNet, and b. incorporation of sustainable energy sources, the Energy transition. The company mission is to become the most sustainable district heating company of the Netherlands.

The network characteristic had become uncontrollable: Network builds out has occurred without a master plan. Effectively SVP had no control on the characteristics of customer installations. Furthermore, hydraulic problems in the grid had been masked with decentralized pumps and control systems.

Heat production capacity was critical, reaching a critical limit under the conditions of the winter of 2008. There was certainly no spare capacity to facilitate the planned expansion of the grid and thus the heat demand as shown in Fig 1.

Network description The 520 km district heating network is fed by a CHP (CCGT) plant of 65 MWth and seven natural gas fired auxiliary boilers with a total power of 131 MWth. During the last 6 years 64% of the total heat production came from the CHP plant. The heat sources are operated by a third party. The production units feed the heat to the network via buffering tanks to the primary network. The heat is then directly transported through substations and a secondary network to the 25.000 customer

Fig. 1 Required heat production 53

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

In 2008 the network showed a heat loss factor of 33,6% (with a Dutch benchmark of 25%). requiring 32.683 m3 of water replenishment in the same year.

Primary network Most substations in the network are provided with a SCADA6 system. This data in combination with a newly developed network model made it possible to calculating annual heat loss at 100.706 GJ.

Parts of the network showed excessive heat loss and repairs, mainly due to high ground water table, exposing the pipes in crawl spaces directly to water for most of the year. Repairs with standard material proofed insufficient and innovation on material and building techniques was needed.

According to [3] about 14% of this heat loss is caused by cross-linked polyethylene (PEX) piping material used in the early 90‘s.

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

Considering that those PEX pipes are applied in only 3,5% of the primary network, these may be referred to as ―hotspots‖.

D. Developing and defining measures E. Quantifying KPI results from SlimNet In the following those phases will be discussed.

Secondary network With four public housing companies, SVP conducted research on failures in the district heating related systems in Purmerend [3]. It became clear that during the period 2006-2008 74% of the unplanned repairs were caused by the high ground water level in the crawl spaces where post-insulated steel pipes with Armaflex insulation are installed. In total research identified areas of 4000 houses, where heat loss was extreme, i.e. ―hotspots‖.

KNOWING WHERE THE HEAT FLOWS For SVP the heat losses are defined as:

Qloss  Q produced  Qsold

(1)

The heat losses in the network, Qloss, were 427.158 GJ (33,6%) in 2008. Causes for those losses5 are: 1.

Losses in buffering tanks

Losses in primary network

Losses in secondary network

Undefined losses

This research confirmed the conclusion of an earlier research [4] that the thermal conductivity k for the wet insulation in the crawl spaces will be close to 0,1 W/mK and 0,2 W/mK instead of the 0,02 or 0,03 W/mK for the current pre-insulated pipes. The total of heat losses in the secondary network are estimated at 304.041 GJ.

None of the above can be determined exactly within the boundary conditions of the network but the following describes the results of the research performed on this matter and the localization of ―hotspots‖, parts of the grid with excessive losses.

Conclusion addressing heat losses Table 1 gives the overall results of the heat loss analysis.

Buffering tanks In [1] an estimated calculation was made for the heat losses due to the buffering tanks, 5.562 GJ annually. There are four buffering tanks with a 4.000 m3 capacity in the network which are used for peak shaving. A check upon this calculation [2], based upon an IR-scan of one of the buffering tanks resulted in an estimate of 14.032 GJ annually which is considered to be a maximum value. 5

Table 1: Overall results of heat loss analysis Main network part Buffering tanks Primary network Secondary network Undefined losses Total

Losses from heat plants are not taken into account.

Loss(GJ) 14.032 100.706 304.041 8.170 427.158

Supervisory Control And Data Acquisition

% of total 3,3 % 23,6 % 71,2 % 1,9 % 100%

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 primary network and post-insulated pipes in the crawl spaces of houses in the areas identified as ―hotspots‖ was the most effective strategy for heat loss reduction.

DEVELOPING ANALYZING TOOLS Research had located the ―hotspots‖ of unplanned repairs and heat loss in an area of 4000 houses. These ―hotspots‖ were responsible for 50% of the unplanned repairs. In order to define and implement a suitable and cost-effective replacement strategy a set of tools was developed.

DEFINING KEY PERFORMANCE INDICATORS The main goal of SlimNet is improving network efficiency as part of the new business plan that sets course for a future proof company which provides sustainable, cost-effective and 80% renewable heat. The Key Performance Indicators (KPI‘s) can be divided in four main criterions: 1. 2. 3. 4.

Upgraded network diagram Analyzing networks requires reliable and comprehensive network diagrams. All required information such as dimensions, age, depth etc. should be available in the diagram. Many network diagrams are drawn using CAD-software. Analyzing from those drawings is costly. It therefore was chosen to revise the diagram completely and apply the possibility to add element attributes to the drawing connected to an integral database system. The upgraded network diagram had a catalytic effect on two other models, the network model and the grid valuation model.

Economics Sustainability Reliability Customer Satisfaction

Economics Every GJ of heat lost in the network cannot be sold and has therefore a negative effect on the balance sheet. Consequently the heat loss in the DH-network is an obvious and important KPI.

Network model In 2009 SVP replaced the outdated and inadequate network with a validated dynamic model (TERMIS), developed by 7-Technologies with COWI as system integrator. With the upgraded network diagram SVP had the first and validated model of the primary network within five months.

Another parameter that has a negative effect on profitability is the amount of water that is replenished. Sustainability The avoided CO2-emissions are and should be an important driver for DH grids. According to subsequent directives in the Netherlands for assessing energy performance of buildings NEN 7120, the avoided CO2-emissions has to be determined on the required primary energy sources and by referring to common state-of-the-art technologies. The HR-107 type (107% LHV efficiency) is the required and accepted common state-of-the-art reference technology.

In combination with a new CRM system, operational since 2010, SVP will soon be able to tap into the information on customer behavior and consumption. This will allow SVP to dynamically calculate the current state of flow, pressure and temperature throughout the network at a configurable cycle time. Additionally, every real-time model calculation cycle will include a forecast simulation for a given period. This allows SVP to be abreast of demands, enabling optimization of operations and planning of the future.

Reliability The condition of the network in terms of reliability presents itself in the amount of times that mechanics have to deal with unplanned repairs. It was apparent that SVP was facing an increasing trend curve. The actual deprecation of the replaced piping provided a another criterion for assessing system degradation.

Valuation model The upgraded network diagram supplied database information on lengths, dimensions, age and type. With the following equations added to the database it was possible to develop a valuation model, that could help to prioritize and direct renovation efforts.

Customer satisfaction Reducing off time, during replacement important element of the SlimNet approach.

was an

X 

KPI summary 1. 2. 3. 4. 5. 6.

network

R x 1

Heat loss Water replenishment Avoided CO2 emissions Unplanned repairs Network degradation Off-time during replacement

Y

network

 x 1

Lx

 D  Ax     R x  Lx  D 

(2)

(3)

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Z Y 

network

 x 1

 D   Ax  1     Rx  Lx D  

Strategic metering (4)

It was concluded that actual data on heat loss on smaller scale (houses and clusters of houses) would facilitate decision making on future renovation projects and grid management. To get hold of this information SVP installed heat meters with radio transmission modules on strategic positions in the network. Together with the metering data from heat meters in customer installations this firstly gives accurate data on the heat loss in the corresponding part of the grid. This setup will also provide us with empirical data on the long term results of network improvement measures.

X = value network in new state (€) Y = current network value (€) Z = required annual maintenance costs (€) x = pipe Rx = construction costs per meter pipe dimension x (€) Lx = length of pipe x (m) D = lifetime expectancy (year)

In order to make the data comparable, two areas where chosen. One with the new SlimNet approach (Using polybutene pipes and new construction techniques) and one with conventional material and construction techniques. First comparative results will be available by the end of 2010.

Ax = age of pipe x (year) The network degradation is defined as factor :



Y X

(5)

Leak detection

From consultation with amongst others COWI, it was concluded that networks with a < 0.5 are in a critical stage.

Most producers of pre-insulated pipe systems offer the possibility of leak detection wiring. Using a master plan with proper zero and recurrent measurements this would be a reliable method of leak detection. Unfortunately this is not applicable to the situation in Purmerend.

For the entire grid the  was above the threshold. Discriminating the  for separate grid sections helped to identify the hotspots and monitoring  will help to determine the effect of SlimNet.

With one of its partners SVP developed a method using tracer gas to detect leakages. The detection devices proofed to be very sensitive and with this method almost 2.500 houses have been inspected this year and last year. Leakages were detected in 3% of those cases, mostly in an early stage, that otherwise would only have been detected through visual sighting of damp.

Sustainability assessment model To assess the current sustainability results of the network, SVP developed a sustainability assessment model in accordance with Dutch law and guidelines, resulting in Fig. 3 [5]. This model can also predict the effects of optimization in the chain from production, distribution and delivery to customer installations

DEVELOPING AND DEFINING MEASURES It became clear soon that the only way to improve network performance was to rigorously renovate the hotspots and to start implementing a structural maintenance program in accordance to Z, Eq. 4. In sum the challenge was: a.) cost effectively renewing the steel pipes with wet insulation in narrow crawl spaces while b.) improving network efficiency. To meet (a), SVP started the first two pilots in 2008 with pre-insulated steel flex piping material, using two different construction methods. Both pilots met the technical requirements but were too time consuming, costly and, because access to the crawl spaces had to be gained by digging in the gardens, meant huge inconvenience for customers.

Fig. 3 CO2 reduction DH-network Purmerend in past

It appeared that the ratio of CHP operation to the total of heat produced and the heat loss factor have the biggest impact on the sustainability results.

Parallel to this SVP had challenged pipe manufactures to come up with innovative material construction methods, suitable for the Dutch situation (groundwater 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 solution came from Flexalen of Thermaflex, using flexible polybutene (PB) carrier pipes. The producer of the PB material offers a 50 years plus life guarantee [6] for the pressures and temperature profiles of the SVP network.

SlimNet part I: Renovation and smart redesign Applying Flexalen means an improvement of k from 0,1 of the wet post-insulated steel pipes to theoretically 0,031 W/mK (manufacturer information, at 50 C). Key to the SlimNet approach was smart redesign. Calculations in TERMIS showed that many parts of the DH-grid in Purmerend are generally oversized, and that the common circular grid can easily be changed into a star shaped grid, whilst reducing pipe lengths. Using TERMIS redesign focused on reducing radial dimensions and pipe lengths by deleting obsolete pipes.

A pilot with Flexalen was conducted in September 2009. The pilot used prefabricated joints of Flexalen, called Flexalinks, which were under research and development at that time. The pilot did meet all the requirements. Costs were reduced by 30% compared to the steel flex pilots, 16 houses were overhauled within a week and access could be gained by the crawl space hatches.

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 currently started at the first 309 houses, at a speed of 30 houses a week.

Table 2: Results from redesign 2010 area

The second part of the challenge (b): improving network efficiency, is furthered by SlimNet through optimizing pipe dimensions and lengths (smart grid redesign)

Heat demand

Heat loss

Current situation

100,0 %

New dimensions

93,0 %

76,3 %

Finger system

91,0 %

69,5 %

Heat losses can be reduced by optimizing: 1.

Thermal conductivity

Pipe lengths

Radial dimensions

Fluid temperature

These elements are captured in the following equation for heat loss in a pipe [7]:

Qlo ss _ p ip e  2    k  L 

(Tin  To u t) r ln o u t rin

(6) Fig. 4 Existing network part to be renewed

k = thermal conductivity (W/mK) L = length of pipe (m) Tin = temperature of inside layer pipe (K) Tout = temperature of outside layer pipe (K) rin = inner radius (mm) rout= outer radius (mm) The first three of the above heat loss parameters can only be changed by renewing pipes. The last can only be changed by chain modification, i.e. production and customer installations. SlimNet addresses both.

Fig. 5 Redesigned and renewed network 57

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 closely monitored in the grid, through the strategic metering project.

houses that have a 90–50 C characteristic during design conditions (-10 C). In most areas before that time SVP found return temperatures that are structurally higher than the required 50 C. Hence the flows in those areas are also much higher than necessary.

SlimNet part II: Smart chain management The last heat loss parameter, fluid temperature (Eq. 6), can only be changed by modification of the complete chain.

The high return temperatures and corresponding high flows are caused by absence of pressurizing valves in the customer installations and defective control valves in the hot tapping water installations. By the end of 2010 SVP starts a campaign to encourage house owners to improve or renew their installations, also for their own benefit. This campaign will make use of local approved installers of customer installations. Research indicated that in certain areas the peak flow can be reduced with 60% [9].

To start at the production side, the current supply temperature is dependent on the ambient temperature, 95 C at Ta=-10 C and 75 C at Ta=15 C, Fig. 6. Lowering this curve, while still meeting the requirements of customer installations, would reduce the average network temperature hence the heat losses. It was calculated through the network model that the alternative temperature curve in Fig. 6 solely would reduce the heat losses with 4%. Further research will focus on matching the most effective temperature curve with production characteristics.

QUANTIFYING KPI RESULTS FROM SLIMNET Summarized, the measures that SVP takes before 2014 to improve network efficiency:

Fig. 6 Existing and alternative temperature curve

This research will also look upon the possibilities of implementing demand-driven heat production. This is achieved by using a real time network model connected to the substations and production SCADA. The model uses the weather forecast with customer information to adjust the temperatures and pressures just to meet the requirements of customer installations. It is expected that this will reduce the average fluid temperature even more.

Renewing the distribution pipes and house connections in the crawl spaces of 4000 houses, while optimized to dimensions and lengths.

Replacing 4,0 km PEX-pipes in the primary network, while optimized to dimensions and lengths.

Doing this with a minimum of off-time for customers

Implementing demand-driven heat production

Implementing cascaded heating installations

Encourage house owners to improve or renew their installations in accordance with SVP guidelines.

Eliminating arrears of maintenance an implementing a structural preventative maintenance program.

Heat losses will reduce from 33,6% in 2008 to 22,1% in 2015. While heat consumption prognoses stays the same, the corresponding required heat production falls, Fig. 8. This results in a energy saving of 227.000 GJ that year. In Fig. 9 the results of the sustainability assessment model are shown regarding CO2 savings.

Further research is done to implement cascading heating services, i.e. using the latent heat in the return pipes of the network with temperatures between 45 C and 60 C to the customer installations. This is however only possible to implement in new houses with low temperature heating installations. This research will focus on further reducing the heat losses. At the other end of the chain are the customer installations. Since 1996 the district heating company in Purmerend has only accepted installations in new 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 034APD-2009-0021)‖, TNO Bouw en Ondergrond, Apeldoorn, the Netherlands (2009)

Fig. 7 Required heat production with SlimNet

[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, ―Introduction to thermal and fluid engineering‖, John Wiley& Sons, Hoboken, USA (2005), pp 103

Replacing the post-insulated steel and PEX pipes together with a maintenance program including leak detection will have a positive effect on the water replenishment. The leak detection actions have already resulted in a 30.285 m³ replenishment in 2009, which is a 7% reduction compared to 2008.

[8] T.A. Østergaard, ― New dimensions for O16‖, COWI A/S, Aarhus, Denmark (2010) [9] A. B. Zitoony and B. E. Roukema, ―Rapport inregelstatus onderstations Stadsverwarming Purmerend 10.001.V2‖, Roukema B.V., Groningen, the Netherlands (2010)

It is expected that al measures will result in a 50% reduction in 2015. Unplanned repairs will also reduce 50% and consequently  is expected to improve significantly.

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

Marek Brand , Jan Eric Thorsen , Svend Svendsen and Christian Holm Christiansen

Ph.D. student, Technical University of Denmark Senior project manager, Danfoss District Energy, Nordborg, Denmark 3 Professor, Ph.D., Technical University of Denmark 4 Danish Technological Institute, Denmark will form areas with lower heat demand than nowadays. Currently used DH networks will not be able supply these areas in economical way, because the ratio between network heat losses and heat consumption in buildings would be unacceptable and thus cost of heat for end users will increase and DH systems will loose concurrency with other solutions, e.g. heat pumps. Recently, research in DH is focused to find the way how to use DH in areas with low energy buildings and how to increase ratio of heat produced by renewable sources of energy as solar heat plants or heat pumps driven by electricity from renewable sources.

ABSTRACT The increasing number of new and renovated buildings with reduced heating requirements will soon make traditional District Heating (DH) systems uneconomic. To keep DH competitive in the future, the heat loss in DH networks needs to be reduced. One option is to reduce the supply temperature of DH as much as possible. This requires a review of the behaviour of the whole domestic hot water (DHW) supply system with focus on the user comfort and overall costs. This paper describes some practical approaches to the implementation of this Low Energy District Heating (LEDH) concept. It reports on the testing of the dynamic behaviour of an Instantaneous Heat Exchanger Unit (IHEU) designed for DHW heating and space heating in detached family houses supplied by LEDH ensuring an entry-to-substation temperature of 51 °C. We measured the time it takes for the IHEU to produce DHW with a temperature of 42 °C and 47 °C when the tap is opened. Measurements were made for control strategies using internal and external by-pass and no by-pass. Our results show the importance of keeping the branch pipe warm if comfort requirements are to be fulfilled, but this involves higher user costs for heating. To increase user comfort without increasing costs, we propose the whole-year operation of floor heating in bathrooms, partly supplied by by-pass flow.

One of interesting application of renewable energy in DH is use of decentralised heat sources as e.g. solar collectors installed on roofs of individual buildings, supplying heat to DH network, but it still needs more time and work to develop new substations and new concept of DH networks to be able to handle these new features. The solution for future development of DH is to reduce heat losses of DH networks by means of pipes with better insulation properties e.g. twin pipes, use better concepts of network design (circular network configuration, possibility of using circulation line for main pipes) and to reduce the supply temperature of district heating water to lowest level as possible. The District Heating Systems designed due to this philosophy are called Low Energy District Heating Systems (LEDH). The main focus in LEDH system is to reduce heat losses from network as much as possible, exploit more sources of renewable energy for heat supply and still maintain or improve level of comfort for users, because without high level of comfort this concept can‘t be successful. LEDH concept was reported e.g. in project ―Development and Demonstration of Low Energy District Heating for Low Energy Buildings [2], where theoretical case study documented, that LEDH concept is a good solution for future and even in sparse housing areas is fully competitive to heat pumps. This article is focused on application of LEDH for DHW heating. Considerations related to use of LEDH for space heating will be reported in future in another article.

INTRODUCTION District Heating (DH) is a well known concept of providing buildings with heat for space heating (SH) and Domestic Hot Water (DHW) heating in economical and environmentally friendly way. Nowadays, building regulations have been introduced worldwide and are pushing to reduce energy consumption in buildings, because 40% of all energy consumption takes place in buildings. The energy policy of European Union is recently focused on energy savings, reducing production of CO2 and increasing the ratio of renewable energy [1]. DH is one of the most suitable solutions to achieve these goals for building sector and it gives high priority for further development of DH. But recenlty used traditional high and medium temperature DH systems are not optimal solution for the future. Sooner or later, energy consumption of all buildings will be in accordance with low energy building regulations and it 60

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

investigation of Legionella in DHW system using IHEU, producing DHW with temperature below 50 °C and reduced volume of the system below 3L.

LOW TEMPERATURE DISTRICT HEATING CONCEPT Reduced risk of Legionella by use of system with minimal volume of DHW

For single family houses with appropriate close location of tapping points, volume of DHW in IHEU and pipes will be lower than 3 L and thus temperature of 50 °C on primary side will not cause Legionella problems. For multi-storey buildings, district heating substations for each flat is a state of the art solution [8]. In this case, each flat has own completely separated DHW system (with volume of water below 3 L) and thus has increased users comfort and no huge DHW systems with circulation, where Legionella is forming and spreading [9]. The other advantage of using flat station in multi-storey buildings is individual metering of each flat and complete control over space heating and DHW preparation, which is positively affecting energy savings. With properly designed DHW building installations, supply temperature of LEDH will be defined by requirements for users comfort. These requirements are discussed in following text.

Since LEDH is mainly developed for low energy buildings already designed with low temperature space heating, the lowest acceptable forward temperature of LEDH system is defined by requirement for DHW supply temperature. The hygienic requirement for heating of DHW is due to recent standards 50 °C for single-family houses and 55 °C for multi-storey buildings [3] where DHW circulation is used. In case of using circulation, temperature of recirculated water should never fall below 50 °C. These requirements are based on need to avoid Legionella growth in DHW pipes and storage tanks. It is widely believed, that Legionella grow in temperature range between 46 °C – 20 °C, in systems with high volume of water. Mentioned temperature levels are made in order to assure comfort and hygienic requirements in furthest tap away from a heat source. It is important to say, that there is high level of discrepancy among different results and national standards focused on Legionella.

Users comfort in DHW supplied by LEDH Another important question, when concerning DHW systems is level of user comfort. From comfort point of view, requirements for temperature and waiting time for DHW can be specified. Due to Danish Standard DS439 ―Code of Practice for domestic water supply installations‖, [10] temperature of DHW should be 45 °C in kitchen and 40 °C in other taps, provided with nominal flowrate and desired temperature reached within ―reasonable‖ long time, without significant temperature fluctuations. It is a question, if requirement of 45 °C degrees for kitchen tap is not too high, but argument of problems with fat dissolving from dishes can be objected and should be investigated. Based on mentioned standard, desired temperature of DHW flowing from fixture is 45 °C. But in order to define desired forward temperature of LEDH system, we should be aware of temperature drop in DH network, in user‘s substation and in DHW installations in building. The temperature drop in DH network is not in focus of this paper, so our goal is to find needed temperature level at the entrance of substation to produce 45 °C from tap in building. Desired temperature will be found by experimental measurement of LEDH substation later in article.

Due to German Standard W551 [4], temperature of DHW can be below 50 °C and not cause Legionella promotion, if total volume of DHW system connected to one heat source is lower than 3 L. From literature studied, it can be concluded that requirements to produce DHW with temperature higher than 50 °C are defined for an old fashion DHW building installations, which can be characterized as systems with vertical riser, branched pipes with bigger diameter (increasing water volume of the system), using DHW circulation. For new and renovated buildings, DHW installations are designed in much better manner, with individual connection of DHW pipes between each tap and source of DHW and with maximally reduced pipe diameter, defined by requirements for noise propagation and pressure drop. Due literature, danger of Legionella growth in DHW system is influenced by temperature of DHW, nutrients in DHW, laminar or turbulent flow in the DHW pipes and water stagnation [5]. Several on site measurements were performed in buildings using DH for DHW heating. From results of Martinelli [6] and Mathys [7] can be concluded, that Instantaneous Heat Exchanger Unit (IHEU) tend to have much less problems with Legionella than traditional units with DHW storage tank. Both studies concluded, that these findings are caused by the fact that in IHEU, DHW is produced with temperature 60 °C, while in case of storage units only with temperature 50 °C. But is necessary to mention, that in case of traditional DHW storage tanks, overall volume of DHW in a system is much higher than in case of IHEU system. Due to our knowledge, there is not reported

Beside temperature requirements, users comfort is influenced by time needed for DHW to reach a fixture after tapping was started. This waiting time is in following text called ―tap delay‖. Due to DS439, suggested value for tap delay is 10 sec and it is defined in order to avoid wasting of water and to protect users against too long waiting times for DHW. In large multistorey buildings with centralised preparation of DHW, 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 properly designed or maintained DHW circulation is quite often responsible for increased risk of Legionella [11]. Another disadvantage of DHW circulation is big heat losses, sometimes even bigger than net energy needed for DHW heating [8]. The 10 sec waiting time is not rule and for some people it is a long time, for some people short, but this value is used to evaluate tested concepts if they are fulfilling requirements for high level of users comfort or not. An overall tap delay can be studied from different angles. From dynamic point of view, tap delay consists of transportation time needed for ―new volume‖ of water travel to tap and dynamic thermal behaviour of passed components, i.e. pipes and substation. From point of view related to location, it consists of three parts, tap delay in branch pipe (pipe from DH pipe in street to users substation), in DH substation and in DHW system in building. A tap delay in branch pipe and substation are related to DH network and substation‘s control system strategy, while tap delay in DHW pipes in buildings without DHW circulation are defined only by thermal capacity of pipes, volume of water in individual pipes, nominal flow and to some extend also by their insulation.

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

Tap delay in DHW system in building For DHW systems with individual feeding pipes and overall volume of pipes lower than 3 L, DHW circulation is not needed, because waiting time for DHW with desired temperature is not critical. In Table 1, transport delays for individual fixtures in typical house built in pilot LEDH project in Larch Garden - Lystrup, Denmark [11] are presented. It should be mentioned, that data are only transport delay, without dynamic behaviour of cooled pipe. From Table 1 can be seen, that reasonably designed close locations of fixtures, not so far away from substation, lead to maximal transport delay around 6 sec, for basin. The total volume of DHW system consists of 0.99 L in pipes and 1.1 L in HEX (type XB37H-40). It means, that it is possible to install longer pipes or more fixtures and still fulfil requirement of DHW system with volume lower than 3 L. The velocity of flowing water is below 2 m/s and thus problems with noise propagation during tapping are avoided.

Table 2 – Transport delay for nominal flows for individual fixtures due to DS439, in branch pipe, 10 m long, for typical house in Lystrup, data simulate using proportional flow controller without by-pass inner nom. pipe fixture .flow Ød (L/min) (mm) basin 3.4 15 kitchen 6 15 shower 8.4 10 shower 8.4 15 bath 12.6 15

Table 1 – Transport delay for nominal flows for individual fixtures due to DS439, in DHW system in typical house in Lystrup, for pipes with inner diameter 10 mm length volume nominal transp. to in velocity fixture flow delay fixture pipes (m/s) (L/min) (s) (m) (L) shower 8.4 2.2 0.17 1.8 1.2 basin 3.4 4.1 0.32 0.7 5.8 kitchen 6 6.3 0.49 1.3 4.9

volume in pipes (L)

velocity (m/s)

transp. delay (s)

1.77 1.77 0.79 1.77 1.77

0.3 0.6 1.8 0.8 1.2

31.6 17.7 5.6 12.6 8.4

Full opening from beginning of tapping leads to much higher flow rate on primary side than on secondary and time delay is decreased substantially. This solution can be used for short branch pipes with reduced diameters. But it should be mentioned, that transport time in branch pipe will be always limited by maximal allowed flow on 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 restrictor or by available differential pressure in DH network. To reduce tap delay on primary side, control concepts with by-pass, avoiding cooling of DH water in branch pipes and substations, and thus reducing substantially waiting time for DHW are available. There are two concepts of by-pass in relation to the heat exchanger: external and internal by-pass (see Fig. 1). In

case of external by-pass, DH water enters substation, but not enters heat exchanger and is sent back to DH return pipe and thus branch pipe is kept on desired temperature. Desired temperature is controlled by thermostatic valve situated in by-pass loop. Increased level of comfort expressed by reduced tap delay can be adjusted independently on temperature of DHW on secondary side.

Fig. 1 Different by-pass strategies for IHEU: left - no-by pass; middle - external by-pass (cold HEX); right - internal by-pass (hot HEX)

The set-point temperature of external by-pass is always compromise between insufficient cooling of DH water and additional heat consumed by customer and reduced waiting time for DHW. In case of operation of space heating system, the function of by-pass is to some extend overtaken by space heating loop and thus heat for ―by-pass‖ operation is not wasted and temperature of DH water returning to DH network is cooled sufficiently..In case of internal by-pass, by-pass flow is passing through heat exchanger and keep it warm (see Fig. 1). The benefit of this solution is even more reduced tap delay than in case of external by-pass, but on the other hand, since heat exchanger is kept warm, internal by-pass solution has additional heat losses. If substation is installed in room with need of space heating, heat losses are considered only outside of heating season.

Fig. 2 Combined by-pass concept, with possibility of use by-pass flow in space heating loop

In order to run by-pass without drawback of insufficient cooling of DH water and wasted heat also outside of heating season, it is proposed to use by-pass flow for floor heating, installed in bathroom and operate it all year. From preliminary calculations it looks, that flow needed to keep bathroom floor surface temperature on 24°C will be enough as by-pass flow. Considering the use of renewable sources of heat, the problem of insufficiently cooled DH water is related to reduced efficiency of these sources and whole year using of floor heating for comfort in bathroom is reasonable.

Contrary to external by-pass solution, where it is not so important if space heating loop is installed in series or in parallel to DHW heat exchanger, in case of internal by-pass it is in importance. If space heating loop is connected in parallel to DHW heat exchanger in traditional way, by-pass water just pass through DHW heat exchanger and is sent back to DH network with still high return temperature, without any other use. If space heating loop is connected in series to DHW heat exchanger or in parallel but with possibility to sent bypass water flown through internal by-pass to space heating loop (see Fig. 2), this solution provides high level of comfort for users as well as proper use of heat needed for by-pass operation.

Supply – supply recirculation As an alternative solution for customers who don‘t want to use whole year bathroom floor heating, solution called supply-supply recirculation is a possibility how to use benefits of by-pass without whole year heating of bathroom. In this case, district 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, circulated through HEX or external by-pass (see Fig. 3) and then sent back to district heating network (DHN) supply by pipe no.3. This concept is in early stage of investigation but it looks promising. The main question will be related to flow of DH water in branch pipe in order not cool it down too much before will be sent back to DH supply pipe in the street.

lower heat loss. On site measurements were started in Lystrup to evaluate performance of both types of DH substations, but no detailed measurements requiring short time steps are performed to evaluate level of users comfort. The measurements more focused to user‘s comfort are planed to be performed this year in Danish Technological Institute and Technical University of Denmark (DTU) on DH systems simulating the conditions in Lystrup. The DH systems will consist of branch pipes, substation and DHW building installations and different control approaches (external or internal by-pass, different set up by-pass temperatures, possibility of supply-supply recirculation, etc.) will be studied for DH substations supplied by LEDH. Measured data will be used for evaluation of performance of different control concepts, level of users comfort and lately also for validation of numerical model which is aimed to be developed for optimization LEDH systems. TEST OF TEMPERATURE PERFORMANCE

Fig. 3 Supply – supply recirculation with external by-pass

As a first part of measurements planed to be performed at DTU, the time needed for IHEU to produce DHW with temperature of 42 °C and 47 °C was measured, after tapping of DHW was started. The tap delay was investigated for two control strategies, one using internal and second using external by-pass. The measurements were performed for different initial conditions before tapping was started to simulate in realistic way users behaviour. Finally, the period between two by-pass flow operations was measured.

This solution is expected to be favourable mainly for circular shapes of DH networks, but it should be mentioned, that re-heating stations will be probably needed in point of DH network, where temperature of DH water decrease bellow defined value. Full scale demonstration of LEDH Full scale demonstration of LEDH is recently running in Larch Garden in Lystrup, Denmark [11], where 40 low energy houses class 1 and 2 are connected to LEDH system, with designed forward temperature from heat plant 52 °C. For primary side of substation, forward temperature of 50 °C and return temperature of 25 °C are designed. The DH network is built from highly insulated single pipes (for main pipes) and main pipes with smaller diameter, distribution and branch pipes are built from twin pipes. Two types of district heating substations providing houses with DHW and space heating are tested by customers in real conditions. The first concept is 29 Instantaneous Heat Exchanger Units (IHEU), second is 11 District Heating Water Units (DHWU). IHEU is classical concept of substation with instantaneous heat exchanger, only with enlarged number of plates. IHEU units have external by-pas, with set point temperature of 35 °C for customers situated not at the end of street and 40 °C for customers situated at the end of the street. DHWU is new concept of DH substation, reported e.g. by Paulsen [13]. DHWU consist of buffer tank for district heating water and when DHW is needed, DHW is heated in instantaneous heat exchanger as in previous case. Advantage of concept with buffer tank is peakshaved demand of DH water during charging and use of branch pipes with lower diameter, connected with

Experimental setup and instruments Tested DH substation was prototype of Instantaneous Heat Exchanger Unit (IHEU) developed specially for LEDH pilot project in Larch Garden – Lystrup, Denmark. The IHEU is a type of district heating substation consists of a heat exchanger (HEX) without storage tank. DHW is heated instantaneously in HEX only when tapping is performed and then supplied directly to DHW taps by individual feeding pipes, while space heating is using direct connection without heat exchanger, i.e. concept typical for Denmark. Substation is same concept as regular IHEU for traditional DH. The difference is in increased number of plates in heat exchanger assuring better heat transfer. Water volume of primary and secondary side is 1.1 L each and the heat exchanger is not insulated. The experiments were focused only on dynamic behaviour of substation related to DHW heating and thus space heating loop wasn‘t connected and space heating valves in substation closed. Desired temperatures of DHW were chosen in accordance with requirements in DS439 for temperature of DHW for kitchen sink and other fixtures. Required temperatures mentioned in DS 439 are 45 °C and 40 °C. In order to 64

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

cover additional temperature drop in building DHW installations, 2 °C were added. This addition is based on experience from previous measurements. During the experiments, temperatures of four different flows passing through the DH substation were measured. On primary side it was temperature of DH water supplied to substation (T11) and temperature of DH water returning back to DH network (T12) and on secondary side it was temperature of cold potable water entering substation (T21) and temperature of heated DHW (T22). All temperatures were measured by thermocouples type T, installed directly in pipes, in flowing water, so they do not have any practical time delay for the measurements. The time constant to reach 90% of step change was less than 1 second. The distance of thermocouples from substation flanges was 5 cm and thermocouples were previously calibrated. We also measured surface temperature of HEX in upper (HEX-UP) and bottom part (HEXDOWN) and temperature of air in the testing room. Temperatures were measured and collected by multifunction acquisition unit every second. For authentic simulation of DH network, DH water with constant temperature of 51 °C was necessary. It was solved by connecting of IHEU to source of DHW in laboratory of DTU, where DHW is supplied by DH system. DHW system of DTU is big enough, to assure stable temperature 51 °C without any fluctuations. In order to prevent cooling down of pipes supplying DHW to laboratory in periods when there was not flow through substation (stopped by by-pass controller), small guard flow, just before entrance to substations was kept to maintain DHW always on 51 °C and drained to sink.

temperature of 51 °C started to flow in the substation and flew through external by-pass, until closing temperature was reached and by-pass flow stopped. Then we wait until by-pass was opened again. Time between two by-pass openings as well as volume and temperature of DH water passed through by-pass was written down and after by-pass was closed again, we waited a little bit shorter time than was needed to open by-pass flow again and we start tapping on secondary side with flow rate 8.4 L/min. In this way, most unfavourable condition for substation with by-pass, i.e. highest recovery time, was measured. After tapping of DHW was finished, we wait 5 minutes and we performed one more tapping to simulate short time step between two subsequent tapping of DHW. 2. For measurement of internal bypass concept, IHPT controller was used. In case of IHPT, by-pass set point temperature can‘t be adjusted independently and is defined by desired temperature of DHW, i.e. 47 °C for our measurements. IHPT controller was developed for traditional DH networks operating with forward temperatures around 70 °C. For traditional DH, bypass opens when temperature in HEX falls 5–7 °C below set point of DHW, but in case of LEDH with forward temperature 51 °C, by-pass opens 1 °C below DHW set point temperature, i.e. 46 °C in our case. The testing procedure was similar to measurements with external by-pass. After supply valve on primary side of substation was opened, DH water with temperature of 51 °C started to flow in the substation and temper HEX, until by-pass closing temperature was reached. Then we wait until by-pass was opened again and we performed tapping of DHW just before next by-pass opening was expected. In following steps was procedure same as in case of external by-pass.

Experimental procedure As a first step, both controllers were adjusted to provide 47 °C on DHW side with supply temperature of DH water 51 °C. Then we measured time delay in the substation, i.e. time needed for substation to produce DHW with temperature 42 °C and 47 °C on secondary side outlet from the moment when DHW tap is opened.

Moreover, we also performed measurements of time delay in IHEU for control concept without by-pass. RESULTS Time delay for IHEU with PTC2+P controller and external by-pass adjusted to 35 °C to start supply DHW water with temperature 42 °C and 47 °C after long idling period just before opening of external bypass was expected, can be seen from Fig. 4 and is 11 and 22 seconds, respectively. This measurement represents condition with the longest time delay for PTC2+P controller. Temperature of room, where IHEU was installed was 22.2 °C. For this case, temperatures of produced DHW in first 10 sec after tapping was started are listed in Table 3.

The measurements were performed for different initial conditions and secondary flowrate was always 8.4 L/min, which is nominal flow for shower. 1. For measurements of concept with external by-pass, substation was controlled by PTC2+P controller with by-pass set point temperature adjusted to 35 °C. This setup is exactly the same as is installed in Lystrup pilot project. The testing procedure was made in following steps. Substation was left idle for long time in the testing room, so all components and water in HEX were on room temperature. Than we opened the valve on DH supply in substation and DH water with

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 long idling just after by-pass flow was stopped, time delay decreased to 8,5 and 16,5 seconds. In this measurement, temperature of substation and thus water standing in the HEX was little higher than ambient air temperature. It is expected that time delay will be slightly longer, if substation will have real ambient temperature but still shorter than in case 2. We also performed measurement of tap delay five minutes after previous DHW tapping was finished.

In this case, tap delay in substation to produce DHW with temperature 42 °C and 47 °C was shorter, 7 and 14 seconds. For room temperature around 22 °C, external by-pass was opened roughly every 30 minutes. The by-pass was in average opened 2.5 minute and volume of DH water needed to close the by-pass was in average 3 L, i.e. when substation is idle, by-pass uses 6 L of DH water per hour.

Table 3 – Temperatures measured for PTC2+P controller in first 10 sec after tapping was started for situation after long idling, just before by-pass was expected to run again (sec)

T22 (°C) 21.6 22.3 26.0 29.7 32.6 35.0 36.9 38.7 39.9 41.2 42.2 42.8 43.5 44.2 44.7 45.1 45.5

Time delay in IHEU equipped with IHPT controller with internal by-pass adjusted by requirement of DHW to 47 °C was 6 and 14 seconds to reach 42 °C and 47 °C on outlet for situation when tapping was performed just before by-pass was expected to open. The internal bypass opens 3 minutes after previous tapping is finished and when is once opened never closes, only when another tapping is performed, but again only on 3 minutes.

The average flow of internal by-pass was 24 L/hour and average return temperature to DH network was 45 °C. When internal by-pass is once opened, the time delay in substation decrease substantially to 1.5 and 7 seconds to produce DHW with temperature 42 °C and 47 °C. The condition with expected longest time delay was solution without by pass. In this case time delay to produce DHW with temperature 42 °C and 47 °C was 12 and 25 sec. All measured results are summarized in Table 4.

Table 4 – Overview of time delays for all measured cases case number and description NO BY PASS

1 – after long idling, no by-pass (BYP) 2 – after long idling, just before BYP was EXTERNAL expected to open again BY-PASS 3 – after long idling, just after BYP closed 4 – 5 minutes after previous tapping finished 5 – just before BYP was expected to open (3 INTERNAL min after prev. tapp. finished)) BY-PASS 6 – anytime, when BYP was already in operation

T11 (°C)

τ42 τ45 (sec) (sec)

τ47 (sec)

T12 (°C)

T12AVG THEX-UP THEX-DOWN (°C) (°C) (°C)

50.1

16.2

19.5

20.4

49.6

30.1

19.3

21.5

21.4

50.6 50.8

8.5 7

12 10

16.5 14

42.6 25

19 19.1

29 22.3

26 37.4

50.5

19.5

19.1

22.6

49.3

1.5

3.5

47.3

18.4

45.5

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

only difference is that in case 5 (internal by-pass), tap delay is again reduced because tapping was performed 3 minutes after previous (to prevent influence of by-pass) and thus HEX was warmer.

DISCUSSION Focused on level of users comfort and proper cooling of DH water during idling, time delay of LEDH substation to supply DHW with temperature 42 °C and 47 °C was measured. Three different control strategies related to tap delay were investigated. Obtained results represent case of IHEU used in single-family house in period when space heating is not in operation. Explored concepts can be evaluated from two different points of view, due to highest advantages for customer and for DHN.

If the requirement is to fulfil 10 sec tap delay for less favourable fixture, i.e. in our case basin (see Table 1), DHW should leave DH substation with temperature 42 °C in 4 sec after tapping was started, because it will take 6 second to reach the tap. This requirement was reached only by concept with internal by-pass and only when by-pass was already opened. On the other hand from Table 3can be seen, that even for concept with external by-pass and tapping after long idling and just before expected bypass opening, DHW at a temperature 26 °C leaving substation in 3 sec. DHW with this temperature is not sufficient for taking a comfortable shower for which temperature 37±1 °C is preferred, but for washing hands this temperature should be enough. The values in Table 3 are for flow rate used for shower, but it can be used to explain that it is time to rethink the suggested value of tap delay from 10 sec to another value and consider also nominal flows and use of tapped water. The different standards for the different use of DHW based on new solutions in DHW supply systems and results from test panels are needed, because it may have some influence on design of optimized DHW systems. Nevertheless, for customers requiring DHW in very short time e.g. continuously or discontinuously (only during rush hours) operated trace heating elements can assure almost no tap delay by keeping DHW staying in pipes on desired temperature.

The solution without by-pass is from energy savings point of view very interesting because doesn‘t need any DH water for idling, but from users comfort point of view is very poor because of reduced comfort and problems with wasting of water during waiting for DHW with desired temperature. Solution without by-pass can be probably used for substations equipped with combined thermostatic and proportional flow controller, for customers with short branch pipes or for customers with low requirements for level of users comfort. If solution without by-pass will be used for substation controlled only with proportional flow controller, even transport delay in 10 m long branch pipe for nominal flow for basin will be 32 sec. For period when space heating is operated, branch pipe will be kept warm from flow needed for space heating and time delay for solution without by-pass will be very similar to solution with external by-pass. Anyway, in non-circularly shaped DH networks, by-pass should be installed at least at the end of a street, so it is better to find solution how to use by-pass flow in useful way than sent it directly back to DH return. Considering this, it is suggested to use by-pass flow for whole year operation of floor heating in bathrooms to increase comfort for customers and at the same time solve problem with by-pass flow which otherwise increasing return temperature to DH network.

CONCLUSION Based on literature study it can be concluded that hygienic requirement of DHW with 50 °C on outlet of DHW heater is not needed for systems with a total volume of the DHW lower than 3 L.

From user comfort point of view, better solution than solution without by-pass, but consuming more energy, is substation equipped with external by-pass. By comparison of results of concepts without by-pass (case 1) and solution with external by-pass, for case when tapping is performed after long period of idling just before by-pass opens again (case 2), we can see that time delays are almost the same (see Table 4). Difference is only that for external by-pass are pipes in DH substation kept on higher temperature and it made slightly faster reaction. In the case 3, time delay is even more reduced since pipes in substation were warmer by just finished by-pass flow. For control concept with external by-pass and tapping repeated 5 minutes after previous one, time delay is again reduced, since HEX is still hot from previous tapping. The time delay for case 4 and 5 are almost the same,

From results of our measurements and evaluation of IHEU supplied by LEDH, only substation with external by-pass with set point 46 °C is able to produce 47 °C DHW in time bellow 10 sec. The easiest step how to decrease waiting time also for other concepts is to insulate HEX. This measure will reduce time delay for DHW tapping and also will decrease heat losses from DH substation. The lower waiting times for DHW can be also achieved by further optimisation of HEX in way of decreased number of plates reducing volume of water in HEX and thus transport delay, and by increased thermal efficiency of HEX (followed on the other hand by higher pressure loss). These modifications can lead for higher temperature of DH water returning to DH network, but during all our experiments, average return temperature was below 20 °C, what is 5 °C less than is designed for LEDH. 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 always trade-off between users comfort and reduced cooling of DH water during idling and thus customer should have to some extent possibility to choose which solution prefers. In case of traditional concepts, decision is between longer waiting time for DHW and energy savings or vice versa, if by-pass in substation is used. In non-circularly shaped networks, by-pass should be used anyway at least at the end of a street line. The one of possible solutions how use by-pass flow in better way can be proposed innovative concept of whole year operated floor heating in bathrooms or supply-supply recirculation. Both solutions will increase level of user comfort and at the same time also energy efficiency of DH system.

[4] DVGW, ‖W551 - Trinkwassererwärmungs- und Trinkwasserleitungsanlagen‖ ,1993, Bonn, (in German) [5] Z. Liu, ―Effect of flow regimes on the presence of Legionella within the biofilm of a model plumbing system‖, 2006, Journal of Applied Microbiology, Vol. 101, pp 437-442 [6] F. Martinelli, ―A Comparison of Legionella pneumophila Occurrence in Hot Water Tanks and Instantaneous Devices in Domestic, Nosocomial, and Community Environments‖, 2000, Current Microbiology, Vol. 41, pp. 374-376 [7] W. Mathys, J. Stanke, et. al.,‖ Occurrence of Legionella in hot water systems of single-family residences in suburbs of two German cities with special reference to solar and district heating‖, 2008, Int. J. Hyg. Environ. Health, Vol. 211, pp. 179-185

LEDH is a promising solution for providing buildings with DHW and space heating regarding fulfilling requirements of modern society with reduced CO2 emissions and energy consumption. More detailed investigations by testing of different parameters and numerical simulations are needed in order to optimize LEDH concept.

[8] H. Kristjansson, ―Distribution Systems in Apartment Buildings‖, Published at the 11th International Symposium on District Heating and Cooling, August 31 to September 2, 2008, Reykjavik, ICELAND

Future work It will be very interesting to compare time delay of substation for traditional DH with time delay for DHW produced by LEDH substation. It is expected that timed delay for LEDH will be higher because dynamic response is slowed down by lower temperature difference between DH water and desired temperature of DHW, but on the other hand, lower temperature difference is in some extend compensated by bigger HEX. It is also suggested to rethink ―10 sec tap delay suggestion‖ for different taping flows and purposes of DHW use.

[9] T. Persson, ―District Heating for Residential Areas with Single-Family Housing, paper IV‖, 2005, Doctoral Thesis, Lund Institute of Technology, Lund [10] Dansk Standard, ―DS 439 Code of Practice for domestic water supply installations‖, 2009 [11] P.K. Olsen, ―Low-Temperature District Heating System for Low-Energy Buildings‖, 2009, http://www.fbbb.dk/Files/Filer/Peter_Kaarup_Olsen _-_COWI_29-10_2009.pdf

REFERENCES

[12] H., Boysen, J.E. Thorsen, ―Control Concepts for DH Compact Stations‖, Published in Euroheat and Power IIII 2004

[1] S. Froning, ―Low energy communities with district heating and cooling‖, PLEA 2008 – 25th Conference on Passive and Low Energy Architecture, Dublin,

[13] O. Paulsen, ―Consumer Unit for Low Energy District Heating Net‖, Published at the 11th International Symposium on District Heating and Cooling, August 31 to September 2, 2008, Reykjavik, ICELAND

[2] Hovedrapport, ―Udvikling og Demonstration af Lavenergifjernvarme til Lavenergibyggeri‖ 2009, (in Danish) [3] EUROHEAT & POWER, ―Guidelines for District Heating Substations‖, 2008, downloaded from www.euroheat.org in October 2009, pp 8

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

CHALLENGES ON LOW HEAT DENSITY DISTRICT HEATING NETWORK DESIGN 1

M. Rämä and K. Sipilä 1

VTT Technical Research Centre of Finland PB 1000, FI-02044 VTT, Finland heating. The expansion of mature and large scale systems take place in areas with lower heat consumption. This transition to more demanding operational environment both technically and financially represents challenges to district heating network design. This is also true in small scale systems of limited consumption separated from a larger system. A careless network design in these circumstances can lead to deterioration of the advantages of district heating; efficiency and reliability. An annual heat loss of 5% in district heating distribution is considered a good result, but the case in question the heat losses can easily reach 10% or even tens of percents if the characteristics of low heat density areas are not taken into account in design.

ABSTRACT While district heating is an energy efficient solution to provide heating to areas with high heat consumption, mature systems extending out to more demanding operational environment face challenges maintaining competitiveness over alternative heating systems. As the heat density falls below a certain level, district heating is no longer economically feasible. Studying the possibilities of extending this threshold by means of district heating system design and pointing out the operational challenges while approaching it are the main topic of this paper. The problem is investigated in a representative case of a low heat density area bordering a more extensive district heating network. A node-and-branch type network simulation model is used study the operation of the network and a simulation period of one year is used to get a realistic view of the system in a normal operational cycle.

LOW HEAT DENSITY AREA A detached house area consisting of 56 identical 150 m2 houses with energy consumptions in compliance of today‘s building standards is studied. Dedicated heat exchangers between the network and the consumer exist for both heating and domestic hot water. Total energy consumption for the houses is 18.75 MWh/year of which domestic hot water has a share of 20 percent.

Not taking into account the characteristics of a low heat density area in network design can result in inefficient distribution system. Operational problems, especially maintaining the temperature level in summertime, must be solved. Only concentrating on minimizing the heat losses will not result in best possible design.

The district heating network studied is presented in Figure 1. The detached house connections are marked as green dots and the connection to the main district heating network as a red rectangle. The connections have 1, 2 or 6 detached houses as consumers, indicated by the size of the dot.

The temperature level issue can be solved with a bypass valve, auxiliary heating or accumulators, but in overall more efficient system requires steps to be taken in the houses. Floor heating and a heat pump coupled with an accumulator enables the use of low temperature design where the heat losses can be cut significantly. INTRODUCTION

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 heating systems. From the consumer point of view, district heating is considered as a reliable and carefree source of heating energy and is also often an economically sound choice.

Figure 1. District heating network studied.

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

Areas with high heat consumption i.e. economically the 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

The consumptions for both heating and domestic hot water use were given as hourly time series as well as the radiator supply and return temperatures on the secondary side.

As the pipe diameters are quite small, twin pipes with insulation class IV are used in the area as recommended by Energy Industry [1], [2] in Finland. The pressure drop design principle used here is roughly ~1.5 bar/km.

The heat exchangers were modelled with logarithmic temperature principle in a design point (described in Table 1) after which the conductance in W/K is assumed to be constant. When heat demand, both supply and return temperatures on secondary side and supply temperature on primary side are given as input, the primary return temperature and district heating mass flow can be calculated.

700

Pipe lenght (m)

600 500 400 300 200 100

Table 1. Design point for heat exchangers.

0 15

Description

Value

Pipe size (DN)

Primary side temperatures

Figure 2. Pipe size distribution

115/45 °C

Radiator heating

70/40 °C

The linear heat density is 0.44 MWh/m which makes the area a low heat density area by definition [3].

Domestic hot water

55/10 °C

Design heating load

8 830 W

The heat demand around the year is presented in Figure 3. The peak demand for the area is 507 kW. As expected, in the summertime the load consists almost solely of domestic hot water consumption.

Design DHW load

2 060 W

The design loads for domestic hot water are low compared to a real life design load of a heat exchanger in normal detached house in Finland, 50 kW is a common choice. This is due to the simulation model taking hourly data originally calculated for a multifamily house as input so the domestic hot water demand is also flatter than it really is. However, from the network design point of view hourly data is considered accurate enough.

Total heat demand (kW)

600 500 400 300 200

Other input data used were the undisturbed ground temperature of 5 °C, assumed to be constant, and the supply temperature from the main district heating network as a function of outdoor temperature. The outdoor temperature time series used described a typical year in Southern Finland. The supply temperature reaches its maximum value of 115 °C in an outdoor temperature of -26 °C and its lowest value of 75 °C in 5 °C. Between these two points, the relation is linear.

100 0 0

100

150

200

250

300

350

Days

Figure 3. Heat demand of the simulated area. SIMULATION MODEL A node-and-branch type simulation model [4] was used to study the case in hand. The model calculates temperatures and pressures for the nodes and flows and heat losses for the pipes, i.e. the branches. From these results pumping power can also be calculated, although a constant efficiency of 0.5 is used for the pump. The pressures are calculated separately from temperatures. The temperature calculation is dynamic while the flow and pressure calculation is not. A minimum 0.6 bar pressure difference over a consumer is assumed.

SIMULATION RESULTS The most interesting results concern the heat losses and the temperature variations within the network. The pumping needed (less than 1 MWh) in a network of this size is quite low and thus negligible. In the initial simulation runs it was noted that the system was struggling to maintain high enough temperature level in the summertime when the load consist solely of domestic hot water demand. This problem was met by defining a flow through valve at the consumer, opening

When defining the network, each pipe is given a start and an end node, a pipe type (twin, single), an insulation standard (class I to IV) and length. 70

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

relative heat losses in the heating season are acceptable, they reached 47 % in the summertime. The high heat losses are partly because of the by-pass valve letting hot water past the heat exchangers. The by-pass valve is also responsible for small cooling, i.e. the difference between supply and return temperatures, within the system in summertime (Figure 6).

when the supply temperature on primary side dropped too low (< 65 °C). The valve allowed a constant mass flow (0.015 kg/s) to go past the heat exchanger on the primary side. This solution helped the situation significantly although not without ill effects as can be seen from the heat losses presented below. The use of a by-pass valve to ensure the appropriate temperature level for domestic hot water also mean higher heat losses and pumping power and effectively lower cooling; all of which are undesirable outcomes. One possibility to solve the problem is just to accept the flaw and to use additional electrical heating element to raise the temperature of domestic hot water to the required level. As the temperature boost needed is for most of the time quite small and is only needed in summertime, the increase in electricity consumption is reasonable.

Relative heat losses (-)

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 I

Because of the high capital costs of district heating, the pipes should basically be sized as tight as possible while keeping in mind the future demand for the pipeline in question. As the pipes are small, the volume of water contained is also low. This leads to water cooling more rapidly than in larger pipes. The Figure 4 illustrates this with a simplified example by showing the temperature on supply side service pipes if there is no flow for three different pipe sizes. The temperature drop of 15 °C, for example, takes 5 times longer with a pipe size DN 50 than with a small DN 15 pipe. The calculations assume a constant return side temperature of 30 °C and a ground temperature of 5 °C.

III

VII

VIII

XII

Month

Figure 5. Monthly relative heat losses.

90 80

Cooling (°C)

70 60 50 40 30 20 10 0

DN 15

DN 25

DN 50

100

150

200

250

300

350

Days

Temperature (°C)

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

50 40 30

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

20 10 0 0

Time (h)

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

The use of smaller pipes reduces the heat losses in W/m and this is accentuated if the temperature level drops as described above. As a result, looking solely on heat losses when designing a low heat density area network on common design principles can lead to reliability issues as the system cannot supply the heat required by the consumers. The relative heat losses (that is, heat losses per needed production) for the simulated case are 13.8 % in a year. The monthly values can be seen in Figure 5. While the 71

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

1.25 MWh with an average COP of 3. With this setup, supply temperature would need to be just 40 °C.

REFERENCES [1] Lappeenranta University of Technology, Kaukolämpöjohtojen optimaalisen eristyspaksuuden tarkastelu / Investigation of the optimal insulation thickness on district heating pipes, Energy Industry, 2009, 36 p.

CONCLUSIONS The use of traditional district heating network design principles can lead to an inefficient area heating system in areas with low heat density. Special attention must be paid on operation of the system to ensure reliability, one of the advantages of district heating.

[2] Preinsulated district heating pipes, Recommendation L1/2010, Energy Industry, 2010, 44 p.

When aiming for an efficient system, one goal is to minimize the heat losses. However, concentrating solely on this can make another problem, maintaining high enough temperature level for domestic hot water in summertime, even worse. The problem can be solved using a by-pass valve, but this causes unwanted effects; worse cooling and an increase in heat losses and pumping power. Other solutions are auxiliary heating (electrical heating or a heat pump) or the use of an accumulator and with it, aiming for a steady domestic hot water load.

[3] Zinko, H., Bøhm, B., Kristjansson, H., Ottoson, U., Rämä, M., Sipilä, K., District heating distribution in areas with low heat demand density, IEA DHC Annex VIII, 2008, 117 p. [4] Ikäheimo, J., Söderman, J., Petterson, F., Ahtila, P., Keppo, I., Nuorkivi, A., Sipilä, K. 2005. DO2DES – Design of Optimal Distributed Energy Systems, Design of district heating network. Åbo Akademi. Report 2005-1.

Another approach is lower the supply temperature significantly and to use floor heating and heat pump with an accumulator for domestic hot water demand. This is not suitable for existing areas with a heating system already designed, but for new areas it is a reasonable and, compared to the traditional district heating design, an efficient way to provide heating.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

DESIGN OF LOW TEMPERATURE DISTRICT HEATING NETWORK WITH SUPPLY WATER RECIRCULATION 1

Hongwei Li , Alessandro Dalla Rosa , Svend Svendsen

Civil Engineering Department, Technical University of Denmark technologies become barriers to further increase the market share [2]. In order to sustain the economic competiveness and realize the long term sustainable development, the concept of design and operation of DH system needs to be re-examined under the new energy regulation and development trends. This is the main impetus for the development of the new generation DH system. Based on previous studies, in a properly designed in-house substation system, the network supply temperature at 55oC and return temperature at 20oC can meet the consumer space heating and domestic hot water demand [3].

ABSTRACT The focus on continuing improving building energy efficiency and reducing building energy consumption brings the key impetus for the development of the new generation district heating (DH) system. In the new generation DH network, the supply and return temperature are designed low in order to significantly reduce the network heat loss. Meanwhile, the low network operational temperature can make a better utilization of renewable energy and further improve the CHP plant efficiency. Though the designed return temperature is low, it may increase considerably when the heating load becomes low and the by-pass system starts to function. The aim of this paper is to investigate the influence of by-pass water on the network return temperature and introduce the concept of supply water recirculation into the network design so that the traditional by-pass system can be avoided. Instead of mixing the by-pass water with return water, the by-pass water is directed to a separated circulation line and returns back to the plant directly. Different pipe design concepts were tested and the annual thermal performances for a selected residential area were evaluated with the commercial program TERMIS. The simulation program calculates the heat loss in the twin pipe as that in the single pipe. The influence of this simplification on the supply/return water temperature prediction was analyzed by solving the coupled differential energy equations.

The low return temperature has the advantages to reduce the network heat loss, increase CHP plant power generation capability, and utilize direct flue gas condensation for waste heat recovery. However, the return temperature can become much higher than the designed value when the heating load becomes low and the by-pass system at the critical user starts to function. In this paper, the influence of by-pass water on network return temperature was examined for a reference residential area. The concept of supply water recirculation was introduced to avoid the mixing of bypass water and the return water. Three network design methods were tested. The annual thermal performance was evaluated with the commercial district heating network hydraulic and thermal simulation software TERMIS [4]. The simulation program calculates the heat loss in the twin pipe as that in the single pipe. The influence of this simplification on the supply/return water temperature prediction was analyzed by solving the coupled differential energy equations.

INTRODUCTION In European Union, one of the major energy development targets is to reduce the building energy consumption and increase the supply of renewable energy. The introduction of European Energy Performance of Building Directive (EPBD) poses stringent requirement for the member countries to effectively reduce their building energy consumption. According to the national energy policy, the building energy consumption in Denmark will drop to 25% of current level by the year 2060, while the renewable energy share will increase from 20% to 100% at the meantime [1]. District heating (DH) benefits from economic of scale with mass production of heat from central heating plants. The significant reduction of building energy consumption and wide exploitation of waste heat and renewable energy, however, makes the current DH

SUPPLY WATER RECIRCILUATION The solution to overcome the excessive temperature drop along the supply pipe due to reduced flow rate is to install by-pass system at the critical user in the network. Figure 1 shows the principle of supply water by-pass. Extra flow is called based on the temperature measurement at the critical user until the minimum supply temperature requirement is met. This extra flow is then â&#x20AC;&#x2022;by-passedâ&#x20AC;&#x2013; and sends back to the return pipe. As the by-pass flow rate may be considerable and its temperature is high, the mixing with return water will significantly increase the return water temperature which causes both increased heat loss in the return pipeline and decreased power generation capability in the CHP plant. 73

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

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

NETWORK SIMULATION Heating Load The simulation was performed for a reference area with 81 low energy demand houses. The house was designed based on the building standard Class 1, following the Danish Building Regulation. The domestic hot water draw-off profile was designed similar to the Danish standard DS439 [5]. Detailed space heating and domestic hot water heating load simulation can be found from [6, 7]. Figure 4 shows the averaged heating load and the corresponding duration hours. The annual heating load is divided into 8 intervals, varying as a function of undisturbed ground temperatures which ranges from 0 to 15 ºC. The summer season lasts 3281 hours and the heating load comes only from the domestic hot water demand. The space heating is required for the rest of the year.

Fig. 1 Schematic for hot water by-pass system

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 ºC. On the other hand, simultaneous factors which are the probabilities for multiple users‘ concurrent use of hot water are considered for the design of street pipes and main pipes, as shown in Table 1 [3]. Fig. 5 shows the domestic hot water storage tank (DHWS) in the DH system. The DHWS design load is 8 kW per house. To avoid the legionella problem, the design temperature for DHWS is higher than HE, at 65 ºC /30 ºC for supply and return respectively.

Fig. 2 Schematic for by-pass water recirculation

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

District Heating Network The DH network and the connection to the end users are shown in Fig. 6. The house is designed to connect to the plant directly through different diameter pipes which were optimized with the simulation program. The direct connection allows the primary DH network to circulate water directly into the end user installation. It is suitable for a moderate pressure level network and the differential pressure of DH network is sufficient to circulate water to the house installation. The networks and house installations are assumed to withstand maximum pressure 10 bar. The consumer differential pressure is set as 0.5 bar. It is controlled at the end user along the network critical route which is shown in green color.

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

Three network design scenarios were investigated for each house installation:

Table 1 Simultaneous Factors



Case 1: It is the reference case. The total network length is 3080 m and the network line heat density is 177 kWh/year. Network was designed in the traditional way for two pipes with one supply and one return, respectively. The differential pressure is controlled at user A. Twin pipes were selected for the DH network. They are called ―reference pipe‖ in this paper.



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



Case 3: Double pipeline supply. The main pipe (from plant to the junction point at each street) in the third pipeline which was sized in case 2 functions all year round. It acts as supply pipe during winter season and functions as supply water recirculation pipe when there has bypass water demand. In this case, the main pipe in the reference case was resized as a portion of supply water is shared by the recirculation pipe. The connection of recirculation pipe to the reference pipe is shown with red color.

The thermal by-pass temperature was set as 50 °C for HE and 60 °C for DHWS with dead band 2 °C. The bypass is placed on the end user at each street in case 1, while at the virtual point adjacent to the end user in case 2.

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

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

coefficients corresponding to the temperature difference between the flow and the ground.

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

The temperature variation along the pipeline was calculated as internal flow with isothermal boundary condition. The downstream temperature in the pipe is expressed as [4]: [3] Td, Tu and Ta represent the downstream fluid temperature, upstream fluid temperature, and ambient temperature respectively. M and K are parameters include the overall heat transfer coefficient. As the overall heat transfer coefficients have to be calculated beforehand, the influence of flow temperature variation on Us and Ur along the pipeline is neglected. It is a reasonable assumption when the thermal by-pass temperature is set close to the plant temperature, however, may cause appreciable errors if the temperature drop along the network is high.

The heat loss in the twin pipe was calculated according to the reference [7,9]

[1]

It is worth to be noted that though the design return temperature (22 oC) is higher than ground temperature, the net heat transfer in the return pipe may absorb heat from surrounding which makes Ur negative. However, negative Ur has to be set to zero as the simulation program cannot handle negative heat transfer coefficient.

[2]

RESULTS AND DISCUSSION Heat Exchanger Network simulation starts from proper selection of pipe dimension, based on the design condition and the design criteria introduced in the previous section. Table 2 shows the selected pipe types and corresponding length for three different cases. Case 1 is the reference case. Flexible twin pipe Alx 20 to 32 and steel twin pipe DN32 and DN 40 were selected. The third recirculation pipe was designed in case 2 based on the summer by-pass flow rate. Pressure gradient 1500 pa/m for street pipes and 500 pa/m for main pipes were set as the dimension criteria. Though smaller pipe was suggested by the program, the Alx16 single pipe was selected as the minimum diameter pipe available on the market. It was assumed that the recirculation pipe can be used as water supply in winter in case 3. Therefore, the main pipes in the reference line were re-designed with considering that a portion of supply water goes through the recirculation line. It can be seen that the supply pipe has smaller diameter than return pipe in some sections in the twin pipe line.

Fig. 6 District heating network

The supply and return pipe are assumed identical and placed horizontally in the same depth from the ground. The linear thermal transmittance Uij reduces to U11=U22=U1 and U12=U21=U2. In addition, the thermal conductivity of insulation foam was assumed constant. U1 and U2 were then calculated with the analytical solution developed from the multi-pole method [10]. The simulation program cannot handle two heat transfer coefficients in the same pipe, Us and Ur were derived to represent the overall heat transfer 76

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Case 2 has higher return pipe heat loss comparing with case 1 due to the introduction of recirculation line. At constant supply temperature 55 ºC, 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 º 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.

Table 2 Selected pipe types and length in Case 1–3

Supply water in the recirculation pipeline in winter increases the supply pipe heat loss in case 3. The concept of double pipe supply may not economical feasible, according to the simulation results. However, it may be used as an alternative solution to supply water in the 3rd pipeline under extreme whether condition, which otherwise has to raise the plant supply temperature to meet the increased heating demand. Furthermore, results in table 4 were limited to fixed recirculation pipe diameters. The double pipe supply concept may be economical feasible by free selection both reference pipe and recirculation pipe diameter with the objective to minimize the annual network operational cost or exergy consumption. This study is out of the scope of current paper due to the limitation of the simulation program.

Figure 7 shows the pressure profile along the critical route. The network is designed for a 10 bar system. The minimum network static pressure is 2 bar and the minimum differential pressure at consumer is 50 kPa. The plant static supply pressure is 853 kPa in case 1 at design condition. In case 3, the designed plant supply pressure head rise to 917 kPa, which is due to the increased flow rate indicated in Table 4. The pressure drop along the reference line during summer is quite low due to the reduced flow rate. However, extra pressure head has to be applied to overcome the pressure loss along the recirculation line in Case 2. The required static supply pressure is 800 kPa during summer as a result of small dimension recirculation line. Table 3 shows the simulation results for case 1. By-pass is required when the heating load is smaller than 1.53 kW. The return water temperature increases along with the increase of by-pass water flow rate. In summer, the amount of by-pass water flow rate exceeds the actual flow rate passing through the consumer, and the return temperature at the plant increases up to 35.5 ºC. The heat loss in the return pipe is accounted when the plant return temperature is raised to higher than 30 ºC. Simulation results for case 2 and case 3 are shown in Table 4. They were put in the same table as case 2 operates when there has by-pass requirement, while case 3 operates in the rest seasons. Italic is used for case 3 to distinguish the two cases. Thanks to the recirculation line, the return temperature at the plant in the reference line remains low at 22 ºC, while the return temperature in the recirculation line can reach 44 ºC in the summer, after deducting the single pipe heat loss. The low plant return temperature can help extract more power in the CHP plant or be used in other circumstance like direct flue gas condensation. On the other hand, high temperature return water in the recirculation pipe can be re-heated by an additional heat exchanger or boiler with minimum energy input.

Fig. 7 Pressure profile on the critical route in Case 1–3 77

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia Table 3 Simulation results in Case 1

Table 6 Simulation results in Case 1

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

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

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

Further Discussion on Heat Transfer As shown in Eq. 1â&#x20AC;&#x201C;3, the simulation program simplifies the calculation of the heat loss in the twin pipe as that in the single pipe. The influence of the adjacent pipe was accounted through converting the linear thermal transmittance Uij to the overall heat transfer coefficients Us and Ur , with pre-assumed constant network supply/return temperatures. To assess the influence of this simplification on the temperature predication, the thermal interaction between the supply and return pipes was calculated by solving the coupled pipe heat transfer differential equations. The governing equations for supply and return pipes can be expressed as:

Table 5 Selected pipe types and length in Case 1â&#x20AC;&#x201C;3

[4] [5] The boundary conditions can be expressed as: [6] The dimensionless temperature is introduced with: [7]

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 The boundary conditions change to : [9] The system linear ordinary differential equations can be solved with Eigen value method or with Laplace transformation. The Laplace transformation was applied in this study. Eq. 8 is transformed to: [10]

The by-pass water temperature in this study was set in a conservative way. In many practices, the by-pass water can be set 10 °C lower than the supply water temperature. Even lower by-pass temperature is proposed for the low temperature district heating network [3]. Table 9 shows the simulation results based on a10 °C temperature drop along the supply pipe. It shows the prediction errors increase in both supply and return pipes. The heat transfer was predicted in a reverse trend in the return pipe at 4 °C. Considerable prediction error was found in the return pipe at high ground temperature.

The final solutions are given as: [11] [12] Where : , ,

[13]

It is worth to be noted that the increase of supply temperature drop has more influence on the return pipe temperature prediction than that of supply pipe. The reason can be explained from the expression of Us and Ur in Eq. 1–2. As the magnitude of Ts-Tg is higher than Tr-Tg, the same amount of return water temperature variation will have more influence on Ur than Us, therefore causes a larger prediction error in the return pipe than in the supply pipe.

[14]

Tws- DN32, which is the longest main pipe in HE of case 1, is selected for the assessment with U1=0.141 and U2=0.0523. The pipe length is assumed 500 m. Ground temperature ranges from 0 to 15 oC. The inlet of supply and return temperatures are known as 55 oC and 22 oC respectively. The outlet temperature of supply pipe is controlled as 50 oC and 45 oC, respectively.

Table 9 Pipe temperature predication comparison (supply o outlet temperature is controlled at 45 C)

Table 8 shows the temperature prediction based on single pipe simplification and the coupled pipe equations. T_Difference represents the coupled solution minus the single pipe solution. When the temperature drop along the supply pipe is controlled at 5 oC, the prediction between the single pipe and the coupled pipe is very close. The prediction errors increase with increase the ground temperature. The single pipe approach predicts lower supply water temperature and higher return temperature than those of coupled pipe solutions. It was also observed that when the ground temperature is higher than 4 oC, the net heat transfer effect in the return pipe is to absorb heat to the surrounding.

CONCLUSION In this paper, a preliminary study was conducted on the influence of by-pass flow on the network return water temperature in a designed low temperature DH network. The concept of supply water recirculation was 79

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

introduced to avoid the mixing of by-pass water to the return water. Double pipe water supply concept was tested to use the recirculation pipe supply water during winter season. Two different house installation modes were considered in the analysis.

s = Supply u = Upstream d = Downstream Abbreviation DH = District heating

The by-pass water significantly increases the return water temperature in the traditional design. The mixed return temperature can reach 35.5 oC for HE and 45.6 oC for DHWS. With applying the by-pass water recirculation, this return temperature can be maintained at 22 oC, while the re-circulated by-pass water can be kept as high as 44 oC and 53.5 oC for HE and DHWS at the plant, respectively. It was found that the double pipe supply leads to the highest network heat loss. However, the conclusion that whether the concept of double pipe supply is inferior to other network design methods can only be drawn after further network thermal-economic optimization.

HE = Heat exchanger DHWS = Domestic hot water storage tank REFERENCE [1] H. Lund, B. Moller, B. V. Mathiesen, A. Dyrelund, ― The role of district heating in future renewable energy systems‖, Energy, 35, pp. 1381-1390, 2010. [2] Charlotte Reidhav, Sven Werner, ― Profitability of sparse district heating‖, Appliced Energy, 85, pp. 867-877.

The simulation program simplifies the twin pipe heat transfer prediction as a single pipe, and neglects the return pipe heat loss when the return pipe absorbs heat from the surroundings. The temperature prediction errors due to the single pipe assumption were analyzed through solving the coupled supply/return pipe differential energy equations. The prediction errors increase with increase the allowable temperature drop in the network. Considerable error was found for the return pipe at high ground temperature.

[3] ―Udvikling og Demonstration af Lavenergifjernvarme til Lavenergibyggeri‖, EFP 2007. [4] TERMIS Help Manual, 7-Technologies A/S.

Version

2.093,

[5] Dansk Standard DS 439, 2000. Norm for vandinstallationer, Code of Practice for domestic water supply installations, 3. udgave, www.ds.dk. [6] Otto Paulsen, Jianhua Fan, Simon Furbo, Jan Eric Thorsen, ―Consumer Unit for Low Energy District Heating Net‖, The 11th International Symposium on District Heating and Cooling, 2008, Iceland.

NOMENCLATURE cp = specific heat capacity [ J/kg.K]

[7] P. K. Olsen, et.al, ― A new low-temperature district heating system for low energy buildings‖, the 11th International Symposium on District Heating and Cooling, Iceland, 2008.

q = Heat transfer rate [kW / m] s = Laplace transform variable T = Temperature [ K] U = Overall heat transfer coefficient [ kW /m.K]

[8] Logstor. http://www.logstor.com/

Uij = Linear thermal transmittance [kW/m.K]

 = Dimensionless temperature

[9] Benny Bohm, Halldor Kristjansson, ―Single, twin and triple buried heating pipes: on potential savings in heat losses and costs‖, International Journal of Energy Research, 29, pp. 1301-1312, 2005.

Subscripts

[10] P.Walleten, ―Steady-state heat loss from insulated

= mass flow rate [ kg/s] Greek Letter

pipes‖, Thesis, Lund Institute of Technology, Sweden, 1991.

g = Undistributed ground r = Return

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

STEADY STATE HEAT LOSSES IN PRE-INSULATED PIPES FOR LOW-ENERGY DISTRICT HEATING 1

A. Dalla Rosa , H. Li , S. Svendsen 1

Technical University of Denmark insulation series to reduce heat losses and thus saving operational costs; however, this option would increase investment and installation costs. The design principles for DH networks could instead be changed towards the use of media pipes with small nominal diameters, with a higher permissible specific pressure drop. All-year around lower supply temperature and return temperature constitute an effective option to reduce heat losses [3]. These principles have a big potential for heat supply to low-energy buildings, as explained in [4] and they are investigated in this paper.

ABSTRACT The synergy between highly energy efficient buildings and low-energy district heating (DH) systems is a promising concept for the optimal integration of energy saving policies and energy supply systems based on renewable energy (RE). Distribution heat losses represent a key factor in the design of low-energy DH systems. Various design concepts are considered in this paper: flexible pre-insulated twin pipes with symmetrical or asymmetrical insulation, double pipes, triple pipes. These technologies are potentially energyefficient and cost-effective solutions for DH networks in low-heat density areas. We start with a review of theories and methods for steady-state heat loss calculation. Next, the article shows how detailed calculations with 2D-modeling of pipes can be carried out by means of computer software based on the finite element method (FEM). The model was validated by comparison with analytical results and data from the literature. We took into account the influence of the temperature-dependent conductivity coefficient of polyurethane (PUR) insulation foam, which enabled to achieve a high degree of detail. We also illustrated the influence of the soil temperature throughout the year. Finally, the article describes proposals for the optimal design of pipes for low-energy applications and presents methods for decreasing heat losses.

The total length of branch pipes can be significant in proportion to the total length of the network, above all in areas with a low-energy demand density. Moreover the temperatures in the critical service lines affect the temperature level in the whole network, so that the heat losses and the temperature decay in building connection pipes are decisive for the overall performance of the system. In this paper particular focus was given to branch pipes. State-of the art of district heating pipes At present time DH distribution and service lines are based either on the single pipe system, where the supply/return water flows in media pipes with their own insulation, or on the twin pipe system, where both pipes are placed in the same insulated casing, or in a mixture of them. All plastic pipe systems are characterized by having the water medium pipe made of plastic (crosslinked polyethylene (PEX) or polybutylene (PB)). They are covered by insulation, usually polyurethane foam, but in some cases of PEX foam or mineral wool; the outer cover is formed by a plastic jacket. Durability of plastic pipes is not a real issue, since it has been proved that the expected life of PB pipes and PEX pipes is, respectively, more than 40 years and approx. 100 years [5]. As consequence of even lower average operational temperature, longer lifetime can be predicted according to Annex A in [6]. Studies have indicated that cross-linked polyethylene (PEX) pipes have a cost advantage over steel pipes at pipe dimensions less than DN60, due to their greater flexibility since the joints do not require welding [7]. Alternative design concepts must be considered in branch pipes from street lines to consumersâ&#x20AC;&#x2DC; substations: a pair of single pipes, twin pipes or triple pipes. Traditionally most DH branch connections have been built with two single steel pipes: one supply pipe and one return pipe. Twin pipes can be made of steel,

INTRODUCTION The energy policy on energy conservation poses stringent requirements in the building energy sector, so that the entire DH industry must re-think the way district energy is produced and distributed to end-users [1, 2]. This is a requirement to be cost-effective in low heat density areas. Low-energy DH networks applied to lowenergy buildings represent a key technology to match the benefit of an environmentally friendly energy supply sector and the advantages of energy savings policy at the end-usersâ&#x20AC;&#x2DC; side. Future buildings with a high performance envelope will lead to reduced space heating load and therefore to a lower required distribution temperature for heating. The introduction of low-energy DH networks is an appropriate and natural solution to enhance energy and exergy efficiencies. Distribution heat losses represent a key-point for designing low-energy DH systems, due to the critical role they have in the economy of the system. The industry could meet the requirements of higher

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

demand, although a non perfect cooling of DH water occurs when tapping of DHW starts. The concept based on twin pipes and a substation with instantaneous production of DHW in a heat exchanger is an optimal solution, if certain conditions are respected. The first requirement is that the control method gives priority to DHW preparation over space heating; the second condition is that the space heating load during summer, to keep a high level of comfort in bathrooms for example, has to guarantee a sufficient cooling of the return water. As a result media pipes with inner diameters as small as 10 mm can be applied in the primary loop and the water return temperature can be kept sufficiently low, even in summer conditions.

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

The triple pipe system is applicable in three different operational modes. The first one (mode I) occurs in case of DHW demand, when pipe 1 and pipe 3 both act as water supply pipes; the second operational mode (mode II) is activated when an idle water flow is supplied by pipe 1 and pipe 3 acts as re-circulation line to the supply distribution line, while the return line (pipe 2) is not active: this is often the case when there is no demand for space heating, but a small amount of water circulates in the DHW heat exchanger, keeping the loop warm to satisfy the instantaneous preparation of DHW in the required time. This system avoids an undesirable heating of the water in the return distribution line. The third operational mode (mode III) occurs during the heating season when there is only demand for space heating and no tapping of DHW: pipe 1 and pipe 2 operate as a traditional supply-return system, while there is no water flow in pipe 3. The different modes are summarized as follows:

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



Operational mode I: DHW tapping, pipe 1, 2, 3 active.



Operational mode II: supply-to-supply re-circulation, pipe 1, 3 active; pipe 2 not active.



Operational mode III: space heating demand, pipe 1, 2 active; pipe 3 not active.

METHODS Theory of steady state heat loss in buried pipes In order to calculate steady-state heat losses in DH buried pipes there are analytical methods [8] and explicit solutions for the most common cases [9]. A complete review of the available literature about steady-state heat losses in district heating pipes has been carried out in [10]. Here the methods are presented with reference to the present status of the technology in the district heating sector. Furthermore key-points and critical aspects are discussed; finally, improvements in the methodology of how to calculate steady-state heat losses are proposed, with particular focus on low-temperature and medium-temperature

A simple and cost-effective configuration is composed of the control system and two heat exchangers for, respectively, space heating (SH) and domestic hot water (DHW). The main disadvantage of such type of substation unit is that only rather short lengths of service pipes can usually be applied; otherwise it would not be possible to assure the required DHW temperature at tapping points in the required time, due to the unsatisfactory transportation time. A modified unit is therefore proposed and it is equipped with a booster pump which assures quicker response to DHW 82

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

thermal coefficient, which is function of the temperature in this case. U-values are dependent both on temperature and time. If the time-dependency due to the ageing of the foam can be restrained by introducing effective diffusion barriers, that is not true for the intrinsic dependency on temperature. It is practice to evaluate the steady state heat loss applying a thermal conductivity value that corresponds to a hypothesized mean temperature of the insulation. Nevertheless we need models based, for example, on the finite element method (FEM) when complex geometries or a high degree of detail are requested.

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

Supply pipe:

(2)

Return pipe:

(3)

Temperature dependant thermal conductivity of PUR insulation foam In this paragraph the authors want to explain and demonstrate the importance of taking into account the temperature-dependency of the thermal conductivity of the insulation (lambda-value). The temperature gradient in the insulation foam in the radial direction is often higher then 10 °C/cm, meaning that the thermal conductivity of the material locally varies remarkably. In the example shown Figure 2, it varies more than 10% of the prescribed mean value. This affects the magnitude of the heat transfer. Considering a life cycle assessment of a DH system, the main impact to the environment is represented by heat losses [13]. The thermal conductivity of the insulation material in pre-insulated DH pipes is usually stated at a temperature of 50 °C. The lambda-coefficients were chosen according to the available data at the end of 2009; the lambda-value at 50 °C for straight pipes, axial continuous production was set to 0.024 W/(mK) and for flexible pipes to 0.023 W/(mK). Since April 2010 new results are available [14]. It is preferable to have a model that takes into account the temperaturedependency of the thermal conductivity of the insulation foam. The calculations in this paper use the following expression, if not differently stated. It derives from experimental data [15]:

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

(4)

Return pipe:

(5)

λ(T) = 0.0196734 + 8.0747308.10-5.T [W/(mK)]

(1)

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

Equations (4) and (5) show how the heat transfer from each pipe can be calculated by use of only one linear 83

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Combined heat and moisture transfer is disregarded. The material properties are homogeneous and phase changes, i.e. freezing and thawing were not considered. Table 1 lists the material properties, used as input values also for the following models; a sketch of the slab-model, where the boundary conditions are described, can be seen in Figure 3.

Temperature field in the soil around the pipe In this paper we address the question of how to create a simple yet detailed FEM model for steady state heat loss calculations. The overall heat transfer resistance between the DH water and the environment is mainly composed of the thermal resistance of the insulation and the thermal resistance of the soil; compared to these two factors, the thermal resistances of the pipe wall and the convective resistance at the surface waterpipe are in practice negligible. The insulation foam always offers the greatest share in the overall insulation effect. The contribution of the soil is smaller on small-sized pipes than on large-sized pipes. The share is smaller in Insulation Series 2 and 3 [3]. The heat conductivity coefficient of the soil is the main parameter affecting the thermal resistance of the soil itself, and its value is often unknown in practice.In the calculations we chose a value of 1.6 W/(m.K). The soil temperature influences heat losses from DH pipes. The soil layer around the heating pipes slightly warms up around the pipes. The evaluation of the temperature field in the soil is a prerequisite to create a realistic model for calculations of heat losses. Finite Element Method (FEM) simulations were carried out and temperature conditions in the soil around a typical DH service pipe, suitable for low-temperature applications were evaluated over a 10-year period.

Figure 3: Sketch of the model. Dimensions are in [mm].

FEM model A rectangle representing a semi-infinite soil domain (width: 10–20 m, height: 20–40 m) is the most used geometry to model the ground in heat loss calculations [18, 19]. In this paper a finite, circular soil domain was applied, instead. Its diameter is 0.5 m and it is equal to the distance between the surface and the centre of the casing pipe. Calculations show that the introduced simplification hardly affects the accuracy of the results. The mesh model and an example of the temperature field in a small size twin pipe are shown in Figure 4.

Table 1: Thermal properties of materials. λ

[W/(m∙K)]

[kg/m³]

[J/(kg∙K)]

λsoil

1.6

ρsoil

1600

Cp_soil

2000

λPE

0.43

ρPE

940

Cp_PE

1800

λPUR

0.023

ρPU

Cp_PUR

1500

λPEX

0.38

ρPE

938

Cp_PEX

550

λSteel

ρSte

λCu

400

ρCu

8930 8930

Cp_Stee l

Cp_Cu

480 385

The simulation calculated the soil temperature at various x-coordinates from a commercial branch pipe. The selected pipe was the Aluflex twin pipe 16-16/110. Temperatures were set at 55 °C and 25 °C, respectively for the supply pipe and the return pipe. The heat transfer coefficient at the ground surface was assumed to be 14.6 W/(m2K), including convection and radiation [16]; we set the outdoor air temperature during the year according to the harmonic function valid for the Danish climate [17]:

M  Tair  8.0  8.5  sin 2   12  

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

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

(6)

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 twin pipe (DN 20 and DN 80) and by [10]. For twin pipes of even smaller size, such as in branch connections, the heat losses occurring in case of vertical layout are only slightly more favorable than the losses occurring on horizontally arranged pipes; this result is shown with an example in the results section.

FEM model: geometry of the ground and of the pipes We considered the geometric model of the preinsulated Aluflex twin pipe type 16-16/110; the temperatures of supply/return/ground are 55/25/8 °C. We calculated the heat losses for vertical or horizontal placement of the media pipes inside the casing, which was embedded in a rectangular or a circular model of the ground. The same calculations were repeated for other twin pipe size, up to DN 32 and other medium pipe materials, i.e. steel and copper. The results confirm that the vertical placement of the media pipes inside the insulation barely affect the heat transfer, being the difference between the two configuration less than 2% for the considered cases.

RESULTS AND DISCUSSION In this section we discuss the influence of the soil temperature on heat losses; next, we present the validation of the FEM models; finally we apply the method to show the potential for energy saving in the case of asymmetrical insulation of twin pipes, in the case of double pipes and triple pipes.

Table 2: Heat loss for various placements of the media pipes and various model of the ground.

Temperature field in the soil Temperature conditions in the soil around a typical twin pipe, type Aluflex 16–16/110, were evaluated over a 10-year period. Figure 5 shows the all-year temperature profiles of the outdoor air and of the ground at depth equal to 0.5 m, at three horizontal distances from the centre of the casing, during the first year of operation. No notable differences in the yearly profile were noticed in longer periods of time.

Ground Media pipes model layout

Heat loss Heat loss Heat loss supply return total [W/m] [W/m] [W/m]

Vert.

3.79

-0.17

3.62

Horiz.

3.80

-0.18

3.62

Vert.

3.84

-0.18

3.66

A: Semi-infinite, rectangular (width x depth: 40 m x 20 m) B: Finite, circular (diameter: 0.5 m)

We found that in state-of-the-art well insulated twin pipes (series 2 or 3) a certain amount of soil is slightly heated up by the warm twin pipe; nevertheless the level of such heating can be neglected because its effect is not noticeable in comparison to the fact that the uncertainties about the thermal properties of the soil usually have a bigger impact. Considering yearly average temperatures, the magnitude of the soil heating is about 1 °C for distances of around 0.2-0.3 m from the centre of the casing, and less than 0.5 °C by 0.5 m. The temperature raise is considered in comparison to the undisturbed temperature of the ground at a distance of 10 m.

Steady-state heat loss in commercial pipes The model was validated by comparing the results from FEM simulation to the analytical calculation for preinsulated pipes embedded in the ground [14]. Calculations were carried out for four different sizes of Aluflex twin pipes (size 14–14, 16–16, 20–20, 26–26) and for chosen sets of supply (50, 55, 60 °C), return (20, 25, 30 °C) and ground (8 °C) temperatures. The selected pipes are suitable to be used as branch pipes in low-energy demand areas. There is a good accordance between the two methods, the deviation being lower than 1%. Figure 6 gathers the values of total heat loss for the Aluflex twin pipe category; four different approaches are reported. The term ―standard‖ is used when the effect of the temperature on the thermal properties of the insulation is neglected and the thermal conductivity of the PUR foam is thus constant. This is in accordance with [21]. The term ―advanced‖ is used when the calculation method takes into account that the thermal conductivity of the insulation depends on the temperature. Based on the temperatures calculated for a number of points in the insulation the program calculates an average temperature for the material; the lambda-value of the insulation is then calculated as a function of such temperature. An average temperature of the ground is similarly calculated. The calculation is repeated until the mean temperature difference for the insulation material, pipe

Figure 5: All-year temperature profiles of the outdoor air and of the ground at depth equal to 0.5 m and 3 horizontal distances from the centre of the casing.

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 consecutive calculations. The‖standard‖ and ―advanced‖ model are available online [14]. In the ―FEM advanced‖ model we directly implemented equation (1) in the insulation domain, instead. The results indicate that in case of low-temperature operation, lower total heat losses are calculated if the temperature-dependency of the insulation lambdavalue is taken into account. Moreover the heat transfer between the pipes in twin or triple pipes can be properly evaluated.

(Aluflex: ≤ DN 26, steel: ≤ DN 50) the best design is to put the supply pipe in the centre of the casing, assuring the best possible insulation for the supply pipe. This strategy guarantees also the lowest temperature drop in the supply side, which is a critical figure in low-temperature applications. For bigger sizes (Aluflex: ≥ DN 26, steel: ≥ DN 50) the best design is achieved by ―moving up‖ the media pipe layout and at the same time by keeping the same distance between the media pipes as in the symmetrical case.

Total Heat Loss [W/m]

7.0

Double pipes

6.0

A double pipe consists of a pair of media pipes of dissimilar size, co-insulated in the same casing. It is a further development of the twin pipe concept. A sketch of a possible application of the double pipe concept is shown in Figure 7. Though these measures, network heat loss reduction is possible, in case of operation during low heating load periods.

5.0 4.0 3.0

2.0 1.0 0.0

Standard FEM Standard Advanced FEM Advanced

DN 14 3.34 3.19

DN 16 3.61 3.68 2.86 3.51

DN 20 4.24 4.33 3.36 4.14

DN 26 4.62 4.80 3.69 4.59

DN 32 5.71 6.02 4.55 5.75

DN 40 6.45 6.76 5.10 6.47

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

Asymmetrical insulation in twin pipes The results show that improvements are possible, thanks to asymmetrical insulation (see Table 3). We proved that a better design leads to lower heat losses from the supply pipe (leading to a lower temperature drop); next, the heat loss from the return pipe can be close to zero, maintaining isothermal conditions in the return line. If commercial available casing sizes are kept, we suggest two design strategies, depending on the size of the pipes. For small pipe sizes

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

The space heating demand in summer is diminished, except for the energy requirement in bath room heating. According to the energy balance, the reduced heating load requires less Table 3: Comparison between asymmetrical and symmetrical insulation in twin pipes. network flow rate as far as the The centre of the casing is the origin of the Cartesian system. designed building temperature drop is sustained. However, the Coordinates Heat loss asymm.reduction of network flow rate (x; y) [mm] [W/m] symm. [%] will increase the supply water Size Mat. Sup. Ret. Sup. Ret. Tot. Sup. Tot. temperature drop along the (DN) pipeline due to heat loss. As a 14 (0; 0) (0; 27) 3.24 0.01 3.25 -7.6 2.0 consequence, the supply 16 (0; 0) (0; 28) 3.56 -0.01 3.55 -5.1 1.1 temperature at the end user 20 (0; 0) (0; 30) 4.16 -0.04 4.12 -4.2 -0.3 may lower down below the Alx. minimum requirement. This 26 (0; 0) (0; 36) 4.67 0.00 4.67 -5.1 1.9 problem is relevant to towenergy DH systems with an 32 (0; -16) (0; 28) 5.54 0.00 5.54 -5.8 -2.5 already low supply temperature. 50 (0; -25) (0; 55) 5.69 -0.03 5.66 -7.7 -2.4 This design is based on the fact Steel that the supply line acts also as 65 (0; -36) (0; 60) 6.70 -0.02 6.68 -7.8 -3.2 re-circulation line during low 86

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

heating load periods; hence by-pass at the critical consumers are not necessary and the exergy loss due to the mixing of warm water into the return line is avoided. Furthermore the water flow in the return line has the same direction as in the supply line (clockwise in the example), so that the smallest size for the return pipes are expected in correspondence to the biggest size for the supply size, and vice versa. This results in lower local pressure differences between supply and return lines and savings in operational costs, thanks to lower heat losses. This is shown in Table 4 and Table 5, by means of two examples: the first one refers to a small to medium-size distribution network, the second one to a bigger one, being capable to supply four times more energy than the previous one.

Triple branch pipes The development of an optimized triple pipe solution for low-energy applications is reported to show the potentiality of utilizing detailed models for steady-state heat loss calculation. In this survey focus was given on the choice of media pipes diameters as small as possible. The triple pipe geometry is based on modifications of the 14-14/110 (outer diameters in [mm] of respectively supply pipe, return pipe, casing) twin pipe design which has been reported in [18]. Four geometrical variations have been considered (see Figure 8) and the Cartesian coordinates describing the placement of media pipes inside the casing are listed in Table 6.

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 째C. Heat loss [W/m] Size (DN)

Sup.

Ret.

Tot.

40-40

-6.24

0.04

-6.20

80-80

-7.66

0.07

-7.59

40-80

-5.55

0.05

-5.58

80-40

-7.41

0.05

-7.36

Total (system)

[%]

Twin: -13.79 6.1 Double: -12.94 Figure 8: four different geometries for a triple service pipe type Aluflex 14-14/110.

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

Table 6: placement of media pipes inside the casing for four triple pipe geometries, type Aluflex 14-14-20/110. Coordinates (x, y) [mm]

Heat loss [W/m] Size (DN) 100-100

Sup.

Ret.

Tot.

-7.83

-0.55

-8.39

200-200

-8.92

0.24

-8.68

100-200

-6.4

0.08

-6.36

200-100

-8.07

-0.03

-8.69

Total (system) Twin: -17.06

[%]

11.8

Double: -15.05

Variation

Pipe 1 (Sup.)

Pipe 2 (Ret.)

Pipe 3 (Sup. or re-circ.)

A B C D

(14;-14) (10;-14) (3;-14) (0; 0)

(0;20.5) (0;20.5) (0;20.5) (0;25)

(-14;-14) (-21;-7) (-21;-7) (0;-28)

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

We considered an optimal placement of the media pipes in case of double pipes, thus asymmetrical insulation is applied. The total amount of insulation is used both in the twin pipe-based distribution network and in the double pipe-based one, so that the investment costs are equal in both cases. Results show that the heat loss can be reduced by 6% by means of double pipes instead of twin pipes for the low to medium-size distribution network. Even higher energy savings (around 12%) are possible in the case of the large-size distribution network.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

mode II is the most likely outside heating season, the conclusion is that geometry D is preferable.

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

Mode

I (DHW tapping)

II (supply-tosupply recirculation)

III (space heating)

Geo m.

CONCLUSIONS

Heat loss [W/m] Pipe 1

Pipe 2

Pipe 3

Tot.

2.67

-0.08

2.67

5.30

2.91

-0.29

2.75

5.38

2.52

-0.22

2.74

5.06

2.46

0.05

2.74

5.24

2.67

5.34

2.69

2.85

5.55

2.48

2.70

5.18

2.49

2.75

5.25

3.46

0.48

3.95

3.39

0.43

3.83

3.41

0.35

3.76

3.53

-0.01

3.53

The soil temperature at 0.5 m below the surface varies between 2 °C in January-February and 14 °C in July–August, for Danish conditions. This knowledge can be used to better predict the winter peak load and the temperature drop in the distribution line during summer. The slab-model for steady state heat loss calculations can be replaced, in case of small size distribution/service pipes, by a model where the effect of the soil is represented by a circular soil layer around the district heating pipe. The results confirm that the vertical placement of twin media pipes inside the insulation barely affects the heat transfer, in comparison to the horizontal placement; the difference between the two configurations is less than 2% for the considered cases. We proposed a FEM model that takes into account the temperature-dependency of the thermal conductivity of the insulation foam; in this way we enhanced the accuracy of the heat transfer calculation among pipes embedded in the same insulation.

Table 8: Steady state heat losses of triple pipes type Aluflex 14/14/110 for 4 geometries and operational mode II. Temperature supply/recirculation/ return/ ground: 55/55/25/14 °C. Heat loss [W/m] Geom. Pipe Pipe Pipe Tot. 1 2 3 II (supply-tosupply recirculation)

2.35

4.70

2.37

2.51

4.88

2.39

2.63

5.02

2.20

2.42

4.62

We conclude that an absolute best design for the service triple pipe does not exist, but it depends on the operational mode that is chosen as critical. In fact the results reported in Table 7 and Table 8 show that geometry C gives the lowest total heat loss for operational modes I and II, while geometry D has the best thermal performance for operational mode III and for operational mode II, if a temperature of the soil of 14 °C is considered. It has to be underlined that, considering the operational mode III, geometry D shows no heating of return water; this is a situation always desirable, although it has a slightly higher heat loss from the supply pipe than the other geometries. It is proved that usually operational mode I occurs for less than 1 h/day [20]. Moreover the temperature drop in the supply pipe to the DHW heat exchanger is critical in low-temperature applications, so that it is strongly recommended to minimize the heat loss from this media pipe. Considering all this and the fact that mode III is the most likely during the heating season and

We applied the model to propose optimized design of twin pipes with asymmetrical insulation, double pipes and triple pipes. We proved that the asymmetrical insulation of twin pipes leads to lower heat loss from the supply pipe (from -4% to -8%), leading to a lower temperature drop; next the heat loss from the return pipe can be close to zero. It is possible to cut the heat losses by 6–12% if an optimal design of double pipes is used instead of traditional twin pipes, without increasing the investment costs. The development of an optimized triple pipe solution was also reported. It is suitable for low-energy applications with substations equipped with heat exchanger for instantaneous production of DHW. REFERENCES [1] S. Froning, ―Low energy communities with district heating and cooling‖, 25th Conference on Passive and Low Energy Architecture, Dublin (2008). [2] S. F. Nilsson et al., ―Sparse district heating in Sweden‖, Applied Energy 85 (2008), pp. 555–564. [3] F. Schmitt, H.W. Hoffman, T. Gohler, Strategies to manage heat losses – technique and economy, IEA-DHC ANNEX VII, (2005). [4] P.K. Olsen, B. Bøhm, S. Svendsen et al., ―A new-low-temperature district heating system for 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 on district heating and cooling, Reykjavik (2008). [5] M. Klompsch, H. Zinko, Plastic pipe systems for DH, handbook for safe and economic application, IEA-DHC ANNEX V (1999). [6] DS-EN 253:2009, District heating pipes – Preinsulated bonded pipe systems for directly buried hot water networks - Pipe assembly of steel service pipe, polyurethane thermal insulation and outer casing of polyethylene. [7] H. Zinko, GRUDIS-tekniken för värmegles fjärrvärme (The GRUDIS technology for low heat density district heating), Swedish District Heating Association, Stockholm (2004). [8] J. Claesson, J. Bennet, Multipole method to compute the conductive heat flows to and between pipes in a cylinder. Department of Building Technology and Mathematical Physics, Lund (1987). [9] P. Wallenten, Steady-state heat loss from insulated pipes, Lund (1991). [10] B. Bøhm, ―On transient heat losses from buried district heating pipes‖, International Journal of Energy Research, 2000, Vol. 24, pp. 1311-1334. [11] Terminology of HVAC, ASHRAE, Atlanta (1991). [12] I.B. Kilkis, ―Technical issues in low to mediumtemperature district heating‖, International Journal of Global Energy Issues, 2002, Vol. 17, pp. 113-129.

[13] J. Korsman, S. de Boer and I. Smits, ―Cost benefits and long term behavior of a new all plastic piping system‖, DHC ANNEX VIII (2008). [14] www.logstor.com (March 2010). [15] Udvikling og demonstration af lavenergifjernvarme til lavenergibyggeri (development and demonstration of low energy district heating for low energy buildings), 2007. [16] B. Kvisgaard, S. Hadvig, Varmetab fra fjernvarmeledninger (Heat loss from pipelines in district heating systems), Copenhagen (1980). [17] DS418:2002, Calculation of heat loss from buildings. [18] H. Kristjansson, F. Bruus, B. Bøhm et al., Fjernvarmeforsyning af lavenergiområder (District heating supply of low heat density areas), 2004. [19] T. Persson, J. Wollerstrand, ―Calculation of heat flow from buried pipes using a time dependent finite element model‖, 45th International Conference of Scandinavian Simulation Society, Copenhagen (2004). [20] B. Bøhm, H. Kristjansson, ―Single, twin and triple buried heating pipes. On potential savings in heat losses and costs‖, International Journal of Energy Research (2005), Vol. 29, pp.1301-1312. [21] EN 13941:2003, Design and installation of preinsulated bonded pipe systems for district heating.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

TRANSIENT THERMAL CONDUCTIVITY OF FLEXIBLE DISTRICT HEATING TWIN PIPES C. Reidhav and J. Claesson Department of Civil and Environmental Engineering, Division of Building Technology, Chalmers University of Technology, Göteborg, Sweden. The standardised method used for determining the steady-state thermal conductivity of district heating pipes, the guarded hot pipe method, is only applicable on straight single pipes. The method is based on [1], described in [2] and [3]. A heater pipe is placed inside the service pipe and the heat transferred through the insulation is measured. The measurements must be conducted with a constant distance between the heater pipe and the service pipe along the test specimen which can not be achieved with flexible pipes. An alternative method was presented in [4] and applied in [5] where the thermal conductivity of flexible district heating single and twin pipes can be determined. The temperature decline of hot water pumped in a flexible pipe coil is measured. A long pipe coil is needed to have a sufficient temperature decline along the pipe. The Danish method is based on steady-state measurements at different temperatures to get the temperature dependence of the decline. The Finite Element Method is used to determine the thermal conductivity λ(T) of twin pipes.

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

A transient method to determine the temperaturedependent thermal conductivity of flexible single district heating pipes was presented in [6]. A pipe coil is immersed into cold water and the temperature decline of hot water inside the coil is measured. The measured temperatures are compared to numerically calculated values to characterize λ(T). A Kirchoff transform is used to simplify the calculations. Finally, the mean square difference of the measured and calculated temperatures are minimized which gives λ(T). In this paper, a similar experimental set-up is used for a flexible twin pipe. The numerical and mathematical model developed in [6] cannot be used for twin pipes due to the complicated geometry of twin pipes. In [7] and [8] a method was presented where heat losses from district heating twin pipes were calculated with conformal coordinates describing the twin pipe geometry. In this paper, the conformal coordinate model is used to calculate the temperature decline in a flexible twin pipe. The calculated temperatures are compared to experimentally measured temperatures. This gives the temperature-dependent thermal conductivity of semi-flexible polyurethane foam of the studied flexible twin pipe.

Fig. 1 Cross-section of single (left) and twin (right) district heating pipes 90

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The twin pipe studied in this paper has two copper service pipes, semi-flexible polyurethane foam and a slightly corrugated LDPE casing. The pipe is of dimension DN 20 with the pipe dimensions described in Table 1. The pipe producer declares a thermal conductivity of the semi-flexible polyurethane foam of λ50 = 0.0255 W·m-1·K-1of a newly produced pipe of this kind. The pipe has no diffusion barrier. The density of the polyurethane foam was ρ = 60 kg/m3.

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

Table 1. Dimensions of the twin pipe studied Casing outer diameter (mm)

Casing thickness (mm)

2.2

Service pipe outer diameter (mm)

Service pipe thickness (mm)

1.0

The initial coil temperature was T0 = 81.3ºC. The water temperature at the service pipe (Tw,meas(t)) decreases during the 16 hours of measurements. The pool temperature was initially T1 = 17.4 ºC and increased slightly to T1 =17.9 ºC during the 16 hours. The measured coil and pool temperatures are showed in Fig. 3. A sawtooth disturbance of about T = 0.07 ºC and a small noice of about T = 0.015ºC can be seen in the pool water measurements in Fig. 4. A detailed study of the coil temperatures Tw,meas(t) shows that the temperature of some thermocouples decreases abrupt occurred at about t = 5.5 hours. The marked chosen interval in Fig. 3 and Fig.4 is chosen to minimize the errors.

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

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Chosen interval 70

Temperature (C)

Tw, meas

T1, meas 10

0 0

Time (h)

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

18,0

17,9

Chosen interval

Temperature (C)

17,8

17,7

17,6

17,5

17,4 0

Time (h) Fig. 4 Measured pool temperature T1, meas

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

water and the right-hand boundary against the poll water. The heat flux in the vertical v-direction is zero on the horizontal boundaries due to symmetry.

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

 c

T  T  T  ( (T )  )  ( (T )  ) t x x y y

Fig. 6 Initial temperature distribution in the crosssection of a pipe quarter in the u, v-plane. (1) In the numerical solution, the region is divided into a rectangular mesh. The area factor is now the area of each of the cells shown in Fig.5. They are shown in Fig. 7. The largest cell is the one in the lower left corner in Fig.5 near the stagnation point (usp). The areas are used to calculate the heat capacity of each cell in the uv-plane.

The (x,y)-coordinates ( z  x  i  y) are transformed to suitable conformal coordinates (w  u  i  v) with the aid of line sources and so called multipoles.

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

The initial steady-state condition for a twin pipe with coil water temperature Tw = 81.3ºC immersed into pool water at T0=19.7ºC is showed in u-v coordinates in Figure 6. Then, the temperature decline of stagnant water in the twin pipes are calculated

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

The density ρ and the heat capacity c of the polyurethane foam are assumed constant in the temperature interval studied. The boundary temperatures at the casing are given by the pool temperature.

The heat equation in the conformal coordinates is:

  c  A(u, v) 

T  T  T  ( (T )  )  ( (T )  ) t u u v v

(2)

The thermal conductivity λ(T) of the polyurethane foam is determined by the thermal conductivity at 50ºC λ50 (W/m·K) and a coefficient λ‟ to account for a linear temperature dependence.

Here, A(u,v) is the area factor in the conformal transformation. The considered region shown in Fig. 5 is transformed to a rectangular region in the u, v-plane, see Fig. 6. In the figure, the left-hand boundary lies against the coil

 (T )  50  1   ' (T  T50 ) 93

(3)

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The difference between the calculated and measured temperatures for the optimal parameter values of λ50, λ‟ and c are showed in Figure 9. The saw tooth disturbance from the measurements is seen. The difference giving the best fit lies in the interval -0.20 to 0.25 (ºC). The error is small.

EVALUATION OF MEASUREMENTS The temperature-dependent thermal conductivity of the polyurethane foam is obtained by calculating the temperature decline of the coil water. Certain values of the thermal conductivity of the polyurethane are chosen, λ50 and λ‟. The actual λ(T) are obtained by minimizing the difference between the measured and calculated coil temperatures, (4). The heat capacity c (J·kg-1·K-1) of the polyurethane foam is input to the calculations. Literature references for the heat capacity of polyurethane foam varies, 1300 J·kg-1·K-1 at 50ºC in [9], 1400 J·kg-1·K-1 in [10], 14001500 J·kg-1·K-1 for rigid polyurethane foam in [11]. The densities and heat capacities of water, service pipe and insulation were assumed to be constant in the temperature interval studied. The optimal parameter values of λ50 λ‟ and c were obtained by minimizing the difference D (ºC) between the calculated Tw (ºC) and measured coil temperatures Tw,meas (ºC).

D  max Tw, calc (t )  Tw, meas (t )

for t1  t  t2

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

(4)

A certain time interval, 0.5<t<5.5 h, was chosen for the optimization of the parameters λ‟ λ50 and c. Outside this interval, the optimization was unstable. This is mainly due to the disturbances in the measurements probably caused by penetration of pool water at the position of three thermocouples. The precision decreases also with time due to the decreased difference between the coil temperature and pool temperature. The problems at the start can be an effect of not having steady-state conditions before starting the temperature decline.

The optimal thermal conductivities λ50 and λ‟ give the final result for λ(T), as shown in Fig.10. 0,031

λ(T) (W/mK)

0,03

The maximal values of D for the interval studied are compared for different combinations of the parameters. A single optimum point is obtained with the lowest value of D as illustrated in Fig. 8.

D = 0.23

D = 0.43

λ50 = 0.43

o D = 0.69 λ' = 0.0036

c = 1500

DISCUSSION AND CONCLUSIONS In this paper, the twin pipe method is evaluated only from the experiences of one experiment. However, this experiment is part of a greater context with experiences from previous experiments on single pipes.

D= 0.28

c = 900

Fig. 10 Estimated thermal conductivity of the studied semi-flexible polyurethane foam

o D = 0.60

D= 0.43

Temperature (ºC)

λ50 = 0.029

λ' = 0.0072

0,027

0,025

λ50 (W/m·K) o

0,028

0,026

λ' (W/m·K)

D = 0.55

0,029

c (J/kg·K)

λ50 = 0.0285 W/m·K λ’ = 0.0054 W/m·K c = 1200 J/kg·K

The experimental procedure can be improved. In this experiment, the reattachment of the polyurethane foam at the positions of the thermocouples was probably insufficient. Penetration of pool water lead to a certain change of slope at about t = 5.5 h. The problem with penetration could have been solved by measuring the coil temperature inside the copper pipe instead.

Fig. 8 The maximum difference D for different combinations of λ50, λ‟ and c in the chosen interval.The optimal parameter values giving the lowest D are showed in the box. 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] is seen in this measurement as well and should be further investigated.

[5] District Heating Association (2008), Heat plan Denmark, Ramboll Danmark A/S and Aalborg University, (2008), In Danish, available at Dansk Fjernvarmes F&U-Konto, www.danskfjernvarme.dk

A large difference between the pool and coil temperature is desirable to minimize the relative errors.

[6] C. Reidhav and J. Claesson, A transient method to determine temperature-dependent thermal conductivity of polyurethane foam in district heating pipes, Building Physics 2008 - 8th Nordic Symposium, Copenhagen, Denmark, (2008)

It is also important to assure that steady-state conditions are established before starting the temperature decline. The final result, the obtained thermal conductivity:

 (T )  0.0235  10 105  T (ºC),

50  0.0285

[7] C. Persson and J. Claesson, Prediction of heat losses from district heating twin pipes, The 11th International Symposium on District Heating and Cooling, August 31 to September 2, Reykjavik, Iceland, (2008)

(5)

is in reasonable agreement with the declared λ50 = 0.0255 W·m-1·K-1 for a newly manufactured pipe. This pipe piece had been in store for some time and had no diffusion barrier. The temperature-dependent part of the thermal conductivity is in well agreement with [12].

[8] C. Persson and J. Claesson, Numerical solution of diffusion problems using conformal coordinates. Application to district heating pipes, Report Department of Civil and Environmental Engineering, Chalmers University of Technology, Göteborg, Sweden (2008)

REFERENCES [1] U. Jarfelt, Test apparatus of pipe insulation. Doctoral thesis. Chalmers University of Technology, Göteborg (1994)

[9] S. Peng, P. Jackson, V. Sendijarevic, K.C. Frisch, G.A Prentice, A. Fuchs, Process Monitoring and Kinetics of Rigid Poly(urethane-isocyanurate) Foams, Journal of Applied Polymer Science, (2000) Vol 77, 374-380

[2] European standard EN 253:2009, District heating pipes - Preinsulated bonded systems for directly buried hot water networks – Pipe assembly of steel service pipe, polyurethane thermal insulation and outer casing of polyethylene, Brussels, Belgium. (2009)

[10] R. Zevenhoven, Treatment and disposal of polyurethane wastes: options for recovery and recycling, Helsinki University of Technology, Report TKK-ENY-19, Espoo, Finland, June (2004).

[3] European committee for standardization. European standard EN ISO 8497:1996, Thermal insulationDetermination of steady-state thermal transmission properties of thermal insulation for circular pipes, Brussels, Belgium. (1996)

[11] BING, Federation of European Rigid Polyurethane Foam Associations, Thermal insulation materials made of rigid polyurethane foam (PUR/PIR),Report No1 October (2006)

[4] Danish District Heating Association. Development of an experimental set-up for measuring the heat conduction properties of flexible pipes, Project nr. 2006-05, Århus, Danmark. (2006), In Danish, available at Dansk Fjernvarmes F&U-Konto, www.danskfjernvarme.dk

[12] U. Jarfelt and O. Ramnäs, Thermal conductivity of polyurethane foam – best performance, th 10 International Symposium on District Heating and Cooling, Sept 3-5, Hanover, Germany, (2006) .

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

DISTRICT HEATING PIPES 200 MM BELOW SURFACE IN A STREET WITH HEAVY TRAFFIC 1

Anders Fransson and Sven-Erik Sällberg

Göteborg Energi AB, Sweden Building Technology and Mechanics, SP Technical Research Institute of Sweden sales (i.e. less district heating, d.h. to be sold). To connect new district heating customers in the future, with the competition of other heating suppliers, it is not enough to use just smaller pipes because of the smaller demands. Building the grid and maintaining the grid needs to become more cost efficient.

ABSTRACT This article reports the results from a field experiment initiated by Göteborg Energi AB with an extreme shallow burial of district heating pipes 162/76,1 (DN 65) casaflex under a street with heavy traffic designed for an average of 2000–4000 passes of vehicles a day and line. The pipes were laid only 200 mm below the surface. The backfill was of 0–40 mm particle size.

The purpose of this article is to inspire and if possible help whoever is interested in making district heating and cooling in the world more cost effective using the ideas or test results from this article.

Several consecutive measurements were done to study the effects from instant and long term loads from the traffic. The tests were done on a test pipe prepared with displacement gauges and on operating pipes.

1.2 Cost-cutting due to shallow burial in roads When reducing costs, it is important to maintain the qualities that are required. The road owner needs the road to be functional and has its standards. The district heating supplier is responsible for its pipes and deliveries of heat and has its standards. Finally there are workers (contractors and maintaining staff) who need acceptable working conditions.

The aim is that the results will inspire and give input for making district heating and cooling more cost effective. The tests showed that both the instant and long turn deformation of the pipes are small at the actual laying depth and also that the acceleration in the ground as heavy vehicles passes does not seem to be alarming. The conclusion is that shallow burial is technically possible if the road and backfill is done properly. 1. INTRODUCTION 1.1 New conditions for district heating The branch of district heating is in need of a new generation of district heating pipes. The conditions for selling district heating are slowly changing due to new legislation, harder competition, new technique and climate changes. Since 2003 Göteborg Energi AB is connecting more and more customers but is selling less and less energy. New legislation from 2006 allowed new buildings in Gothenburg to use a maximum of 110 kWh/m2 externally supplied energy for heating (,cooling) and producing domestic hot water. Today the municipality of Gothenburg wants new buildings to use 60 kWh/m2 at most.

Fig. 1 Standard shaft section

In a standard shaft section the drainage may be taken away in roads. A properly built road has a hard top and is drained as it is. You do not need to drain it any more. It is also possible to make the shaft more narrow and maintain acceptable working conditions if either long pipes with no joints are used or if the joints are welded on top of the shaft.

These changes are not unique. New houses are using less and less energy per square meter. There are already households that are not using but producing energy. The former energy suppliers in Europe are finding themselves not as suppliers but distributors, buying and selling energy. Climate changes are global and have already measurable effects on district heating

Less coverage is also an alternative. Earlier studies [1]–[3] shows that the pipes are solid enough to be placed with very little coverage (180 mm) and in rough materials. It is also shown that there is less settling in 96

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

Fig. 4 Marking of the location of the gauges.

The test shaft Three displacement gauges were placed in a test pipe containing air (see section 2.1) beside two operating pipes (see Fig. 3–5). Two accelerometers were placed in separate boxes near the test pipe. The gauges were monitored through wires at the bicycle path beside the street.

Fig. 2 Left; Supporting leg with no cover, Right; Rock curb.

If the existing fraction is used as backfill transports can be reduced which lower the costs and the environmental influence. Normally the district heating pipes have no problems coping with the traffic load. The extra pressure and the movement in the soil are making extra loads that are quite negligible compared to the thermal load, the inner pressure load and the load from the outer pressure from the soil.

The test was done with two single district heating pipes 162/76,1 (DN 65) casaflex buried in a fraction with grain size 0–40 mm. (Normal standard is a fraction of sand 0.2–16 mm.) The distance from the top of the pipes to the top of the fraction was 60 mm. (Normal standard is 460 mm.) The distance from the top of the pipes to the top of the asphalt was 200 mm. (Normal standard is 600 mm.)

To get an idea of if the graphite gaskets used in the pipe joints (for casaflex) can stand the traffic load 200 mm below the surface in a street with heavy traffic the traffic load was empirically measured in the test area.

As extra protection, the operating pipes were wrapped in a grid of polyethylene, PE.

EXPERIMENTAL The tests were done during 2009–2010. The test site An industrial street classified as a street with 2000–4000 passes of vehicles for every lane and day as an average through the year was chosen as the test site. The extension of the test area was 8 meters as the pipes crossed a street. The gauges were placed in one of the lanes close to the centre of the street. The location of the gauges were visualised with a cross on the asphalt (see Fig. 4).

Fig. 5 Shaft section in the test area including the test pipe.

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

Fig. 3 Drawing over the test area. 97

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

length 140 m of a single pipe means that there is no need to work with the pipe in the shaft).

100 100

Indicator 1

Casaflex is a type of pipe that differs from ordinary district heating pipes in several ways. It is compared with ordinary d.h. pipes in earlier studies [4]. This type of pipe is not particularly common in the Nordic countries. The pipe is supposed to be used as ordinary pipes with sand as backfill.

Indicator 3

Indicator 2

Fig. 8 Locations of displacement gauges inside test pipe.

During the test period that lasted for one year indications from the displacement gauges were measured twelve times. During the test period the temperature varied between summer temperatures to winter temperatures.

The media pipe is made of corrugated stainless steel and surrounded by CFC-free polyisocyanurate foam. The foam is wrapped in a multi layer barrier foil at the outside covered with a corrugated low density polyethylene casing. Inside the insulation along the pipe there are three surveillance wires. The casaflex pipe can be delivered in very long lengths. The pipes used in this test were 140 m.

2.2 Instant deformation of the pipe and accelerations from traffic load Two accelerometers were placed 200 mm respectively 600 mm below the asphalt surface (see Fig. 5), close to the test pipe, to measure the vibrations in the road structure when heavy vehicles pass over the test pipe.

To connect different casaflex pipes a system with flanges, bolts and gaskets are used. The gaskets are made of graphite.

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 4â&#x20AC;&#x201C;6 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 km/h) while the vibrations in the road structure were registered with the two accelerometers. The weight of the lorry was 26 400 kg.

2.1 Deformation of the pipe over time The test pipe was 1.66 m long and prepared with three displacement gauges inside to measure the radial deformation in three directions. The displacement gauges were installed at the half length of the pipe with a distance of 100 mm in between. One displacement indicator measured on the upper side of the pipe casing, the second on the underside of the pipe casing and the third at the side of the pipe casing. The displacement gauges were fixed to the media pipe to measure the changes in the pipe casing.

To investigate the instant deformations in the test pipe when heavy vehicles pass over the pipe the indications from the displacement gauges (see section 2.1) were measured at the same time as the vibrations in the road structure were registered. 2.3 Radial and axial stiffness of pipe In laboratory the physics of the test pipe were tested concerning radial stiffness and axial stiffness. The tests were done on a 165 mm long test specimen from the same pipe as the test pipe. The arrow in Fig. 9 shows the direction of the applied load during the test.

Before the test pipe was installed reference measurements were done at the laboratory to create zero values for the displacement gauges.

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

Fig. 9 Arrangement for test of radial stiffness. 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. Fig. 10 describes the three arrangements for applying the load in the tests, (a) the applied load acts on the whole cross section, (b) the applied load acts on only the steel pipe and (c) the applied load acts on the outer steel net including the pipe casing. The arrow in the figures indicates the direction of the load. (a)

(b)

2.5 Leak test of the pipe casing To discover moisture or even water in the insulation, there are different indicators on the market. The typical indicator system used in Gothenburg is the so called Nordic System. The Nordic System is a system which is using two naked cupper wires inside the insulation along the pipe at 10 am and 2 pm. The casaflex pipe uses the Hagenuk System. That system uses three wires a) Ni Cr, b) Cu, insulated and c) Cu, not insulated.

(c)

In this test different pipes and different systems were connected. The Ni Cr wire in the Hagenuk System was left disconnected.

165

The resistance was measured with an ordinary ohm meter, BM 400.

140

Fig. 10 Three types of arrangement for applying the load for test of axial stiffness.

The pipes were also three times tested with a, State meter, Time Domain Reflectometer (TDR) from Stateview.

2.4 Pipe prolonging while pressurizing A casaflex pipe does not expand because of the thermal load. It is self compensating. But because of the geometry of the media pipe it expands when it gets pressurized. On the other hand the multi layer barrier foil in the pipe holds the expansion back. Because the pipe is flexible, it will still be able to expand, but only until the multi layer barrier foil stops the expansion.

2.6 Test of degree of compaction of the street To get the permission from the road owner to do this test in the street there were certain standards to follow [5] and [6]. Before the asphalt could be put on the shaft there were to be some tests of the degree of compaction of the street with certain limits. It is a German test that is also used in Sweden [7]. Basically the soil gets compressed with a known load over a known area and one measures the Youngâ&#x20AC;&#x2DC;s modulus Ev for the soil two times. The demands were that;

To see how much the pipe expands because of the pressurization, a distance indicator, Hilti PD4, was fixed on the pipe before it was installed and pressurized while it was still on the ground. The distance was measured three times against an iron angle which also was fixed on the pipe. After the pressurization the distance was measured again three times.

a) b) c)

Ev2 / Ev1 < 2,8 Ev2 > 50 MPa At least 4 out of 5 tests should be correct.

Fig. 11 Left; Fixed distance indicator, Right; Fixed iron angle.

To see with which force the casaflex pipe was expanding, the following equation was used: Fp = PA

Fig. 12 Test of degree of compaction of the street

(1)

2.7 Visual control of the surface of the street

Where Fp is the prolonging force [N], P is the internal over pressure [Pa] and A is the maximum inner area of the pipe [m2].

As an extra precautionary measure, the street was optically inspected every month through a year. During the first month, the street was inspected every week. And there was an extra inspection in spring in order to 99

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

In Fig. 16 and 17 the diagrams show the vibration velocity (m/s) in the ground when a heavy lorry pass over the test area at a speed of 40 km/h. The vibration velocity is calculated from the acceleration signal by integration.

find potential frost action damages. The inspections were documented with photos. 3. RESULTS 3.1 Test results from deformation of the pipe over time

The diagrams in Fig. 18 and 19 show the maximum amplitude of the acceleration in the ground as a function of the speed of the lorry when it passes over the test area in 20 km/h and 40 km/h, respectively the maximum vibration velocity as a function of the speed of the lorry.

The measured pipe deformations during the test period turned out to be very small. The diagram in Fig. 13 describes the measured changes in the casing since installation and average air temperatures during the test period. All three displacement gauges were set to zero before the installation. The diagram shows that the casing of the test pipe during the installation was squeezed out up to 0.5 mm at the three measurement points. The deformations in the casing are most likely caused by the packing of the backfill surrounding the pipe.

200 mm below surface, vehicle speed 40 km/h

After the installation during the test period the results indicate that the upper side (violet curve in the diagram) of the test pipe casing have been pressed in 0.2 mm. The side of the test pipe casing have squeezed out approximately 0.1 mm. The under side (red curve) was squeezed out approximately 0.1 mm during the period between the first and second measurement results. During the rest of the test period the casing have been pressed back in 0.1 mm.

Time (s)

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

600 mm below surface, vehicle speed 40 km/h

It is to be observed that these measured changes are very small relative to the test pipe casing diameter. Compared to the zero values in the laboratory the measured changes are not more than 0.3 % relative to the casing diameter. 0,9

Average temperatures (째C)

0,8

0,7

Time (s)

0,6

-20

Under Side

0,5

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

-40

0,4 0,3

-60

Upper Side

0,2

-80

Side

0,1

200 mm below surface, vehicle speed 40 km/h

-100 0

100

200

300

400

Velocity (mm/s)

Days since installation

Fig. 13 Average air temperatures and changes in casing at installation and during test period.

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

Time (s)

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

100

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

over the test area. At next instant it is squeezed together approximately 0.07 mm at the same time as the blue curve indicates that the pipe goes eccentric approximately 0.04 mm.

Velocity (mm/s)

600 mm below surface, vehicle speed 40 km/h

0,12

0,10

Side (mm)

0,08

Up (mm)

0,06

Down(mm)

0,04 mm

Time (s)

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

0,02 0,00 -0,02 -0,04

-0,06 0,00

1,00

2,00

3,00

4,00

time (s)

Fig. 20 Instant deformation 200 mm below the road surface when lorry passes at 40 km/h.

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

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

1600 1400 1200

Load, N

1000 800 600 400

Test 1

200

Test 2

0 0

0,5

1,5

Deformation, mm

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

Fig. 21 Diagram radial stiffness of a casaflex pipe.

Feed speed 3 mm/min The acceleration of the ground increases with the speed of the traffic. And the effect is more sensitive the closer you are to the surface (see Fig. 18).

Load case (a)

18 16 14

Axial load, kN

The vibration velocity also increase with the speed of the traffic. The effect is not as sensitive as for the acceleration when it comes to the coverage (see Fig. 19).

Load case (c)

12 10 8

Load case (b)

6 4 2

In Fig. 20 the diagram describes the instant deformations in the test pipe when the heavy lorry passes over the test area at a speed of 40 km/h. It can be seen from the red and the violet curves that the pipe casing is squeezed together approximately 0.17 mm from top to bottom at the instant when the lorry passes

0 0

0,5

1,5

Axiell compression %

Fig. 22 Axial stiffness of a casaflex pipe 101

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

3.4 Test result of pipe prolonging while pressurizing The test results are as follows: Table 1. – Test result of pipe prolonging while pressurizing Tests 1

Average

[mm]

Distance before pressurizing

4 360

Distance after pressurizing

4 365

The pressure that was used was approximately 1400 kPa. The test results show that a single pipe casaflex prolongs itself 100*5/4360 = 0,11%.

Fig.23 TDR graph for the supply pipe.

his could be compared to the more common steel pipe for district heating. If that pipe would be loaded with a thermal load of 100 ºC it would prolong itself 0,12%.

3.6 The results from the test of degree of compaction of the street The different tests were plotted in diagrams and gave the different Young‘s modulus Ev1 and Ev2 for different places. The places were documented in a photo. The results can be read in the table below.

The force with which the casaflex pipe is expanding because of the inner pressure would for 11400 kPa be 7,3 kN according to the supplier. That would mean that the diameter would be 81,5mm. In real life the diameter was measured to be 83,9 mm. The corresponding force for the diameter 83,9 mm would be 7,7 kN.

Table 2. - Results from test of degree of compaction Spot

If the pressure would have been 1 600 kPa and the diameter would have been 81,5 mm then the corresponding force would have been 8,3 kN.

(MN/m )

This could again be compared to the steel pipe with the thermal load of 100 ºC. This pipe would prolong itself with the force of 164,9 kN. So the casaflex pipe expands with a force that is approximately 100*8,3/164,9 = 5,0% of the force from a steel pipe when heated 100 ºC. Looking at Fig. 22, case a), one sees that the inner force (axial load) that the expanding force has to overcome is negligible.

There are no leaks in the test area, neither in the supply pipe nor in the return pipe.

38,75

91,44

2,36

18,23

25,25

1,39

61,29

126,4

2,06

51,53

100,65

1,95

29,30

70,08

2,39

The strength of the road There was no change in the surface of the street due to the shallow laying of the district heating pipe what so ever the first eleven months. In spring after an unordinary cold winter one could see a small crack (approximately 12 cm) in the street along the pipes extension. As this article is getting written it is not investigated why the crack has appeared nor of the importance of it. The street has much worse injuries from frost action damages outside the test area.

Through metering the resistance and making TDR graphs it is proven that:

Ev2/ Ev1

3.7 Results from the visual control of the surface of the street

Different TDR graphs have been made in May 2009, in June 2009 and in April 2010.

It can be done to connect the two different systems (The Nordic System and the Hagenuk System).

Ev2 2 (MN/m )

For every test the division Ev2/Ev1 are approved. Spot 2 Ev2 is to low but the other four spots are approved so overall the test is ok.

3.5 Test results from the leak test of the pipe casing

Ev1

102

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The acceleration and vibration velocity are also negligible under the traffic load from e.g. a heavy lorry. Probably there will be no problems using graphite gaskets also with only 200 mm of coverage.

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

The casaflex pipe is prolonging itself if it may but the force with which it is prolonging itself is but a fraction of what comparable ordinary steel pipes uses. This effect makes it more suitable for shallow shafts. It is possible to combine different leak indicator systems and still get the TDR-graphs. The graphs done in this test indicates that there are no leaks in the operating pipes after one year. The demanded levels for the degree of compaction of the street are possible to reach also with a d.h. pipe 60 mm below the surface as it gets compressed without hurting the pipe.

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

There are still other issues that can be considered that are not included in the tests presented in this paper, aspects as e.g. heat losses. 5. ACKNOWLEDGEMENT The authors would like to express their appreciation to a couple of key persons. There had been no test of this kind without their support and permitting. The persons are:

Fig. 25 Left Test area in March, Right; D. H. chamber revealing itself in February.

However the heat loss is of course bigger compared with normal standard because of that the pipes are placed closer to the air. Frost action damages In theory one could imagine that the street on both sides of the district heating pipe would erect during the winter if there were soil that could frost heave. This could of course damage the asphalt. But streets are not supposed to be built with soil that could frost heave. So there should not be any problem.

–

Mr. Bo Andersson Planing Manager at Trafikkontoret Göteborgs Stad,

–

Mr. Lars Ljunggren Manager at Göteborg Energi AB and

–

Mr Göran Johnsson Technical Manager at Powerpipe Systems AB.

6. REFERENCES

There was no notable difference in the height of the street over the district heating pipes compared to the street beside the test area during the winter.

[1] Molin J., Bergström G. and Nilsson S. (1997). Kulvertförläggning med befintliga massor, Swedish District Heating Association FOU 1997:17, (in Swedish)

4. CONCLUSION

[2] Bergström G., Nilsson S. and Sällberg S-E. (2001), Täthet hos skarvar vid återfyllning med befintliga massor, Swedish District Heating Association FOU 2001:58, (in Swedish)

The article probably describes the first operating d.h. pipes placed in backfill of 0-40 only 200 mm below the surface in a street with heavy traffic. As expected, the pipes are working nicely. The loads that have been measured are acceptable or even low for the d.h. pipe. As it seems also the street is satisfactory working even though there are d. h. pipes close to the surface.

[3] Nilsson S, Sällberg S-E, Bergström G, (2006) Grund förläggning av fjärrvärmeledningar, Swedish District Heating Association FOU, FOU Värmegles, 2006:25, (in Swedish)

The pipe deformations are negligible with respect to the pipes function both over time and under an instant traffic load.

[4] Gudmundson T. ÅF-Processdesign AB, (2002), Casaflex-rör i Malmö 2001,. Swedish District 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

[5] Trafikverket, (2005) Allmän teknisk beskrivning för vägkonstruktion ATB Väg, http://www.vv.se/Startsida-foretag/vagar/Tekniskadokument/ATB-Allmanna-tekniskabeskrivningar/Vagteknik/Aldre-versioner/ATB-Vag2005/, visited 2010-04-27, (in swedish)

9AJyMvYwMDSycXA6MQFxNDPwtTo2Anc_2CbE dFABCTfUM!/?WCM_GLOBAL_CONTEXT=/wps/ wcm/connect/goteborg.se/goteborg_se/Foretagare/ Upphandling_staden%20som%20kund/Specifik%2 0upphandlingsinformation/art_N400_FOR_Up_SU _Trafikkontoret, visited 2010-04-27, (in swedish)

[6] Göteborg Stad Trafikkontoret, Bestämmelser för arbeten inom gatu- och spårområden i Göteborg, http://www.goteborg.se/wps/portal/!ut/p/c0/04_SB8 K8xLLM9MSSzPy8xBz9CP0os3gjU-

[7] Trafikverket, (1993) Publikation 1993:19 Bestämning av bärighetsegenskaper med statisk plattbelastning Metodbeskrivning 606:1993, (in swedish)

104

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

STUDY ON THE HEAT LOSS REDUCTION METHOD FROM THE SECONDARY PIPELINES IN THE APARTMENT COMPLEX 1

Byung-Sik Park , Yong-Eun Kim 1 1 1, 3 Sung-Hwan Park , Yong-Hoon Im , Hyouck-Ju Kim , Dae-Hun Chung Mo Chung 1

Building Energy Research Center, Korea Institute of Energy Research, 102 Gajeong-ro, Yuseong-gu, Daejeon 305-343 KOREA, bspark@kier.re.kr 2 Energy System Engineering, University of Science and Technology, 113 Gwahangno, Yuseong-gu, Daejeon 305-333 KOREA, rainyday@ust.ac.kr 3 Dept. of Mechanical Engineering, Yeungnam University, 214-1 Dae-dong Gyeongsan-si Gyeongsangbuk-do 712-749 KOREA individual heating, central heating and district heating. At the moment there is a lot of potential for district heating and cooling. Korea has seen about a 10% supply of DHC among total residential houses which is very low compared to that of European countries which supplies over 50% DHC.

ABSTRACT This study aims to suggest better methods for reducing heat losses from the pipelines installed as secondary heating pipes in the apartment complex in which hot water is being supplied for space heating and hot water by a district energy supply company. Right now the district heat supplier is responsible only for the primary district heating pipelines just before the substations in the apartment complex. That is why the heat loss reduction becomes more important in the secondary pipelines after the substation in the Korean apartment complex.

If we are to increase green growth with low carbon, it is crucial to supply DHC, which has higher energy efficiency than any other method, in dense regions of population. The recent Korean government has shown effort in making a point of energy efficiency throughout main energy consuming sectors including building area. However, the supply policy of DHC is now being crippled due to various reasons. Makers or consumers of individual heating devices do not have positive attitudes toward DHC. Therefore it is important to draw attention to the multitude of benefits and merits of DHC.

Several methods to reduce the heat loss from the secondary pipelines were set up and compared by a simulation technique. One of the methods is to combine the hot water heating pipes and space heating pipes. Another method is to install a small heat exchanger in each house to supply hot water from the single space heating pipeline. In this case we can easily change the means of heat supply and the right choice of end users can be ensured for the means of heat supply.

There is certainly some heat loss from the pipelines installed under the ground to supply the district energy from the power plant to the consumers. To reduce the heat loss from these primary pipelines many innovations and advancements have been made for a long period since the district energy was supplied in the northern European countries. The heat loss generally differs according to the network type of pipelines. The more compact that the network is, the less heat loss occurs. But the type of network cannot be made arbitrarily by the designer. The designer can simply optimize the network in view of geological and environmental conditions, such as population density, not the type of network. Although heat loss exists with the primary pipelines, it can be controlled and maintained effectively by the district heat supplier. On the other hand, the heat loss from the secondary pipelines cannot be controlled properly by the building owners who are responsible for.

In this study the preferable method to reduce the heat loss in the secondary pipelines has been suggested. The simulation result has shown about 30% heat loss reduction compared to the existing scheme for the simple change of methods and much more reduction for the optimization of pipe diameter and insulation thickness or surface enhancement by low emissivity. INTRODUCTION Korea is characterized as having four distinct seasons. Apartment complexes became a typical type of residence in urban areas after the recent rapid industrialization of last 30 years. At the moment over half of the population chooses to live in apartments rather than in individual houses and the trend will increasingly continue in the future. There are three typical heating methods for apartment complexes -

The apartment complex is a unique housing system in Korea. It contains many high rise buildings of over 10, often over 20, stories high. In many cases it has over 105

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

to be measured and evaluated on the heat loss from the secondary pipelines. Many thermocouples and flow-meters were installed in the region of the pipelines to collect information on heat demand pattern, temperatures and heat loss from certain regions to be evaluated. 208

202

205

206

203

204

HWR HWR HWR HWR HWS

207

HWR HWR HWR HWR HWS

201

HWR HWR HWR HWR HWS

one thousand homes. However, the basic structure is almost the same as that of western apartment buildings except for the pipeline network between buildings and substation. In the past there was one substation in one apartment complex. The substation has a minimum of two heat exchangers which are in general shell and tube type or plate type. Nowadays the number of substations grows bigger and bigger. That means that the designer plans to install the heat exchanger separately and respectively according to the buildings which stand nearby each other. The secondary pipelines have been said to have much heat loss in Korea. There have been a few studies related to heat loss from the secondary pipelines. It is very hard to distinguish between positive heat gain and heat loss from the pipelines installed within the buildings. If the pipelines are installed in the center of the building, the heat loss from the heating pipes or hot water supply pipes can be regarded as positive heat to the consumer. But if the pipelines are installed near the building surface, the heat from the pipes can be regarded as loss.

Supplementary Water

Calorimeter

City water

DHWR DHWS

Fig. 2 Secondary pipeline network from the substation

ANAYSIS OF HEAT CONSUMPTION PATTERN

Measurement and analysis of the heat loss from several apartment complexes in Korea has been tried. The heat loss data from the several sites has been stored and accumulated throughout the year. A simulation method has been set up and the accuracy of the simulation has been investigated. Some alternatives to reduce the heat loss have been prepared from the existing scheme. The simulation method and results have been presented in this paper.

1) Space heating water flowrate Space heating amount is being measured daily. Toom temperature does not differ much between the homes in the apartment complex. Thus the temperature difference (Î&#x201D;T) between inlet and outlet of the pipeline of individual homes remains fairly constant except in the summer season. Therefore the heating water flow rate can be estimated from the following equation Q=CmÎ&#x201D;T. In other word, the flow rate could be evaluated from the measured calorific amount. A good example of this is shown in Fig. 3.

TYPE OF APARTMENT COMPLEX The apartment complex was built and opened in November 2007. It has 8 buildings which are comprised of 518 homes. Each home has 112 m2 of heating area. Fig. 1 shows the location and overall shape of the apartment complex which was chosen

Fig. 3 Temperature difference between supply and return North KOREA

In this study the flow rate was measured for the two months of November and December 2009 using the flow meter installed in the space heating water pipeline. And the hourly heating water flow rate of individual homes for the year 2009 was extracted from the comparison of the total measured amount and the individual house measurement. Fig. 4 represents the annual heating water flow rate. Some differences exist during the cold winter season.

South KOERA JAPAN CHINA

Fig. 1 The location and the shape of the apartment complex

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

T out

7th floor 6th floor 5th floor 4th floor

T low

3rd floor

Primary pipe line

Secondary pipe line

2nd floor 1st floor

DHWS

T base

T sb

DHWR

Fig. 4 Heating water flow rate

Fig. 6 Category of temperature charicterization.

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

Tout:

Temperature outside the building

Tbase:

Temperature of the underground space frequently open to the outside surrounding.

Tsb:

Temperature of the underground space closed to the outside surrounding.

Tlow:

Temperature of low-rise region in the building

Thigh: Temperature of high-rise region in the building

Fig. 5 Hot water flow rate

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

Fig. 7 Varous temperature of the surroundings of pipelines.

SIMULATION METHOD For the heat loss simulation commercial tool, â&#x20AC;&#x2022;Flowmasterâ&#x20AC;&#x2013; of 1 D system analysis has been used. Flowmaster is a program which can analyze the thermo-hydrodynamic characteristics of pipe systems if the following items are given such as the physical properties of pipes, flow rate and outside temperature through the following equations. Annual heat loss can be simulated by using the information such as measured temperatures, flow rates and various physical properties of pipes and insulation materials according to the drawing of all the pipelines which are 107

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

VARIOUS SCHEMES FOR SIMULATION 1) Present Scheme 12th floor 11th floor

HWR

Hot water return

9th floor 8th floor 7th floor 6th floor

HWR

Heating water return Hot water supply

HWR

HWR HWR

10th floor

HWS

(2)

Heating water supply

HWR

(1)

HWS

5th floor 4th floor

(3)

3rd floor

Primary pipe line

Secondary pipe line

2nd floor

HWS HWR HWR HWR

DHWS DHWR

1st floor

Fig. 8 Pipeline configuration of present scheme

fluid temperature, 째C

The present scheme is composed of 4 pipelines, two of which are for heating water supply and return and the other two of which are for hot water supply and return. Heating water is supplied to each home and returned from each home and resultantly the flowrate of supply and return are equal. On the other hand hot water has a certain amount of recirculation in order to keep supply water hot. Hence the same amount of consumed hot water in each individual home should be supplied to the hot water heat exchanger to guarantee instant hot water supply.

ambient temperature, 째C convection heat transfer coefficient, radiation heat transfer coefficient, internal heat transfer coefficient, insulation thermal conductivity, pipe thermal conductivity, external pipe diameter, external insulation diameter,

internal pipe diameter,

Alternative A

Alternative A is a scheme which removes the hot water pipelines and combines with the heating water pipelines. Thus there are only two pipelines of supply and return from substation to each building. These supply and return pipelines have two functions of heating and hot water supply and return. By reducing from 4 to 2 pipelines the heat transfer surface area can be decreased. However, this scheme has disadvantages in summer when the heating water supply has been closed. If the pipelines should be used for the supply of hot water, the resultant water speed in the pipelines would be very small.

For accurate simulation, individual flow rate was used respectively and differently based on total measured flow rate and read amount of individual flow meter of 518 homes throughout the year. By doing this, the flow rate in the individual pipelines can be determined according to the usage amount of heating water and hot water. This similarly leads to actual flow rate in the pipelines. From this complicated process, the number of individual flow rate of hot water and heating water comes to 378,140. Macro which was combined with Excel and Flowmaster was used for a 365 day analysis. This process requires 32 hours for 8 buildings for only one case.

12th floor 11th floor

HWS

Heating water supply

HWR HWR

Heating water return Hot water supply

9th floor

HWR

Hot water return

7th floor

10th floor

6th floor 5th floor 4th floor 3rd floor 2nd floor

Primary pipe line DHWS

Secondary pipe line P-1

HWS

1st floor P-9

P-4 P-6

E-2

DHWR

E-1 P-2

HWR

Fig. 9 Pipeline configuration of Alternative A 108

P-3

E-3

HWR

HW R

HWS

8th floor

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

For hot water, this scheme is effective in heat exchange because the heat exchanger acts for the consumed hot water only and can avoid extra recirculation pipelines. But it has the drawback of supplying cold water or non heated water when intermittently using hot water. 3)

5) Use of PEX PIPES PEX is being used in western countries as district heating pipes for low temperature service from renewable resource application. For this reason PEX can be used as secondary pipelines which normally are under service of low temperature.

Alternative B

VALIDATION OF SIMULATION RESULT To validate the simulation result, the measurement value was compared with the prediction result by the present simulation method. The measured value of December 2009 was used for heat loss reference value. As seen in Table.1 the simulation result fairly agreed with the measured data. And the simulation method can be used without much modification. For more accurate prediction it needs slightly more supplementation in the numerical modelling of heat transfer phenomena of outer pipe surface and environment.

12th floor E-10

11th floor

HWS

Heating water supply

E-11

10th floor

E-15

HWR

Heating water return

9th floor E-14

8th floor

HWS

6th floor

E-13

HWR

E-12

7th floor

E-7

5th floor E-8

4th floor E-9

3rd floor E-6

2nd floor

Primary pipe line

Secondary pipe line

E-5

P-1

DHWS

HWS

1st floor E-4

E-1

DHWR

HWR

P-4

Fig. 10 Pipeline configuration of Alternative B

Table1 Comparison of measured and predicted values[unit: MWh]

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

Measurement

heat

Hot water

Heat loss rate

loss

20.2

11.2

9.44%

supply

264.7

67.8

heat

Hot water

Simulation Heat loss rate

loss

17.0

14.29

9.90%

supply

252.6

63.53

SIMULATION RESULT 4) Alternative C

Fig. 12 shows typical heat supply and heat loss for the 24 hours of 11.11.2009. The simulation result of each scheme is shown from Fig. 13 to Fig. 17. Each graph shows similar patterns and the heat loss comparison of each scheme is summarized in Table 3.

12th floor 11th floor

HWS

Heating water supply

HWR HWR

Heating water return Hot water supply

9th floor

HWR

Hot water return

7th floor

10th floor

HWR

HW R HW R HWS

8th floor

6th floor 5th floor 4th floor 3rd floor 2nd floor

Primary pipe line DHWS

Secondary pipe line P-9

HWS

1st floor P-7 P-14

DHWR

E-5

HWR

P-8

Fig. 11 Pipeline configuration of Alternative C

Alternative C is a variation of Alternative A to make up for the defect of cold water supply when intermittently supplying hot water. In this alternative the recirculation pipelines are equipped in the buildings.

Fig. 12 Hourly heat supply and heat loss of 11.11.2009

109

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

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

The heat loss comparison of each scheme can be summarized in Table 2. Table 2 Heat loss comparison of each scheme[unit: MWh]

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

Supply

Loss

Heat loss rate

Present scheme

4863.7

681.4

14.01%

Alternative A

4477.6

456.3

10.19%

Alternative B

4317.1

516.3

11.96%

Alternative C

4929

523

10.61%

PEX Pipe

4691.2

508.9

10.83%

HEAT LOSS COMPARISON DUE TO THE CHANGE OF INSULATION THICKNESS,NOMINAL DIAMETER AND INSULATION MATERIAL

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

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

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

Fig. 16 Annual heat supply and heat loss (Alternative C) 110

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

More enhancements in heat loss can be extracted from heat loss reduction by the selection of optimum pipe diameter, good insulation material, increasing insulation thickness and changing surface emissivity of insulation material.

Table 3 Heat loss comparison due to various parameters Heat loss comparison Present pipe of insulation thickness 40 mm

100%

Downsizing nominal diameter by 1 level

88.7%

Insulation thickness 50 mm

86.1%

Insulation thickness 60 mm

75.5%

Closed-Cell Elastomeric thermal insulation

89.3%

Pipe diameter downsizing + insulation thickness 50 mm

76.7%

Pipe diameter downsizing + insulation thickness 60 mm

67.6%

ACKNOWLEDGEMENT The financial support from KDHC made this work possible. This paper is based on the results of an ongoing research project which will be completed at the end of 2010. REFERENCES [1] W. F. STOEKER, DESIGN OF THERMAL SYSTEMS 3rd edition, Mc Graw Hill, pp. 53-160 [2] Incropera, HEAT TRANSFER 5th,WILEY [3] Byung-sik Park et al, Study on the Reduction method of Heat Loss from the Secondary Pipelines installed in the Apartment Complex, 2008

CONCLUSION Present scheme for the secondary pipelines is evaluated to have 14% annual heat loss based on the total heat supply to the apartment complex. This is a very large amount when we consider that the primary district heating pipeline has only about 4 to 5% annual heat loss in dense population urban areas.

[4] Byung-sik Park et al, Study on the Reduction method of Heat Loss from the Secondary Pipelines installed in the Apartment Complex, 2009 [5] Flowmaster, Flowmaster heat transfer manual [6] Manfred Kl psch, Plastic pipe system for DH, Handbook for safe and economic application, IEA R&D Programme on District Heating & Cooling

Heat loss by Alternative A can be reduced about 30% compared to that of the present scheme which has been widely adopted in Korea until now. However, Alternative B has more heat loss compared to that of Alternative A, which was not the common expectation. The main reason was the increase of the pipe insulation surface area which acts as a heat transfer area. Alternative C and PEX system can be replaceable when they have merits in the point of initial cost.

111

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

HEAT LOSS OF FLEXIBLE PLASTIC PIPE SYSTEMS, ANALYSIS AND OPTIMIZATION 1

EJ.H.M. van der Ven , R.J. van Arendonk 1

Thermaflex International Holding B.V 2 Liandon B.V. for a small-sized pipe to about eight hours for the largest sized diameter.

ABSTRACT A newly developed, in-house, test rig for measuring heat loss of pipe systems allows the user to analyse various systems in a short timeframe. This allows quick insights into heat loss variables and mathematical analyses. The effect of alternative compositions of insulation and other layers can be evaluated within a short time span.

The novelty of the test rig is described in the paper ‗Verification of heat loss measurements conducted on (semi) flexible pipe systems‘ [3]. The novelty for product improvement is that due to the reduced time required for a test run the effect of alternative systems can be mathematically analysed and evaluated in a short time. In this way the analyses of alternative production methods has a short feedback. Optimization of the product can be effected in a short time.

This already led to improvements of the production process and of the insulating foam. INTRODUCTION

In the near future the test rig will be used for quality control of the production process. This test will partly replace other currently applied standard tests, such as density and cell size measurements.

Liandon developed a heat loss testing rig for Thermaflex to test their produced flexible pipe systems. Within a short time the pipe system, undergoing a heat loss test, tends towards the controlled temperature in the sections of the sample, the added power reaches equilibrium and the test results can be collected. Due to the short time required for testing, the results of alternative production methods are easily available. Due to the short response time the test is a great help in the search for product and production improvements.

METHOD DESCRIPTION In the Flexalen 600 pre-insulated pipe a PB medium pipe is encapsulated in insulating foam, which is protected against wear and tear in a corrugated hard cover pipe. The pipe product has a solid bonding between the insulation and cover and no bonding between the insulation and the medium pipe.

The objective of this paper is to present the results of the research to the overall heat loss performance of a flexible plastic pipe product, Flexalen 600.

According to EN 15632, the European Standard for pre-insulated flexible pipe systems, this pipe system is classified part 3: Non bonded system with plastic service pipes. The Flexalen 600 plastic pipe system differs in some areas significantly from most other systems in this class:

The objective of the research is: 1 Find correlations between heat loss and other parameters of the pipe system such as outer diameter, inner diameter, foam surface and foam structure. These correlations are determined by the mathematical analysis of practical heat loss measurements.

1 Physical bonding between foam and outer casing, 2 One layer of foam, filling the complete space between service pipe and cover,

2 Find possibilities for the improvement of the pipe 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 and test methods give rules for calculation of the heat flow to ambient (heat loss) from measured values, making the heat flow of various parameters comparable.

The actual heat loss of pre-insulated pipe products is determined under controlled, similar conditions for an entire diameter range. This range comprises various outer diameters, various inner diameters and various compositions in materials and pipe systems. The time required for one single test run varies from half an hour

The heat loss calculations of annex D are based on the thesis of Wallentén 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

systems. Test spools can be extracted directly from production. In this way tests can be executed with fresh product.

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

Also testing of cured piping and piping that is aged and has been degassed during storage or in high temperature aging is possible. Knowing the relationship between heat loss and various parameters, the most prominent parameters for heat loss can be evaluated. The most prominent parameters may lead to the improvement of the flexible pipe system to ensure optimal performance with minimal heat loss. The influence of several prominent parameters is determined and recommendations are given in order to optimize the insulation performance. Reliability and reproducibility of the test rig is discussed in [3] Verification of heat loss measurements conducted on (semi) flexible pipe systems (van der Ven et al).

Figure 1, Thermaflex heat loss equipment

In the new developed test rig (figure 1) a test spool (figure 2) is put in a slim fitting sleeve.

MANUFACTURING PROCESS

The test spool is heated internally in three sections. The middle section of the spool is the test section. Heating in this section is controlled to obtain the required test parameters. The two ends are heated to compensate for the heat loss from the ends of the middle test section. In this way an endless pipe is imitated.

The Flexalen 600 product has been developed by Thermaflex, located in The Netherlands. The development started in 2002 and resulted in a first small-scale commercial production in 2005. During the production of Flexalen 600 four different production techniques are combined, partly simultaneous and partly sequential:

The outer side of the sleeve is water-cooled to obtain heat transport from the test spool.

1 Production of PB service pipes optionally covered with an EVOH oxygen barrier layer.

During the start of the test, heat is lost into the heating of the pipe system and into the surrounding cooling water. When heat losses have reached equilibrium, the steady state heat transfer can be measured.

2 Production of LDPE insulation foam to fill the area between medium pipe and outer casing. 3 Production of outer casing of HDPE. 4 Assembly of the different elements (1, 2 and 3) with a full bonding of the foam and the outer casing, while corrugating the casing. These techniques are based on extrusion technology. The production line consists of purchased equipment combined with technology developed in-house. The complete production is a (semi)-in-line production. All pipe systems are produced at Waalwijk in the Netherlands. Unique for the process is the ability to produce continuous lengths. For practical reasons the lengths produced depend on the outer casing of the product and the size of the reel. The maximum length produced can reach up to 2000 meters.

Figure 2, Longitudinal section guarded end heating probe

The time span required for testing in the test rig is rather short. The time to reach equilibrium lies in the order of hours, depending on the diameter and insulation thickness. Comparable tests often require time spans in the order of days.

The complete Flexalen 600 pipe system includes preinsulated pipes, couplings, sleeves, pre-insulated Tconnections, etc. The production range is described by Engel and Baars. [5]

Containing various diameters, which are based on the standard production outer diameters, the test rig enables heat loss tests for various diameters of piping 113

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

extruder. This foaming agent is a hydrocarbon that causes the expansion of the LDPE.

Production of PB service pipes All service pipes are made of Poly-butene. This is a plastic with a special combination of properties. Polybutene has excellent heat and creep resistance, flexibility and strength at a long lifespan and is fully recyclable. In accordance with the temperature duration profile mentioned in the BRL5609/EN15632, PB is suitable up to a maximum temperature of 95 °C.

The quality of the foam, moreover the insulation properties of the foam, depends on parameters such as density, cell size and chemical composition. All parameters are measured and adjusted within limited tolerances to meet the specifications. The best efficiency is further improved when the space between the service pipe and the foam is filled better.

All PB-pipes are weldable by socket fusion, electro fusion and butt welding, which allows for an all plastic distribution system without metal parts that are prone to corrosion. The service pipes are produced via state-ofthe art extruders.

As there is no bonding between the service pipe and foam, there is no risk of damaging the foam and properties by expansion of the pipes due to thermal fluctuations in the applications.

The production line consists roughly of an extruder, calibration tools for adjusting pipe size, cooling baths, and marking and cutting equipment.

Production of corrugated outer casing (HDPE) and assembly of complete product The outer casing is applied by an extrusion process and thermally welded to the foam after the pipe has been inserted. The outer casing is also corrugated to optimize the flexibility of the finished product. Now the product is ready for coiling.

Pipe dimensions are checked inline every second during and after production with ultrasonic measurements. PB-pipes can be produced up to an outer diameter of 225 mm. If desired an outer oxygen barrier layer may be applied via co-extrusion up to an outer PB pipe diameter of 90 mm.

After production the final product must cure for 5 days. During this curing period the degassing of the foaming agent starts, while the insulation foam is still stabilizing

After production the PB-pipes are stored for a minimum period of 5 days and cured to create the correct crystalline polymer structure. After curing the pipe is used for the Flexalen 600 production.

Production testing and controlling During the production, parameters are checked and controlled, such as:

Every batch produced is verified by the in-house QC department according to Dutch directives

1 Chemical composition of the foam. 2 Settings of all extruders involved (foam extruder and extruder for corrugated outer casing).

- BRL-K5609, for PB pipes with oxygen barrier or - BRL-K17401 for PB pipes without oxygen barrier. Product and manufacturing processes are checked 6-8 times a year and certified by independent agencies such as Bureau Veritas, KIWA and CSTB and le Centre Scientifique et Technique du Bâtiment. The quality of the QC department is validated by these checks and by internal and external audits.

3 Density and cell size of the foam. 4 Dimensions of the foam (outer diameter and inner diameter). 5 Line-speed of all involved products (Foam / PB-pipe / End-product). 6 Thickness of the corrugated outer casing and the connection of the corrugated outer casing to the foam.

Production of LDPE insulation foam Thermaflex has now approximately 35 years of experience in the production of LDPE foam via extrusion techniques.

7 Since a few months: in-line production control of Heat Loss of the pre-insulated pipe system.

Most of the raw materials that are used are tailor-made mixtures according to Thermaflex specifications. Through these specifications and proper production quality control the company‘s philosophy related to core business is also realized for raw materials.

The most prominent factors to influence heat loss are parameters one to four. The foam lambda is directly affected by these factors as represented in equation 1.

 total

During the heating, melting and mixing of the raw materials the foaming agent is injected into the

114

 convection  conduction       radiation  blowingagent 

(1)

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

RESULTS AND CORRELATIONS

DISCUSSION OF PARAMETERS

It is again the company´s philosophy that paves the way for innovations. One of these innovations is the development of an in-house device for testing pipe systems, directly from production, cured and degassed. This proved to be a suitable device for production control, but also for gaining more insights into product parameters.

Table 1 summarises the results of measurements and calculations of tests on Flexalen 600 pipes directly from production. Various diameters are tested and calculated according to EN 15632 for a surface temperature of 10 °C and a common medium temperature of 70 °C (Instead of the maximum medium temperature of 95 °C).

The objective of testing is to improve the knowledge of the produced pipe systems in order to optimize:

The table indicates the relationships between product, cross sectional area of the foam, foam density, cell size of the foam, remaining foaming agent, calculated thermal conductivity and the calculated heat loss of a buried piping system.

Production methods: Machine data can be adjusted based on test results. Test results may lead to new production methods with new equipment.

The products 50A25, 63A32, 75A40 and 90A50 are newly developed. These products are not necessarily District Heating products. However, they are produced using the same process and have their application in the connection between the district heating network and the building or house. It is also applicable in case of low temperature differences, cooling or in-house heating or cooling.

Chemical and physical composition of layer material: Knowledge of various composition materials may lead to an improvement of insulating values. Cell structure and gap between insulating foam and medium pipe: Cell size influences values, test result based improvements are possible. The gap is a bad insulator. The quest for a minimal gap started with testing. Improvement and minimization of this gap was an achieved challenge in the testing period.

Table 1: Test results of fresh, uncured piping systems Foam 50,calc Heat Loss* section density cell size agent mm² kg/m³ (mm) % mW/m.K W/m 50 A 25 50,0 0,47 52 39 15,3 1.473 63 A 32 34,0 0,50 52 38 15,2 2.313 75 A 40 38,0 0,40 46 44 17,8 3.044 90 A 40 42,0 0,80 64 51 17,1 5.105 90 A 50 39,0 0,80 62 55 23,0 4.398 125 A 63 39,0 0,88 70 56 22,0 9.155 160 A 75 40,3 1,20 81 54 21,0 15.688 160 A 90 35,0 1,30 85 61 25,2 13.745 200 A 110 21.913 45,0 1,60 81 68 27,4 *) calculated heat loss of buried system at temperature difference of 60 K Product

Thermaflex is a lean and mean organisation that responds quickly to new insights. Therefore new insights were applied even before the complete range of production testing was performed. For the company, improvements of product and production have the highest priority. Although the production range is wide, the insight into specific and general parameters increased considerably. The research provides the prominent variables to improve insulation performance. Practical heat loss determination, in combination with analytical studies, results in a clear understanding of heat loss behaviour in single and twin flexible pipe systems during their entire lifetime.

In Graph 1 foam density and cell size are related to the cross sectional surface. In Table 1 the foam density varies from about 35 kg/m³ to about 50 kg/m³. Graph 1 shows hardly any relationship with the surface of the cross section.

As a result of the tests the manufacturing process is improved in two steps.

Table 1 shows that cell size varies from 0.47 to 1.60 mm. Graph 1 shows that cell size is directly related to the cross sectional surface, however less than 1 to 1. This relationship is influenced by physical production parameters.

The emphasis of the first step was to diminish the cell size of the foam. This succeeded in a decrease of cell size by some 20%. The latest step is altering production such that the content of anti-radiation agent increases. The initial results are promising but are not yet conclusive as the anti-radiation agent is also a good heat conductor.

115

1,8

1,5

1,2

0,9

0,6

0,3

0,0 25.000

5.000 10.000 15.000 20.000 Cross section of foam [mm²]

Graph 3 shows the relationship between outer pipe system diameter size and the percentage of foaming agent directly from production. With increasing diameter the foaming agent increases, possibly to an asymptotic value.

Density [kg/m³]

Average cell size [mm]

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

An additional interesting factor is the ongoing process in the foam during and after production. As described before the final step in the production is a 5-day curing stage. During the curing stage the foam expands and part of the foaming agent releases from the foam. As the foam is locked by a hard outer shell, expansion is directed inwards. By this the gap between the foam and the PB medium pipe, typical for our production method, is decreased.

40,0

35,0

30,0

25,0

20,0

15,0

30 200

100

150

Table 2 shows the effect of curing and degassing on both the contents of foaming agent and the calculated heat loss.

Conductivity [mW/m.K]

Heat loss of buried piping [W/m]

Graph 1: Density and cell size in relation to foam cross section

Even when forced, degassing takes time. The number of degassed samples manufactured in the same way as the fresh samples is therefore limited. Table 2 is short due to a lack of adequate and comparable samples. Table 2: The effect of time on curing and degassing

Outer diameter pipesystem [mm]

Graph 2: 50,calculated and calculated heat loss of buried pipe in relation to outer pipe size

63 A 32 75 A 40 90 A 40

Graph 2 shows the influence of outer pipe size to calculated conductivity 50 and heat loss of buried pipe systems. It also shows that part of the increase of the heat loss with the diameter is caused by increase of conductivity.

6 days curing Heat  Loss % W/m % 23 17,5 15 40 16,9 -5 53 18 5

Agent

Agent % 0 0 0

Degassed Heat  Loss W/m % 17,1 12 20,1 13 18,5 8

Heat loss of buried piping [W/m]

Percentage foaming agent

Fresh Heat Loss % W/m 52 15 46 18 64 17

Product Agent

80 70

19 18 17 16 15

Percentage of foaming agent 50

Graph 4: Relationship between heat loss and foaming agent

40 50

100

150

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

200

Outer diameter pipesystem [mm]

Graph 3: Foaming agent content in relation to outer diameter size 116

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia 31.0

The tests have led to improvements to the flexible plastic pipe products produced by Thermaflex. In the near future we expect improvements in:

29.0

Heat loss

27.0

Heat loss Degassed

Heat loss of buried piping [W/m]

RESULT BASED FUTURE DEVELOPMENTS

Production methods: Starting June 2010 a change in the machine configuration will be implemented as an extra step. The new configuration improves temperature control in the extruder, which leads to better cell structure.

110

130

150

170

190

Examples of product improvement Further research has led to product improvements. Based on these improvement proposals Thermaflex has been able to produce new pipe system samples. As represented in graph 5 the new samples have a heat loss decrease up to 16 percent compared to the previous results.

Heat loss of buried piping [W/m]

40.0

CONCLUSION The results of testing are reliable. Knowledge of the product and production has led to promising improvements of both. to

17.0

Graph 5: Degassed heat loss values of a buried system at a temperature difference of 60K

In this paper only the heat loss of the Flexalen 600 preinsulated pipe product has been handled. Information about the system can be read in [4] Heat loss optimization of flexible plastic piping systems, life time heat loss performance (Korsman et al) and [5] New economical connection solutions (Engel).

lead

19.0

Outer diameter pipe system [mm]

Up till now the heat loss performance on single pipes has been measured and analysed. This has resulted in an understanding of the heat loss principles in district heating systems. Twin pipe systems will soon be tested, analysed and evaluated.

certainly

21.0

Cell structure and gap between foam and medium pipe: In the tests we see variations in cell structure and gap width. Future research will aim at acquiring more detailed knowledge of these phenomena.

will

23.0

15.0

Chemical and physical composition of layer material: Up to a certain degree the anti-radiation agent improves isolating values. With trial and error the antiradiation agent content is increased. Up till now the maximum content has been limited by production methods. Research is required to investigate maximum desired value for insulating effects.

Further research developments.

25.0

Heat loss

35.0

Heat loss New

30.0

25.0

20.0

15.0 50

110

130

150

170

190

Outer diameter pipe system [mm]

further

Graph 6: New heat loss values of a buried system at a 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 5â&#x20AC;&#x201C;13 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 dedication to all the heat loss measurements performed during this research.

[1] EN 15632 District heating pipes, Pre-insulated flexible pipe systems, Requirements and test methods [2] P. Wallentén; Lund Institute of Technology, Sweden; 1991, Steady-state heat loss from insulated pipes

FURTHER INFORMATION Questions concerning the paper may be addressed to:

[3] E. van der Ven, F. Duursma, H. Korsman, I. Smits; Paper on DHC, Tallinn; 2010, Verification of heat loss measurements conducted on (semi) flexible pipe systems

Thermaflex International Holding B.V, Veerweg 1 5145NS Waalwijk The Netherlands www.thermaflex.com

[4] H. Korsman; Paper on DHC, Tallinn; 2010, Heat loss optimization of flexible plastic piping systems, life time heat loss performance

Liandon B.V. Dijkgraaf 4 6920AB Duiven The Netherlands www.liandon.com

[5] C. Engel and G. Baars, ―New economical connection solution for flexible piping systems‖, 12th ISDHC 2010.

118

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

COMPARISON OF COMPETITIVE (SEMI) FLEXIBLE PIPING SYSTEMS BY MEANS OF HEAT LOSS MEASUREMENT 1

I.M. Smits , J. Korsman , J.T. van Wijnkoop and E.J.H.M. van der Ven

Liandon B.V. Thermaflex International Holding B.V. Notice that this study does not compare entire district heating systems. For system comparisons see „Heat loss analysis and optimization of a flexible piping system‟ by J. Korsman et al. [2].

ABSTRACT Different types of pre-insulated pipes are tested on their heat loss values. Three flexible pipes and a rigid pipe are tested. The different heat loss values are compared not only on absolute heat loss, but also on their performance relative to the insulation surface. The heat loss values are measured according to EN 15632 and published as ―declared values‖. The declared values are calculated according to EN 15632 Annex D1-D3.

NOVELTY AND MAIN CONTRIBUTION Where most studies only focus on one product this study compares different types of flexible pre-insulated pipes on their practical heat loss values and gives an explanation of the practical heat loss values. It also compares flexible pipes with rigid pre-insulated pipes on an equal basis.

The flexible pre-insulated systems, with PE and PE-X foams, show a variance of up to 5 W/m in the heat loss values. These absolute differences in the system are caused by the outer casing dimensions of the preinsulated pipes. Recalculation to the same outer casing diameters shows a slight advantage for the PE system in service pipes of 32 and 63 millimetres

BRIEF METHOD DESCRIPTION First a brief description of different types of flexible preinsulated pipes and a rigid pre-insulated pipe is given. This chapter highlights the differences and similarities. The different types of foam for plastic pre-insulated pipes are described in a separate paragraph.

The flexible piping system with the PUR insulation foam on the other hand performs better compared to equally dimensioned flexible PE and PE-X insulation foams.

Secondly, the method of testing is briefly addressed. Thirdly, the different types of flexible pre-insulated pipes are tested on their absolute heat loss just after production.

Flexible pre-insulated pipes have a higher heat loss compared to rigid pre-insulated pipes. Recalculation to the same transport capacity [kg/s] and the same outer casing diameter also shows that rigid pre-insulated pipes perform better. However the fact that smaller diameters show a smaller heat loss difference between rigid and flexible pre-insulated pipes is interesting.

Since heat loss of pre-insulated pipes can increase over time due to degassing of the insulation foam, a gas analysis is performed on all test samples. In the second paragraph the absolute heat loss values of the different types of flexible pre-insulated pipes are compared on the basis of service pipe dimensions. The third paragraph defines a comparison on the basis of insulation surface and service pipe dimension.

INTRODUCTION & OBJECTIVE The objective of this research is to compare different types of competitive (flexible and rigid) pre-insulated pipes on their differences in heat loss values. The comparison is based on an overall heat loss measurement under similar conditions. Overall heat loss is determined for different samples of pre-insulated pipes, by using newly developed heat loss testing equipment as described in ‗Verification of heat loss measurements‘ by J.T. van Wijnkoop et al. [1]. The heat loss data of these flexible pipes will be compared with practical measurement on a rigid pre-insulated pipe.

In the fourth paragraph the comparison of flexible pre-insulated pipes versus a rigid pre-insulated pipe is described. The comparison in the third and fourth paragraph is based on ―declared values‖. The defined conditions are: (1) thermal conductivity of soil: 1.0 W/m.K, (2) thermal transmittance factor of earth to ambient air: 0.0685 m2.k/W and (3) soil covering: 0.8 m. In the first paragraph absolute heat loss values are compared on the basis of corresponding service pipe dimensions. The second paragraph gives a comparison based on equal transport capacity for flexible and rigid 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 second comparison by adding a heat loss value based on insulation surface.

barrier (EVOH) is placed in Polyurethane (PUR) insulation foam with a corrugated outer casing of high density poly-ethylene (HDPE). This product shows a tight connection between the service pipe, the foam and the outer casing. This product therefore cannot be re-used once it is formed.

Finally, both flexible and rigid pre-insulated pipes are compared, resulting in conclusions concerning flexibility versus heat loss behaviour. PRE-INSULATED PIPES This paper compares different types of pre-insulated pipes and highlights their mutual similarities and differences. The flexible pre-insulated pipe systems are;  Two different types of Cross linked Polyethylene (PEX) service pipe with Cross linked Polyethylene (PEX) insulation;  One type of Cross linked Polyethylene (PEX) service pipe with Polyurethane (PUR) insulation; 

Fig. 2. Section view of PEX/PUR pipe

PB service pipe with PE insulation The third type of pre-insulated pipe is a flexible PB/PE pipe. Figure 3 shows the cross section view of the PB/PE/PE pre-insulated pipe. A Polybutene (PB) service pipe with anti-oxygen barrier (EVOH) is placed in a low-density poly-ethylene (LDPE) insulation foam with a corrugated outer casing of high density Polyethylene (HDPE).

One type of Polybutene (PB) service pipe with Polyethylene (PE) insulation.

The rigid pre-insulated pipe system is; 

One type of Steel (ST) service pipe with Polyurethane (PUR) insulation.

The PB service pipe makes it possible to use electro fusion welding with a PB coupling. This makes a strong bond. Corrosion is not an issue, because PB is inert with water.

Firstly all types of pre-insulated pipes are functionally explained. Secondly the different kinds of foam production methods are described. All types of preinsulated pipes described are commonly available products used for district heating purposes in Europe.

There is no connection between the PB and foam plus outer casing. Therefore it is possible to re-use both elements in own production. The complete product can be re-used.

PEX service pipe with PEX insulation A short description of the PEX/PEX systems is given. Figure 1 shows the cross section view of the PEX/PEX/PE pre-insulated pipe. A cross linked Polyethylene (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.

Information regarding the use of this product is given in ‗‘New economical connection solutions for flexible piping systems‘ (Engel) [7].

Fig. 3. Section view of the PB/PE pipe

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

Fig.1. Section view of the PEX/PEX pipe

PEX service pipe with PUR insulation This paragraph describes the flexible PEX/PUR preinsulated pipe. Figure 2 shows the cross section view of the PEX/PUR/PE pre-insulated pipe. A cross linked poly-ethylene (PE-Xa) service pipe with anti-oxygen 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 connection between the service pipe, the foam and outer casing. Once it is formed, this product cannot be re-used in own process.

extruder. Once the material flows out of the extruder, the pressure drop causes the expansion of the hydrocarbon. The aggregation state of the molecule changes from liquid into gas. Examples of hydrocarbon gases that can be used are: LPG, Butane or Isobutane. And just like the PUR foam there is a degassing effect: the exchange of the blowing agent with air will increase the heat loss of the product. This effect is shown in ‗Heat loss of flexible plastic pipe systems, analysis and optimization‘ by E.J.H.M. van der Ven et al. [4].

Fig. 4. Section view of Steel/PUR pipe

Cross linked Poly-ethylene foam (PE-Xa)

Foam production processes

Although cross linked PE foam (x-PE) is also made of PE, there is a big difference compared to PE foam: the type of Blowing agent.

The insulation foams described in this paper are made of Poly-urethane (PUR) foam, Poly-ethylene (PE) foam or cross linked Poly-ethylene foam (PE-Xa). These foams have different properties. Some of these properties influence the heat loss properties of the complete product.

The foaming process to make x-PE foam is called the ―chemical foaming process‖. In this case a chemical is mixed into the PE matrix. The blowing agent can for instance be Azodicarbonamide. While heating the matrix, the chemical starts decomposing and gases are released. These gases are Carbon dioxide and Nitrogen. The thermal conductivity of these gases is more or less equal to the thermal conductivity of air. So the aging effect of this product in relation to the heat loss is less.

Polyurethane (PUR) foam PUR foam is a thermo-set foam. It is made out of two chemicals, a Poly-alcohol and an Iso-cyanate. These materials react and the Polyurethane is formed. This reaction is irreversible, so the material can never return into its original chemicals. The blowing agent for this kind of foam can be Carbon Dioxide, Nitrogen or Hydrocarbon molecules, for instance Cyclopentane or Butane.

To make this foaming process possible, it is necessary to connect the Poly-ethylene chains with each other. This is called the cross link. The complete process to make x-PE foam is called the ―chemical foaming process with cross link‖. To make the comparison with PE foam complete: this process is called the ―physical foaming process without cross link‖.

If Hydrocarbon gases are used, these gases strongly influence the heat loss performance of the preinsulated system. These gases have different thermal conductivities compared with air. After production of the foam an exchange with air starts. A product that is freshly made contains a high percentage of Hydrocarbon gases. At this point in time the product will have the lowest heat loss possible. If the same product is for instance three years old it contains more air and less Hydrocarbon gases due to gas diffusion. And so the product will have a higher heat loss compared to the fresh product. The process of degassing is described in research papers ‗Long term heat loss of plastic Polybutylene piping systems‘ by S. de Boer et al. [3].

The blowing agent is not the only additive that influences the thermal conductivity of the foam and therefore the heat loss properties of the pre-insulated system. Also other additives can influence the thermal conductivity of PE foam. Nucleating agents will influence the cell structure of the foam. As a basic rule: the finer the foam the lower the thermal conductivity. With this additive the convection part of the insulation material will be influenced. Another additive that influences the thermal conductivity is an anti-radiation additive. By using this special kind of additive it is possible to create a reflection of radiation energy.

Poly-ethylene (PE) foam PE foam is a thermoplastic foam. Once it is formed, it can go back to its original state by heating it above its melting point. Because of this property, it is possible to re-use these kinds of foams.

HEAT LOSS TEST METHOD This chapter briefly describes the test rig and test method used to determine the absolute heat losses of the different types of pre-insulated pipes.

The foaming process to make PE foam is called the ―physical foaming process‖. A Hydrocarbon molecule is mixed into the PE matrix under high pressure in an 121

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Test rig

Measured samples

The test rig is used to determine the absolute heat losses. The test rig has been designed in compliance with EN 15632 and the tests are carried out according to ISO 8497 and EN 15632.

The table below contains all products and dimensions of outer casing and service pipe that are treated in this paper. Table 1: Measured products

The physical part of the Thermaflex heat loss equipment consists of three sections. The first is the water cooled compartment in which all the tests are performed. This compartment is kept at a constant 23 °C during each measurement.

ST/PUR Dc/Ds

Heating probes are used as a heat source. These heating probes are custom made by preparing a two meter Thermaflex piping segment of all available diameters. The third part of the heat loss equipment is the control unit. This unit powers the probes and regulates the temperature and reads out the temperature and power values.

Dc/Ds

62A32

90DN25

90A32

125DN50

125A63

160A63

175A63

162DN80

162A110

200A110

Soil covering 0.8 m

o 

Thermal transmittance factor of earth-air 0.685 m2.K/W

o 

Thermal conductivity of the soil 1.0 W/(m.K)

The heat loss is calculated using the following formulas:

For more information concerning the test rig and method of testing see the paper ‗Verification of heat loss measurements‘ (J.T. van Wijnkoop et Al. [1])

All products that are involved in this paper have been analyzed on quantity of blowing agent and type of blowing agent. The following results were found: The samples of PEX/PEX I and PEX/PEX II did not show any amounts of hydrocarbon blowing agents;

 Tflow  Tsurrounding   Rsoil  Rflow  

(1)

Q  

(2)

Z  

 d4  H  2

(3)

Zc  Z  R0 soil

(4)

 4 Zc  1 Rsoil   ln  2    soil d4

Blowing agent analysis

The samples of ST/PUR and PEX/PUR products contained a mixture of hydrocarbon gases. These gases were analyzed. Both product types contained approximately 95% of blowing agent.

Dc/Ds --



Information



Dc/Ds

PEX/PEX I

This standard describes in Annex D a method to present the results of testing in end-use condition. This means: the product is buried in soil. According Annex D.3 the following general values are used for the calculation:

The heat loss measurement is done by measuring the energy required to keep the probe at a constant temperature, by measuring the current at constant voltage in the heating coils and calculating the power consumption. Since the middle/testing coil is exactly one meter in length the required energy represents the exact heat loss through one meter of piping and insulation in W/m. For this paper the heat loss is determined for multiple probe temperatures.

The PB/PE samples contained a quantity of hydrocarbon blowing agent over 50 percent;

Dc/Ds

PEX/PEX II

All results have been extracted from measurements carried out by the Thermaflex testing rig. The new European standard EN 15632 has been used.

Different heating probes are used for the testing. The probe with the appropriate diameter is inserted in a test sample and inserted in the cooled test section.



PB/PE

Results of testing

Method of testing



PEX/PUR





m K W

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

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 Celsius).

The 32, 63 and 110 millimetre service pipes First the absolute heat loss is displayed, followed by the insulation area analysis.

Table 2: Results according to EN 15632 at a temperature difference of 60 Kelvin. Product

Type

Heat Loss

Λ System

Buried system

(W/m.K)

Absolute heat loss In this paragraph all absolute heat loss values are compared for the 32, 63 and 110 millimetre service pipes.

W/m ST/PUR

90DN25

11.6

0.042

ST/PUR

160DN80

16.0

0.033

PEX/PUR

162A110

22.3

0.049

PB/PE

63A32

15.2

0.038

PB/PE

90A32

12.8

0.044

PB/PE

125A63

22.0

0.056

PB/PE

200A110

27.4

0.068

PEX/PE II

90A32

16.6

0.057

PEX/PE II

160A63

17.6

0.055

In Graph 1 the results are displayed for temperature difference of 60 Kelvin. PB/PE

PEX/PEX I

PEX/PEX II

PEX/PUR

35.0

Heat Loss [W/m]

30.0 25.0 20.0 15.0 10.0 5.0 0.0 90/32

160/63

200/110

Diameter service pipe [mm] PEX/PE II

200A110

31.1

0.073

PEX/PE I

140A32

12.5

0.057

PEX/PE I

175A63

17.6

0.059

PEX/PE I

200A110

28.8

0.051

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

The products based on PE or PE-x foam show higher heat losses for the 110 mm service pipe than the system based on PUR foam. The difference is approximately 20 percent. The different test samples show a wide variance in the diameter of the outer casing.

COMPARISON OF FLEXIBLE PLASTIC PRE-INSULATED PIPES

Therefore, only the results for the 32 millimetre service pipe are comparable for PB/PE 90A32 and PEX/PEX II 90A32. For the 110 mm service pipe, a comparison can be made between the PEX/PUR 200A110, PB/PE 200A110, PEX/PEX II 200A110 and PEX/PEX I 200A110.

This chapter compares the flexible pre-insulated pipes. The comparison is based on three diameters representing the entire diameter range for plastic preinsulated pipes. The comparison is expanded by evaluating the heat loss in correlation to the outer casing diameter (resp. the foam area).

Another difference in this comparison is the use of a PB pipe or a PE-x pipe. PB and PE-x have different thermal conductivities (0.19 W/m.K versus 0.40 W/m.K). However, this effect is already corrected by using the Wallentén equation [5], as shown in (1).

In Table 1 the flexible plastic pre-insulated pipes are defined. These are the products PEX/PEX, PEX/PUR and PB/PE. For more information concerning the PB/PE preinsulated pipes see „Heat loss of flexible plastic pipe systems, analysis and optimization‟ by van der Ven et al. [4].

Insulation area To compare the different kinds of flexible pre-insulated pipes on their performance, all outer diameters are altered towards 90, 160 and 200 millimetres. The 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 of Wallentén [5], as in (5).



2  Tprobe  Tcasing

 

  i 

1 s

is expanded by evaluating the heat loss in correlation to the foam area.



Rigid Pre-insulated pipes

 d2  1  d3  1  d4     ln    ln  d1    i  d2   c  d3 

 ln

The different systems and their corresponding dimensions are represented in Table 1.

(5)

The rigid pipe product that has been tested according to EN 15632 was the ST/PUR product.

Where:



Tprobe/Tcasing =

Probe / Casing temperature



d1 to d4 = inner/outer diameters of service pipe and casing



λs, λi, λc = heat coefficient of service pipe, insulation and casing

First the absolute heat loss is displayed, followed by a recalculation towards transport capacity and finally the insulation area analysis. Absolute heat loss In this paragraph all absolute heat loss values are compared. The rigid DN25 pipe service pipe is compared with a flexible PB/PE-x service pipe with an outer diameter (OD) of 32 mm. DN50 is compared with OD 63 mm and DN80 is compared with OD 110 mm.

Graph 2 represents the comparison on the basis of the same outer casing. PB/PE

PEX/PEX I

PEX/PEX II

PEX/PUR

In Graph 3 the results are displayed for temperature differences of 60 Kelvin.

35.0

25.0

PB/PE

20.0

35.0

15.0

30.0

10.0

25.0

Heat Loss [W/m]

30.0

5.0 0.0 90/32

160/63

200/110

Diameter casing/service pipe [mm]

PEX/PEX I

PEX/PEX II

PEX/PUR

ST/PUR

20.0 15.0 10.0 5.0

Graph 2 Relative Heat Loss 32, 63 and 110 mm service pipe, all with an equal outer casing (dT=60K).

0.0 DN25-PB32

Result analysis The flexible pre-insulated systems, with PE and PE-x foams, show a variance in heat loss values. The absolute differences in the system are caused by the dimensions of the pre-insulated pipes and the quantity and type of blowing agent that has been used. Also the recalculation to the same outer casing diameters shows an advantage for the PE foamed system in PB service pipes of 32, 63 and 110 millimetres.

DN50-PB63

DN80-PB110

Diameter service pipe [mm]

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

The different test samples show a wide variance in the diameter of the outer casing. The heat loss for ST/PUR 160DN80 is much lower compared to the heat loss of the 200A100 flexible piping products.

COMPARISON OF FLEXIBLE PLASTIC PREINSULATED PIPES VERSUS A RIGID PIPING SYSTEM

Even the difference with the PUR based PEX/PUR system is high (28 percent). For the PE and PE-x foam based products the difference is even higher (42 percent)

In this chapter the flexible pre-insulated pipes are compared with a rigid piping system. The comparison is based on diameter. The comparison

The heat loss for ST/PUR 90DN25 is more or less comparable with the heat loss for PB/PE type 90A32 (9 percent). So it seems that for smaller sizes the 124

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

difference in absolute heat loss is lower, compared to the absolute heat loss difference for larger sizes.

PB/PE

PEX/PEX I

PEX/PEX II

PEX/PUR

ST/PUR

40.0

Transport capacity

35.0

Heat Loss [W/m]

When comparing rigid steel service pipes with flexible plastic service pipes there is a difference in transport capacity for comparable diameters. This paragraph calculates the amount of heat loss when transporting water of 70 degrees Celsius through one meter of steel DN25, DN50 and DN80 (k-factor = 0.07 mm, velocity = 1.0 m/s). Subsequently the same amount of heat loss is used as a reference for calculating the amount of water that can be transported through a plastic pipe 32 and 63 (k-factor = 0.007 mm, velocity = 1 m/s). For these calculations the thesis of Colebrook and White [6] is used. The results of this calculation are displayed in Table 3.

Velocity [m/s]

DN2 5

PB32

DN5 0

PB63

DN8 0

PB110

1.0

1.072

1.0

0.81

1.0

0.76

Head Loss [Pa/m] Flow [kg/s]

425

0.64

188

0.57

2.33

5.35

20.0 15.0 10.0

0.0 90/DN25-PB32 160/DN50-PB63

200/DN80PB110

Diameter service pipe [mm] Graph 4 Relative Heat Loss, all with an equal outer casing and transport capacity.

Result analysis Flexible pre-insulated pipes have a higher absolute heat loss compared to rigid pre-insulated pipes. Recalculation to the same transport capacity [kg/s] and the same outer casing diameter shows that rigid preinsulated pipes perform better.

112

1.68

25.0

5.0

Calculation results transport capacity

Table 3:

30.0

The reason for this difference is the relative small inner diameter of the plastic service pipes. The low k-factor can not compensate for the smaller diameter. Table 4 shows the steel versus plastic service pipe diameter dimensions.

4.86

Next the absolute heat loss is recalculated to an equal flow per diameter. The basis is 0.57 [kg/s] for the DN25/PB32, 1.68 [kg/s] for the DN50/PB63 and 4.86 [kg/s] for the DN80/PB110.

Table 4: Service pipe diameter dimensions

ID [mm]

Insulation area To compare the flexible pre-insulated pipes and the rigid piping system on their performance, all outer diameters are altered towards 90 and 160 millimetres and compared on the same transport capacity. The corresponding heat loss is calculated using the thesis of Wallentén, as in (5). The steel DN80 and PB 110 has the same outer casing and is not recalculated.

DN25

PB32

DN50

PB63

DN80

PB110

28.5

26.0

54.5

51.4

82.5

90.0

However the fact that smaller diameters show a smaller heat loss difference between rigid and flexible preinsulated pipes is interesting. CONCLUSIONS This chapter briefly addresses each chapter and outlines its conclusions.

Overall comparison

The comparison on basis of the same outer casing and transport capacity is shown in Graph 4.

Test samples This paper compares different types of pre-insulated pipes that have been randomly taken from the market. The flexible pre-insulated pipes compared in this paper;  PB/PE,  PEX/PEX I,  PEX/PEX II,  PEX/PUR. 125

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The flexible piping system with the PUR insulation foam on the other hand performs better compared to flexible PE and PE-X insulation foams with equal dimensions.

The rigid system in this paper;  ST/PUR. Method of testing All heat loss tests are performed on a test rig that has

been designed in compliance with EN 15632. The tests are carried out according to ISO 8497 and EN 15632.

Comparison of flexible piping system versus the rigid pre-insulated pipes Flexible pre-insulated pipes have a higher absolute heat loss compared to rigid pre-insulated pipes. Recalculation to the same transport capacity [kg/s] and the same outer casing diameter shows that rigid preinsulated pipes perform better.

Blowing agent analysis All measured products are checked on type of gas and gas content. The ST/PUR and PEX/PUR products contain approximately 95 percent of blowing agent.

However the fact that smaller diameters show a smaller heat loss difference between rigid and flexible preinsulated pipes is interesting.

The PB/PE product range has a quantity over 50% of blowing agent. In the products of PEX/PEX II and PEX/PEX I no Hydrocarbon gases were detected.

To be comparable in heat loss some dimensions of the flexible piping systems range need to be optimized. However, other advantages of flexible pipe systems, for instance the potential decrease of service meters because of a curved layout-design, can partly compensate the higher heat loss compared to the rigid system (see ‗Heat loss analysis and optimization of a flexible piping system‘ by J. Korsman et al. [2]).

Comparison of flexible pre-insulated pipes A fair comparison is difficult because of differences in outer casing and other dimensions. These conclusions are therefore only valid for the products that have been tested for this paper. In a buried condition the PB/PE pre-insulated pipe shows for equally dimensioned pipes 90A32 and 200A110 the lowest absolute heat loss values for all pre-insulated pipes based on PE or PE-x foam.

ADDENDUM Significant product improvement of the PB/PE/PE pipe system has led to a decrease in heat loss [4]. Graph 5 is updated with these improvements resulting in the comparison displayed in Graph 6. The new samples are displayed under the name of PB/PE II.

As mentioned before, the absolute differences in the system are caused by the dimensions of the preinsulated pipes. Recalculation of the same outer casing diameter shows also an advantage for the PB/PE system in service pipes of 32, 63 and 110 mm. See Graph 5.

PB/PE PEX/PUR

PB/PE II ST/PUR

PEX/PEX I

PEX/PEX II

40.0 PB/PE

PEX/PEX I

PEX/PEX II

PEX/PUR

ST/PUR

35.0

Heat Loss [W/m]

40.0

Heat Loss [W/m]

35.0 30.0 25.0 20.0

30.0 25.0 20.0 15.0 10.0

15.0

5.0

10.0

0.0 90/DN25-PB32 160/DN50-PB63

5.0 0.0 90/DN25-PB32 160/DN50-PB63

200/DN80PB110

Diameter service pipe [mm]

200/DN80PB110

Graph 6 Relative Heat Loss 32, 63 and 110 mm service pipe, all with an equal outer casing and transport capacity (dT = 60 K)

Diameter service pipe [mm]

Graph 5 Relative Heat Loss 32, 63 and 110 mm service 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, ―Heat loss analysis and optimization of a flexible piping system‖, in Proc. of the 12th International Symposium on District Heating and Cooling, Tallinn, Estonia (2010).

Questions concerning the paper can be addressed to: 



Thermaflex International Holding B.V. Veerweg 1 5145NS Waalwijk The Netherlands www.thermaflex.com

[3] S. de Boer, J. Korsman and I.M. Smits, ―Long term heat loss of plastic Polybutylene piping systems‖, in Proc. of the 11th International Symposium on District Heating and Cooling, Tallinn, Reykjavik (2008).

Liandon B.V. Dijkgraaf 4 6920AB Duiven The Netherlands www.liandon.com

[4] E. J .H. M. van der Ven and R.J. van Arendonk, ―Heat loss of flexible plastic pipe systems, analysis and optimization‖, in Proc. of the 12th International Symposium on District Heating and Cooling, Tallinn, Estonia (2010).

ACKNOWLEDGEMENT We would like to thank all involved employees of Thermaflex Isolatie B.V. who made this research possible (especially H. Leunessen and M. van Doorn).

[5] P. Wallentén, ―steady-state heat loss from insulated pipes‖, Lund Institute of Technology, Sweden, 1991

Special thanks go to P. Blom and P. van Rijswijk for the dedication they showed in carrying out all the heat loss measurements during this research.

[6] C. F. Colebrook, "Turbulent flow in pipes, with particular reference to the transition region between smooth and rough pipe laws", February 1939

REFERENCES

[7] C. Engel and G. Baars, ―New economical connection solution for flexible piping systems‖, in Proc. of the 12th International Symposium on District Heating and Cooling, Tallinn, Estonia (2010).

[1] J. T. van Wijnkoop and E.J.H.M. van der Ven, ―Verification of heat loss measurement‖, in Proc. of the 12th International Symposium on District Heating and Cooling, Tallinn, Estonia (2010).

127

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

EFFECTIVE WIDTH â&#x20AC;&#x201C; THE RELATIVE DEMAND FOR DISTRICT HEATING PIPE LENGTHS IN CITY AREAS 1

Urban Persson , Sven Werner

School of Business and Engineering Halmstad University, PO Box 823, SE-30118 Halmstad, Sweden effective width becomes the width of an analogous rectangle with the trench length as the length and where the rectangle area is equal to the given land area.

ABSTRACT One key concept when assessing network investment cost levels for district heating systems is the linear heat density. In contrast to a traditional way of expressing this quantity entirely on the basis of empirical data, a recently developed analytical approach has made it possible to estimate linear heat densities on the basis of demographic data categories. A vital complementing quantity in this analytical approach is the concept of effective width.

The concept was introduced by Werner [3] and has been further elaborated recently in model estimations of distribution capital cost reactions to decreased heat demands in four north European countries [2]. Essential for calculations of anticipated investment cost levels for future district heating systems, the effective width constitutes an important model parameter indicating levels of network extensions in given land areas.

Effective width describes the relationship between a given land area and the length of the district heating pipe network within this area. When modelling distribution capital cost levels by use of land area values for plot ratio calculations, there is a potential bias of overestimating distribution capital cost levels in low dense park city areas (e < 0.3).

Since the concept of effective width itself is rather new, with no previous analytical or statistical use, data on effective widths are in principal non attainable within national statistical sources. Effective width might be regarded as an innovative model quantity with no previous representation in the field of district heating research.

Since these areas often include land area sections without any housing, avoiding overestimations of network investment costs demand some kind of corrective mechanism. By use of calculated effective width values, a compensating effect at low plot ratio levels is achieved, and, hence, renders lower anticipated distribution capital cost levels in low dense park city areas.

AIM The aim of this paper is to describe the concept of effective width and outline the basic properties of this quantity. On the basis of, although sparse, empirical observations, preliminary statements concerning the properties of effective width are made. The aim is further to enlighten the theoretical environment in which effective width contributes when applying demographic quantities for estimations of district heating network investment costs.

INTRODUCTION One key concept when estimating investment cost levels for district heating systems is the linear heat density, i.e. the quota of annually sold heat in a district heating scheme and the trench length of the piping system in this scheme (Qs/L) [1]. In contrast to a traditional way of expressing this quantity entirely on the basis of empirical data, a recently developed analytical approach has made it possible to estimate linear heat density on the basis of demographic data categories [2]. A vital complementing quantity in this analytical approach is the concept of effective width.

LIMITATIONS Due to a limited amount of empirical data, in principal less than 100 observations, the specific result values and relationships accounted for in this paper must be considered as preliminary. Although thorough in theory, the concept of effective width needs to be supported further by extended empirical data gathering. In order to be able to produce solid and reliable estimations of effective width values in different kinds of city areas, such information is considered vital for future use of the concept.

BACKGROUND Effective width is a stand alone concept within district heating theory, describing the relationship between a given land area, AL, and the length of the district heating pipe network, L, within this area. Hence, the 128

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

EFFECTIVE WIDTH

p = P/AL

[number/m2]

(7)

Effective width is a measure indicating the district heating network extension level within a given land area. The quantity effective width, which is symbolised by use of the letter w, with the unit metres, expresses the ratio between land area and the total trench length of the distribution network within a district heating system [3]

α = AB/P

[m2/capita]

(8)

P = Total population

[number]

AL= Total land area

[m2]

AB= Total building space area

[m2]

w = AL / L

The concept of effective width hereby plays a key role in the reformulation of the traditional expression for linear heat density, and hence, constitutes a central quantity in model estimations of the feasibility and viability of future district heating network. If linear heat density can be said to indicate the level of district heat distribution system utilisation, the effective width indicates the distribution system coverage of the land area at hand.

[m]

(1)

Being in this way the result of explicit area and grid properties, effective width can be used to describe typical district heating properties in different population density areas and hence, give information on prerequisite conditions for future district heat establishments. THE CONCEPT In order to introduce the concept of effective width, it is necessary to first understand some basic principals regarding the linear heat density. The concept of linear heat density, being the division of total annually sold heat in a district heating system and the total length of the district heating piping network, indicates the level of district heat distribution system utilisation. Furthermore, linear heat density is a denominator parameter when calculating district heating network capital costs.

LinearHeatDensity 

Qs L

[GJ/m]

THE PROBLEM From a district heating distribution point of view it is relevant to distinguish between two kinds of land area low plot ratio situations. The land areas can, principally, consist of either a wide dispersion of households spread out over the whole area (A), or households can be closely limited to only a fraction of the land area (B), see figure 1.

(2)

As has been put out in [2], this traditional presentation of the concept of linear heat density offers ―no entrance for estimations of future district heating systems, since none of the two quantities can be known for yet not built systems‖, which is the fundamental reason for reformulation of the expression by use of demographic quantities. If combining the two concepts of population density (p) and specific building space (α) into the city planning quantity plot ratio (e), which is suggested in [2], the concept of linear heat density can be alternatively expressed as;

Qs  qew L

[GJ/m]

Figure 1. Low plot ratio land areas, scenario A with wide dispersion of buildings and scenario B with high concentration of buildings.

In the first case (A), a district heating distribution grid would have to cover all of the land area at hand in order to deliver heat (at very low linear heat density), while in the latter case (B), the grid could be narrowed down to the limited area fraction. If, when conducting district heating feasibility model analysis, plot ratios are extracted by means of (5), it would be relevant and recommended to somehow adjust the land area magnitude in order not to include non-targeted area fractions. An adjustment to reach this purpose can be achieved in several different ways, of which Effective Width compensation suggested in this paper is one option.

(3)

The three new parameters, specific heat demand (q), plot ratio (e) and effective width (w), are defined as: q = Q/AB

[GJ/m2a]

(4)

e=pα

[1]

(5)

w = AL/L

[m]

(6)

where 129

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

plot ratio values above 1. This would indicate that the relationship between high dense inner city land areas and the length of the required piping grid in such areas is constant.

DATA AND VALUES In the spring of 2009, the authors, both being lecturers at Halmstad University in Sweden, initiated a pre-study to be carried out by two Bsc-students at their department [4]. The study was two-fold in regard of gathered data. Partly it delivered previously assembled and crucial data on plot ratios, land areas and trench lengths in 39 detached house districts heating schemes in Sweden [5], allowing estimations of effective widths in these districts, see Figure 2, and partly own collected data.

Still, if plotted explicitly, the function does not converge at any effective width value, no matter how far the plot ratio value is extended, but the rate of divergence decreases with higher plot ratio values. Since plot ratios values above 3 are considered extremely rare, effective width values within high dense inner city areas (plot ratio values above 0.5) can be anticipated to be found in the interval of 50 < w < 60 meters.

Effective width (w)[m]

Effective width (w) [m] 400

250 200

Detached houses

300

150

250

Power (Detached houses)

100 -0,3731

y = 27,802x 50

Detached and MF houses

350

Power (Detached and MF houses)

200 150

y = 61,838x-0,1495

100 0

50 0,0

0,1

0,2 Plot ratio (e)

0,3

0,4

0 0

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

Effective Width (w) [m] 140 130 120 110 100 90 80 70 60 50 40

Multifamily houses

250 Power (Multifamily houses)

200 150

y = 56,622x-0,41

100

1,5

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

Effective width (w) [m] 400 300

1 Plot ratio (e)

The own collected data of the study refers to data from 34 district heating schemes in multi-family housing districts in the Swedish cities of Halmstad and Gothenburg, see Figure 3.

350

0,5

1,5

2,5

Plot Ratio (e)

0 0

0,5

Plot ratio (e)

1,5

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 Sweden. Source [4]

For plot ratio values below 0.5, on the other hand (outer city area and park areas), the relationship is by no means constant, but diverges rapidly with increased effective width values as a consequence. At a plot ratio value of 0.04 the effective width reaches a value of 100 meters, and the curve reveals that the increase of effective width values at even lower plot ratio values below 0.04 renders values above 100 meters and beyond.

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

w  61.8  e 0.15

[m]

(9)

The graph characteristics of Figure 5 has significance for estimations of district heat distribution capital cost

As can be seen in Figure 4, the graph suggests a convergence at effective width values at 60 meters for 130

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

levels in park areas, since these areas often also include land area fractions without any housing, i.e. not to be targeted by district heating networks. When using crude statistical land area values for plot ratio calculations, there is a potential bias of overestimating distribution capital cost levels in these suburban areas, since actual habitations plausibly only occupy parts of the land area at hand. In these occasions, effective width values arrived at by use of eq. (9). have a compensating effect by rapidly increasing it‘s value at low plot ratio levels, and, hence, rendering lower anticipated distribution capital cost levels.

REFERENCES [1] Frederiksen S. and Werner S, Fjärrvärme – teori, teknik och funktion (District Heating – theory, technology and function). Studentlitteratur, Lund 1993. [2] Persson U. and Werner Competitiveness of District published.

S, The Heating,

Future to be

[3] Werner S, Fjärrvärme till småhus – värmeförluster och distributionskostnader (Sparse district heating – heat losses and distributions costs). Report 1997:11, The Swedish District Heating Association. Stockholm 1997.

CONCLUSION The main conclusion from this analysis is that the concept of effective width offers a new simple shortcut for quick estimations of capital investments for heat distribution in virgin urban areas.

[4] Netterberg H and Isaksson I, District Heating in

Slough. BSc thesis from Halmstad University, Halmstad 2009.

[5] [Andersson S et al, Nuläge Värmegles Fjärrvärme

This conclusion is especially valid if the effective width has almost a constant value over a plot ratio of 0.5 as preliminary stated from Figure 4. Further data collection will show how true this new finding will be.

(The current situation for sparse district heating), The Swedish District Heating Association, research report FoU 2002:74. Stockholm 2002.

131

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

INTEGRATING RENEWABLE ENERGY INTO LARGE-SCALE DISTRICT HEATING SYSTEMS Peter Begerow, Dr. Stefan Holler MVV Energie AG, Mannheim, Germany If the used technology just produces heat, a financial aid from the BAFA (Federal Office of Economics and Export Control, Germany) or KfW (bank under control of the Federal Republic, Germany) is possible. There are different regulations which have to be fulfilled by the project in order to be eligible for those subsidies [8]. Main criteria are the size and type of the investor and planner, the type of technology and the size of the heat plant and of the storage tank.

ABSTRACT Renewable energy for heating is mostly used in small systems for single-family houses. The existing district heating networks are generally run by large heating plants or combined heat and power plants fired with fossil fuels. To combine these two systems, a feasibility study was completed with a focus on the district heating grid in Mannheim, Germany, and with a focus on solar thermal heat. Other renewable energy heat sources, geothermal heat and heat from biomass, are included for a comparison.

Previous studies [2], [12], [16], [18] showed that solar thermal energy is mostly used in single-family houses and smaller heating grids combined with seasonal heat storage systems. Those systems are still in development and need financial support to be realized. Most of these heating grids run with a lower flow temperature and use either fossil fuels or heat from biomass for an auxiliary heat generation. The largest solar thermal district heat system in Germany is located in Crailsheim. It covers an area of approx. 7300 m² of solar collectors with two buffer tanks with a combined volume of about 500 m³ as thermal storage. In addition a seasonal geothermal storage has been built which will cover 50 % of the heat demand for about 2000 residents. This research project has heat production costs without any financial support of about 19 ct/kWh. This sum will be reduced depending on the possible subsidies. [12] Reported technical difficulties were mostly in the thermal storage technology. There were little problems with the collectors as common flat plate collectors were used which are commercially available and used in large numbers in smaller systems.

The study focuses on the heat price as a key figure to analyse the economic feasibility. The technical feasibility has been evaluated by using a simulation model of a secondary district heating grid, which is operated on a low flow temperature level of 70 °C and which is connected to a central solar thermal energy plant. The paper describes which technical and economic framework conditions are necessary for implementing renewable energy into large-scale district heating systems. The calculations show that in comparison with other renewable heat sources solar heat has the highest heat costs ranging from 7,7 ct/kWh to 14,5 ct/kWh depending on the plant size, the solar fraction and the use of a storage system. The major technical problems for integrating solar heat into a heat grid are the pressure difference between the flow pipe and return pipe and the low temperature the flat plate solar collectors are working with.

The project in Crailsheim has shown the technical feasibility of a system with a seasonal thermal storage, but it also shows that considerable costs are involved.

INTRODUCTION Based on the protocol of Kyoto and European regulations a high reduction of CO2 emissions in Germany is necessary. To achieve those goals, an expansion of renewable energy in the heat market is required. In Germany the major aim to reach is a share of 50 % renewable energy in the heat market by 2050. Furthermore, 50 % of the renewable heat is supposed to be contributed by a heating grid. [13]

Furthermore, if a single-family house will install a seasonal storage to get a solar fraction above 50 %, it needs more than 10 m³ of hot water storage (depending on the building type and planned solar fraction). However, in common buildings there isn‘t enough room for that size of storage [17]. Those two aspects show that the use of a heating grid could significantly reduce the costs of the solar thermal systems and could save space otherwise necessary for a storage tank.

To achieve these goals, different governmental as well as local support mechanism and financial subsidies are available. If the heat production is combined with an electricity production, the major financial support is based on the EEG (German law for renewable energy).

The paper will give an overview of how the expansion of renewable energy in the heat market will be possible 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 network. A detailed simulation for a solar thermal integration was done by using RETscreen [14] as simulation software.

is not part of the simulated area and on the reduction through population, which has not a direct effect on one special housing area. The reduction through influence of temperature has a share below 5 % within 15 years and is therefore not included within the simulation.

MATERIAL AND METHODS

Existing Systems Between newly built and existing heat networks there exist some main differences which have to be considered. If the network is designed especially for the renewable energy source, it can be technically specialized (e.g. forced low return temperature for building owners; special isolation of the used pipes). Older heating grids on the other hand are normally constructed for the heat production with fossil fuels and are normally designed for higher temperatures. Furthermore, in some heating grids a high temperature is necessary either for thermal cooling systems (e.g. absorption chillers) or for the heat transfer stations within the houses which are built for high temperatures (low flow temperatures need optimized heat transfer stations [12]. In the following, the main aspects for the integration of different sustainable heat generation technologies are described.

Evaluation of heat demand For planning a new heat production facility, the heat demand of the connected consumers is necessary. If those are existing households, the heat demand from the past can be used for calculations. For newly built houses the heat demand should be exactly calculated with the standards named in DIN V 4108-6. If this is not possible, the yearly heat demand can be assumed by the given figures: Table 1: heat demand [15] Building size

Heat demand (room heating)

heating demand (hot tap water)

[kWh/mÂ˛a]

1-2

72,3

More than 3

55,3

[housing units]

Heat grid for renewable energy For the integration of renewable energy into heat grids, different possibilities for the connection exist. Especially for the solar thermal energy production it is assumed, that more than one heat plant will be connected.

These figures can be realized in buildings constructed between 2011 and 2020. [15] Another factor for the planning of a heating grid is the outlook into the future, because the payback period of a renewable heat production facility is very long.

The three options are:

The following graph shows the expected change in heat demand for Germany focusing on different factors of influence:

Taking water from the return pipe, heat it and return it into the return pipe

Taking water from the flow pipe, heat it further and return it into the flow pipe

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 can be done with equation (1). It gives the pump work W depending on the necessary heat flow â&#x2C6;&#x2020;Q, the

Within the following simulation this development is not further regarded. The major reduction within whole 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

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

Used Software RETscreen is a program to make first feasibility studies of all kind of green energy projects. In terms for solar thermal heating, it uses an included weather database to calculate the expected heat production. Furthermore a product database is included with the necessary technical parameters for many different solar thermal collectors. The needed amount of heat can either be calculated other ways or assumed by the software depending on the amount and size of buildings. Combining those input factors with others, the simulation tool gives a recommendation of the used number of solar collectors and the size of a thermal storage system. If all input factors are included the program calculates the yearly heat production and the solar fraction. Beyond that, the program can be used to include a second heating system for the remaining needs to get the final payback period and the total emissions. For the simulation of this paper the version 4 (November, 2009) of the named software was used.

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

Financial Calculation The calculation of the heat costs is based on the net present value method. For the internal rate of return the given value was used, all other costs included and the heat costs varied to get a net present value of zero. This method gives the current heat price and a further increase during the next years is included. This makes it possible to compare the actual heat price to the given values of other systems. For the economical calculation in the conclusion of this paper, a competitive heat price from now on was realized.

Description of selected site For modeling the integration of solar thermal energy into a district heating network, a yet to be built housing estate was selected. This housing area is planned with a district heating grid running at a flow temperature of about 70 °C. This area is connected with a heat exchanger to the central heating grid of the city, which is run with flow temperatures between 90 °C and 130 °C.

The named financial support which is included in the calculations are subsidies on the capital cost. They depend, like mentioned in the introduction, on different aspects. A research project like the one in Crailsheim, can get a higher support than commercial ones run by large companies. [1]

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

Number of buildings

Solar thermal heat production

Total heating demand [MWh/a]

For the heat supply of the given housing area, different scenarios based on solar thermal energy were developed. Using the RETscreen software tool the technical parameters of the flat plate solar thermal collectors, weather data from a climate database and the given heat demand of the area was combined for each scenario.

(room heating + hot tap water) 1-2

135

1561

More than 3

111

585

Sum

246

2146

The scenarios differ in the necessary amount of collectors needed to achieve a solar fraction of the total heating demand of 50 % [scenario 1], a 100 % solar heat production of the used hot tap water (which stays constant throughout the whole year) [scenario 2] and a

Table 2 gives an overview of the planned houses and their heat demand. The whole heating grid will have an length of about 1,3 km and the total heat demand will be 2146 MWh/a. 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 a thermal storage system [scenario 3].

For a comparison of the different scenarios the heat cost per kWh were calculated.

For the thermal storage a hot water system is assumed, because those are state of the art and can be used in most applications. Other systems have more specific requirements to the geological situation of the area. A geothermal heat storage for example does not work in an area with a flow of the ground water. For the simulation of scenario 2 a smaller thermal storage compared to scenario 1 was assumed, because there is no necessity for a seasonal heat storage system.

The calculation of the emissions is based on the operation of the system and not on its total life cycle. For solar thermal heat the CO2 emissions only arise from the used electricity for the necessary pumps. Included in the calculation is only the pump energy for the solar thermal collectors and, if necessary, to increase the pressure for the integration into the heating grid flow pipe. The CO2 emissions for the German electricity grid are given with 506 g/kWh. For the calculation without a thermal storage system [scenario 3] it was assumed that the produced solar heat can directly be distributed throughout a district heating network. This would make it possible to save the investments of a seasonal heat storage system and also reduce the losses within the thermal storage system.

For the simulation model a commercial solar collector was taken (s. Table 3). It is a flat plate collector with a anti-reflection glass and a gross area of about 2,6 m². Its efficiency is 84,4 % (calculated according to EN 12975). The simulation uses specific given parameters. Those are shown in Table 3.

For those calculations the same heat amount was used than in scenario 2. But in this case it is not possible to cover 50 % of the heat demand of the total grid. Just a small amount, for example the losses of the grid and the base load, can be produced with solar thermal technologies without a thermal storage.

Table 3: used input parameters for solar simulation

Flow temperature

Scenario 1: 1076 MWh Scenario 2: 493 MWh Scenario 3: 1071 MWh Scenario 4: 11,6 MWh Scenario 5: 11,6 MWh 67 °C

Return temperature

45 °C

Slope of collector

55°

Azimuth of building

-45° (southeast)

Type of collector

WagnerSolar L20 AR

Storage capacity

Scenario 1: 1000 l/m² Scenario 2: 100 l/m² Scenario 3: 1 l/m² Scenario 4: 100 l/m² Scenario 5: 10 l/m² 80 %

Annual heating energy (calculated with given method)

Heat exchanger efficiency Miscellaneous losses

Pump efficiency (for grid integration) Time period Internal rate of return

Increase of heat price per year Financial support

Another option would be to integrate small systems into the district heating grid. In this case the operator of the grid would not run the facility by itself. The heat producer could use a solar thermal collector for its own heat demand but without a thermal storage system. Instead of using an in-house thermal storage (what is getting very large if a seasonal heat storage system is used) the heating grid could be used. For the single house technology an internal rate of return of 5 % was used for the economic calculation (average percentage of building credit [4]). Furthermore the financial support is a little different because of different regulations for large and small systems. In the following those two calculations are named ―scenario 4‖ for the heat production of a single-family house with a thermal storage and ―scenario 5‖ for the calculation without a thermal storage.

5 % if storage is used (smaller grid) 8 % if integrated into large grid 40 %

Summary of different scenarios:

20 year

Scenario 1

Scenario 1: 8,5 % Scenario 2: 8,5 % Scenario 3: 8,5 % Scenario 4: 5,0 % Scenario 5: 5,0 % 2%



Solar fraction of 50%



Seasonal thermal storage included

Scenario 2

30 %

135



100% heat production of hot tap water



Buffer heat storage included, but no seasonal thermal storage

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Scenario 3 

Same amount of heat produced than in scenario 1



No storage; connected to large district heating grid

Heat production with biomass The heat production from biomass is technically very similar to the fossil fuel powered heating plants. Therefore the integration into existing district heating grids is the easiest way compared to the other renewable energy sources.

Scenario 4 

Solar fraction of 100% for a single family house



Seasonal thermal storage included

The exact technology depends on the used fuels and therefore the economic calculation is mainly based on the price development of the biomass.

Scenario 5 

Same amount of heat produced than in scenario 4



No storage; connected to large district heating grid

The emissions of such a system are by way of calculation zero, because the emitted CO2 was firstly bound by the biomass during its growing period. If the biomass is planted in an area which was deforested for that, the emissions are not zero any more. The former forest was a CO2 sink which does not exist anymore and should be included in the calculation. Furthermore the transport and processing of the biomass should be included. [19]

Geothermal heat production The geothermal heat can be used in various ways for room heating. Using the shallow geothermal heat is only possible in combination with a heat pump. Therefore, for a large integration into heating grids the deep geothermal energy is the favoured one. Furthermore in the upper valley of the river Rhein (Oberrheingraben) the geothermal heat can be used for a combined heat and power production because of its high temperature. In Germany this gives the possibility to get a payment for the electricity based on the EEG which grows for 3 ct/kWh if the heat is used as well.

Fossil fuels for comparison In our days the district heating grid in Mannheim is fed with heat from a fossil fuel fired CHP plant. The heat prices from that system are much lower than the renewable heat. Looking into the future it mainly depends on the price development of CO2 emissions and the coal price. [8]

For a comparison to the solar thermal heat a geothermal power plant in Landau, Germany is used as a reference.

The emissions of such a system are very high, even if the used heat is more or less waste heat. To reduce those, a CCS technology can be implemented in the future.

This project began in 2004 and at the end of 2007 the power plant started its first electricity production. The first heat output was planned for 2009. The power plant uses the ORC process (Organic Rankine Cycle) to generate electricity. A drill hole with a depth of 3000 m connects to thermal water with a temperature with up to 160 °C which is cooled down during electricity production to 70 °C. The whole yearly energy output of the power plant is planned to be 22.000 MWh electricity and 9.200 MWh heat. One of the major benefits of the geothermal heat production is the base load which is always available. On the other hand this gives the problem that the heat is also available in the summer time and needs to be cooled down in other ways.

RESULTS The results of the simulation are shown in Tab. 4 and Fig. 2 and 3. In conclusion the heat price is lower if the collector area increases (economy-of-scale). Furthermore the use of a district heating grid instead of a thermal storage lowers the heat cost extremely. For scenario 1 it is necessary to install a gross area of 3080 m² solar thermal collectors. 1076 MWh heat can be produced in combination with a 2820 m³ hot water storage. The heating costs calculated with the given framework conditions are 11,2 ct/kWh. To operate the collector area, pumps are needed which consume electricity. The emissions of that electricity are, based on the produced heat, 7,9 g CO2/kWh.

The calculated emissions of the power plant are 0 g CO2/kWh because the electricity production has no emissions and for the pumps the own electricity can be used. [6] Currently the power plant runs with a limited output due to small earthquakes in the area of the drilling hole and does not deliver heat until now. Additional geological studies are done right now and a heat output should start after they are finished.

In scenario 2, 1916 m² solar thermal collectors need to be installed. Combined with a hot water buffer storage with a volume of 175 m³, 494 MWh of heat can be produced. The financial calculation over 20 years lead 136

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

to heating costs of 14,5 ct/kWh. The emissions of such a system are 8,9 g CO2/kWh. If the heating amount of the 50 % scenario is used but without a storage [scenario 3], and therefore without those losses, a much smaller collector area is calculated. However, the produced heat has to be used directly within a large heating grid. For such a system 2542 m² of solar thermal collectors are needed which produce 1071 MWh/a. The smaller collector area and the elimination of a storage system give heating costs of 7,7 ct/kWh. On the other hand the emissions of such a system are higher because of the necessary pump energy for the pressure compensation. The total specific emissions of that system are 12,2 g/kWh.

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

Under consideration of scenario 4, the heat costs are 13,8 ct/kWh within a single family house. If a heating grid would be used for storage and therefore no large thermal storage is necessary, the heat costs can go down to about 11 ct/kWh. If the losses of the storage system are included in the calculation, a smaller gross area of collectors can be used. Combining all those savings, the heat cost for a single-family house can go down to 7,2 ct/kWh (scenario 5). This shows, that there is a wide margin and a high potential of cost reduction if a heat grid is used. But it has to be said, that those heat costs are still much higher than from other heat generating systems. Table 4 shows the technical results and parameters for each calculated scenario. Based on those figures the financial and ecological calculation where made. Those results are shown in Figure 1 and 3. For the renewable technologies the increase of 2% of the heat costs can easily be included in the calculation. For the fossil (and the biomass) use, the heat price is highly dependent on the fuel price development. Therefore a price range is given on those systems. The reason for the range for fossil CHP heat CO2 emissions is that different references are used.

Fig 3: heat costs of different system [1], [20]

CONCLUSION To make the solar thermal heat production economical compared to the other systems, different aspects have to be changed. In order to show the potential of cost reduction for solar thermal heat generation a sensitivity analyses has been carried out. The following parameters have been varied in order to reach a heat prices of around 3,5 ct/kWh in the beginning year. This is the actual heat price for private customers in Mannheim.

Table 4: output parameters of simulation scenario

Collector gross area [m²]

Produced heat [MWh/a]

Storage size [m³]



Future heat price development



Change of investment

3080

1076

2820



Amount of financial support

1916

493,6

175



Internal Rate of Return

2542

1071

11,6

137

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia Table 5: parameters for economical operation Scenario 1

Scenario 3

Heat costs

3,5 ct/kWh

3,4 ct/kWh

IRR

Capital cost

70 %

90 %

Financial support

40 %

30 %

Heat price development

8 % p.a.

For the near future it might get more interesting to look on the biomass and geothermal heat, particular if the heat is needed in a region where high temperatures in the depth could be exploited or cheap biomass sources are available. Further research in the solar collector technology is needed to lower the capital costs and equally within the thermal storage technology, as it might get interesting in the future to include those even in district heating grids with fossil fuels as heat source to cover peaks in the demand and transfer a surplus heat production from the summer into the winter season.

Fig 4: influence on the heat costs of different factors

Figure 4 shows the influence of the different factors, if the others stay the same. But it also shows, that by changing just one aspect, a reduction of the heat cost to 3,5 ct/kWh is only possible with lowering the capital costs by 50% of the scenario without a thermal storage [scenario 3]. Therefore a combination of different factors was done. The capital costs are also influenced by the financial support und were calculated separately. In order to reduce the heat costs down to about 3,5 ct/kWh in scenario 1, a reduction of the capital cost by 40% combined with financial support of 50% is necessary, if the heat price will rise with 8% per year. This is higher than shown in figure 4 and is based on a high price assumption as reported in [10].

NOMENCLATURE W [J] work of pump

If a lower IRR is assumed (5 %), the capital costs have to go down to 70 % and a financial support of 40% of the investment is necessary. For scenario 3, a rise of the heat price and the lower IRR (5 %) just need a reduction of capital costs of 10 % to achieve heat costs of 3,4 ct/kWh. In this case the assumed financial support of 30% stays the same. This shows that an economical use of solar thermal energy within district heating could be achieved. The financial support of 30% can be possible based on a KfW program, the capital costs of 90% of the base investment is possible within a feasibility study and the 5% IRR is an average figure for building loans.

∆Q [J]

heat flow

∆p [Pa]

pressure difference (flow / return)

∆T [K]

temperature difference (flow / return)

cp [J/(kg*K)]

heat capacity

ρ [kg/m³]

density

pump efficiency

REFERENCES [1] Begerow, P.; Integration von erneuerbaren Energien in Fernwärmenetze – Eine technische und wirtschaftliche Analyse aus Sicht eines Fernwärmeversorgers, Diplomarbeit an der Universität Flensburg, MVV Energie AG, Mannheim; 2010.

The high requirements to make the solar heat profitable show, that this technology is not advisable if other renewable energy sources are available. Furthermore for the technical integration a low temperature heating grid is necessary.

[2] Bodmann, M.; Mangold, D.; Nußbicker, J.; Raab, S.; Schenke, A.; Schmidt, T.: Solar unterstütze Nahwärmeversorgung und LangzeitWärmespeicher; Forschungsbericht zum BMWA Vorhaben; Universität 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.; Neuhäuser, M.; Meissner, E.; Streicher, W.; Halmdienst, C.; Dezentrale erneuerbare Energie für bestehende Fernwärmenetze; Bundesministeriums für Verkehr, Innovation und Technologie; Wien, 2005. Page 15-16

[12] Mangold, D.; Riegger, M.; Schmidt, T.: Solar Nahwärmeversorgung und LangzeitWärmespeicher; Forschungsbericht zum BMU Vorhaben; Solites; Stuttgart; 2007. Page 14, 20 [13] Nitsch, J.; Wenzel, B.: Langfristszenarien und Strategien für den Ausbau erneuerbarer Energien in Deutschland; Leitszenario 2009; BMU; Berlin; 2009. Page 53-57

[4] Deutsche Bank; Zinslandschaft; https://www.deutsche-bankbauspar.de/de/media/Zinslandschaft.pdf; 2010.

[14] RETscreen Version 4; Natural Resources Canada; http://www.retscreen.net; 2009

[5] Fielenbach, H.; Ohl, G.; Schwarzburger, H.: Effiziente Wohnwärme und hoher Komfort; GBG – Mannheimer Wohnungsbaugesellschaft mbH; 2009.

[15] Smolka, M.: Ökologisch-technische Auswirkungen dezentraler Energieversorgungsszenarien mit Blockheizkraftwerken in elektrischen Verteilungsnetzen; Verlagshaus Mainz GmbH; Aachen; 2009. Page 18

[6] Frey, M.; Milles, U.: Geothermische Stromerzeugung in Landau; BINE Projektinfo 14/07; Karlsruhe; 2007.

[16] Solarge: Marstal district heating Plant; Project Summary; http://solarge.org/index.php?id=1235& no_cache=1; 14.03.2010

[7] Heidemann, W.: Solare Nahwärme und saisonale Speicherung; FVS LZE Themen; Berlin; 2005. Page 36

[17] Sonnenhaus-Institut e.V.; http://sonnenhausinstitut.de/wohnhaeuser.html; 2010.

[8] Kaltschmitt, M.; Streicher, W.; Wiese, A.: erneuerbare Energien; Springer Verlag; Berlin; 2006. Page 29

[18] Ulbjerb, F.: Large-Scale Solar Heating; Hot|Cool; 3/2008; DBDH; Frederiksberg; 2008

[9] KfW: Programm erneuerbare Energien; http://www.kfw-mittelstandsbank.de/DE_Home/ Service/Kreditantrag_und_Formulare/Merkblaetter/ KfW-Programm_Erneuerbare_Energien_ 270_ 271_272_281_282.jsp; 2010.

[19] Watter, H.; Nachhaltige Energiesysteme; Vieweg+Teubner; Wiesbaden, 2009. Page 168 [20] Voß, A.: Das Wachstumspotential der Nah- und Fernwärme - wirtschaftliche und gesetzliche Voraussetzungen für den Ausbau; aus: Forschung und Entwicklung Heft 10; AGFW; Frankfurt, 2005.

[10] Klöpsch, M.; Besier, R.; Wagner, A.: Reicht für Kunststoffmantelrohre die Standarddämmung?; Euroheat&Power 38. (2009); issue 12 [11] Lutsch, W.: Neue Wege zur Marktumsetzung solarer Nah- und Fernwärme; Fernwärme-, Kälteund 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

SOLAR DISTRICT HEATING (SDH): TECHNOLOGIES USED IN LARGE SCALE SDH PLANTS IN GRAZ – OPERATIONAL EXPERIENCES AND FURTHER DEVELOPMENTS 1

M. Schubert , C. Holter and R. Soell 1

S.O.L.I.D. Solarinstallationen und Design GmbH, Puchstr. 85, A-8020 Graz, m.schubert@solid.at First solar thermal plants for district heating were built in the 1970‘s in Sweden. Since then, various plants have been built mainly in Austria, Denmark, Germany and Sweden.

ABSTRACT S.O.L.I.D. installed three large scale solar plants for feeding into the city‘s district heating in Graz in recent years. These three solar plants have an annual heat production of 15,8 PJ, the city‘s grid delivers 2800 PJ per year. Therefore the integration of solar thermal in a technical and economical feasible way has to meet the requirements of Graz‘ existing district heating grid, which is one of the largest in Austria.

Most of these solar plants feed into rather small heating grids or sub-grids with an annual heat delivery below 50 GWhth(180 TJ). In Denmark, this market was growing rapidly in recent years and is now bigger than the market for small-scale solar systems for singlefamily houses.

The first plant, at stadium Graz-Liebenau with 1.420 m², has been now for seven years in reliable operations, with very good power output data.

In Graz, Austria, solar thermal plants feed into a large scale heating grid with an annual heat delivery of 830 GWhth(2,99 PJ) and a maximum power of 382 MW th. Technical parameters and operation strategies in large scale heating grids are different to those in small scale grids and solar thermal technology has to adopt to these circumstances.

AEVG Graz, the largest plant in Graz at 4.960 m², feeds into the gas power station (maximum power of 250 MW) and from there the heat is distributed through the district heating grid. The latest plant, at Wasserwerk Andritz with currently 3.860 m², has a buffer storage of 60 m³ and the planning for installation of a heat pump is completed. The plant feeds into the district heating grid and supports the room heating of a large office building.

Three solar thermal plants in Graz are presented and the way they are integrated into the city‘s heating grid. SDH PLANT DESIGNS IN GRAZ

This paper presents operational experiences about three different ways for feeding solar thermal energy into a large city‘s district heating grid. Recent developments like buffer management for combined district heating and room heating and integration of a heat pump are outlined.

1. Feeding directly into the district heating grid – plant at stadium Graz-Liebenau This plant is located on the roof of an ice-skating hall next to the city‘s football stadium (Fig. 1).

INTRODUCTION For reasons of energy security and environmental protection, the European Union has set a target of 1% solar fraction in district heating in 2020 and of 5% in 2050 [1]. Solar thermal technology is widespread in the single family house sector in most European countries. Mainly for domestic hot water preparation (DHW), but also for room heating (RH). In multi-family houses and for heating grids, there are not yet as many solar thermal plants and the market begins to develop.

Fig. 1: Aerial view of solar plant Stadion Liebenau

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 transferred to the flow (Fig. 2) [2]. The adaption of solar thermal technology for the temperature and pressure levels of the district heating grid were challenging. This project was realized with standard large scale collectors (1420 m² collector area) of the Austrian manufacturer Ökotech and temperature levels in the district heating flow of above 70 °C have to be reached dependant on the ambient temperature.

During first operation years, detailed monitoring was done on the plant‘s performance. Dependant on climate condition, the annual yield of the plant was between 521 MWh/a and 569 MWh/a. This corresponds to a specific yield of 370–404 kWh/a per square meter collector area. Also the return temperature of the heating grid is of great importance for the performance of the solar plant.

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

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

Fig. 3: Solar plant AEVG Graz 141

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Even lower temperature levels in shoulder seasons and in winter can be raised by a heat pump. The installation is planned for the end of 2010. COPs above 4 are expected, i.e. when heat from the collectors of 26 °C is heated up to 55 °C for room heating.

3. SDH for combined room heating and district heating with buffer and heat pump – plant Wasserwerk Andritz As solar thermal systems can‘t always generate the high temperatures as required for the district 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 60 m³ buffer tank and later used for room heating of an office building (low temperature floor heating). The buffer is also fed by district heating and thus decreases the required connected load of the office building.

This work is supported by the EU in the project ―SDHtake-off‖ (IEE - Intelligent Energy Europe). REFERENCES [1] ongoing EU-funded project ―SDHtake-off‖ [2] Bucar, G., Schweyer, K., Fink, Ch., Riva, R., Neuhäuser, M., Meissner, E., Streicher, W., Halmdienst, Ch. (2005), FEEt – Bestehende fossile oder teilfossile Fernwärmenetze – Einbindung von dezentraler Energie aus Erneuerbaren Energieträgern – Chancen und Hemmnisse, Endbericht zu „Energie der Zukunft― Forschungsprojekt No 807718 im Auftrag des BMVIT, publisher: Grazer Energieagentur Ges.m.b.h.

Fig. 4: solar thermal plant Wasserwerk Andritz

142

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

BIOENERGY COMBINES IN DISTRICT HEATING SYSTEMS: PROSPECTS FOR A FUTURE GROWTH INDUSTRY? 1

E. Axelsson , A. Sandoff , C. Overland

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

One industrial branch that shows promising prospects in this respect is bioenergy production, i.e. production of various kinds of biofuel, biogas and solid biofuel. Integration of bioenergy production to district heating production eventuates in a bioenergy combine were the residual heat from the bioenergy production can be utilised for district heating. Moreover, the integration can, in many cases, offer additional positive synergies, e.g. regarding the use of steam and combustible by-products.

ABSTRACT District heating offers opportunities for integration of bioenergy production (e.g. of biofuel). The aim of this paper is to assess the environmental benefit and the economic value of such integration, in order to evaluate the prospect for bioenergy combines in district heating systems. Since the detailed characteristics of the district heating system are crucial for the feasibility for integration of bioenergy production, the assessment is based on four real district heating systems. The environmental evaluation shows that the decrease in green house gas emissions from a combine are in proportion to the increase in output of CO2 neutral energy products. However, the CO2 reduction per used quantity of biomass is higher in conventional combined heat and power production as long as marginal electricity is related to high CO2 emissions. Also the economic evaluation show ambiguous results: two cases had negative net present value even for low discount rates, while the two other cases showed to be more economically robust. In addition to this, a more detailed analysis of the industrial conditions for the integration shows a need for achieving a fit regarding several operational, strategic and economic circumstances for this type of business ventures. Two important conclusions that can be drawn from this is that: 1) not all district heating systems are suitable for bioenergy combines 2) there are many barriers for a wide spread adoption of bioenergy combines.

The fact that worldwide bioenergy production as well as the number of bioenergy products offered is increasing is a result of changing demand, which in turn offers new business opportunities. However, one of the great issues with large-scale production of bioenergy products is the growing concern over the negative externalities (social and environmental aspects as well as resource efficiency). Since energy production and consumption shows strong path dependence [1], there is an urgent need to develop and establish production technologies that help minimize the negative externalities. Utilizing the taiga and deciduous forest resources in the Northern hemisphere for this purposes is, arguably, a promising alternative. The majority of these natural resources exist in harvested forests, typically found in regions with, or suitable for, district heating. This paper investigates the prospects of using district heating production as a base for bioenergy production and its potential to become a wide spread technology. For this purpose, we use data from four existing district heating companies to which a bioenergy production unit is fitted. By acknowledging the complexity of this integrative business venture, it is possible to get credible assessments of the magnitude in energy efficiency, environmental gains and economic profits. Equally important is the possibility to detect potential limitations for bioenergy combines to become a complement to district heating. Finally, conclusions are made to acquire clues to important restrictions to a wide spread adoption.

INTRODUCTION District heating is a technology that receives increasing interest as it has great potentials in several ways. One unique characteristic of the district heating technology is the use of low temperature energy flows for large scale energy distribution. In contrast to other energy transformation technologies (e.g. condensing power or distributed gas heating), district heating can interact with energy flows that otherwise do not have any alternative use (e.g. industrial residual heat). Although this is one of the competitive advantages of the technology and a fundamental platform for its business model, this can further enhance the scoop of the business: by backward integration it is possible to increase profitability in other industrial processes with waste heat as a by-product.

RESEACH DESIGN We argue that prospects for becoming a future growth industry are dependent on the environmental benefits, economic attractiveness and fit with existing business context. Hence, these three aspects of joint production 143

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 with a system perspective on greenhouse gases (GHG) emissions, taking into account both on and off site consequences of introduction of an energy combine; see Environmental evaluation below. Moreover, the resource efficiency in the form of CO2 reduction per used quantity of biomass is evaluated for each combine.

Below follows a description of the environmental and economic evaluation procedure. It is important to stress that the input data for these assessments only include the change resulting from the integration of the bioenergy production. One implication of this approach is that the environmental benefit of the heat produced (for district heating) is not included, since one base condition is that the heat deliveries are the same with and without bioenergy production. Another implication is that production units in the district heating system that are not affected (e.g. base load and peak load production units) are not included. This system boundary is also pervading for the Description of the cases to follow.

The economic benefits of the ―joint production‖ set up are analyzed through both a short and long-term commercial lens. By using discounted cash flow techniques as a base for this analysis, it is possible to account for both the yearly consequences as well as long term economic value; see Economic evaluation below.

Description of the cases

Fit with existing business context is analysed with respect to input/output markets, production and system configuration and general business conditions dominant in the host industry. The analysis focus on restrictions for short term fit; see Business context evaluation.

The four district heating systems with reference and combine cases, respectively, are presented in brief below. The four objects for the evaluation are also summarized in Table I. A more comprehensive description can be found in ref. [3]. Table I. Overview of the reference and combine cases in the four district heating systems. Economic and energy data are given for both the reference and combine case, separated with a slash (ref./combine).

Since the detailed characteristic of the district heating system is paramount to the feasibility for integration of bioenergy production, we base our investigation on four real district heating systems in Sweden with different compositions. The chosen systems are all of equal size (500-600 TWh of yearly heat deliveries) established in towns with 40 000 to 80 000 inhabitants. These systems are in turn equipped with a bioenergy production unit that best suits ruling company strategy as well as operational characteristics and maximizes energy efficiency. In order to capture the additional values of these investments, evaluation of each combine configuration is made in relation to a reference case consisting of the existing system (complemented with investments to maintain a comparable level of production quality). The reference and combine cases are further described in the Description of the cases below.

CONFIGURATION 1

Heat deliv. (TWh/y)

500

530

560

620

Ref. inv.

Bio CHP

None

Bio CHP

Combine technology

Pyrolysis

Enzymatic hydrolysis

Acid hydrolysis

Gasification

Products

Bio oil

Ethanol1

Ethanol

FTdiesel2

ECONOMIC DATA, reference/combine 1

Inv. (M€)

74/60

0/144

116/310

146/473

O&M (M€/y)

2.3/2.8

0/8.8

3.6/15.8

6.1/11.1

ENERGY CONSUMTION, (GWh/year), ref./combine

Much effort was put into indentifying efficient technical solutions that best take advantage of the site-specific conditions in each system. This work included everything from choice of equipment, appropriate size of the integrated production unit and production strategies over the year regarding output of heat, electricity and other energy products. To identify efficient technical solutions an integrative computerized process was applied, including both the district heating simulation software MARTES [2], and detailed spread sheet calculations. In order to guarantee high quality input data, representatives from these four companies gave access to technical, environmental as well as economic data.

Biomass

397/244

730/1537

470/1271

362/2970

Others

74/1353

ENERGY PRODUCTION (GWh/year), reference/combine 1

Electricity

125/0

218/209

145/55

99/78

Biofuel

0/90

0/444

0/294

0/1336

Others 1

0/384

Besides ethanol also biogas and pellets is produced. Also kerosene and nafta is produced. Fuel oil (21/15) and industrial waste heat (53/120). 4 Biogas (0/114) and Pellets (0/270) 2

144

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

benefit in production.

System 1 In the current configuration of this system 15-20% of the energy demand is covered with fuel oil, which needs to be reduced. One interesting option could be to convert biomass into bio oil by pyrolysis and then use the bio oil in the existing oil boilers. Bio oil that is not used within the system can be sold (e.g. summer time). If no pyrolysis reactor is built, a conventional biofuel fired combined heat and power plant (bio CHP) will be invested in, building up the reference case.

accordance

with

marginal

GHG

Production, distribution and use of biomass

Power system

DH system with or without bioenergy production

System 2 In this system, there is no need for new production units, rather there is a high production capacity, allowing for integration of a bioenergy production unit. System 2 has good access to biomass, but might have difficulties to find a market for large quantities of byproducts. Based on these prerequisites, a suitable combine technology could be cellulose ethanol production with enzymatic hydrolysis aiming at high yield and in-house use of energy by-products. Regarding the O&M cost for the enzymatic process in Table I, future enzyme price are assumed [4], With today‘s prices, the enzymatic process will not be profitable.

Direct GHG emissions

electricity

GHG

Production, distribution and use of transportation f uel

Fig.1. Illustration of the applied system approach for assessing the changes of GHG‘s.

In the assessment, all GHG‘s of significance are included [3]: carbon dioxide (CO2), dinitrogen oxide (N2O) and methane (CH4). For all energy carriers, life cycle emissions are considered, i.e. both combustion emissions and well-to-gate emissions such as emissions from fuel extraction, processing and transportation. Also leakages are considered when applicable. How the GHG‘s for the relevant energy carriers are assessed are described in brief below, a more thorough description can be found in [3].The adopted life cycle GHG emissions associated with changes in consumption/production of the energy carriers are summarized in Table II.

System 3 In System 3 there is a need for new production capacity, which is represented by a bio CHP in the reference case. This system has good access to a large energy market, which enables output of other energy products. Hence, a cellulose ethanol plant based on acid hydrolysis can complement the reference case investment to build up the combine case.

Table II. Emission factors for included energy carriers.

System 4 This system is in many aspects similar to System 3, but ethanol production is not in line with company strategy. Moreover, System 3 has good access to peat, which could supplement biomass for a large scale production unit. Hence, gasification of biomass for production of synthetic biofuel is evaluated for this system.

ENERGY CARRIER

LIFE CYCLE EMISSION (kg CO2 eq./MWh)

Biomass

14-17

High emission elec. (E1)

800

Low emission electricity (E2)

260

Pyrolysis oil

292

Ethanol

307

Environmental evaluation

FT diesel

277

The assessment of the environmental implication of introducing a bioenergy production in an existing district heating system focuses on changes in emissions of green house gases (GHG). A system approach for analysing the changes of GHG‘s is applied. This means that besides changes of the direct emissions on site, also the changes of emissions in affected parts of the energy systems are included; see Figure 1. For instance, production of biofuel in the combines ads to the environmental benefit since fossil fuels can be replaced, while reduced electricity production has a negative impact to the environmental

Fuel oil

312

Biogas

207

Pellets

286

The lifecycle emission of biomass is dependent on how the biomass is used in the energy combines (e.g. hydrolysis for fermentation or gasification)

Biomass The energy input in all four combines is in the form of biomass. Production, distribution and use of biomass is related to GHG emissions. The GHG emission from the use of biomass differs depending on how the biomass is used. Combustion raises emissions of both methane 145

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 from a climate perspective), while hydrolysis and fermentation is not assumed to raise these emissions. Hence, the net lifecycle emission of biomass differs between 14-17 kg CO2 eq./MWh fuel.

Biogas and pellets In the energy combine of System 3, also biogas and pellets are produced. The biogas is assumed to be used as a transportation fuel to replace both petrol and diesel. The net GHG reduction for replacing fossil transportation fuel with biogas is set to 207 kg/MWh including life cycle emission and gas leakage in the production. The pellets are also assumed to replace fossil fuel, in this case oil with a net GHG reduction of 286 kg/MWh pellets.

Electricity In all district heating systems, the electricity production decreases as a consequence of introducing the combine (see Description of the cases). Any change in electricity production is assumed to be compensated by changes in marginal electricity production. For instance, if the electricity production decreases by 85 GWh/year, it is assumed that other producers will increase their production by 85 GWh/year. To assess the environmental impact of this, the decrease has to be multiplied with a emission factor for marginal electricity.

Resource efficiency With the emission factors in Table II and the energy flows of the reference and combine case in Table I, the environmental benefit of the energy combine can be assessed. However, if biomass is assumed to be a limited resource from a sustainability point of view, it makes sense to evaluate the use of biomass from an efficiency perspective. Hence, the resource efficiency is assessed as the net GHG reduction potential (in kg CO2 eq.) per used quantity of biomass (in MWh). By comparing this key figure for the reference case with the combine case for each system, the resource efficiency of the combines can be evaluated.

There are many opinions regarding the emissions of marginal electricity. Here we have used a high and a low level, based on dynamic response for electricity production with two different developments over a long time period [5]. By using a high and low figure, the impact and importance of changes in electricity can be illustrated in a clear way. For the high figure, the reference case in [5] is used where lifecycle emissions of marginal electricity are about 800 kg/MWhel. This marginal electricity is denoted E1 hereon. With more stringent environmental targets the electricity production can be carbon lean [5] implying that the long term lifecycle emissions would be about 260 kg/MWhel, denoted E2 hereon.

Economic evaluation In order to analyze whether an investment adds financial value we rely on a standard discounted cash flow (DCF) model estimating the net present value (NPV) for each project so that: n



NPV   CFt / 1  r  t 0

Biofuel As seen in Table I, the evaluated bioenergy combines have various biofuel products as output. In System 1 pyrolysis oil is produced. The pyrolysis oil is assumed to replace fossil fuel oil (but is categorized as an biofuel herein). If lifecycle emissions are regarded according to the approach in ref. [6] for both pyrolysis oil and fossil fuel oil, the net GHG reduction for replacing fuel oil with pyrolysis oil is 292 kg per MWh of pyrolysis oil exported from the combine. Also the amount of fuel oil used differs in the combine case from the reference case in System 1 (see Table I). The net life cycle GHG of this fuel oil is set to 312 kg/MWh.



(1)

where CFt denotes the net cash flow in year t, r is the future weighted cost of capital and n is the number of years included in the cost-/benefit analysis. The cash flow at year 0 indicates the initial outlay. Concerning r, the weighted cost of capital (WACC), we do not predetermine a specific hurdle rate; instead we analyze value added for three different levels of discount rates. We do so because any statements on the actual riskiness of the project or an estimation of the WACC for the companies are outside the reach of this study. As stated before, when estimating cash flows the point of departure is a reference object. That is, our NPV calculations only address the differences in cash flows between the reference and the bioenergy combine; this for two reasons. First, only the incremental cash flows are relevant in a DCF analysis. For instance, in the case of System 3 they already decided that they would at least build a combined heat and power (CHP) facility, and the question is if they gain from making additional investments in a bioenergy production unit. Second, by focusing on the differences we do not need to consider the cost structure in the reference case, it is treated as a given. Besides simplifying the analysis,

In systems 2 and 3 ethanol is produced, which is assumed to replace gasoline with net GHG reduction of 307 kg per MWh of ethanol reaching the market. In System 4, three biofuels are produced: Fischer Tropsch (FT) diesel, nafta and kerosene. All three products are assumed to replace fossil transportation fuel with the net GHG reduction of 277 kg/MWh. The possible leakage of methane from the gasification process is assumed to be negligible. 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 reveal sensitive information.

can see for what potential price changes extra concern is warranted. Certainly, a drawback with the sensitivity analysis is that it is just a ceteris paribus analysis and does not take into consideration the potential covariance of variables, for instance between ingoing biomass and outgoing biofuel.

Table III. Assumptions made for non-site idiosyncratic input and output prices (€/MWh). Ethanol

Biomass

FT-diesel

Fuel oil

Kersone

Pellets

Business context evaluation

Nafta

Electricity

Biooil

Electricity excise

The environmental and economic analyses of a joint production operation act as a starting point for the business context analysis. A wide-spread adoption demands not only indications of environmental benefits and economic profits, but must also offer a fit with the existing business context. Even though the degree of fit is defined on company level we will not analyze it as such. Rather we use the business context of the studied systems in order to put together a compilation of restrictions and barriers to a wide-spread adoption. The magnitude and importance of these will give important indications of the short term possibilities of realizing environmental benefits and economic profits in making bioenergy combines a future growth industry. The restrictions and barriers are identified through the fit with existing input/output market situation, production and system configuration and general business conditions, (i.e. strategic focus and capacity to absorb additional risk) dominant in the host company.

Biogas 1

Electricity certificate

0.5 1

Premium paid to producers of renewable electricity.

Cash flows The initial outlay is assumed to take place in full at year 0. Yearly operational cash flows are projected by first estimating an operational cash flow for the first year. As cash flows are the products of price and quantity, this estimation is based on the technical analysis in order to obtain energy flow estimates (see Table I), and then multiply them with price estimates, to which we add out-payments for operation and maintenance. We extrapolate this operational cash flow over the 20 year long investment horizon with a three percent yearly growth rate (adjusted for the fact that green certificates are obtained for fifteen years only). All cash flows are conservatively assumed to occur at the end of each year. Next, we add tax payments (assuming an effective tax rate of 26,3%), tax discounts from depreciation (according to Swedish tax code), changes in working capital (approximated by dividing the difference between in-payments and out-payments of year t by 12 and subtracting the corresponding value from year t-1, save for the last year where the difference is set to zero) and a terminal value (5% of the initial outlay). Initial outlays are determined by consulting [7]– [19]. Our price assumptions for non-site idiosyncratic inputs and outputs are presented in Table III. For translation between different currencies the following exchange rates were used: 9.6 SEK/€ and 6.5SEK/USD.

ENVIRONMENTAL BENEFITS As already stated in the Research design, the environmental benefit from integrating bioenergy production into an existing district heating system is assessed as the reduction of GHG‘s from a system perspective. As also explained, the net difference depends on the reference case as well as the composition of the energy combine. In Figure 2, the GHG reduction for the included parts of the reference case and energy combine case of System 3 is displayed. In the reference case (left bar in Figure 2) – a combined heat and power (CHP) plant – biomass is converted into heat (for district heating) and electricity. The amount of heat is the same in both the reference and combine cases and, hence, not considered in the evaluation of GHG reduction. However, the production of electricity will change and the system consequences of that is, as stated, considered by including two different assumptions for marginal electricity. Assuming that marginal electricity is related to about 260 kg CO2 eq./MWhel (E2), the electricity produced in the reference case results in a yearly reduction of 38 Mtonne (dark blue bar to the left in Figure 2). If the emissions of marginal electricity instead is assumed to be 800 kg/MWhel (E1), the emission reduction would increase by 78 Mtonne/year (light blue bar) to be in total 116 Mtonne (dark + light blue bar = E1). The handling of the biomass is related to GHG emissions

Sensitivity analysis We then control the robustness of the NPV estimates through sensitivity analysis; that is, we examine how the cost-/benefit analysis is affected when changing a variable at the time, holding all else equal. We do this in two steps for each system. First, we illustrate the changes in estimated NPV by changing yearly inpayments, yearly out-payments, initial outlay and terminal value respectively. Second, we show how yearly in-payments and out-payments respond to price changes. By this sensitivity analysis, we can to some degree compensate for the uncertainty that surrounds our estimates of initial outlays and terminal value, and we 147

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

reduction from introducing an energy combine in System 3 is 158 or 109 Mtonne/year depending on the assumption for marginal electricity (E2 and E1, respectively).

(see Environmental evaluation) and, hence, there is a negative bar of 8 Mtonne for biomass. To sum up, the net GHG reduction in the reference case is 30 or 108 Mtonne CO2 equivalents depending on assumptions for the marginal electricity.

The equivalents to the right hand bar in Figure 2 for all four systems are shown in Figure 3. As can be seen, the reductions of GHG‘s are significant in systems 2-4, especially if the electricity is associated with low emissions (E2, dark blue bar only). In System 1, the environmental benefit is negative, even if the marginal electricity is CO2 lean.

300

Net reduction (E2/E1): 30/108 188/217 158/109

200

Significant environmental benefits, as displayed for systems 2-4, are expected since the combines in these systems use more biomass, which eventually replaces fossil fuel in the system approach applied (in system 1 less biomass is used which explains the negative results for this system). However, if biomass is assumed to be a limited resource from sustainability point of view, the use of biomass should also be evaluated from an efficiency point of view. As explained in the Environmental evaluation, one measure of resource efficiency is the GHG reduction potential per used quantity of biomass. This key figure is presented in Figure 4 for both the reference case and the combine case for the four district heating systems evaluated.

Electricity, E1-E2* Electricity, E2 Pellets Biogas

100

Ethanol

Biomass

GHG reduction (Mtonne)

GHG reduction (Mtonne CO2 eq./yr)

The combine case of System 3 has lower electricity production than in the reference case (see Description of the cases). Consequently, the GHG reduction from the electricity production is also lower, which is seen as lower dark and light blue bars for the combine case; middle stacked bar in Fig. 2. Moreover, the negative bar for biomass is larger for the combine since more biomass is used in this case. In the energy combine, however, bioenergy products such as biofuel (ethanol in this system), biogas and pellets are produced. As already explained, these energy products are assumed to replace fossil fuels and the resulting GHG reduction from the combine is significant: 188 or 217 Mtonne CO2 eq. with carbon lean (E2) and carbon intense (E1) electricity production, respectively.

* additonal emission reduction/change if electricity is related to high CO 2 emissions

-100 -200

Reference Combine Difference

Fig. 2. GHG reduction in System 3 for the reference case, combine case and the net difference for converting to the combine.

400 350 300 250 200 150 100 50 0 -50 -100 -150

Others*

Net reduction (E2/E1): -2/-69

124/119 158/109

321/309

Biofuel Elec., E1-E2 Elec., E2 Biomass * biogas and pellets

System 1 System 2 System 3 System 4

The dark blue bars are related to marginal electricity associated to low GHG emission (E2). The additional emission reduction/change if electricity is related to high GHG emissions (E1–E2) is indicated by the light blue bars. The total emission/change for E2 is given by the sum of light blue and dark blue bar.

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

As seen in Figure 4, the energy combines are less resource efficient than the reference cases (generally a biomass fired CHP plant) if the marginal electricity is associated with high CO2 emissions (E1, dark + light blue bar). However, if the marginal electricity is associated with low CO2 emissions (E2, dark blue bar only), the combines are more resource efficient than the reference cases. As also can be seen, the resource efficiencies do not differ dramatically between systems 2–4. System 1, however, shows lower resource efficiency, which can be explained by the fact that a major part of the produced pyrolysis oil is consumed internally in the system instead of replacing fossil fuel off site.

The implication in terms of GHG‘s of integrating bioenergy production in System 3 can be visualised by moving from the left bar in Figure 2 to the middle bar. Consequently, the difference of the two bars shows the GHG implication of converting to an energy combine in System 3, which is presented in the right hand bar in the figure. The change from the reference to the combine case gives rise to GHG reduction from the fuel products (green bars) However, the electricity production decreases, implying decreased reduction (emission increase) and, hence, negative bars for electricity. As can be seen in the figure, the net GHG 148

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

E1-E2 E2

200 150

100 50

Change in NPV (M€, 10% disc. rate)

100

Comb.

Ref.

Comb.

Ref.

Comb.

Ref.

Comb.

Ref.

Resource efficiency (kg CO2 eq./MWh biomass)

250

System 1 System 2 System 3 System 4 Fig. 4. Resource efficiency of biomass quantified as GHG reduction per used quantity of biomass.

ECONOMIC VALUE

Initial outlay (M€)

- 13.9

144

194

327

Cash flow (M€y)

-3.4

18.8

15.7

-50 -100 -150 -200

-250 -300

-20%

-10%

10%

20%

30%

Change in cash flows

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

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

Table IV. Summary of cost/benefit analyses for adding a bioenergy combine to the reference investment in the studied systems.

Out-payments Terminal value

-30%

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

In-payments Initial outlay

NPV (M€) for different discount rates -40

-62

362

-27

-89

207

10%

-19

-4

-108

101

25%

Change in marginal in-payments

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

20% 15%

Ethanol

Biogas

Pellet

10% 5% 0% -5% -10% -15% -20% -25%

-30%

-20%

-10%

10%

20%

30%

Price change

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

149

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Change in marginal out-payments

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

20% 15%

Biomass

Electricity

biomass, since increased use of biomass implies increased output of CO2 neutral energy products. However, from a resource efficiency point of view, biomass should not be used to replace transportation fuel as long as the marginal electricity is related to high CO2 emissions. One important explanation to the coherent environmental profiles of the different bioenergy combine solutions is similar resource efficiency for the four technologies evaluated. Hence, our results suggest that it is possible to find different energy combine with similar resource efficiency. However, these similarities in resource efficiency do not indicate similarities in economic attractiveness. In fact, the economic evaluation seems to suggest that some bioenergy production technologies are not currently economic viable for integration with district heating system. Furthermore, the results indicate that not all district heating systems are suitable for integration with a biofuel production unit. Despite being of the same size, use the same raw material and being evaluated only on marginal effects on the economic situation, differences in district heating system characteristics have a profound impact on the economic possibilities of energy combine integration. In this study we have matched every system with a combine solution in order to maximize the site-specific opportunities in each system. This opens of course the possibility that there exist other matches with less resource efficiency but higher economic profitability. Even if this can be the case, we would like to point out that one of the starting points of this study was to base in-data on the conditions of real systems. This includes taking various kinds of restrictions into consideration. Even though these restrictions vary, the ones prominent in this study can be grouped into four different categories:

O&M

10%

5% 0%

-5% -10%

-15% -20%

-30%

-20%

-10%

10%

20%

30%

Price/unit cost change

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

Similarly, Figure 7 shows how out-payments vary with input prices. Inputs included in the figure are biofuel, operations and maintenance (O&M) and electricity7. Not surprisingly, biofuel is the key input, where a 20% price change results in a 10% change in out-payments, which in Figure 5 translates to a € 42 million change in NPV. The sensitivity analyses of System 3 show that minor changes in underlying factors can result in significant changes in the NPV estimates. However, a not insignificant part of the indicated variability in cash flows should be hampered by the offsetting effects driven by the probable covariance between prices for biomass and bioenergy products. To be noticed is that the order of importance of the inputs in the other three systems show a similar ranking, where biofuel and biomass price being the two most important ones.

   

Proximity to input resources Proximity to customers or infrastructure for transporting the finished products Existing production and system configuration Dominant business conditions

Proximity to input resources Some combine solutions (such as the one for System 4) demand huge amounts of biomass. This requires large areas of regional biomass recourses and little or no competition over it. Import by sea is an alternative but it requires production sites close to a harbour.

FIT WITH EXISTING BUSINESS CONTEXT The environmental and economic evaluations indicate that the integration of bioenergy production into medium sized district heating systems can be associated with both environmental and economic benefits, but the picture is mixed and ambiguous. From an environmental point of view, the results are coherent across all systems: the absolute environmental benefit of bioenergy production is in proportion to the use of

Proximity to market for the finished product The production of biogas is one example of both the importance of proximity to customers and to

150

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.

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 is considered too challenging at present time.

Yearly Cash Flows (Mâ&#x201A;Ź) 100

Existing production and system configuration Investments in bioenergy combines are seldom green field but, as we have shown earlier, have to be adapted to suit existing heat volumes, demand curves, system configurations and also production site layout. In one of the systems, the production site was too small to house the large amounts of biomass necessary for achieving an economic profitable size of an ethanol operation.

Biofuel 80 Biogas Free cash flow 60

Pellets In-payments

Out-payments Biomass

Industrial waste heat Electricity

Electricity certificate

Dominant business conditions The results of the study show that two business areas have an evident influence on the type of bioenergy combine investments the companies carry out: 1) the strategic framing of the district heating company and 2) the risk that these investments innate. Concerning the first, many of the municipally owners use the utilities to enhance and to some extent even realize the environmental visions that are formed and expressed on the political level. Examples of these found among the companies represented in this study include; phasing out fossil fuels, use of local waste resources and visions of a fossil free cities based around locally produced bioenergy fuels. When present, strategic framing has a visible effect on limiting the number of available alternatives for integrates production.

O&M System 1

System 2

System 3

System 4

-20

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

The considerable positive free cash flow of system 2 from its existing operations is explained by the companyâ&#x20AC;&#x2DC;s sell of hydropower. Although irrelevant for the value of this investment, it could function as a general safeguard against negative results, due to unfavourable relation between biofuel and biomass prices. The investment in system 1 was not profitable according to the valuation earlier. Despite this, it is worth pointing out that the risk of this investment should be low since it uses its own products as input. It too has, relatively speaking, a strong free cash flow from its current operation that will decrease the risk of ending up in the red.

As stated, the second area that has an significant influence on the type of bioenergy combine that these companies consider is the risk that these investments innate. Due to the municipal ownership, these companies are inherently dependent on stable business conditions. The ability to absorb negative results is strongly limited. The added business risk of bioenergy production must, if needed, be able to be absorbed by cash flows from existing operations or a strong capital base. In principle, this can be done in two ways, either by keeping the investment relatively small, or by only accepting business propositions with cash flows that can be made relatively stable.

CONCLUSIONS The results of the bioenergy combine analyses show that there are indications for both environmental gains and added economic value of such investments. However, these benefits seem to be limited by several operational, environmental and economic circumstances present in these systems. First, these investments are dependent on the need for making major changes in current production layout, typically the need for new or altered production plants. This limits the available window of opportunity. There are also several limitations related to operational characteristics, availability of input resources and suitable product markets. A closer investigation of existing governance situation also shows that these investments often are made to fit owner strategies regarding environmental goals of the local energy system. Finally, the municipally ownership typically limits the risk appetite which also limits available investments. The doubtful short term environmental benefit is a more general objection based on the valuation of the current marginal power production.

In Fig. 8, the operational risk of the investment can to some extent be visualized by the size of the marginal cash flows of the different investments. The investment in system 4 stands out not only because it is the largest one but also because its in-payment comes from one source only. If the price correlation with biomass is high, this might not be a large problem. However, it is interesting to note the relatively small positive cash flow available from existing operations in Systems 4, and also for System 3. If the company carries through with the evaluated investment, it will dramatically change its operational risk profile and over-all business focus.

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 widespread adoption of bioenergy combines.

[6] IVL, ―Miljöfaktabok för bränslen‖, IVL Rapport B 1334B-2 (2001).

These circumstances lead us to conclude that not all biofuel production technologies are suitable for all district heating system. Our economic analyses also indicate that not all district heating systems are suitable for bioenergy combine production. In fact the barriers are so many that it is reasonable to assume they will effectively reduce the number of systems adopting this operational design in the near future.

[7] Svebio, ―Kraftvärmeutbyggnad 2007-2015‖, Svebio repport 2008-03-31.

ACKNOWLEDGEMENTS

[10] A. Hang and S. Ilic, ‖En förstudie för bioetanol produktion i Borås‖, Master‘s thesis at Institutionen Ingenjörshögskolan, Högskolan i Borås (2008).

[8] H. Hansson, S-E. Larsson, O. Nyström, F. Olsson and B. Ridell, ―El från nya anläggningar - 2007‖, Elforsk repport no 07:50 (2007). [9] M. Zakrisson, ―Internationell jämförelse av produktionskostnader vid pelletstillverkning‖, Master‘s thesis no 29 2002, SLU.

The main funding for this project is provided by Fjärrsyn, which is a research program organized by the Swedish district heating branch agency. Additional funding is also received from the project ―Pathways to Sustainable Energy Systems‖.

[11] M. Lantz, ―Drivmedelsproducenters betalningsförmåga för energigrödor‖, Miljö- och energisystem, LTH (2006).

We kindly thank the representatives from each district heating system for a good cooperation and for providing us with technical and economic data of their systems. Without these inputs, the work would not have been as solid as it is.

[12] J. Benjaminsson and A. Dahl, ―Uppgradering av biogas‖, Presentation at ―Temadag uppgradering av biogas‖, Göteborg (2008). [13] I. Granberg, Project leader at Jönköping Energi, Personal commication (2008).

We also thank Karolina Nilsson and John Jonsson (both at Profu) for their valuable contribution to the work.

[14] M. Tijmensen, A. Faaij, C. Hamelinck, and M. van Hardeveld, ―Exploration of the possibilities for production of Fischer Tropsch liquids and power via biomass gasification‖, Biomass and Bioenergy 2002, Vol. 23.

REFERENCES [1] M. Odenberger, F. Johnsson, ―Pathways for the European electricity supply system to 2050‖, Int. J. of Greenhouse Gas Control, 2010, Vol. 4:2, pp 327-340

[15] I. Johansson, S. Larsson and O. Wennberg, ―Torkning av biobränslen med spillvärme‖, Värmeforskrapport 881 (2007).

[2] J.Sjödin and D. Henning, ―Calculating the marginal costs of a district-heating utility‖, Applied Energy, 2004, Vol. 78:1, pp 1-18.

[16] E. Sandvig, G. Walling, R. Brown, R. Pletka, D. Radlein, and W. Johnsson, ―Integrated Pyrolysis Combined Cycle Biomass Power Systems‖, Repport of Alliant Energy, Iowa, USA (2003).

[3] E. Axelsson, C. Overland, K. Nilsson, and A. Sandoff, ‖Bioenergikombinat i fjärrvärmesystem‖, Fjärrsynsrapport 2009:11.

[17] H. Thunman, F. Lind, and F. Johnsson Delstudie energikombinat, Elforskrapport, 2008. [18] NREL, Research Advances Cellulosic Ethanol, NREL (2007).

[4] T. Brandberg, Senior researcher at SEKAB Etechnology, Personal communication, 2009.

[19] P. Sassner, M. Galbe, and G. Zacchi, ―Technoeconomic evaluation of bioethanol production from three different lignocellolosic materials‖, Biomass and bioenergy 2008, Vol 32.

[5] H. Sköldberg and T. Unger, ‖Effekter av förändrad elanvändning/elproduktion‖. Elforsk report (2008).

152

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

SEA WATER DISTRTICT COOLING FEASIBILITY ANALYSIS FOR TALLINN 1

A. Hani , I. Britikovski , H. Voll and T.-A. Kõiv 1

Tallinn University of Technology, Department of Environmental Engineering, Estonia also important to locate the district cooling station near to energy source.

ABSTRACT In this paper sea water district cooling feasibility analysis for Tallinn is presented. It has become more and more interesting to study alternative solutions for public buildings A/C cooling due to relatively high electrical energy prices. Besides economical aspects technical and environmental sides must be considered.

Typical SW district cooling system principle is indicated in Figure 1. The system consists of three main sections:   

INTRODUCTION

Cold sea water pumping; Cooling plant with heat exchangers; Standard cooling distribution network.

The first large district cooling systems were developed during the 1960‘s in Hartford (1962) and California (1965) in United States [10]. The first systems in Europe were La Defense (1967) in France and in Hamburg (1968) Germany [1]. In the beginning of the 70`s the first system in Japan (Shinjuku) was built [3]. However, because of the energy crises in the end of 70`s, the District Cooling development was slow and no new large systems were built. Until the end of 80`s when many new large systems were opened for example Kioi-cho, Nishi-Shinjuku in Japan and Trigen Trenton in United States. Also the first district cooling system of the Nordic countries was installed in Norway. Operation started in 1989 in Baerum, near Oslo. The first system in Sweden was built in 1992 in Västerås [2] and since then the district cooling in Sweden has developed rapidly. Since the 1990‘s the establishment of commercial district cooling systems has increased rapidly worldwide. Nowadays, more than 20 countries have a commercial district cooling system and this is expected to increase rapidly [4].

Fig. 1 SW district cooling principle schematic

Heat exchangers allow usage of the soft water in distribution network while problematic salty sea water handling will be done in open central circuit. Environmental impact study is required before any of the projects will be executed. Large sea water quantities have to be available to minimize pumping impacts. In addition to evaluation of the deep zone cold water pumping, the analysis of recycling the sea water back to lower sea water zone with higher temperatures should be made.

The sea water (SW) district cooling is based on large natural cold water source. Enough cold water is accumulated in lakes, seas, oceans, rivers, etc [8]. Lowering the coolant temperature with sea water is an alternative to conventional electrical energy consuming chillers [5]. The system working principle is quite similar to geothermal energy production which is used in heating systems [6]. Until now the sea water district cooling is quite conservatively expanded around the World.

Following factors shall be considered before system design [7]:  Minimum altitudes between heat exchangers and water resource level should be designed;  Centralized district cooling plant (heat exchangers, pumping station and chillers) is less expensive than decentralised system;  Centralized system has less maintenance problems.

SW DISTRICT COOLING PRINCIPLE The temperature in conventional cooling water network is between +4…+7 o C so applicable the sea water temperature should be below +5 oC. Despite that compressor based cooling can be used in case cooling water temperature is higher than mentioned [9]. It is

DESIGN PARAMETERS Temperature of the sea water varies during different seasons and distance from the coast. 153

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

cooling network with total capacity of 19 MW. Project is interesting to public buildings which have lower balance temperature and due to that higher cooling demand.

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

Study was carried out to construct:  Cooling plant with 4 water chillers;  Sea water pumping station (free cooling, precooling) 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 Tab. 1. SW parameters Dist. from coast, m

Build. height, m

Storeys above ground

Public 2 area m

8764

1052

18870

2264

1458

175

3564

428

5780

694

5198

624

2340

281

8775

1053

2268

272

2430

292

10260

1231

5049

606

4860

583

24500

2940

4250

510

11200

1344

37221

2233

10500

1260

2200

264

5250

630

3700

444

Build. no

Depth (sea), m

Annual aver. o temp, C

Min temp, o C

Max temp, o C

500

7,5-8,5

2,5

17,5

1500

5,5-6,5

1,5

3200

4-5

14-16

4000

3,5-4

8-15

5500

3,5

6-8

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

Fig. 3 Ambient temperature and cooling power relation

CASE STUDY – TALLINN COSTAL AREA The study was carried out to research feasibility to build to the Tallinn costal area cooling plant and district 154

Cooling demand kW

100%

19,2 MW

85%

16,0 MW

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

case sea water temperature is below 5 oC. Maximal pressure drop in both circuits is selected 0,85 bar. Heat exchanger parameters are indicated in Table 5.

Tab. 3 Main technical parameters for design the system Max cooling demand

19,2 MW

Ambient temp.calc.

27 C

Simultaneous factor

0,85

Cooling station capacity

18 MW

Annual average cooling consumption

21600 MWh

Supply water temp

6 C

Return water temp (max consumption)

16 C

Tab. 5 Free-cooling heat exchanger parameters

o o

Return water temp (min consumption)

13 C

Water chiller cooling Centralized cooling plant contains up to 4 water chillers to gain flexibility of the system. Also it is possible to construct the cooling plant step by step according to consumers‘ interest and cooling energy demand. System contains four 4500 kW water chillers with centrifugal compressors. It is possible to adjust the cooling power of the unit between 300–4500 kW which makes the system more energy (el) efficient during the partial load period. The condenser has to be produced from titan or similar resistant material due to fact that it is being cooled with sea water. In the following Table 4 are indicated technical parameters for water chillers.

Heat exchangers capacity

5x3600 kW

Sea water (SW) supply temp.

4,5 C

SW return temp.

10 C

SW flow

130 l/s

District cooling supply temp.

6 C

District cooling return temp.

16 C

District cooling flow

72 l/s

Max pressure drop

0,85 bar

Tab. 6 Coolant parameters Sea water (SW) o

SW temp

1,5-18 C

Max pressure

6 bar

District cooling liquid

Tab. 4 Water chiller parameters

Temperature

10-18 C

Max pressure

10 bar

Sea water cooling The sea water is supplied through insulated 800mm pipes to pumping station using sea water gravity. Three pumps (max 1080 m3/h) with frequency converters are installed using parallel scheme to suction pipe. Sea water pressure is ca 1,5 m and pumps will add 2 bars to overcome self-cleaning filters, heat exchangers and condensers pressure drop. Frequency converters are used to lower energy consumption during partial load.

Cooling power

4x4500 kW

Refrigerant

R-134a

Condenser temp.

28 C

Seawater (SW) supply temp.

18 C

SW return temp.

24 C

SW flow (each unit)

215 l/s

Evaporator temp.

3 C

Cooling plant operation modes

COP full load

COP partial load

Cooling plants are designed to have three different operation modes:  SW temperature < 5 oC. Completely free-cooling;

District cooling supply temp.

6 C

o o o

District cooling return temp.

16 C

District cooling flow (each unit)

115 l/s

Free-cooling When sea water temperature is lower than return temperature from the network free-cooling through heat exchangers can be used. Optimum logarithmic temperature difference shall be app. 1,5 oC. Five heat exchangers with capacity of 3600 kW are selected, which assure whole cooling plant capacity (18 MW) in



SW temperature 5–12 oC. Pre-cooling with SW + compressor cooling;



SW temperature > 12 oC Only compressor cooling (free cooling heat exchangers are equipped with bypasses).

District cooling network Supply (forward) water temperature is designed 6 oC. Return water temperature between 13–16 oC (see Figure 4).

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 1,5 m depth is 10 oC it is not necessary to insulate the return pipe of the district cooling network. Supply pipe is insulated with 10 cm nowadays heat insulation material.

The cooling plant shall have three operational modes:   

Free-cooling; Pre-cooling + compressor cooling; Compressor cooling.

Optimization of the proposed system should be carried out in further studies. REFERENCES [1]

Vadrot, A. and Delbes, J, (1999). District Cooling Handbook a Survey of Techniques, equipment and Choice of System. European market Group. Number of pages 208.

[2] Feldhusen. H, Francesc. M. R, (2001). "District Cooling-Present Market Assessment," Master, Kungl Tekniska Högskola, Stockholm division of Applied Thermodynamics and Refrigeration. pp 52. Stockholm. [3] Euroheat and Power, (2003). District Heat in Europe Country by Country/2003 Survey. Brussel Belgium.

Fig. 4 District cooling network temp

CONCLUSION

[4] Mildenstein, B. S. P, (1999). District Heating and Cooling Connection Handbook.

The sea water (SW) district cooling has until year 2000 quite modestly developed among different countries around the World. Due to the fact that energy prices have raised rapidly more and more researches for free energy resources are carried out. Wind power, heat pumps, solar energy and sea water have obtained huge attention.

[5] Gosney. W.B, (1982). Principles of Refrigeration. Cambringe University Press. Published by the press syndicate of the University of Cambridge. [6] Westin, P. E. H., (1999). Production Technologies in District Cooling Systems and the Importance of Local Factors. New Energy Systems and Conversion-NESC 99.). pp 6.Osaka.

SW district cooling is centralized and will have 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 the Capacity of District Cooling Systems.20th International Conferenss of Refrigeration, IIR/IIF.). pp 1-8. Sydney.

Current feasibility analysis was done in Tallinn costal area to define possible cooling plant load, potential consumers and technical possibilities. Due low costal area it is possible to locate the cooling plant near to sea water. Further studies should add some more economic aspects to the technical solution. Problematic is to develop the district cooling network in Tallinn area (existing tunnels and subways will ease the process).

[8] Nordell, B., and Skogsberg, K, (2002). Snow and ice storage for cooling applications.Winter Cities 2002.Japan Aomori. Luleå University of Technology [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 internal heat loads. In present research 21 buildings with only public area were included (total cooling demand 19,2 MW). The cooling demand rises considerably when ambient air temperature exceeds 16oC. Sea water temperature 5 oC can be found in depth of 35–40 m.

[10] Morris, A.P, (1995). The Road to Lockport: Historical Background of District Heating and Cooling. Ashrae Transactions: Symposia. [11] Arvidson, J, Asplund, A-L, Birgerrson, E, (1997), Cold production uning low temperature waste heat,. Kungl tekniska högskolan Kemisk apparatteknik. Pp 54, Stockholm

156

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

Yong Hoon Im , Hwa-Choon Park , Byung-Sik Park and Mo Chung 1

Cogen. & Boiler Research Group, Building Energy Research Center, Korea Institute of Energy Research, Korea 3 Mechanical Eng. Dept., Yeungnam Univ., Korea 12.3% on the basis of the total number of households at the end of 2008 [1]. The annual heat sales, via DHC network, in 2008 have reached 16,676 thousand Gcal and it increased by about 5% on average after 2001. Considering the trend of new-town development in metropolitan areas and newly developing residential areas on a large scale, it is generally expected to show a clear increasing trend of DHC systems on the market for the time being. Furthermore, the relevant changes of circumstances such as the long-term expectation for high prices of fossil fuels and the imminent realization of UNFCCC around the world will help the CHP and DHC system tighten its grips on the forthcoming heating and cooling market [2]-[3]. Among the several merits of DHC systems against separate heat & power (SHP) or central heating system, the distinctive feature of being able to construct the networking system with the neighbouring DHC systems certainly deserves to receive attention from the view point of efficient use of energy resources and operation costs reduction [4]. However, the effectiveness of networking operation of CHP and neighboring DHC systems is strongly influenced by the conditions of energy consumption behaviours and corresponding operation scenarios on both sides. The different pattern of energy consumption in new demand areas is highly desirable for creating synergy effects by networking operation. In addition, the different operation strategy of CHP system with that of DHC network can also improve the effectiveness of networking operation. The optimal system configuration of the CHP system with networking operation certainly differs from that of stand-alone CHP system not to mention the operation characteristics. Since the heat flows in the network are bi-directional, the appropriate modelling for the mutual effects on each system is highly required for the accurate estimation of the networking operation.

ABSTRACT A simulation program for analyzing the effects of the networking operation of existing DHC system in connection with CHP system on-site is to be discussed in this study. The practical simulation for arbitrary areas with various building compositions is carried out for the analysis of operational features in both systems, and the various aspects of thermal network operation are highlighted through the detailed assessment of predicted results. The intrinsic operational features of CHP prime movers, gas engine, gas turbine etc., are effectively implemented by realizing the performance data, i.e. actual operation efficiency in the full and part loads range. For the sake of simplicity, a simple mathematical correlation model is proposed for simulating various aspects of change effectively on the existing DHC system side due to the networking operation, instead of performing cycle simulations separately. The empirical correlations are developed using the hourly based annual operation data for a branch of the Korean District Heating Corporation (KDHC) and are implicit in relation between main operation parameters such as fuel consumption by use, heat and power production. In the simulation, a variety of system configurations are able to be considered according to any combination of the probable CHP prime-movers, absorption or turbo type cooling chillers of every kind and capacity. From the analysis of the thermal network operation simulations, it is found that the newly proposed methodology of mathematical correlation for modelling of the existing DHC system functions effectively in reflecting the operational variations due to thermal network operation. The effects of intrinsic features of CHP prime-movers, e.g. the different ratio of heat and power production, various combinations of different types of chillers (i.e. absorption and turbo types) on the overall system operation are discussed in detail with the consideration of operation schemes and corresponding simulation algorithms. The various aspects of system configuration in terms of CHP system optimization are also discussed.

The main purpose of this study is to examine the feasibility of the network operation of the CHP system on-site with the existing DHC system in terms of efficient use of primary energy and reduction of the operation cost. In this study, a simulation program is developed for analysing the thermal networking process between the existing DHC system and the CHP system for the newly developing area. The effects of thermal networking on the existing DHC system operation are implemented using mathematical

INTRODUCTION In Korea, the district heating and cooling (DHC) system gains share of the market steadily and it amounts to 157

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 parameters. The intrinsic features for the CHP prime movers is modeled using the actual performance data of operation efficiency in full or part load conditions.

Fig. 2 and Fig. 3 show examples of the daily unit energy load model of heating for the apartment and hourly unit energy load model of electricity for the office building respectively.

The specific features of the newly developed program in simulation of thermal networking process in district heating is described in terms of the energy load prediction and operation simulation of various system configurations with CHP prime movers and types of cooling chillers. The unit energy load model for various buildings by use, e.g. apartment, hotel, hospital, buildings for business and commercial use etc, is introduced for the accurate prediction of energy loads for newly developing area. The effects of intrinsic features of CHP prime movers, e.g. the different ratio of heat and power production, various combination of different types of chillers (i.e. absorption and turbo types), on the overall system operation are also discussed in detail in the following.

The annual hourly unit energy model can be obtained by synthesizing the daily and hourly unit energy load models [5]. The final annual hourly energy consumption for given building compositions and corresponding scale is to be predicted with the input of the total areas for respective buildings since the unit energy load models have been developed by normalizing the statistical energy consumption measurement data with the corresponding building areas. The example of annual hourly energy consumption for the apartment is shown in Fig. 4.

MODELLING FOR NET-WORKING OPERATION 1. Modelling of CHP system operation In the previous studies [5]â&#x20AC;&#x201C;[9], a simulation tool for the optimal design of the CHP system had been developed, which is composed with three different modules of energy load prediction, operation simulation, and economic analysis modules as shown in Fig. 1. The main goal of the simulation is to draw an optimized system configuration for a given target area by the systematic analysis of the physical and mechanical behaviour of the CHP system and corresponding operational cost structure. In principle, the analysis is performed on hourly basis for a year. The unit energy load model for a variety of building types (e.g. apartment, commercial building, office building, department store, hospital etc.) has been developed for different types of energy loads, i.e. heating, cooling, electricity and hot water [10]-[13]. In energy load prediction module, the hourly, annual energy demand for a target area is predicted using the unit energy load models.

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

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

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

Fig.1. Relationship between load, operation and economic analysis modules [9] 158

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

the number of units, and the load factor in terms of unit capacity. When an option is selected by the user as described above, its corresponding technical data for CHP product will be linked automatically in the subsequent operation simulation procedures. The settlement of the system configuration for the cooling system can also be performed in a similar manner by providing the data for the ratio of being in charge of turbo or absorption type chillers.

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

2. Modelling of DHC system for networking operation In contrast with small cogeneration or CES system, the DHC system is not authorized to sell the electricity to the customer directly in Korea [6]. As a result, the operation mode differs from that of cogeneration or CES system, i.e. the facilities are operating depending on the heat loads, and CHP facilities stop operating during summer to reduce waste heat production. Instead, the hot water load during the summer season is usually supplied from incinerators nearby, or heat only boilers (HOB). However, the operation schemes of DHC system for stand-alone operation are bound to be modified to some extent by networking operation with CHP system on-site and the appropriate modelling for such an effect of networking operation on DHC system is a key element for a reliable prediction of the operation behaviours due to thermal network operation.

(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 production in terms of energy consumption are given as follows,

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

The required data for the establishment of the mathematical correlation is given by, â&#x20AC;&#x201C; Annual, heat and electricity production and the sales per day according to the facilities of heat production (CHP, HOB, Incinerator) Fig. 5. Parametric entry of option for CHP system product

The user is to select the most desirable one among the list of options by referring to the technical specification for each option such as the unit capacity of the product,

â&#x20AC;&#x201C; Annual, fuel consumption per day according to the facilities of heat production (CHP, HOB)

159

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

correlations is a certain time, not a specific time during the year as in the original data. For example, if the DHC system is requested to produce more heat according to the request from CHP system to the amount of Q , the heat load of DHC system can be

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

F  f ( H , P) Where,

(1)

regarded to be changed from Q1 to Q2, i.e. Q2=Q1+ Q . Then, the operation behaviour for DHC system at the moment can be estimated simply from the 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.

F: Fuel consumption H: Heating load P: Electricity load

(a) Time vs. events

(a) Electricity production vs. fuel consumption

(b) Events vs. events Fig. 6. Illustrative diagram for the correlation between

energy production and fuel consumption

(b) Fuel consumption vs. Heat production

Fig. 6 shows the illustrating diagram for the mathematical correlation between energy production and consumptions. For any time t1, an optimized operation scenario already exists and corresponding heat and electricity production, and fuel consumption has been fixed according to the operation scenario and for any time t2, it is the same as above. On the basis of the operation data for a year, the behaviour of system operation can also be described between dependent variables (e.g. F: Fuel consumption, H: Heat production, P: Electricity production). In the correlations between dependent variables, the time t is reflected with implicit manner and the meaning of time t in the

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

SIMULATION OF THE THERMAL NETWORKING OPERATION 1. Operation Conditions and Schemes The operation of the overall system should be carried out by the order of priority of operation for the various heat sources. In this study, the basic schemes in order of priority for supplying the energy demands in newly developed area are established as shown in Fig. 8, 160

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

CHP system operation in A

Thermal networking operation using CHP in B

HOB operation in A

Thermal networking using HOB in B (a) Case A

(b) Case B Fig. 10. Comparison of energy load prediction: elec. Load Fig. 8. Schemes of the networking operation

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

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

For case A, the area is only comprised of residential purpose buildings, i.e. apartments, whereas for case B it is comprised of non-residential purpose buildings such as commercial buildings, offices, hotels, and hospitals. The annual hourly energy load data is estimated by using the energy load prediction module. The comparison of predicted energy loads, in the form of the annual distribution and the cumulative curve, are given as shown in Fig. 9 to Fig. 11.

(a) Case A

3. Operation Simulation Results For the test case comprised of only residential buildings, the cooling load is reflected on the electricity load by assuming that it is covered by the air conditioner or electric fan in individual houses. Consequently, the aspect of efficient utilization of the recovered waste heat during the summer is supposed to be a decisive factor in the optimization of the CHP system.

(b) Case B Fig. 9. Comparison of energy load prediction: heating load

161

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

(a) Gas engine

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

(b) Gas turbine Fig. 13. Annual thermal energy supply and demand operating condition for newly developing area

It is easy to see the typical consumption pattern of heating and hot water for residential houses in Korea as shown in Fig. 12. A large variation of heating load is observed in heat consumption rate and the optimal design of CHP system with such a large variation is more difficult than with a relatively regular consumption pattern. The typical annual operation results of the respective CHP prime-movers, gas engine and gas turbine, is also shown in Fig. 12. There is a large difference in the recovered waste heat prediction for respective CHP arising from the intrinsic feature for gas turbine, i.e. higher heat to power ratio of gas turbine against gas engine.

(a) Gas engine

Figure 13 shows the annual thermal energy supply and demand operating conditions for two distinct CHP prime-movers. First of all, the quantity of recovered waste heat from CHP is not large enough to cover the whole heat demand in the winter, so that most of heat demand is covered by HOB operation on-site. It is noted that the heat supply from DHC network seldom occurs during the winter. This is mainly due to the fact that it is also short of heat energy in existing developed areas during the winter. Of course, it is a probable scenario to operate the HOB in existing DHC system to produce the required amount of heat energy for newly developing area. However, it does not actually happen because the operation of the HOB on -site has priority over that of the HOB in existing DHC system according to the operation schemes.

(b) Gas turbine Fig. 14. Annual electricity supply and demand operating condition for newly developing area

The thermal network operation is observed to take place mainly during the intermediate seasons. It shows that most of heat demand is covered by the thermal networking heat supply and it results in bringing down 162

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. From the view point of system operation efficiency, it has a very positive impact in that the rate of operation of CHP in DHC system increases to some extent. However, in case of supply of surplus heat to existing DHC system as shown for gas turbine, it is vice versa. It is noted that the heat flow of thermal network can be bi-directional for the gas turbine as shown in Fig. 13.

Fig. 15 shows the variations of electricity production on existing DHC system side due to thermal networking operation. It is interesting to note that a minor increase of the electricity production for existing DHC system is observed during the intermediate seasons. This is caused by the increased rate of operation of CHP in existing DHC system due to thermal networking operation.

The annual supply and demand operating conditions for electricity are shown in Fig. 14. A comparatively good electricity-tracking operation is observed for both CHP prime-movers and the supply from the grid tends to increase during the summer due to the peak of the electricity demand.

The detailed variation of electricity production on the existing DHC system side is given as shown in Fig. 16. The net increase of electricity production for gas engines is larger than that of gas turbines. This is because of the intrinsic feature for gas engine CHP system of smaller heat to electricity ratio than that of gas turbine, which induce that more heat is supplied to on-site by the thermal network and consequently increase the rate of operation of CHP in DHC system.

Fig. 15. Variation of electricity production on existing DHC system side due to thermal networking operation (a) Gas engine

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

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

(b) Gas turbine Fig. 16. Detailed variation of electricity production on the existing DHC system side according to CHP prime mover

163

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

In case B, the analysis of operation characteristics for drawing optimal system configuration becomes much more complex due to the existence of cooling load. The simulation results for gas turbine with 50% of absorption type cooling (i.e. 50% turbo type cooling) are given in the following. Fig. 18 and Fig. 19 show the annual heat load and operating conditions of thermal energy supply and demand for on-site.

(a) Heat

As shown in Fig. 9, where the heating loads for case A and B are compared, the heating load for a group of non-residential building composition is much smaller than that of residential building composition. As a result, the waste heat recovered from gas turbine operation is sufficient enough to encompass the whole heat loads in case B as shown in Fig. 18. In terms of thermal networking operation, there is a great change in the pattern of system operation in that a large amount of surplus heat energy is available even in winter not to mention the intermediate seasons. This means that a large amount of heat is flowing toward the existing DHC system side as shown in Fig. 19, and there will be serious effects on the operation of existing DHC system.

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

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

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

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

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

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

164

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

is covered by operating HOB. In other words, only the operation of HOB on the DHC system side is affected by the thermal networking operation in winter. As shown in Fig. 20 (b) and Fig. 21, the production of electricity on the existing DHC system side during the intermediate seasons is certainly decreasing due to the supply of surplus heat from CHP on-site, which results in the diminution of the rate of operation for the CHP on the existing DHC system side. (a) Absorption type 80%

Fig. 22. Heat balance of operating the absorption chillers

(b) Absorption type 20% Fig. 24. Heat balance of operating the absorption chillers for different responsibility by absorption type cooling

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

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

Fig. 23. Annual LNG consumption rate by use for newly developing area of a grope of non-residential buildings

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

(a) Absorption type 80%

(b) Absorption type 20% Fig. 25. Annual electricity supply and demand operating condition for newly developing area for different responsibility by absorption type cooling 165

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The different operation characteristics in terms of electricity demand and supply is given in Fig. 25. It is noted that the electricity demand during the summer increases considerably as the ratio of absorption type cooling is decreasing. This peak of electricity during the summer is due to the consumption of electricity for operating turbo type chillers. From the view point of design of the CHP system configuration in the simulation, the cooling load is an important parameter to be considered carefully, because the capacity of CHP system is given in the form of any percentage of the peak value of electricity, i.e. the maximum value of the annual electricity consumption rate per hour. Therefore, the criteria for defining the CHP capacity is to be varied depending on the amount of cooling load assigned to turbo type chillers. The respective LNG consumption patterns depending on the ratio of absorption cooling load are compared in Fig. 26. It is interesting to note that the composition of fuel consumed by use is substantially changed according to cooling load treatment during the summer. The results confirm that the effects of various aspects of configuration for CHP and cooling system on the prediction of operational parameters (e.g. fuel consumption rate by use) are properly realized in the simulation program.

(b) Absorption type 80% Fig. 26. Annual LNG consumption rate by use for different responsibility by absorption type cooling

CONCLUSION A simulation program that predicts the energy loads for a mix of buildings and estimate the operational characteristics for networking operation of existing DHC system with CHP system on-site is developed. The distinctive features of this simulation approach can be summarized as follows, – The unit energy load models are developed for accurate prediction of energy consumption by use accroding to any combiation of building type and scale. – A simple mathematical correlation for reflecting the variations of the network operation on an existing DHC system side is newly proposed for the sake of simplicity and efficient simulation process.

By using the simulation approach as presented in this study, the optimal design of the CHP system in networking operation with DHC system can be carried out since one can access the detailed physical data regarding the whole operation of the network system such as annual rate of fuel consumption for respective systems (e.g. CHP, HOB, Chiller etc), annual production of electricity, heat, and the amount of heat exchange etc. Along with the appropriate cost structures for fuel, product sales (heat and electricity) and the estimation of capital cost, civil construction, and O&M costs etc, one can also make the assessement for the economic feasibility of various scenarios. However, the detailed economic analysis for the test cases and the procedures to determine the optimized CHP system configuration based on it will not be described in this paper due to the page constraints. These ítems will be discussed in further studies.

– The performance data for the commercial products, operation efficiency in a full and part load condition, has been extensively investigated and the database has been realized successfully on the simulation program. The operational characteristics of thermal networking operation has been assessed in terms of system configurations for the CHP and the cooling facility as follows. – According to the intrinsic features of the CHP prime movers such as gas engine and gas turbine etc, the aspects for the supply of surplus heat is progressing in different manners by and large. For a gas engine, the on-site is almost short of heat so that the prediction results indicate that the additional operation of CHP on the exisiting DHC system side is induced in the intermediate seasons. Whereas, surplus of waste heat recovered from gas turbine CHP is supplied toward the existing DHC system side. As a result, the amount of electricity production is being decreases to some extent. – In case of a group of non-residential buildings, the heating load reduces considerably. Therefore, it is probable that the heating load can be covered by only the recovered waste heat from on-site even in the winter. Due to the heat flow toward the DHC system

(a) Absorption type 80%

166

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

side in the winter the rate of operation of HOB will be decreased.

[4] H. Lund, F. Hvelplund, I. Kass, E. Dukalskis, D. Blumberga, ―District heating and market economy in Latvia‖, Energy, 1999, Vol. 24, pp. 549-559.

– The thermal energy exchanges via the network and the corresponding changes in operation on both sides are prevailing in intermediate seasons in case of similar heat consumption patterns on both sides.

[5] H. C. Park, M. Chung, S. H. Kim, ―Development of system simulator for community energy system‖, Report to Ministry of Industry, 2003.

– The operation of cooling system on the newly developing area is verified not to have much effects in terms of thermal networking operation. However, the significant changes in the LNG consumption patterns by use are observed according to the ratio of responsibility by absorption chillers for the cooling load.

[6] Y. H. Im, H. C. Park, M. Chung, ―A study of optimal heating supply systems for the newly developing area in the vicinity of DHC system supplying area‖, Report to Korea District Heating Corporation, 2006 [7] Y. H. Im, M. Chung, H. C. Park, ―Feasibility study for small size cogeneration systems in the metropolitan areas of Seoul‖, Final Report to SH (Seoul Housing) Corporation, 2008.

The various aspects of system configuration in terms of CHP system optimization are discussed with the development of a simulation program in this study. It is verified that the physical and mechanical mechanisms concerned with the thermal networking operation has been appropriately modeled from the assessment of operational behavior for test cases.

[8] M. Chung, H. C. Park, ―Development of a energy demand estimator for community energy systems‖, Journal of the Korean Solar Energy Society, 2009, Vol 29, pp. 37-44. [9] M. Chung, H. C. Park, ―Development of a software package for community energy system assessment – Part I: Building a load estimator‖, Energy, in press.

ACKNOWLEDGEMENT The author gratefully acknowledges the financial and 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 Journal of Air-Conditioning and Refrigeration Engineering, 2003, Vol. 15, pp. 1088-94.

REFERENCES [1] Korea Energy Management Corporation, ―Statistics for district heating and cooling enterprise in Korea‖, 2009.

[11] H. C. Park, M. Chung, ―Building load models for hotels in Korea‖, Journal of the Korean Solar Energy Society, 2009, Vol. 29, pp. 48-57.

[2] A. Marbe, S. Harvey, ―Opportunities for integration of biofuel gasifiers in natural-gas combined heatand-power plants in district-heating systems‖, Applied Energy, 2006, Vol.83, pp. 723-748.

[12] H. C. Park, ―Development of weighting factors for variables associated with hourly energy consumption pattern for hotels in Korea‖, SAREK (Soc. Air-conditioning, Ref., Engineers of Korea) Winter Annual meeting, 2002, pp. 76-82

[3] C. Weber, I. Heckl, F. Friedler, F. Marechal, D. Favrat, ―Network synthesis for a district energy system: a step towards sustainability‖, Computer Aided Chemical engineering, 2006, Vol. 21, pp. 1869-1874.

[13] H. C. Park, ―Analysis of energy loads for hospital buildings‖, SAREK journal, 2002, pp. 1088-93.

167

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

IMPROVED PRIMARY ENERGY EFFICIENCY OF DISTRICT HEATING NETWORKS BY INTEGRATION OF COMMUNAL BIOMASS-FIRED COMBINED HEAT AND POWER PLANTS WITH BIOMASS PYROLYSIS 1

T. Kohl , N.A. Pambudi , T. Laukkanen and C.-J. Fogelholm

Aalto University, Dept. of Energy Technology, Espoo, Finland Corresponding Author: Thomas Kohl, e-mail: thomas.kohl@tkk.fi 2 Semarang State University, Semarang, IndonesiaAbstract

Furthermore the scarcity of the biomass available demands most efficient use of this resource.

ABSTRACT This paper investigates the influence of the integration of communal biomass-fired combined heat and power plants with wood-pyrolysis on the plant‘s energy balance and product distribution. Further the proposed integration concept‘s influence on the environmental performance of the connected district heating network is pointed out. The environmental performance is evaluated by means of the primary energy factor and the CO2 emission coefficient. For this evaluation, the European standards EN 15603 and EN 15613-4-5 are applied and modified.

As shown in a previous study [1] it looks promising to integrate biorefinery processes, that are linked to transportation fuel production, with CHP plants, since CHP plants can provide both a source for high temperature heat needed for thermal conversion of biomass as well as the district heating network (DHN) as a sink for sensible heat that would usually be rejected in stand-alone biofuel refineries. It has been further worked out that the integrated production of interstage products, such as liquid fast pyrolysis product (often referred to as woodoil) and wood pellets, have several advantages: Firstly, the products are independent from the transportation fuel market developments since they can be seen as a universal input for different upgrading processes to e.g. biodiesel, ethanol, methanol, hydrogen or other chemicals production but they can also be directly combusted for power and heat generation. Secondly, they increase the biomass‘ energy density making it more sustainable for transportation to central plants required for economic fuel production. Thirdly, technologies applied for such pre-processing are relatively simple and robust, thus keeping investment cost and system complexity on a reasonable level and making it therefore also interesting for local small-scale solutions.

The concept comprises the integration of a simple pyrolysis model and of a steam dryer with a base case combined heat and power plant. The yearly plant output is calculated by applying a multiperiod model of the heat duration curve. The work shows that, by co-generation of valuable pyrolysis product, operation hours and electricity production can be considerably improved. The integration also clearly improves the district heating network‘s primary energy efficiency and lowers its carbon dioxide emissions significantly. INTRODUCTION The European Union‘s carbon dioxide mitigation goals and plans to reduce energy import dependency require action towards a more sustainable energy supply that is based on renewable energy sources available in the member states. Biomass is discussed controversially due to its wide range of upgrade possibilities from power, heat, cooling to chemicals and transportation fuels. Among others, EU directives 2001/77/EC (―…promotion of electricity produced from renewable energy…‖), 2004/8/EC (―…promotion of cogeneration…‖) and 2003/30/EC (―… promotion of the use of biofuels…‖) state that the use of biomass for energy purposes should be expanded on a sustainable base.

In this paper, outgoing from a base case, we simulate the retrofit integration of wood fast pyrolysis with an existing wood-fired CHP plant. The aim is highest possible pyrolysis product generation using the free boiler capacity in part loads under the condition that the district heat (DH) demand is still fulfilled. With help of a multiperiod model of the DHN‘s heat duration curve, the work shows the influence of the integration on plant operating hours, electricity production and biomass throughput. In addition the effects on the DHN‘s primary energy factor and CO2 emission coefficients are studied as well. The primary energy factor and the CO2 emission coefficient are calculated according to European standards EN 15603 [2] and 15316-4-5 [3], applying a modified power bonus method. However, no cost estimation is given, since the focus of the work was to find out if this integration

However, the increased use of biomass is expected to raise prices for biomass which will negatively influence, among others, the economy of communal biomass-fired combined heat and power (CHP) plants – a technology that is currently competitive to fossil energy production. 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 CHP plant.

Heat Duration Curve - Multiperiod Model Integrated Case - Lower Loads

District Heat Load [MW]

In the following, first the used multiperiod load model is described. Further a brief introduction to biomass fast pyrolysis is given and it is shown how the process has been simulated and integrated. Then the modification of the European standards is explained, results are presented and finally restrictions of the work and options for further improvement are discussed.

0 0

100

DISTRICT HEATING LOAD

200

250

300

350

Time [d]

The CHP plant chosen has been integrated into a virtual DHN. Therefore yearly data of a real DHN has been scaled so that the CHP plant provides 60% of the hourly peak demand of the DHN when on full load. The CHP plant is assumed to be shut off at 50% load which corresponds with 30% demand in the DHN. As stated in [4], those are common operating parameters for communal solid fuel-fired CHP plants.

District Heat Load

Real CHP DH Load

Multiperiod Model DH Load

Fig. 1b: DH Load Multiperiod Model – Integrated Case

CHP PLANT INTEGRATED WITH WOOD PYROLYIS Wood Pyrolysis Model Biomass fast pyrolysis is the thermal conversion of biomass in the absence of oxygen at temperatures of approximately 500 °C and pressures close to atmospheric [5]. The basic idea of the pyrolysis unit is derived from the bioliq® process developed by the Forschungszentrum Karlsruhe (FZK). There, fast pyrolysis is applied in order to yield a high share of liquid pyrolysis product. Biomass is indirectly heated and pyrolysed with sand in an inert atmosphere at a temperature of about 500 °C. Subsequently, the pyrolysis gases are condensed and the liquid fraction (also referred to as wood oil) is mixed with the coke and forms the so-called bioslurry which leaves the plant as the final product. In this work we use data published by FZK [6] and hence assume that 90% of the biomass‘ energy is converted into bioslurry whereas 10% accrues in gaseous form. The pyrolysis gas is thought to be cofired in the boiler and hence its energy is subtracted from the fuel input into the boiler.

In order to represent the yearly production of the base case plant a multiperiod load model was developed. One full load and five part load levels have been chosen to represent the heat duration curve. The pyrolysis integrated CHP plant is represented by 7 part load levels since lower DH loads can be supplied, as explained later. Operating time periods per part load level are set of equal length and -together with the full load period- match the total operation hours and yearly DH generation of 94.5 GWh as shown in figures 1a and 1b. For each load level, fuel input and pyrolysis yield are then iterated matching the required DH output. DH demand not provided by the CHP plant is assumed to be generated in oil-fired heat-only boilers with a thermal efficiency of 0.85 Heat Duration Curve - Multiperiod Model Base Case 25

District Heat Load [MW]

150

As pyrolysis requires a low fuel moisture content of approximately 10% [5] a dryer must be integrated as well. Indirect steam drying is applied, since this also allows the regulation of the DH load. As explained later, regulation is necessary since the enthalpy of the steam flow after the modification exceeds the demand of the DHN and hence must be adjusted.

0 0

100

150

200

250

300

350

The wood pyrolysis process is modelled as follows: The heat of pyrolysis of wood is set to 1.87 MJ/kg (moisture content 10%) using data for pine derived from [7]. Therewith the pyrolysis yield is calculated from the heat extracted from the flue gases.

Time [d] District Heat Load Multiperiod Model DH Load

Real CHP DH Load

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

CHP Plant – Base Case A base case CHP plant with a bubbling fluidized bed boiler (shown in fig. 1) has been simulated in full and 169

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

Table 1: Base Plant – Input Specification and Performance at Design Load Simulation model input data - design load Wood Fuel Ultimate analysis

50.64

42.22

6.10

0.16

Ash

0.8

0.08

HHV

18.8 MJ/kg

Moisture

50 %

Fuel input

26 MW

Steam Cycle High pressure steam

60 bars 510°C

Condenser pressure

0.69 bar

District heat output

16.5 MW

Electrical efficiency ηel

0.243

Power output

6.3 MW

Power to heat ration α

0.381

Plant Performance

Fig. 1: CHP Plant – Base Case

CHP Plant – Integrated Wood Pyrolysis

CHP plant – Integrated Steam Drying

The modified CHP plant is illustrated in Fig. 2. In order to provide heat for the pyrolysis process, the heat must be extracted from the flue gases leaving the fluidized bed reactor (numbered 3 in Fig.2) boiler at 850 °C. The required amount of flue gas is split off (18) after the fluidized bed reactor. As in the FZK process, those flue gases are thought to heat up sand to 550 °C (which would provide the heat for the pyrolysis process by cooling down to 450 °C) [6]. The flue gas thereby is estimated to cool down to 480 °C. The flue gas is then mixed back (20) into the main flow before the economizer. The heat extraction needed for biomass fast pyrolysis process is modelled by help of an additional evaporator (19). 90% of the biomass energy on a lower heating value base will form pyrolysis slurry whereas 10% accrues as pyrolysis gas. The energy carried by the pyrolysis gas reduces the biomass fuel input as explained above (―Wood Pyrolysis Model‖).

The dryer is modelled as a steam tube dryer. Paying attention to the retrofit situation, live steam is extracted, throttled to 10 bars and further cooled to 190 °C by spraying in the saturated water leaving the dryer. Drying of biomass to low moisture contents requires temperatures far above the saturation temperature at a given pressure due to the hygroscopic properties of biomass. Heat consumption of the dryer has been estimated to 2750 kJ/kg water evaporated [9]. Wood and hot flue gases are led in the dryer (24). If heat is available from the flue gases, those are cooled down to 120 °C and together with the fully condensing steam provide the heat needed for the drying process. Dried wood leaves the dryer at wet bulb temperature. For the drying process live steam is extracted (21), throttled (22) to 10 bars and further cooled to 190 °C by spraying in condensate (23) leaving the dryer. The dryer condensate is throttled to 2 bars (26) and send to the feedwater tank (15). Flue gas temperatures of 120 °C 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 problems, especially not with low-sulphur wood fuels. Certainly this design specification must be reconsidered in case of changed fuel properties.

pyrolysis yield constantly increases with the decrease of the DH levels down to 60%. The maximum flow off the dryer (and thus its capacity) is be restricted by the pressure prevailing in the feedwater tank, which in turn is given by the extraction pressure after the turbine stage (11). The pressure decreases with falling live steam parameters and steam massflow. Hence, there is a pressure dependant maximum enthalpy flow that can be fed into the feedwater tank until saturation state is reached for the mixture of the condensates from the DH exchanger (13) and the dryer (24). In order to overcome this restriction the feedwater tank pressure has been increased load-dependently to a maximum of 2 bars matching its design pressure. However, due to the reason mentioned above, for loads below 60% the heat that would need to be ―dissipated‖ in the dryer (in order to match the DH load) would result in such a high dryer condensate heat flow which again would bring the feedwater beyond saturation state. Hence for those cases the boiler load is gradually decreased, resulting in lower pyrolysis yields. The lowest DH load level that can be represented is 28.6% of the plant‘s full load. Compared to a minimum load of 50% in the base case – which is given by the minimum fuel input required for stable combustion conditions in the boiler-, the integrated process offers possibilities to increase the operating hours of the CHP plant considerably.

The maximum pyrolysis production for each load point is restricted by the maximum steam extraction rate and by the boiler‘s maximum burning power. The maximum possible pyrolysis yield logically requires highest possible fuel input since heat must be provided both, for drying and pyrolysis. Conversely, this means that the steam enthalpy exceeds the demand of the DHN. This is because the boiler temperature is controlled by means of the evaporator- and superheater tubes in the boiler walls. If now, the heat input in the boiler is kept on a higher level as usual the water amount needed to dissipate the heat from the boiler walls is only decreasing to a certain amount (resulting from a reduced temperature after the economizer). Consequently, in order to match the DH load, this heat must now be ―dissipated‖ in the pyrolysis heat exchanger (19) or in the dryer (24). By iteration the DH load is matched by adjusting dryer load, correlated splitoff to the pyrolysis heat exchanger and fuel input. In all cases the boiler load (characterised by the fuel heat input) is restricted to 100%. So, the overload back-up capacity of the boiler is maintained. With this setup the

Fig. 2: CHP Plant with integrated pyrolysis and steam drying

171

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

In the power bonus method fEl. is defined as the PEF of the electricity that is thought to be replaced by the power generated in the CHP plant (for instance, in this study the average power generation efficiency in Finland is used). This allocation pays attention to the fact that the co-generated electricity is more sustainable due the CHP process‘ high overall efficiency. The PEF of the DHN can thus be determined according to;

APPLICATION OF THE PRIMARY ENERGY CONCEPT ACCORDING TO EN 15603 Primary Energy Concept The EU standard EN 15603 [2] handles the energy performance of a building as a whole and gives guidelines how energy use and production of a building shall be calculated. In order to aggregate the different forms of energy produced and used within the building, primary energy (PE) and CO2 emissions are accumulated and expressed by means of primary energy factors (PEF) and CO2 emission coefficients, respectively. PE is energy that has not been subjected to any conversion or transformation process [2]; it is hence not yet extracted from the source. In the PE approach described in EN 15603, all energy carriers involved in the generation process are retraced to their sources and all energy needed to deliver the final energy product are aggregated to the total PE consumption and CO2 emissions. Thus the PE approach applies the holistic principles of life cycle assessment to an energy rating procedure. By retracing energy consumption to the source, the system boundaries automatically include the whole world, and thus depict the real impact of the system concerning energy consumption and CO2 emissions.

f DH 

 E F  P  f El . QDH

f

F ,i

 E F ,i  P  f El .  E Pyro  f Pyro

QDH

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

The total primary energy factor is the sum of all PE input to the energy system divided by the useful energy delivered at the system border. It thus describes how much PE input is needed in order to obtain one unit of energy used and can hence be seen as an inverted efficiency.

Table 2: Primary energy factors and CO2 emission coefficients for fuels and products kg/MWh 2 fBM

In standard EN 15316-4-5 [3] more detailed guidelines for the calculation of PEFs of DH systems are defined. According to EN 15316-4-5 PEFs can be calculated for a certain part of the energy system. In this study the system boundary comprises the power plants and the DHN.

fOil fEl.

fPyro 1

1.09

2 cCO2/BM

1.35

cCO2/Oil

3.11

1 cCO2/El.

1.28

cCO2/Pyro

14 330 270 1

: value is calculated, : value is taken from EN 15603, Annex E

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

The PEF of the DHN has been calculated applying the power bonus method. If yearly demand data of the DHN and the generation data are known, the PEF of the DHN can be calculated by applying the so-called power bonus method. The power-bonus method is derived from the energy balance of the building which can be written as: F ,i

F ,i

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

Primary Energy Factor

f

 EF  f DH  QDH  P  f El . ,

where EF, QDH and P are the heat of the fuels used, DH and power co-generated respectively. fF,i, fDH and fEl. are the PEFs of the fuels used, the DHN and of the cogenerated power.

172

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

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

CO2 Emission Coefficient The CO2 rating is done by calculating CO2 emission coefficients (cCO2) that quantify the total amount of fossil fuel derived CO2, emitted to the atmosphere, per unit delivered energy. As for the primary energy factor the system boundary comprises of power plants and DHN. Also the power bonus method is applied for calculating the DHN‘s specific CO2 emissions. For the sake of completeness it must be mentioned that CO2-equivalent emissions of other greenhouse gases can optionally be included. However this has not been implemented into this study, due to a lack of data. Similarly to the PEF the CO2 emission coefficients for the base case are calculated as:

E

cCO2 , DH 

F ,i

 cCO2 , F ,i  P  cCO2 , El .

QDH

And for the modified plant as:

cCO2 , DH 

E

F ,i

 cCO2 , F ,i  P  cCO2 , El .  EPyro  cCO2 , Pyro

QDH

EF,i, EPyro, P and QDH represent heat in fuels, heat in pyrolysis slurry and co-generated electricity and DH respectively. Accordingly, cCO2,F,i, cCO2,Pyro, cCO2,El. and cCO2,DH are the related CO2 emission coefficients. The corresponding values are given in table 2. RESULTS In table 3, three simulation cases are presented: the base case (case 1), pyrolysis integration with the same operation hours (case 2) and the maximum pyrolysis slurry production (case 3) with prolonged operation

173

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia Table 3: Results – Multiperiod Model Base Case - Case 1 CHP DH Load

[%]

100

Time

[h]

2440

530

212 days

Fuel Input

[MW]

25.90

23.19

20.39

17.47

14.58

11.91

109.56 GWh

Power

[MW]

6.29

5.64

4.91

4.06

3.22

2.54

26.13 GWh

District Heat

[MW]

16.50

14.85

13.20

11.55

9.90

8.25

70.85 GWh

total

CHP + Pyrolysis - Case 2 CHP DH Load

[%]

100

Time

[h]

2440

530

212 days

Fuel Input

[MW]

25.90

36.49

44.24

52.42

60.21

53.88

194.13 GWh

Power

[MW]

6.29

5.54

4.69

3.71

2.88

2.25

25.45 GWh

District Heat

[MW]

16.50

14.85

13.20

11.55

9.90

8.25

70.85 GWh

Pyrolysis Slurry

[MW]

12.21

21.16

30.60

39.58

36.76

74.31 GWh

total

CHP + Pyrolysis - Prolonged Operation Hours - Case 3 CHP DH Load

[%]

100

Time

[h]

2266

633

Fuel Input

[MW]

25.90

36.49

44.24

52.42

60.21

50.97

39.56

28.88

256.62 GWh

Power

[MW]

6.29

5.54

4.69

3.71

2.88

2.27

1.71

1.18

28.17 GWh

District Heat

[MW]

16.50

14.85

13.20

11.55

9.90

8.25

6.60

4.95

81.26 GWh

Pyrolysis Slurry

[MW]

0.00

12.21

21.16

30.60

39.58

33.79

26.11

19.00

115.46 GWh

total 279 days

Table 4: Results - PEF and CO2 Coefficient Base Case Case 1

CHP + Pyrolysis Case2

CHP + Pyrolysis Prolonged Operation Case 3

Required Backup Power

MWh

27.8

15.5

Total PEF

[-]

0.80

0.79

0.68

CO2 Coefficient

kg/MWh

38.6

42.1

-5.3

Another open question is the influence of the real pyrolysis gas on the combustion temperature and flue gas properties. In order to gather more details of the pyrolysis process a simple pyrolysis model is currently under development. Together with the power plant model the integration can be further optimised aiming for highest PEE along with low CO2 emission coefficients.

CONCLUSION AND DISCUSSION The work shows that by integration of a CHP plant with wood pyrolysis operation hours can be increases by 30%, a valuable product can be co-produced and PEE as well as the CO2 emission coefficient of the DHN can be substantially improved. As next steps more comprehensive data of the fuel supply chain should be implemented to get more realistic values that will approve the trend shown with this work. The process can be further improved by integrating heat that is set free during the condensation of the pyrolysis liquid and gaseous product. The heat is available in a temperature range from approximately 500 °C to 25 °C and could hence be used for steam superheating, feedwater preheating, but also for DH generation. This integration is not a simple task since many plant parameters influence each other. The heat integration must be carried out together with a pinch analysis to assure an energy efficient integration.

Further an economic analysis should be carried out in order to show potential economic benefits. The integration itself seems to be viable – a statement that is supported by a press release from June 2009 where boiler manufacturer Metso and forestry company UPM announced the development of a new viable fast pyrolysis process benefitting from the integration with a CHP plant [11]. Concerning the European standards used for evaluation, it can be said that the power bonus method can be easily adapted to a polygeneration concept 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 likely can also be extended to other possible biorefinery products as long as those are ―energy‖ products. This expansion option could be implemented into the standard.

[4] Savola, T., ―Modelling biomass-fuelled small-scale CHP plants for process synthesis optimisation‖, Doctoral Dissertation, Helsinki University of Technology, Espoo 2007 [5] Bridgwater, A.V., 2000, Fast pyrolysis processes for biomass, Renewable and Sustainable Energy Rev., 4(1), pp. 1-73.

However the most difficult question remains how the PEF of other, less common co-products should be determined. In the case of pyrolysis slurry it is not possible to find good average production efficiencies since the technology is not yet on the market. But, how is co-generation of cooling evaluated?

[6] Henrich, E., 2007, ―The status of the FZK concept of biomass gasification‖, 2nd European Summer School on Renewable Motor Fuels, Warsaw. [7] Daugaard, D., Brown, R., ―Enthalpy for Pyrolysis for Several Types of Biomass‖, Energy & Fuels 2003, 17, 934-939

In general it can be said that the implementation of the process will be strongly dependant on investment cost and on the market value of the product. The product value is currently difficult to predict and also its future price development will be strongly dependant on the use of biomass in the future.

[8] http://www.ecn.nl/phyllis: PHYLLIS is a service provided by the Energy Research Centre of the Netherlands – ECN, 17.9.2009 [selected subgroups: untreated wood  birch and fir/pine/spruce].

Summarising it can be said that even though many questions still need to be answered, this works shows clearly that the integration of communal CHP plants with wood pyrolysis is beneficial concerning the connected DHN‘s PEF and CO2 emission coefficient. Vice versa it also shows that CHP plants can play an important role in the sustainable bio-refineries of the future.

[9] Brammer, J., Bridgwater, A., ― Drying Technologies for an integrated gasification bio-energy plant‖, Renewable and Sustainable Energy Reviews 3 (1999) 243-289 [10] Dones, R. et al, 2004, Life Cycle Inventories of Energy Systems: Results for Current Systems in Switzerland and other UCTE Countries, ecoinvent report No. 5, Paul Scherrer Institute Villigen, Swiss Centre for Life Cycle Inventories, Dübendorf, CH, p.170.

ACKNOWLEDGEMENTS This work is part of the Primary Energy Efficiency project of Nordic Energy Research.

[11] N.N., press release on http://www.metso.com/news/newsdocuments.nsf/w eb3newsdoc/C89A8AC3F77ABD29C22575CF003 111F5?OpenDocument&ch=ChMetsoWebEng

The funding of the Graduate School of Energy Science and Technology (EST) is gratefully acknowledged. REFERENCES

[12] H. Lund, F. Hvelplund, I. Kass, E. Dukalskis, D. Blumberga, ―District heating and market economy in Latvia‖, Energy, 1999, Vol. 24, pp. 549-559.

[1] Kohl, T., Järvinen, M., Fogelholm, C.J., ―Gasification and biorefinery in combined heat and power plants‖, Proceedings of the 11th International Symposium on District Heating and Cooling, Reykjavik, Iceland, 2008

[13] H. C. Park, M. Chung, S. H. Kim, ―Development of 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 ratings‖, European Committee for Standardization, CEN, Brussels.

[14] Y. H. Im, H. C. Park, M. Chung, ―A study of optimal heating supply systems for the newly developing area in the vicinity of DHC system supplying area‖, Report to Korea District Heating Corporation, 2006

[3] EN 15316-4-5:2007, ―Heating systems in buildings. Method for calculation of system energy requirements and system efficiencies. Part 4-5: Space heating systems, the performance and quality of district heating and large volume Systems‖, European Committee for Standardization, CEN, Brussels, 2007

[15] Y. H. Im, M. Chung, H. C. Park, ―Feasibility study for small size cogeneration systems in the metropolitan areas of Seoul‖, Final Report to SH (Seoul Housing) Corporation, 2008. [16] M. Chung, H. C. Park, ―Development of a energy demand estimator for community energy systems‖, Journal of the Korean Solar Energy Society, 2009, Vol 29, pp. 37-44. 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 package for community energy system assessment – Part I: Building a load estimator‖, Energy, in press.

[20] H. C. Park, ―Development of weighting factors for variables associated with hourly energy consumption pattern for hotels in Korea‖, SAREK (Soc. Air-conditioning, Ref., Engineers of Korea) Winter Annual meeting, 2002, pp. 76-82

[18] H. C. Park, S. S. Lee, D. J. Kim, ―Development of energy models for department stores‖, Korean Journal of Air-Conditioning and Refrigeration Engineering, 2003, Vol. 15, pp. 1088-94.

[21] H. C. Park, ―Analysis of energy loads for hospital buildings‖, SAREK journal, 2002, pp. 1088-93.

[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

CHP OR POWER STATION? – QUESTION FOR LATVIA 1

D. Blumberga , G. Kuplais , F. Romagnoli and E. Vigants 1

Riga Technical University – Institute of Energy Systems and Environment Kronvalda boulv. 1, LV-1010 – Riga, Latvia In CHP station this dominance is almost total and very high in district heat supply boiler houses. As it well known Latvia is a great consumer of imported fossil from one side but in the same the share of renewable energy resources is one of the highest of Europe.

ABSTRACT This paper presents aspects and problems of the Latvian energy-system connected to the choice of the CHP and/or power stations for the future national energy strategies. In the light of the last EU directive in the subjects of Renewable Energy Sources (RES) the share of electricity produced from RES at the moment is attested on the value of 42.4 % but should be increase to 49.3%. In the same time the share of renewable energy resources in the final energy consumption for 2020 should reach the level of 40% from 30%.

The use of specific energy resource depends on energy supply policy, and total consumption of energy resources depends on development of every type of energy resources in regions. Now there is unjustified high proportion of fossil fuel in state energy balance which is possible to reduce by a beginning of active use of local fuel in regions. The EU directive also requires that Latvia in the year 2015 would generate 49% of electric power from renewable resources (currently it is 45%) [3]. This is supportable, but the power supply of Latvia cannot be let out of the sight and this issue is already problematic.

Dependence on imported energy sources, growth of electricity prices, the need to support local producers are the main reasons for the use of new renewable energy technologies in the Latvian energy sector to implemented in refurbished energy supply system.

Latvia has some electricity production from cogeneration plants and some from hydro-power plants. However, the production of electricity from the hydropower plants fluctuates a great deal from year-to-year. The rest of the electricity for consumption is imported from the neighbouring countries.

Several methods fro the evaluation of the best strategy are explained. This apaper summarizes the application of the Energy Indicators for Sustainable Development (EISD) as good tool for analyzing trends, setting energy policy goals and monitoring progress. The results from the application of a multi-objective optimization regarding the implementation of the landfill biogas in the biogas treatement plant ―Daibe‖ are reported.

In order to understand the role played by CHP and power plant it is fundamental to understand the actual situation in Latvia for thermal energy where more than a half of Latvia district thermal energy is distributed and consumed mainly in Riga.

1. INTRODUCTION

Latvian heating primarily is performed on a centralized basis and after the used of wood energy the natural gas imported from Russia is the main source (see Fig. 1).

The structure of energy user in Latvia is characterized by high energy consumption in households, public and service sectors, comparing with relatively low consumption in rural and industrial sector. In light of this situation, for the power sector development, special tasks are required in connection to the choice of the more adequate energy resources in order to ensure the best energy production and supply. Consequently question on which direction address the main efforts for the energetic national improvement is still actual: CHP or Power station? If fuel, which is used to produce heat and electrical energy in Latvia, is taken into account, the dominant one is gas [1] and consequently appear evident how the Latvian dependence on foreigner energy supplies (mainly from Russian) is not only a weak point in connection to the energetic sustainability but can serve as a convenient way of exerting economic pressure [2].

Fig. 1. Main resources used for local and individual heat supply [3] 177

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

of the gross energy-consumption from renewableenergy sources in Latvia in the year 2007.

This type of heating supply scheme means that consumers are grouped and heating is performed from heat source which is intended for the consumer group.

The most important domestic renewable-energy resource in Latvia is biomass in the form of fuelwood: in fact approximately 45% of Latvia is covered with woods and this substantial area makes wood a significant potential as a resource for energy supplies.

About 70 % out of this thermal energy volume is produced in the cogeneration cycle (only in Latvenergo owned CHP, and Rigas siltums) and around 30% of centrally supplied heat energy is produced in Riga CHP plants and boiler houses. Of course as main fuel in Riga natural gas is used approximately 98% of thermal energy is produced from natural gas (CHP plants and boiler houses together) [3].

Even though the share of renewable is one of the most large Europe the EU directive fixes the target of 40% share of renewable energy resources in the final consumption in 2020.

As for heat supply outside of Riga, the dominant thermal energy is produced in boiler houses with relatively high proportion of local fuel usage (as shown in Fig.1). Outside of Riga CHP heat production rate does not exceed 5%.

This means that the increase is not feasible without the need of refurbishment and/or construction of energetic infrastructures. The fact that Latvia has domestic renewable-energy resources makes it interesting because the utilisation of the domestic fuels would be sustainable both from an environmental and an economic point of view.

2. EXISTING ENERGY SITUATION IN LATVIA: SHORT OVERVIEW

Latvia has comparatively well developed power, natural gas supply and district heating systems, and as a consequence the electricity is basically produced by hydro power plants and by cogeneration plants, which are operated according to district heating demand, and part of electricity is imported (fig. 2). Consequently the main objectives of the Latvian energy policy now are to ensure sustainable accessibility to necessary energy resources and security of supply in order to favourite the economic growth and improve quality of life, to ensure environmental quality retention and meet the objectives set in the Kyoto protocol of UN FCCC and Latvian Climate Change Program.

During the recent past central (large) power plants in Latvia supplied roughly 65% of the total annual power demand - distributed energy resources (DERs) covered 3–6%, but the rest were received as import supplies from Estonia, Lithuania and Russia (mainly) [4]. Regarding fuel sources Latvia has no real fossil-fuels of its own and the consumption must be imported. However Latvia uses the domestic renewable-energy resources hydro-power and biomass. Table I. – primary energy-consumption in Latvia in the year 2007 [1] %

Natural gas

56.92

27.8

biogas

0.32

0.16

Biodiesel

0.07

0.03

Oil products

73.33

35.8

Fuelwood

48.47

23.7

Hydroenergy

9.84

4.8

Import of Electricity

10.80

5.3

Import Coal and coke

4.36

2.1

Wind

0.19

0.09

Biodiesel

0.07

0.03

Total

204.6

Electricity Supply in Latvia Electricity amount, billion kWh

8 7

imported electicity

wind generators

small HPS

small CHP

CHP

HPP

1 0 2000

2001

2002

2003

2004

2005

Year

Fig. 2. Electricity supply in Latvia (Source: state JSC Latvenergo, Ministry of Economics, Central Statistical Bureau)

2.1 Lack of energy sources for electricity The main domestic electricity capacity consists of 1517 MW of hydro and 520 MW [5] of thermal (CHP units in Riga) all of which is controlled by the state company, Latvenergo. The generating potential mainly consists of three hydro power plants (HPP) on the Daugava River, hence directly dependent on the river‘s water flow. Due to small reservoirs, utilization rates are low and the

The use of primary energy for the gross energyconsumption in Latvia can be seen in Table 1. The share of renewable energy in the gross energyconsumption is made up of fuelwood and hydro energy. That means that there is approximately a total of 30% 178

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

production is quite seasonal following the water flows. The amount of power produced by the Daugava river HPP cascade is average 2.6â&#x20AC;&#x201C;2.8 TWh [1] annually, reaching in the years, rich by spring floods and rain even 4.5 TWh [ [5].

equipment with this capacity heat production by cogeneration would be a maximum.

More in detail the three HPPs, located on the river of Daugava, form a sort of cascade with the relative capacity of: Plavinas 870 MW, Kegums 263 MW and Riga 402 MW. Almost two thirds of hydro electricity is produced in the spring month of March, April and May. In this period the supplies are from the hydro plants. In the high demand winter season amount of electricity generated by hydro plants is relatively low. Looking the electricity supply statistics [1] the national production of electricity is around 10.0 PJ where the 9.8 are produced using hydro energy and 0.2 PJ produced by wind energy. The net electricity import (including the amount of energy exported) is around 10.8 PJ approximately the 50% of the national supply. These figures shows the lacks of energy sources in the national system and seems reasonable to foreseen a more large fraction of other energy sources for the production of electricity, the main question is on which methodology base this strategy .

Fig. 3. Heating energy distribution by cities in Latvia

If we are looking at the district heating division of Latvia a huge difference can be seen in quantity of heat supply in Riga and the rest of Latvia (see fig. 3) Two large CHP plants, Riga TPP-1 with an installed electric capacity of 144 MW and Riga TPP-2 (390 MW), are located in Riga [5]. CHP plants are the main heatgenerating sources of heating networks of Latvian capital. Power is produced mainly in cogeneration mode, according to the heatâ&#x20AC;&#x201C;load curve.

2.2 Well organized and developed DH system Latvian heating primarily is performed on a centralized basis consequently consumers are grouped and the heat is supply from heat source which is established for a certain consumer group. The heat source power, depending on type of consumer group, varies from the range of kW to several hundred of MW. In general lower power can correspond to building groups, individual houses or even apartments heating. Residential and separate heating of individual houses belongs mainly from the decentralized heating. One of the benefits of district heating is centralization of heat load, which gives a possibility to increase the heat source power and to form basis for the development of cogeneration power. For large heat consumers in Latvia (mainly heating systems in large cities like Riga) large cogeneration plants are installed. The customers who are not connected to a district heating cannot be provided from this system. In the other regions far from the big cities the heat supply system is mainly based on district heating, consequently it means that that there possibility for a CHP development.

During the heating season, when there is a substantial demand for heating and hot water, Riga CHP plants produce approximately 80% of the total annual production volume, while during summer the volume of production reduces [5]. Nowadays Riga CHP plants cover about 20% of the total annual power demand of Latvia [5] . The main fuel used in Latvia biggest cities is natural gas and the rates of thermal energy are 75% - 85% [3]. In Riga and other cities where most part of the heat is produced in cogeneration cycle, the increase of rates was not so high and currently (in the autumn of 2009) heat rates are lower that in the cities where wood chips are used. From the thermal energy point of view seventy percent of the heat in Latvia is supplied from district-heating systems either from boiler houses or co-generation: 37% of the district heating in Latvia was produced by means of co-generation plants [6]. This means that 63% of the district heating is produced in boiler houses [6]. This means that there is potential to replace some of the heat plants with co-generation units (Eighty percent of the district heating in Denmark is supplied from CHP [6]).

CHP plants cover only a part of the total heat load. The rest of the load is covered by the peak load boilers. This means that following the total heat capacity of the source, the potential heat capacity of cogeneration should be assessed quantitatively. Heat capacity of cogeneration plant has to be selected so that operating

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 Riga CHP heat production rate does not exceed 5% [3] (combined heat and power plant up to 4 MW of power operating in Bauska, Valmiera, Ogre, Vangazi, Daugavpils, Jelgava, Dobele, Grobiņa, Saldus, Ventspils, Ozolnieki, Ādaži, Lielvārde and Cesis).

gas), constructive parameters of cogeneration plant, parameters of heat energy consumers, heat load duration curve, duration of heat energy consumption levels, behaviour of energy end users, installed capacity, energy efficiency of technologies, development of demand side management factor, and other factors.

3. METHOD FOR EVALUATION

3.2. Methodologies: EISD method and MOO method

In connection to achieving sustainable development on global scale the correct and judicious use of resources, technology, appropriate economic incentives and strategic planning at the local and national levels is required. Therefore, choosing energy fuels and associated technologies for the production, delivery and use of energy services, it is essential to take into account economic, social and environmental consequences. The research on criteria and/or indicators in order to understand the best energetic choice for Latvia is the first step for a correct energy planning.

In the following paragraph the algorithm of ISED core set tool, included in the conceptual framework used by United Nations Commission on sustainable development (CED), is shown. After is also shortly reported the MOO methodology The EISD is an analytical tool developed which can help energy decision and policymakers at all levels to incorporate the concept of sustainable development into energy policy. EISD core set is organized following the conceptual framework used by United Nations Commission on sustainable development (CSD).

There are several methodologies that can be chosen to identify the most suitable indicators, and in the same time the choice is related and strictly connected on what the planning and consequently analysis is based on. 3.1 Criteria and indicators The methodologies can be chosen using several methodological tools and approach such us: multicriteria or multi-objective optimization (MOO) [7], energy indicators for sustainable development (EISD) [8], Life Cycle assessment (LCA) [9, 10]. Each of these methodology start from different point of views and bases: MOO methodology is connected to best optimization choice of a certain number of variables that optimize certain objectives, EISD methodology aims to evaluate (and consequently increase) the concept of sustainability based on social, economical and environmental indicators, LCA aims to figure out the global environmental load of a process and/or product taking into account the entire outflows and inflows connected (in terms of energy, substances and emissions), in this last case the indicators change depending on type of Life cycle assessment methods choosen.

Fig. 4. set of core EISD [8]

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

A summary of the factors that can influence CHP development in Latvia has been proposed in previous papers. A. Volkova et al. [2] identify four main factors: political, geographical-climatological, legislative and technological.

    

In general the total amount of electricity produced in a cogeneration regime and condensing mode depends on constructive solutions (e.g. technical solution for the biogas‘ collectors), availability of source used (mainly

The next Fig. 5 shows the linkages among the indicators selected for energy policy analysis in Baltic States. Relevant policy actions based on analysis 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 on targeted indicators.

of produced biogas. The main barriers for improved biogas injection are the high costs of improvement and grid connection. Grid injection is limited by location of biogas production and improvement sites, which have to be close to natural gas grid [12]. Problems are connected with biogas utilization in cogeneration plants (CHP) since there are no possibilities to find heat energy consumers, which in turn resulted with low efficiency landfill power plants almost all over Latvia. Due to high electricity feed-in tariff there is an economical motivation for power plant operation with low efficiency. For electricity produced in renewable energy power plants with nominal capacity of up to 4MW high feed in tariff has been transposed in Latvia‘s legislative acts. The development of Latvia‘s landfill sites is at the crossroads. On one hand it is economically feasible to operate CHP just for electricity production, but on the other – it is important to use natural resources on full value by producing from biogas the maximum amount of heat energy. In first case it means that there is no need for waste sorting in landfills, but in the other it is important to sort both – before waste collection and in landfills.

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

Multi-objective optimization (MOO), also known as multi-criteria optimization, particularly outside engineering, refers to finding values of decision variables which correspond to and provide the optimum of more than one objective. Unlike in single objective optimization (SOO), which gives a unique solution, there will be many optimal solutions for a multiobjective problem. Multi-objective optimization involves special methods for considering more than one objective and analyzing the results obtained [7].

Utilization of landfill biogas in Latvia is based on energy production in power station placed close to landfill for different reasons. One of the most important reasons is financial state support of small scale power stations (4 MWe) from renewable energy resources. Such kind of support prevents both, development of waste sorting and utilization of refuse derived fuel in cement production, and biogas improvement to cover needs transportation sector or to connect to natural gas grid.

Often, the various objective functions conflict with each other (i.e., optimizing one of them usually tends to move another towards undesirable values), for solving such models one needs to know how many units of one function can be sacrificed to gain one unit of another, but this trade-off information is not available. In other words, one is forced to determine the best compromise that can be achieved.

In the following is shortly reported the methodology regarding the optimization model of biogas use in landfills in Latvia in connection to the data collected from landfill ―Daibe‖. After the analysis only two of the independent parameters have been chosen: quality of biogas (characterized by heat value), and technological equipment (characterized by electrical capacity).

In the following paragraph an example of MOO applied to the evaluation of possibilities to utilize landfill biogas for electricity production in one of Latvia‘s 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 low biogas quality (4 kWh/m3) the optimal installed capacity is 2.2MW. In case of biogas quality of 5 kWh/m3, optimal installed capacity is 2.8MW, and 3.4 MW – in case of high biogas quality (6 kWh/m3).

The improved biogas is one of the cleanest fuels with a little impact on the environment and human health [11]. One of the advantages of biogas injection into natural gas grid is the fact that natural gas grid connects a place of biogas production (usually in rural areas) with densely populated areas. It allows new consumers to use gas. In this way it is possible to increase the biogas production in remote areas not being worried about use 181

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

5. COMMENT AND DISCUSSION The use of CHP instead of conventional plant will always improve energy efficiency and will reduce CO2 emissions significantly, in Latvia there is potential to replace some of the heat plants with co-generation units (comparing with Denmark where approximately 80% of the district heating in Denmark is supplied from CHP [6]. Hence promotion of high-efficiency cogeneration (CHP) based on a useful heat demand is a priority with regard to saving primary energy, avoiding network losses and reducing emissions, in particular of greenhouse gases [2].

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

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

Of course the choice of the fuel is fundamental in order to reach the target required from the last EU directive in terms Renewable Energy Sources (RES). More use of energy from biomass in terms of woodfuel, biogas, landfilled gas and biofuels seems to be a good direction in order to displace the part of energy sources given by the imported natural gas. The use of wood in the energy sector (through the production of heating and electricity) must become not only an objective for the development of the energy supply system, but it must also become part of strategies for economic development and for the improvement of the import/export balance of the country. These measures can succeed in not only developing local production and job creation, but if can also stimulate and increase the potential export.

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

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

It is particularly important to conduct engineer-technical and economic analysis of the various technological solutions possible to implement wood use in the cogeneration plants of the larger cities (including Riga TEC 1 and TEC 2) [13]. Any possible choice and/or scenario cannot be complete if it not references to a Life cycle assessment (LCA) that it a good tool in order to understand the environmental load of a certain process strategy and in order to give a comparable common base. Based on the targeted indicators for Latvia the best strategy can be identify in: –

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

–

Model of power production in landfill shows that feed-in tariff stated as financial support today in Latvia allows to reach economically feasible projects even in case if cogeneration unit is operated in power station regime (generates only electricity). Results show that state policy needs corrections to improve energy efficiency of biogas utilization for energy production.

– – – – 182

Enhance the diversity and variety of the energy mix. Improve maintenance of existing energy infrastructure. Eliminate constraints and investment in new facilities. Increase the efficiency of energy supply in electricity generation. Increase the share of electricity produced by combined heat and power (CHP) plants. Increase the share of renewable and domestic energy sources in the energy mix.

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 problems of the Latvian energy-system connected to the choice of the CHP and/or power stations for the future national energy strategies in the light of the last EU directive in the subjects of RES. The dependence on imported energy sources, the growth of electricity prices, and the need to support to local producers are the main reasons for use of new renewable energy technologies in the Latvian energy sector. 2. In this paper has been summarized the results from the application of the Energy Indicators for Sustainable Development (EISD), a good tool for analyzing trends, setting energy policy goals and monitoring progress in order to indentify good policy indicators. Also a testing of landfill gas using MOO method has been reported where only two of the independent parameters have been chosen: quality of biogas (characterized by heat value), and technological equipment (characterized by electrical capacity). Model of power production in landfill shows that feed-in tariff stated as financial support today in Latvia allows to reach economically feasible projects even in case if cogeneration unit is operated in power station regime (generates only electricity), but if is feasible from an economical point view is not the same if reference to environmental impact. 3. In the paper has been discussed how LCA can be a good approach that enables the energy requirements, GHG balance and other environmental impacts of bioenergy production chains to be accounted and accurately compared. Hence LCA is good tool in order to give the possibility to compare different RES usage strategies. 4. Due to high electricity feed in there is an economical motivation for power plant operation with low efficiency. For electricity produced in renewable energy power plants with nominal capacity of up to 4MW high feed in tariff has been transposed in Latvia‘s legislative acts. OF course this is not good from environmental point of view. 5. The use of CHP instead of conventional plant will always improve energy efficiency and will reduce CO2 emissions significantly, in Latvia there is potential to replace some of the heat plants with co-generation units. 6. Only crucial measures such as the reconstruction of energy sources in the larger cities (including Riga TEC 1 and Riga TEC 2) adjusting the use of fossil fuels to biomass and conversion to non-natural gas sources, will produce results. Biogas and landfill gas favorite the environmental impact displacing usage of natural gas, the possibility of the feasibility solution for connected CHP in out-of-city region to heat consumer must be evaluated.

[1] Construction, Energy and Housing State Agency Energy Department, Latvian energy in figures, Riga, 2008. [2] A. Volkova, E.Latõšev, A. Siirde, Small-scale CHP potential in Latvia and Estonia, Scientific Journal of RTU Environmental and climate technologies, Ser. 13, n. 2, Riga, 2009. [3] Latvia‘s district heating association , Heat supply in Latvia, http://www.lsua.lv/en/index.php?option=com_conte nt&task=view&id=4&Itemid=5. [4] D. Streimikiene, I. Roos, J. Rekis, External cost of electricity generation in Baltic States, Renewable and Sustainable Energy Reviews n. 13, 2009, pp. 863–870. [5] D. Streimikiene, I. Roos, J. Rekis, External cost of electricity generation in Baltic States, Renewable and Sustainable Energy Reviews n. 13, 2009, pp. 863–870. [6] L.H. Rasmussen, A sustainable energy-system in Latvia, Applied Energy n. 76, 2003, pp. 1–8. [7] G.P. Rangaiah, Multi-Objective Optimization: Techniques and Applications in Chemical Engineering, World Scientific, 2008, p. 454. [8] D. Streimikiene, R. Ciegis, D. Grundey, Energy indicators for sustainable development in Baltic States, Renewable and Sustainable Energy Reviews, 2007, Vol. 11, pp. 877–893. [9] G. Rebitzera et al., Life cycle assessment - Part 1: Framework, goal and scope definition, inventory analysis, and applications, Environment International n. 30, 2004, pp. 701– 720. [10] D.W. Pennington et al., Life cycle assessment Part 2: Current impact assessment practice, Environment International n. 30, 2004, pp. 721– 739. [11] D.Blumberga, Ģ. Kuplais, I. Veidenbergs, E.Dace, The benchmarking method for an evaluation of biogas improvement methods, Scientific Journal of RTU Environmental and climate technologies, Ser. 13, n. 2, Riga, 2009. [12] G. Kuplais, D. Blumberga, E. Dace, F. Romagnoli, Optimisation model of biogas use in landfills in Latvia, 7th International conference ORBIT2010: Organic resources in the carbon economy, June 29-July 3, 2010, Heraklion, Greece. [13] A. Blumberga et al., Assessment on the use of renewable energy resources in Latvia until 2020: report, LVAF, December 2008, Riga. 183

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

LCA OF COMBINED HEAT AND POWER PRODUCTION AT HELLISHEIÐI GEOTHERMAL POWER PLANT WITH FOCUS ON PRIMARY ENERGY EFFICIENCY Marta Ros Karlsdottir, Olafur Petur Palsson, Halldor Palsson University of Iceland, Faculty of Industrial Engineering, Mechanical Engineering and Computer Science mrk1@hi.is energy‖ become more viable in various locations around the world. It is thus important to investigate their primary energy efficiency and environmental impact for comparison with other energy conversion technologies. These energy performance indicators can be used to help decision making of future developments, policy making and energy rating of buildings.

ABSTRACT The aim of the study is to calculate primary energy factors, fp, stating the primary energy efficiency as well as factors for CO2 emission, K, for geothermal combined heat and power production at the Hellisheidi CHP plant in South-West Iceland. These factors state how much primary energy consumption and CO2 emissions result from the production of 1 MWh of heat and electricity due to geothermal utilization. Methods of life cycle assessment (LCA) are used to calculate these factors by taking into account all energy and material streams to and from the CHP plant during construction and operation. The results show that producing heat and electricity in a combined heat and power plant minimizes the primary energy factor for the electricity generation and produces a relatively low primary energy factor and CO2 production factor for the heat generation process. From the results, it can also be seen that life cycle assessment is a useful method to evaluate the total impacts of the geothermal energy conversion process, especially for the emission of greenhouse gasses during the lifetime of the production facilities. The experience in this study also demonstrates that the method can equally be used for processes as it is commonly used for the analysis of total impact of products.

Countries that have access to geothermal areas and produce power by geothermal utilization within the European Union (EU) are: Austria, France, Germany, Greece, Hungary, Italy, Netherlands, Portugal, Romania, Slovakia and Spain. Other European countries such as Iceland and Turkey, which are not current member states of the EU, also utilize geothermal energy extensively [2]. Also, 32 European countries use geothermal energy directly for various purposes such as district heating [3]. Thus, electricity and heat based on geothermal energy are a part of Europe‘s energy mix. For countries using geothermal based power and/or heat and complying to EU legislation, it is therefore important to have easy access to standardized factors accounting for the primary energy efficiency and CO2 emissions from geothermal based heat and power. The aim of this study is to produce standardized factors for primary energy efficiency (fp) and CO2 emission (K) for geothermal heat and power production.

INTRODUCTION The calculation of primary energy and CO2 production factors for geothermal power production has had little attention while factors for some other types of energy technologies such as hydropower, nuclear and coal fired power plants have been developed during the recent years. The importance of these factors is stated mainly in the new recast of Directive 2002/91/EC of the European Parliament and of the Council on the energy performance of buildings [1]. There it is stated that before the end of year 2010, all new building occupied by public authority should be issued energy performance certificates showing these factors, based on the energy mix used by the building and the buildings‘ energy performance.

ENERGY PERFORMANCE INDICATORS FOR PRIMARY ENERGY CONSUMPTION AND CO2 EMISSIONS The primary energy factor is defined as the ratio between the total primary energy inputs involving energy production to the actual energy delivered to the consumer. According to [4], it should always account for the extraction of the energy carrier and its transport to the utilization site, as well as for processing, storage, generation, transmission, distribution and delivery. There are two primary energy factors defined: 

At present time, geothermal power plants are situated in 24 countries [2] and a total of 78 countries have reported direct use of geothermal energy [3]. With increasing fossil fuel prices and focus on renewable energy sources, these power plants producing ―green



184

Total primary energy factor, accounting for primary energy use of both renewable energy sources and non-renewable sources. Non-renewable primary energy factor, accounting only for the primary energy consumption of non-renewable energy sources. This factor is used when expressing only the use of

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

the power generation to 120 MW. A year later, another 90 MW were added, resulting in a power generation capacity of about 210 MW (213 MW in February 2009). Further developments of the power plant include adding heat production in 2010 for district heating and also increasing the power production if possible. Estimated production capacity for the completed Hellisheidi Plant is 300 MW electricity and 400 MW thermal energy [5].

fossil or other non-renewable or polluting energy sources in the energy conversion system. The CO2 production coefficient, K, shall include all CO2emissions associated with the primary energy used. Furthermore, equivalent emissions of other greenhouse gases, e.g., methane, may be included [4]. According to Directive 2002/91/EC, indicators on the energy performance of buildings shall include the consumption of primary energy and the CO2 emissions resulting from the buildings energy usage. Factors for primary energy consumption and CO2 emissions have been calculated for various energy chains producing electricity, and values for these factors are given in Annex E of the standard EN15603 on the energy performance of buildings. An overview of these factors is given in Table 1.

The plant today is a double flash power plant with highand low-pressure turbines and separators as seen in Figure 1. The heat production facilities are currently under construction with a planned 133 MW thermal capacity at the end of year 2010. The technical complexity is moderate and the plant makes a good basis for a LCA study to evaluate the primary energy efficiency and CO2 emission of this type of geothermal power plant. Since it is fairly newly constructed, access to detailed background data for the inventory modelling is possible, making the study more reliable and accurate. Environmental assessment for the production is available as well as measurements of various environmental impacts of the power plant, providing data for the impact assessment of the LCA study.

Table 1: Energy performance indicators for various sources of electricity [4] Primary energy factors fp Source of electricity

[MWh primary energy / MWh delivered energy]

CO2 production coeff. K [Kg/MWh]

NonRenewable

Total

Hydraulic power

0.50

1.10

Nuclear power

2.80

Coal power

4.05

1340

Electricity mix UCPTE

3.14

3.31

617

In this study, a steady production of 213,6 MW electricity and 121 MW heat is used as a basis for the LCA model. The reason for this choice is that the newest inventory data on the construction phase and mass extraction are built on these production capacities, and that the base thermal load is estimated to be 121 MW and not the full capacity of 133 MW. PRIMARY ENERGY OF VARIOUS ENERGY SOURCES

As seen in the standard EN15603:2008 [4] and Table 1, no indicators are given for geothermal power. The directive is under reconstruction and a recast has been released, as mentioned before. Also, the table does not give factors for sources of thermal energy used by buildings for space heating. Thus, there is clearly a need to calculate these factors for energy chains that involve geothermal energy, since they produce both electricity and heat which is delivered to buildings within the European Union and in countries following EU legislation.

There is a matter of inconsistency in primary energy calculations of various energy sources as many different methods are in use and accepted by different energy authorities [6]. As an example, the primary energy factors for power produced from renewable energy sources such as hydro power, wind energy and solar energy are sometimes calculated by assuming that the primary energy factor for the energy conversion system is one, which is the same as assuming that the energy conversion process is 100% efficient. The reason for this assumption is that the primary energy is defined as the first usable stage of the energy flow, which in the case of wind, solar and hydro is the electricity itself produced from these primary sources [7]. For electricity production from heat sources, the first usable stage of the energy stream is defined as the steam input into the turbine, according to an energy statistics manual from the International Energy Agency (IEA) [8]. The methods

GEOTHERMAL HEAT AND POWER PRODUCTION AT HELLISHEIDI CHP PLANT Hellisheidi geothermal CHP plant is situated at the Hengill geothermal area close to Reykjavik, the capital of Iceland. A 90 MW electricity production started in 2006 after several years of construction and research. In 2007, a low pressure turbine was added, increasing 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 HPS HPC

LPT

LPS Hot water tank

LPC

HX1

HX2

Figure 1 – A simple schematic of the Hellisheidi geothermal CHP plant

used to calculate primary energy demand of power production from renewable energy sources tends to underestimate the primary energy input from the original energy sources into the energy conversion system compared to the assumptions made for the heat conversion processes such as coal, oil and also geothermal.

Geothermal production well

Generator

HPS:

High pressure steam separator

HPT:

High pressure steam turbine

HPC:

Condenser for high pressure turbine

LPS:

Low pressure steam separator

LPT:

Low pressure steam turbine

LPC:

Condenser for low pressure turbine

CT:

Cooling tower

CP:

Cooling water pump

HX1:

Heat exchanger 1 for DH system

HX2:

Heat exchanger 2 for DH system

IW: Reinjection well

system compared to the assumptions made for the heat conversion processes such as coal, oil and also geothermal. Definition of Primary Energy of Geothermal Fluid There is no clear definition of primary energy from geothermal energy sources. Published methods of determining the primary energy consumption in geothermal power plants are the following [6]:

PRIMARY ENERGY OF VARIOUS ENERGY SOURCES



Working Group III (WG III) of the Intergovernmental Panel on Climate Change (IPPC) records electricity from geothermal on a 1:1 basis. This results in a fp factor of 1. The Engineering Information Administration (EIA) uses a factor of 6.16 units of primary geothermal energy for each unit of geothermal electricity. International Energy Agency (IEA) records a fp value of 10 by assuming 10% conversion efficiency of geothermal power plants.

In this LCA study, where the main goal is to calculate an accurate fp factor for a specific conversion technology, the main issue is the primary energy content of the geothermal fluid extracted from the production wells. The primary energy content of the geothermal fluid can be based on different assumptions. The first one is the energy content of the geothermal fluid based on its enthalpy in kJ/kg. Second, the exergy content of the fluid can be used as a basis. However, in this study, the primary energy content of the geothermal fluid taken from the production wells (and utilized for both electricity and heat production) is chosen to be the enthalpy above 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) reference temperature [9], and calculated in the following manner:

production based on geothermal energy will help identify how much effect the construction, collection of geothermal fluid and even the demolition phase of the power plant and the distribution system have on the total primary energy consumption. It can identify the impact of the drilling of wells, manufacturing of power plant components and piping, construction of buildings and roads associated with the power plant, operation of the power plant itself and the primary energy extracted from the geothermal reservoir and even the impacts of constructing and operating the distribution facilities.

(1)

Where kJ/kg,

is the specific primary energy content in is the enthalpy of the fluid and

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 The Directive 2002/91/EC defines the concept of primary energy as energy that has not undergone any energy conversion process [1]. The primary energy factor must thus represent all the primary energy consumed in order to provide one unit of heat or power to the consumer. Primary energy consumption of energy chains is not only based on the consumption of fuel (or other energy source) in the power or heat generation process, but also all the primary energy needed for the construction, operation and possibly demolition of the production facilities. Also, some primary energy is needed for the distribution of the product. To calculate such accumulated primary energy, the method of life cycle assessment is well suited. LCA is a method that has been developing since the earliest performance of such a study in 1969 and standards on the methodology where issued in the late 1990s [10].



Defining the goal and scope of the study



Performing inventory analysis



Performing impact assessment

Goal and Scope of the Study The main goal of this LCA study is to analyze the two energy performance indicators presenting the primary energy efficiency and the CO2 emissions for both the electricity and heat production at Hellisheidi power plant. The LCA calculations and impact assessment where done by using the LCA software SimaPro 7 [12] and using different databases such as the Ecoinvent database [13] for the inventory information on various raw materials and processes used in the geothermal power plant. There are numerous geothermal power plants worldwide using similar technology as the Hellisheidi power plant to produce electricity (double flash power plants produced 23% of the electrical power from geothermal resources in 2007 [14]), so the results for the energy performance indicators for the power production at Hellisheidi could be used to represent these power plants. Other types of geothermal energy conversion systems, such as single flash and binary systems, should be treated individually when calculating energy performance indicators for the electricity production.

LCA has been considered a good tool to achieve a holistic approach on evaluating the environmental impact of products. Today, it is widely used to investigate all kinds of production systems and has given valuable insight on the total impact of products and systems on the environment by not only focusing on the operational aspect [11]. Many interesting results have been achieved by using this methodology and those results form a basis for evaluating and comparing different solutions for production of various products, such as vehicles for transport, soft drink containers and power conversion technologies. On the other hand, LCA in the process industry has had much less attention than for manufacturing products, and research is needed before complete methods for processes are readily available [11]. The application of LCA on geothermal energy utilization can be valuable for LCA developers working on further improvements and adjustments on the LCA methodology for the process industry.

Geothermal combined heat and power plants are not common worldwide, but regarding Europe they can be found in Iceland as well as Austria and Germany. By producing heat as well as electricity in geothermal applications, the utilization of the heat taken from the geothermal reservoir in the form of geothermal fluid is maximized. The heat produced has a variety of useful applications, such as for district heating, agriculture, fisheries, swimming pools, snow melting and heating up greenhouses [3]. The calculations of the primary energy factor of the heat production at Hellisheidi geothermal CHP plant will emphasize this increase in thermal efficiency of the power plant.

Using LCA to calculate the total primary energy consumption and CO2 emission for heat and power 187

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The scope of this LCA study includes making the following choices [11]:      

Functional unit System boundaries Choice of impact categories Method for impact assessment Principles for allocation Data quality requirements 1 kWh electricity

1 kWh heat Geothermal fluid, from electricity production

Electricity, geothermal, at Hellisheidi CHP plant

Power plant equipment

Heat, from condenser

Heat, geothermal, at Hellisheidi CHP plant

Geothermal power plant unit

Geothermal fluid, at power plant

Collection pipelines

Geothermal heat production unit

Power plant structures

Geothermal fluid, in ground

Drilling of geothermal wells

Heating station Structure

Heating station equipment

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

plant unit, the geothermal heat production unit and the geothermal fluid. The geothermal power plant is constructed from the power plant structures and equipment while the fluid is transported in collection pipelines from geothermal wells that need to be drilled for the production. The heat production unit consists of the heating station structure and equipment. The energy input into the heating process is waste heat from the power production process in form of heat taken from the steam in the condenser for preheating of district heating water, and the waste geothermal fluid from steam separators used for final heating of the district heating water. Inventory data on all these different components in the flow model was collected and used for the LCA study of the Hellisheidi CHP plant.

Functional Unit The primary energy and CO2 factors are defined as primary energy usage and CO2 emission per MWh and thus, the functional unit of the study is chosen to be MWh of electricity or heat produced in the Hellisheidi geothermal CHP plant. The functional unit is the reference flow to which all other modelled flows of the system are related. System Boundaries The processes included in this LCA study are mainly the operation and construction of the power plant. The demolition or end-of-life phase is disregarded due to insufficient information at this time. Also, the energy and material flows due to maintenance in the operational phase of the power plant are disregarded but both the demolition and the maintenance will be included in further studies. The time horizon in this study is chosen to be 30 years, which is the technical lifetime of the power plant capital goods.

Impact Categories and Methods for Impact Assessment To calculate the two energy performance indicators, the two main impact categories to be used are the primary energy demand of the production process in MWh and Global Warming Potential (GWP) given in CO2 equivalents.

A flow model of the CHP plant as modelled in the LCA study is shown in Figure 2. The two outputs of the production system are 1 MWh of electricity and 1 MWh of heat. The main material and energy inputs into the energy conversion system are the geothermal power 188

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Two different methods of impact assessment had to be used in the impact assessment calculations. For the primary energy factor, the Cumulative Energy Demand (CED) method [15] was used which is based on a method published by Ecoinvent 1.01 and available in SimaPro 7 impact assessment methods. For the calculation of the CO2 emission factor, the IPCC 2007 GWP 100a V1.01 [16] was used to get the CO2 equivalent total global warming potential for the chosen functional unit of 1 MWh electricity produced.

Data Quality To calculate the energy performance indicators by methods of LCA, reliable inventory information is needed on material and energy flows to and from the geothermal power production facilities during their lifetime.. The inventory in this study is constructed from data provided by Reykjavik Energy, the power company in ownership of the Hellisheidi plant. The data on the construction phase is retrieved from the conditions and specifications in a tender for the construction of the power plant, where quantitative information is collected on all major material flows required for the constructions and machinery. The inventory information for the fluid collection and drilling is retrieved from a report done by Reykjavik Energy, including the power and performance of the geothermal wells drilled for the power and heat production [17].

Principles for Allocation To allocate the impacts of the different products, electricity and heat, produced at Hellisheidi CHP plant, several methods can be used. The method used should reflect the physical relation between the two products, such as how the different inputs and outputs of the process are dependent on the two different products. Simple methods of allocation for an energy conversion process can be:   

For a LCA study, the following data quality indicators must be presented:     

Based on energy content of the products Based on exergy content of the products Based on the monetary value of the products

The abovementioned methods can be used when the physical relation between the two products is unclear. In the case of the Hellisheidi CHP plant, the physical relation between the two outputs (electricity and heat) is mainly the use of waste heat from condensers and the geothermal fluid from the production wells, as shown in Figure 2. The impacts of construction can easily be divided between the electricity and heat production with the detail of inventory data provided. Also, the geothermal fluid used in the heat production is taken from steam separators in the electricity generation process and would otherwise be reinjected back into the geothermal reservoir via reinjection wells.

Time period Region Type of technology and representativeness Allocation System boundaries

In this study, the time period of the data is from 2005 to 2009 and the region is Western Europe. The type of technology is modern and the representativeness is data from a specific company. The allocation, as mentioned before, is by physical connections between the two outputs. The system boundaries are described by three different criteria. First, the cut-off criteria is in general set to be less than 5% which means that all inventory data that does not contribute more than 5% to the overall impacts of the two products is disregarded. Also, the system boundary is chosen to be of the first order, only to account for the materials used in the construction and operation of the CHP plant but not the processing and transportation of these materials. The third system boundary criterion is the system boundary with nature, which in this study is described as unspecified at this stage of the LCA study.

The disposed heat in the condenser is utilized to preheat the district heating water by using it as cooling water. The condenser pressure determines the temperature of the steam output from the turbines and thus, also the final temperature of preheating of the district heating water. If the heat demand is high, the condenser pressure must be higher than the optimum for power production in order to supply high enough temperatures to the district heating water. This limits the electrical power production and requires that more geothermal wells have to be drilled in order to sustain the electrical production under high thermal loads of the district heating system. These limitations on the electrical production imply that the allocation of impacts from the drilling of wells should be related to the number of wells that have to be drilled to sustain both the electricity production and the highest thermal load designed for the district heating system.

RESULTS FOR THE ENERGY PERFORMANCE INDICATORS Energy Performance Indicators for Electricity Production The results for the impact assessment of the electricity production alone, focusing on the two energy performance indicators, is shown in Table 2. The highest value of fp 6.33 MWh primary energy/MWh produced energy, is obtained when no heat production is present at the power plant and the effects of reinjection of waste streams is not taken into account. 189

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 cases, where the waste heat is either reinjected back into the reservoir or used for heating of DH water. In those cases, the primary energy content of the waste stream can be subtracted from the primary energy content of the geothermal fluid used for the electricity production, resulting in lower fp values. The share of non-renewable primary energy sources such as oil and gas used in the construction phase or in the manufacturing of various power plant components, only account for about 0.01 of the total fp value in all cases.

The factor K for the CO2 emissions is the same for all three cases of electricity production as reinjection and utilization of waste stream does not have significant effects on the total emissions due to the power for production. The origins of the CO2 emissions can be seen in Figure 4. The largest contributor to the CO2 emission from the electricity generation over 30 years of production is the geothermal fluid, responsible 88% of the CO2 emissions per kWh of electricity production.

Table 2 – Results for the primary energy factor and CO2 emission factor for electricity based on geothermal energy Primary energy factors fp [MWh primary energy / MWh produced energy]

Source of electricity

CO2 production coeff. K [Kg/MWh]

Non-Renewable

Total

Electricity from Hellisheidi geothermal power plant

0.01

6.33

Electricity from Hellisheidi geothermal power plant, with reinjection

0.01

5.33

Electricity from Hellisheidi CHP plant

0.01

5.33

A small share of 8% originates from the geothermal wells while the construction of plant, along with the manufacturing of components, is responsible for 4% of emissions.

value reduces to 0.69. In both cases, the share of primary energy from non-renewable energy sources is less than 0.01. In both cases, the CO2 production coefficient is 0.98 kg CO2 equivalents per produced MWh.

drilling of the power its main the CO2

The origins of the CO2 emission from the heat generation process can be seen in Figure 4. The largest contributor to the total emissions is the drilling of the geothermal production wells that were needed to sustain the electricity production while the heat production is at maximum load of 133 MWth. The manufacturing of the district heating pipeline from the production area to the rural area of Reykjavík city contributes to 15% of the total emission resulted from the heat generation process.

GWP 100a for Electricity Production in kg CO2 eq 4%

0.5% 8%

87.5%

Geothermal fluid (87.5%) Power plant and components (4%) Geothermal well drilling (8%) Collection lines (0.5%)

Figure 3 – Origins of CO2 emissions from the different processes of the power generation

Energy Performance Indicators for Thermal Production The energy performance indicators for the production of heat for district heating are given in Table 3. Two cases are presented for the heat production; heat production process with or without the effects of reinjection of waste geothermal fluid. The highest value for fp is obtained in the case where reinjection is not taken into account, with the value of 1.78 MWh primary energy/MWh produced energy. With reinjection, the

Figure 4 – Origins of CO2 emissions from the different processes of the heat generation 190

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

process. Reinjection of geothermal brine is recognized to improve heat mining and stabilize the production capacity of geothermal fields, if successfully carried out. It can also counteract pressure draw-down in the reservoir by providing an artificial water recharge [18]. In this study, reinjection of the waste stream is modelled, which decreases the use of primary energy in the energy conversion process, since a part of the primary energy from the geothermal fluid it is returned back to the reservoir. Reinjection is present at the Hellisheidi geothermal CHP plant so the values of the energy performance indicators with reinjection are valid for the power plant.

Table 3 – Results for the primary energy factor and CO2 emission factor for heat from a geothermal CHP plant

Source of heat

Heat, Hellisheidi CHP plant Heat, Hellisheidi CHP plant, reinjection

Primary energy factors fp [MWh primary energy / MWh produced energy]

CO2 production coeff. K [Kg/MWh]

NonRenewable

Total

>0.01

1.78

0.98

>0.01

0.69

0.98

4) Life cycle assessment is especially useful to evaluate the total impact of geothermal power plants with respect to their emission of greenhouse gasses. Figure 3 and Figure 4 show how the different phases in the life cycle of the power plant significantly contribute to the overall emission in CO2 equivalents. If LCA had not been carried out for the process, 12% of the CO2 emissions resulting from the electricity generation would not have been accounted for and no emissions would have been found for the heat production, since the emissions from drilling, construction of buildings, and manufacture of components had not been accounted for.

DISCUSSION The following discussion highlights the most significant results from this study: 1) By comparing the energy performance indicators calculated in this study and shown in Table 3 to the indicators given in Table 3 it can be seen that electricity from geothermal power plants has the highest total fp factor while the share of non-renewable energy sources is the lowest. The main reason for the high fp factor is the low conversion efficiency of geothermal power plants due to low working temperatures and pressures. The CO2 production coefficient is relatively low compared to the other energy conversion technologies and could be lowered even further if measures are taken to control the emissions from the power plant. The results for the Hellisheidi geothermal CHP plant cannot be used to represent all geothermal power plants producing either electricity alone or with a combined production of electricity and heat. Further studies are needed on different types of geothermal power plants, such as single flash and organic Rankine cycles, to be able to produce specific or average factors representing geothermal utilization.

ACKNOWLEDGMENTS Special thanks are given to the following partners: Nordic Energy Research (NER) for funding the study and the Energy Research Fund of Landsvirkjun for their support. To Orkuveita Reykjavíkur for providing data for Hellisheidi Power plant, to Mannvit engineering for discussion and data provision and to Ragnar Gylfason for his contribution in the data gathering phase. REFERENCES [1] EU. (2003, January 4). Directive 2002/91/EC of the European Parliament and of the Council of 16 December 2002 on the energy performance of buildings. Official Journal of the European Communities .

2) The results for the heat production at the Hellisheidi geothermal CHP plant, given in Table 3, show that the energy performance indicators are relatively low and, in the case of reinjection, below unity. This is because the primary energy needed to preheat the DH water is not accounted for in the heat production but rather assigned to the electricity production. This is due to the fact that the preheating of the DH water from 5 °C to 41 °C is done in the condenser for the high pressure steam turbine as seen in Figure 1 and is a necessary step in the electricity production, but a beneficial step in the heat production for the DH system.

[2] Bertani, R. (2010). Geothermal Power Generation in the World 2005 – 2010 Update Report. Proceedings World Geothermal Congress 2010, (April), 25-29. [3] Lund, J. W., Freeston, D. H., & Boyd, T. L. (2010). Direct Utilization of Geothermal Energy 2010 Worldwide Review. Proceedings World Geothermal Congress 2010, (April), 25-29.

3) Values for the indicators for both electricity and heat are calculated with and without reinjection of the cooled geothermal brine from the energy conversion

[4] EN 15603:2008. Energy performance of buildings. Overall energy use and definition of energy ratings. 191

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Geneva: International Standardisation (ISO).

Organisation

for

[12] PRéConsultants. (2009, September 6). SimaPro LCA software. Retrieved October 14, 2009, from SimaPro LCA software: http://www.ecoinvent.ch/

[5] Helisheidi Geothermal Plant. (2009). Retrieved May 2009, from http://www.or.is/English/Projects/ HellisheidiGeothermalPlant/

[13] Ecoinvent. (2009, August 13). Home. Retrieved October 14, 2009, from Home: http://www.ecoinvent.ch/

[6] H. Douglas Lightfoot. (2007). Understand the three different scales for measuring primary energy and avoid errors. Energy, 32, 1478-1483.

[14] DiPippo, R. (2008). Geothermal Power Plants – Principles, Applications, Case Studies and Environmental Impact (2nd edition ed.). Oxford: Butterworth-Heinemann.

[7] Segers, R. (2008). Three options to calculate the percentage renewable energy: An example for a EU policy debate. Energy Policy , 36 (9), 32433248. [8] IEA. (2004). Energy Statistics International Energy Agency (IEA). OECD/IEA.

[15] Klöpffer, W. (1997). In defense of the cumulative energy demand (editorial). International Journal of Life Cycle Assessment , 2, 61.

Manual. Paris:

[16] PRéConsultants. (2009, September 6). Methods. Retrieved October 14, 2009, from SimaPro LCA software: http://www.pre.nl/simapro/impact_assessment_met hods.htm#CML2

[9] ISO 2533:1975. Standard atmosphere. International Organization for Standardization, Geneva, Switzerland.

[17] Gunnlaugsson, E., & Oddsdóttir, A. L. (2009). Helisheidi - Gufuborholur 2008 (Hellisheidi - Steam wells 2008). Reykjavík: Orkuveita Reykjavíkur.

[10] Russell, A., Ekvall, T., & Baumann, H. (2005). Life cycle assessment - introduction and overview. Journal of Cleaner Production , 13 (13-14), 12071210.

[18] Stefansson, V. Geothermal reinjection experience. Geothermics, 26, (1997), 99–130.

[11] Baumann, H., & Tillman, A.-M. (2004). The Hitch 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

FLEXIBILITY FROM DISTRICT HEATING TO DECREASE WIND POWER INTEGRATION COSTS 1

J. Kiviluoma and P. Meibom 1

VTT Technical Research Centre of Finland 2 RisĂ¸ DTU

ABSTRACT

INTRODUCTION

Variable power sources (e.g. wind, photovoltaics) increase the value of flexibility in the power system. This paper investigates the benefits of combining electric heat boilers, heat pumps, CHP plants and heat storages in a district heating network when the share of variable power increases considerably. The results are based on scenarios made with a generation planning model Balmorel [1]. Balmorel optimises investments and operation of heat and power plants, including heat storages. It uses hourly resolution and enforces temporal continuity in the use of the heat storages. Scenarios with high amount of wind power were investigated and the paper describes how the increase in variability changes the profitability and operation of different district heating options in more detail than was described in the article by Kiviluoma and Meibom [2]. Results show that district heating systems could offer significant and cost-effective flexibility to facilitate the integration of variable power. Furthermore, the combination of different technologies offers the largest advantage. The results imply that, if the share of variable power becomes large, heat storages should become an important part of district heating networks.

Wind power is projected to be a large contributor to fulfil electricity demand in several countries. This could take place due to relatively low cost of wind power electricity or policy mechanisms promoting renewable energy. In any case, power systems with a large fraction of power coming from a variable power source will need to be flexible. Flexibility is used to cope with the increased variation in residual load (electricity demand minus variable power production) and with the increased forecast uncertainty in the residual load. On the other hand, lack of flexibility will cause larger costs from increased variability and forecast errors. Therefore, it is prudent to investigate the cost optimal configurations for the combined power and heat generation portfolios. Heat generation could offer significant possibilities for increasing the flexibility of the power system. Currently, part of the inflexibility of the power system comes from CHP plants that are operated to serve the heat load while electricity is a side product. Installation of electric resistance heaters next to the CHP units or elsewhere in the heat network could break this forced connection. During periods of low power prices, which will become more common with high share of wind power, CHP plants could be shut down and heat would be produced with electricity. The dynamics can be made more economic with the use of heat storages. Further option is to have heat pumps in the DH network, but they will require large amount of full load hours to be profitable and will compete with CHP plants for the operating space.

NOMENCLATURE Indices i, I I

HeatSto

Unit, set of units Heat storage units

t, T

Time steps, set of time steps

a, A

Area, set of areas

In most countries heat demand is in the same order of magnitude as electricity demand. For example, in UK the demand for primary energy due to heat is around 40% of total primary energy demand [3]. About 25% of the primary energy demand is due to space and nonindustrial water heating. In the US all kind of heat use accounts for about 30% of the primary energy consumption [estimated from 4].

Variables C

New capacity

Power generation

Heat generation

Charging of heat storage

Parameters c

Inv

Annualized investment cost

Fix

Fixed operation and maintenance costs

Operation

Operation cost function of the unit

Weight of time period

Heat demand

Heat is inexpensive to store compared to electricity. Electricity storage has been seriously considered to alleviate the variability of wind power [5-6]. Therefore, it is apparent that the use of heat storages should also receive serious consideration in the current context. Some work has been done [7-9], but not considering 193

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Loading of heat storage adds to the heat demand. Loss during the heat storage process is not considered. The dynamics of heat networks were not taken into account.

optimal investments in new power plants and heat storages. The study has been restricted to residential and industrial district heating systems. Buildings not connected to district heating systems were not considered, although these also require heat. Cooling demand could also offer similar possibilities, but the problem was not addressed here. Industrial heat demand and water heating do not usually have strong seasonal variation and can therefore be more valuable towards the integration of variable power.

Q iI

The Balmorel model is a linear optimization model of a power system including district heating systems. It calculates investments in storage, production and transmission capacity and the operation of the units in the system while satisfying the demand for power and district heating in every time period. Investments and operation will be optimal under the input data assumptions covering e.g. fuel prices, CO2 emission permit prices, electricity and district heating demand, technology costs and technical characteristics (eq. 1). The model was developed by (Ravn et al. [1]) and has been extended in several projects, e.g. (Jensen & Meibom [10], Karlsson & Meibom [11], Kiviluoma & Meibom [2]).

i ,t

t  T ; a  A

(2)

‗Urban‘ area presents the heat demand in the capital region of Finland. The existing power plants in 2035 cover over half of the required heat capacity. Largest share comes from natural gas, which is a relatively expensive fuel in these model runs. The annual heat demand is smallest of the considered areas: 6.2 TWh. ‗Industry‘ area aggregates the known industrial district heating demand from several different locations. This is a necessary simplification, since Finland has over hundred separate DH areas and the model would not be able to optimise all of these simultaneously. The industrial heat demand in Finland is driven by paper and pulp industry, which produces waste that can be used as energy input. This capacity is assumed to be available in 2035 and as a consequence the model does not need more industrial heat capacity. The annual heat demand is 46.8 TWh.

  min   ciInvCi   ciFix CiEx  Ci   wt ciOperation Pi ,t , Qi ,t  (1) iI tT iI  iI 



Z

iI aHeatSto

In this paper, scenarios without new nuclear power are compared (scenarios ‗Base NoNuc‘ and ‗OnlyHeat NoNuc‘ in article [2]). This meant that wind power had a very high share of electricity production. Accordingly, there was more demand for flexibility in the system.

The model and assumptions used for the analysis are described in more detail in [2]. For convenience, most important sections are referenced below. The heat sector of the model is described more thoroughly here.



 hr ,t 

Analysis is done for the year 2035. By this time, large portion of the existing power plants are retired. Three district heating areas were considered. These have a rather different existing heat generation portfolio by 2035. This helps to uncover some interesting dynamics in the results section.

METHODS AND DATA



i ,t



The optimization period in the model is one year divided into time periods. This work uses 26 selected weeks, each divided into 168 hours. The yearly optimization period implies that an investment is carried out if it reduces system costs including the annualized investment cost of the unit.

‗Rural‘ area aggregates non-industrial heat demand excluding the capital region considered in ‗Urban‘. This is probably the most interesting example, as the existing capacity covers only 20% of the heat capacity demand. Therefore, the model has to optimise almost the whole heat generation portfolio. There are wood resources (limited amount of forest residues and more expensive solid wood) available unlike in the urban area. The annual heat demand is 21.0 TWh.

The geographical resolution is countries divided into regions that are in turn subdivided into areas. Each country is divided into several regions to represent its main transmission grid constraints. Each region has time series of electricity demand and wind power production. The transmission grid within a region is only represented as an average transmission and distribution loss. Areas are used to represent district heating grids, with each area having a time series of heat demand. There is no exchange of heat between areas. In this article, Finland is used as the source for most of the input data.

RESULTS Figures 1–3 give an example how heat production meets heat demand in the different areas during the same 4.5 days in January. Negative production indicates charging of heat storage. Electricity price is on separate axis together with the cumulative content of heat storage. When electricity price is low, storage is loaded with electricity using heat boilers and heat pumps. When electricity price is high, CHP units

The hourly heat demand has to be fulfilled with the heat generation units, including heat storages (eq. 2). 194

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

produce heat and electricity. Fluctuations in electricity price are mainly driven by changes in wind power production, since these are larger than changes in electricity demand (Fig. 4). 450

Heat production (MW)

1000

375

800 600

300

400

225

200 0

150

-200

-400

NG_EX_UR

Electricity price (€/MWh) Heat storage content (%)

1200

NG_BP_UR MW_HB_UR MW_BP_UR EL_HP EL_HB Storage use Stor. content

4000

450

3000

375

2000

300

1000

225

150

-1000

Electricity price (€/MWh) Heat storage content (%)

Heat production (MW)

Elec. price -600 0 Fig. 1. Example of operation in ‗Urban‘ heat area. Negative production indicates charging of heat storage.

MW_HB_RU WR_EX WW_EX NG_BP_RU NG_CC_EX PE_BP_RU WO_BP_RU EL_HP EL_HB Storage use Stor. content Elec. price

-2000 0 Fig. 2. Example of operation in ‗Rural‘ heat area. Negative production indicates charging of heat storage.

Heat production (MW)

Electricity price (€/MWh) Heat storage content (%)

6000 450 PE_BP_IN 5000 375 WR_BP_IN 4000 300 WW_BP_IN 3000 2000 225 EL_HB 1000 Storage use 150 0 Stor. content 75 -1000 Elec. price -2000 0 Fig. 3. Example of operation in ‗Industrial‘ heat area. Negative production indicates charging of heat storage.

17500

Electricity production (MW)

15000

Wind

12500

Natural gas (NG) Hydro

10000

Wood waste (WW) Fig.7500 1. Example of operation in ‗Urban‘ heat area. Negative production indicates charging of heat storage. Peat 5000 Solid wood (WO) Forest residues (WR)

2500

Municipal waste (MW)

Nuclear

-2500

Electricity to heat

-5000 Fig. 4. Electricity production. Negative production indicates the use of electric heat boilers and/or heat pumps. 195

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Effects of heat measures in the three heat areas

EL_HP

3000

EL_HB

2500

NG_HB

2000

CO_EX

1500

WO_HB

1000

NG_EX_UR

500

WW_BP_IN WR_BP_IN

Cap.

PE_BP_IN

Cap.

OnlyHeat

Base

OnlyHeat

MW_BP_UR

Fig. 6. Heat capacity and production in the ‗Urban‘ heat area.

NG_BP_IN FO_BP_IN

The combined utilization of the heat measures was used to shut down existing natural gas based CHP power plants during hours of average or lower electricity prices. During low electricity prices electric heat boilers were used to charge heat storage. Accordingly, during average electricity prices heat was used from heat storage to prevent the use of electric heat boilers. During the highest electricity prices electric heat pumps were also shut down with the help of heat from the heat storages.

Prod. 8

Fig. 5. Heat capacity and production in the ‗Industrial‘ heat area.

waste wood were not easily replaced. However, there were some high wind situations with low power prices where it was beneficial to use electric heat boilers to produce heat and decrease heat production from wood waste in the ‗Industry‘ area. There was an annual resource limit on wood waste on the country level and the wood waste use was transferred to the ‗Rural‘ heat area. It was also profitable to install some heat storage capacity. This enabled the full shut down of wood waste back pressure power plants for the duration of low electricity prices. This decreased electricity production and gave more room for the upsurge in wind power production.

Heat capacity (MW) Heat production (GWh)

The most important difference between ‗Urban‘ and ‗Rural‘ heat areas is the availability of wood residues in the ‗Rural‘ heat area (Fig. 7). For the most part this resource was able to outcompete heat pumps as means to produce heat. Heat measures still helped to replace coal CHP. The combination of electric heat boilers and heat storages was again a large source of additional flexibility to the system.

EL_HP

10000

EL_HB

8000

NG_HB

6000

NG_CC_EX

4000

CO_EX WR_EX

2000

WW_EX

Base

In the ‗Urban‘ heat area heat measures enabled the replacement of CHP coal units with production from heat pumps and to smaller extent from electric heat boilers (Fig. 6). Also wood based heat boilers were replaced. Investment in heat storage was relatively smaller. However, they were cycled more due to faster charging rate.

HEATSTOR

12000

Cap.

WO_HB

OnlyHeat

5000

MW_HB_UR

Prod.

Base

10000

OnlyHeat

15000

OnlyHeat

EL_HB

Base

NG_BP_UR

Base

HEATSTOR

20000

HEATSTOR

3500

OnlyHeat

Heat capacity (MW) Heat production (GWh)

25000

Base

Heat capacity (MW) Heat production (GWh)

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

WO_BP_RU PE_BP_RU NG_BP_RU

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

In the ‗Rural‘ area during winter time, charging of heat storages is mostly based on the use of electric heat boilers. They create large amount of heat in relatively short time during periods of low power prices. During summer time, heat storages are charged by turning on wood waste and forest residue CHP units. During spring and fall CHP units operate more often, since the heat load is larger, but still the heat storage helps to shut them down for periods of some hours.

Dynamics of heat storage Most of the daily fluctuation in heat demand was smoothed with heat storages and electric heat boilers in all heat areas. If CHP units were operated, they were usually operated at maximum heat output. The investment cost for heat storage was assumed to be 1840 €/kWh. With the assumed ratio of 12 between storage capacity and heat capacity this translates to 153 €/kW. In comparison the capacity cost of electric heat boilers was assumed to be 40 €/kW and 50 €/kW for natural gas heat boiler. This means that investment into heat storage capacity was not driven by need for new capacity since heat boilers were cheaper. There had to be operational benefits from the use of heat storage to cover the additional investment costs.

‗Urban‘ area has similar dynamics, but during summer time the adjustment is made by heat pumps instead of CHP. In the winter during high power prices old natural gas CHP units are less expensive to operate than the heat pumps. CONCLUSIONS

Heat storages create operational benefits by moving consumption from more expensive sources of heat to less expensive by shifting demand in time. In all heating areas whole operating ranges of heat storages were extensively utilized. During most 168 hour periods heat storage reached both the minimum and maximum storage capacities. In the ‗Rural‘ area heat storage was 2.1% of the time either full or empty. With a larger storage capacity this could have been reduced, but it was not worth the investment.

District heating systems offer good possibilities for increasing the flexibility of the power system, if the penetration of variable power like wind power increases greatly in the future. According to the results, main vessels to increase flexibility are the use of heat storages, electric heat boilers and flexible operation of CHP units. Investment in electric heat boilers in district heating systems is driven mainly by periods of very high wind power production. The resulting cheap electricity is converted to heat and to some extent stored in heat storages for later use. Investments in heat storage in turn are driven by the same mechanisms, but also to create flexibility in the electricity production when prices are higher. To enable this, the operation of CHP units and heat pumps is altered with the help of heat storages. Heat pumps mainly compete against CHP as a source of heat. They succeed in replacing coal CHP, but are not very competitive against wood residues. This is naturally due to assumed costs where coal has a considerably penalty due to CO2 cost. Heat pumps are not very important as a source of flexibility, since they require lot of full load hours due to their investment cost.

The size of the heat storage in ‗Industry‘ area was larger than in other areas in relation to daily heat demand (Fig. 8). In ‗Industry‘ area charging of heat storages took place over several days during higher power prices, when wood waste CHP units were producing extra electricity. Storing the extra heat required larger heat storage capacity. On the contrary, in ‗Rural‘ and ‗Urban‘ charging and discharging was more balanced and smaller heat storage was enough.

180 160

Heat (GWh)

140 120

Heat storage size Max daily heat Min daily heat Average

While the research has been conducted on district heating, similar dynamics could be achieved in household heating not connected to district heating networks. However, the costs are likely to be larger unless there is an existing hot water tank. Flexibility could also be gained from district cooling or airconditioning units with the addition of a cold storage.

100 80 60 40 20

Further research should also address some of the shortcomings of current study. Sensitivity analysis would be important, especially concerning the cost estimates of the analysed heat measures. Heat storage model was very simple and this should be improved. Heat grade, especially in the industrial environment, can vary and the model should take this into account.

0 Rural

Urban

Industry

Fig. 8. Heat storage size compared to maximum, minimum and average daily heat demands. 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 and this is a crude approximation even if the heat source is groundwater or sea water.

[6] J.K. Kaldellis and D. Zafirakis, ―Optimum energy storage techniques for the improvement of renewable energy sources-based electricity generation economic efficiency‖, Energy, Vol. 32, pp. 2295–2305. Elsevier.

REFERENCES

[7] H. Lund and E. Münster, ―Modelling of energy systems with a high percentage of CHP and wind power‖, Renewable Energy, Vol. 28, 2003, pp. 2179-2193. Elsevier. doi:10.1016/S09601481(03)00125-3

[1] H. Ravn et al. Balmorel: A Model for Analyses of the Electricity and CHP Markets in the Baltic Sea Region. Balmorel Project 2001. See also: http://www.balmorel.com/Doc/BMainReport0301.pdf

[8] H. Lund, ―Large-scale integration of wind power into different energy systems‖, Energy, Volume 30, Issue 13, October 2005, pp. 2402-2412. Elsevier. doi:10.1016/j.energy.2004.11.001

[2] J Kiviluoma and P. Meibom, ―Influence of wind power, plug-in electric vehicles, and heat storages on power system investments‖, Energy, Volume 35, Issue 3, March 2010, pp. 1244-1255. Elsevier. doi:10.1016/j.energy.2009.11.004

[9] H. Lund, B. Möller, B.V. Mathiesen and A. Dyrelund, ―The role of district heating in future renewable energy systems‖, Energy, Vol. 35, 2010, pp. 1381-1390. doi:10.1016/j.energy.2009.11.023

[3] Energy consumption in the UK: overall data tables, 2009 update. Department of Energy and Climate Change - secondary analysis of data from the Digest of UK Energy Statistics, Office of National Statistics and the Building Research Establishment. [4] Annual Energy Review 2008. Information Administration.

U.S.

[10] K. Karlsson and P. Meibom, ―Optimal investment paths for future renewable based energy systems – Using the optimisation model Balmorel‖, International Journal of Hydrogen Energy Vol. 33, 2008, pp. 1777-1787.

Energy

[11] S.G. Jensen and P. Meibom, ―Investments in liberalised power markets. Gas turbine investment opportunities in the Nordic power system‖, Int. J. Electr. Power Energy Syst. Vol. 30, 2008, pp. 113–124.

[5] H. Ibrahim, A. Ilinca and J. Perron, ―Energy storage systems—Characteristics and comparisons‖, Renewable and Sustainable Energy Reviews, Volume 12, Issue 5, June 2008, pp. 1221-1250. Elsevier. doi:10.1016/j.rser.2007.01.023

198

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

DAILY HEAT LOAD VARIATION IN SWEDISH DISTRICT HEATING SYSTEMS H. Gadd and S. Werner School of Business and Engineering, Halmstad University SE-301 18 Halmstad, Phone: +46 35 167757 henrik.gadd@hh.se, sven.werner@hh.se, www.hh.se 

ABSTRACT If daily heat load variations could be eliminated in district heating-systems, it would make the operation of the district heating system less costly and more competitive . There would be several advantages in the operation such as:    



Easier to optimize the operation that leads to higher conversion efficiencies. Less need for maintenance because of more smooth operation of the plants

To do this some questions need to be answered: 

Less use of expensive peak load power where often expensive fuels are used. Less need for peak load power capacity. Easier to optimize the operation that leads to higher conversion efficiencies. Less need for maintenance because of more smooth operation of the plants

 

What input and output capacity to/from the heat storage is needed? What size of the heat storage is needed? Are the daily heat variations in the specific system large or small during a year?

METHOD Nomenclature

There are a number of ways to handle the daily variations of the heat load. Two often used are large heat storages or using the district heating network as temporary storage. If it would be possible to centrally control the customer substations, it would also be possible to use heavy buildings connected to the district heating system as heat storages.

Ph= Present hour value [MWh/h] Pd= Mean hour value for the present day [MWh/h] Pa= Mean hour value for the whole year [MWh/h] Sh = Energy transfer capacity [MWh/h] Sd = Size of heat storage [MWh/day]

To be able to find the best way to reduce or even eliminate the daily heat load variations, you need to understand the characteristics of the daily variations. This paper will describe a way of characterizing daily heat load variations in some Swedish district heatingsystems.

Sa = Total annual daily heat load variation

INTRODUCTION

Variables

 h = Momentary daily variation [h/h]

 d = Total daily variation [h/day]  a = Total annual relative daily variation [h/year] Measured data has been collected from some district heating systems in Sweden. The collected data is the heat power that is generated and fed into the district heating network. It is hour mean power that is used, i.e. 8 760 data points per year. Only whole years is used from 1 of January to 31 of December. To describe the daily variation three variables is defined.

For all heat generation/distribution systems, heat load variations leads to inefficiencies. You need to design your system for the peak load even though you only need the top capacity for a very short period of time of the year. This is of cause expensive. The solution to this problem is heat storage. There are a number of possibilities to store heat in DH systems:   

Large heat storages at the heat generation plants Heat storage in district heating networks Heat storage in heavy buildings in by allowing small variation in indoor temperatures[1].



Momentary daily variation (  h )

Total daily variation. (  d )

Total annual relative daily variation. (  a )

Three system examples are presented in this paper to exemplify the method to characterize district heating daily heat load variation:

If it would be possible to extinguish daily variations it would lead to several profitable advantages such as: 

Less use of expensive peak load power where often expensive fuels are used. Less need for peak load power capacity.

System A: From a city in South of Sweden with an 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

System B: From a city in Southwest of Sweden with an annual heat generation of 64 GWh.

1 24  Ph  Pd 2 h 1 d  Pa

System C: From a city in the middle of Sweden with an annual heat generation of 1550 GWh.

The total daily variation is presented in Fig. 2 for the three example systems. The figure verifies that the variations are more pronounced in the two smaller systems compared to the larger system. Another implication is that the highest day values are very few, giving an incentive to construct heat storages somewhat smaller than the peaks in the figure. Hence, the investment costs will be reduced more the lost benefits from the storage, giving a more optimised heat storage.

Momentary daily variation (  h ) The momentary daily variation is proportional to the amount of heat that needs to be fed in or out to the DH network to extinguish the daily variation. This variable describe the heat power capacity needed for in and out put from and to the heat storage. For each district heating systems you will get 8 760 (8 784 during leap years) values per system and year. The momentary daily variation is defined as the difference of each hourly measured value and the mean value of heat per hour of the same day divided by the mean heat per hour of the year.

Systen A System B System C

4,5 4

Ph  Pd Pa

3,5 Total daily variation, dτ [h/day]

h 

Total daily variation 5

The momentary daily variation is presented in Fig. 1 for the three example systems. The figure shows that the variations are more pronounced in the two smaller systems compared to the larger system.

3 2,5 2 1,5 1 0,5 0

0,6

Systen A System B System C

0,5

Momentary daily variation, hτ [h/h]

0,4

0,2

0 -0,1 -0,2 -0,3 -0,4 -0,5 -0,6 -0,7 3000

4000

5000

200

250

300

350

Total annual daily variation is a variable that is proportional to the total amount of energy that at daily basis divert from the mean value accumulated for a period of one year. It is used to compare different systems between themselves. For each DH systems you will get 1 value per system and year.

0,1

2000

150

Total annual relative daily variation (  a )

0,3

1000

100

Fig. 2 Total daily variation sorted by size day by day for the three different district heating systems.

0,7

Days of the year

Momentary daily variation

6000

7000

8000

Hour of the year

Total annual daily variation is defined as the sum over the year of the difference between each hourly measuring value and the mean value of energy per hour of the same day divided by two times the mean energy per hour of the year.

Fig. 1 Momentary daily variation sorted by size hour by hour for the three different district heating systems. Total daily variation (  d ) Total daily variation is defined for each day and is a variable that is proportional to the amount of heat that divert from the daily mean heat load. If you want to extinguish the daily variation in a system this variable describe the size of the heat storage. For each DH systems you will get 365 (366 during leap years) values per system and year.

1 8760,365  Ph  Pd 2 h 1, d 1 a  Pa The annual daily variation is presented in Fig. 3 for 10 different Swedish district heating systems. Since the annual daily load variation has a magnitude of 250–500 h, only 3–6% of the annual heat load is generated above the daily average heat loads. Hence, it is the seasonal variations that dominate the heat load variations in the Swedish district heating systems.

The total daily variation is defined as the sum over the day of the difference of each hourly measuring value and the mean value of energy per hour of the same day divided by two times the mean energy per hour of the year. 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

An expected conclusion would be that large district heating systems have smaller relative daily variations

500 SYSTEM A

Total annual daily variation, τa [h/year]

450 400

(  a ) than small district heating systems. There are two

SYSTEM B

350 300

reasons for that:

SYSTEM C

250

1. In a large district heating system, the use of heat power is spread on different distances from the heat plant, i e the chilled water in the return pipe return back to the heat generation at different time compared to when the return water left ach substation (geographical diversity)

200 150 100 50 0 1

100

1 000

10 000

Annual heat supply [GWh]

Fig. 3 Total annual daily variation for 10 different district heating systems in Sweden.

2. In large district heating networks, you would expect that the operators have more active operation of the heat distribution network with respect to temporary heat storage.

RESULTS To characterize daily heat load variations in district heating systems three variables have been defined.

But as can be observed in the Fig. 3 there does not seem to be such a trend. One explanation could be that the heat users differ in different systems. e.g. in the system in Fig 3 with an annual heat supply of 9 GWh, mostly single and multi family houses are connected and very few industry or office buildings are connected.

 h = Momentary daily variation

 d = Total daily variation  a = Total annual daily variation Together with the mean annual heat per hour (Pa) and the energy transfer capacity in and out of the heat storage, size of storage to extinguish the systems daily variation and the total daily variation and can be determined according to the expressions below.

Since there is a large diversity among the annual daily variation more data need to be collected to be able to make any further conclusions. REFERENCES

Energy transfer capacity: Sh =

[1] Olsson L, Werner S: ―Building mass used as short term heat storage‖, The 11th International Symposium on District Heating and Cooling Reykjavik 2008.

 h ·Pa [MWh/h]

Size of heat storage: Sd =

 d ·Pa [MWh/day]

Total annual daily heat load variation: Sa =

 a ·Pa

[MWh/year]

201

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

DISTRICT HEATING AS PART OF THE ENERGY SYSTEM: AN ENVIRONMENTAL PERSPECTIVE ON ‘PASSIVE HOUSES’ AND HEAT REPLACING ELECTRICITY USE Morgan Fröling

1,2

and Ingrid Nyström

1 2

Engineering and Sustainable Development, Mid Sweden University, Östersund, Sweden Chemical Environmental Science, Chalmers University of Technology, Göteborg, Sweden 3 CIT Industriell Energianalys, Göteborg, Sweden Increased energy efficiency is in itself a desirable goal for a society – it increases the robustness of the energy system and the possibilities for a resource efficient and more sustainable energy system in the long run. However, it is possible to create a system with higher environmental impacts with energy efficient buildings compared to less energy efficient buildings through choice of less good energy carriers. It is not enough that the individual parts of a system are good and efficient to give a low environmental impact; the parts must be connected into the system in a good way. Thus it is important to identify system solutions that avoids sub optimization and gives us energy efficient buildings and an efficient energy system with good environmental performance.

ABSTRACT Energy use for space heating, hot tap water and other heat use at comparatively low temperature levels represent a substantial part of the total energy use in Sweden and countries with similar climate. It is thus of importance to meet this demand in a way generating as small environmental impact as possible. However, it is possible to create a system with higher environmental impacts with energy efficient buildings compared to less energy efficient buildings through choice of less good energy carriers. It is not enough that the individual parts of a system are good and efficient to give a low environmental impact; the parts must be connected into the system in a good way. From environmental perspective energy efficient buildings and district heating don‘t oppose each other – good parts connected in a good system will give an optimal. The results from the study of the three items of household equipment show possibilities for district heating to be an alternative with good environmental performance, but not under all heat generation regimes.

In a synthesis studies within the framework of Chalmers Energy Center [1] the role of district heating in a future society with more energy efficient buildings have been investigated. Here we report on general findings of this study with a special focus on the environmental performance of the possibility to convert some household electricity use into district heating - for the use in dish washers, washing machines and tumble driers [2]. The environmental performance is studied using life cycle assessment methodology and different assumptions regarding electricity and district heating generation.

INTRODUCTION It is of importance to meet for space heating, hot tap water and other heat use at comparatively low temperature levels in a way generating as small environmental impact as possible. This can be done by increasing the efficiency in the use phase and in the heating systems of buildings as well as through heat generation systems with low environmental impact. During recent years there has been a focus on houses with low need of space heating, low energy houses or ―passive houses‖. In such buildings the heat from the incoming sun radiation together with body heat from people living in the houses and different household equipment will cover the whole or at least substantial parts of the space heating need over a year (extra heating might be needed during the coldest days of a year). Hot tap water still need to be heated. For parts of the year this can be achieved by solar panels, but there is a need for extra heating during winter. This might result in the extra heating demand being covered by electricity, directly or indirectly.

DISTRICT HEATING – DEMAND SIDE There are today several drivers in the direction of lower total heat market for district heating in future [1]. Among possible such drivers in Sweden are:      

Warmer climate (due to climate change) Higher energy prices Increased environmental awareness Increased energy efficiency of existing building stock Limited amounts of new housing New housing more energy efficient

However, there are also possible drivers for a larger heat market in future, e.g.: 

202

Increased wealth giving larger living space per person and higher demands on comfort

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia 

 

Electricity prices might increase faster than heat prices might lead to interest in heat instead of electricity for ―new‖ applications (washer, dishwasher, et.c.) Heat for comfort cooling Increased use of heat for other purposes – e.g. drying of biofuels et c.

With strategic planning the resulting effect for district heating might be a lower total but at the same time more even demand of heat (Fig. 1). DISTRICT HEATING – SUPPLY SIDE

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

A strategic role of district heating in the energy system is the ability to utilize and deliver resources that otherwise would have been lost. Among possible system drivers on the supply side in Sweden are [1]:   

At the same time we can also expect:  

Increased utilization of industrial surplus heat Remaining large potential of waste incineration Increase of CHP power production



Increased competition for bio fuel resources Higher prices on high quality energy carriers (electricity and fuels) might drive towards smaller fraction as heat. Increased energy efficiency in industrial processes.

With strategic planning district heating might utilize residual heat from processes producing combinations of high quality energy carriers (or bio based material production). The focus can probably not be on heat production. Even combined heat and power production from bio fuels might not be efficient enough for competitive district heating (Fig 2). Fig. 1 Possible change for district heating demand in future – decreasing demand but more even over the year.

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

If district heating should continue to be seen in general as an environmentally preferable option it is important that district heating companies continue to develop district heating production in a favourable direction.

THE OVERALL ENERGY SYSTEM The energy system of a country is complex, and it is important to understand how changes in sub systems may affect the whole system. Sub optimizations might easily occur. A simplified example of a situation where a more energy efficient building through sub optimization of the total system gives a larger overall primary energy need is illustrated in Fig. 3. Obviously it is possible to create a system with higher environmental impacts with energy efficient buildings compared to a system with less energy efficient buildings. It is not enough that the individual parts of a system are good and efficient to give a low environmental impact; the parts must be connected into the system in a good way.

Heat for district heating should originate from resources that are otherwise wasted. In the long term that will mean that bio fuelled district heating is not enough, but heat from other primary production like bio energy or biomaterial combines producing transport fuels and/or bio based materials. CONCLUSIONS From environmental perspective energy efficient buildings and district heating don‘t oppose each other – good parts connected in a good system will give an optimal. It is not enough that the individual parts of a system are good and efficient to give a low environmental impact; the parts must be connected into the system in a good way. The results from the study of the three items of household equipment show possibilities for district heating to be an alternative with good environmental performance, but not under all heat generation regimes. Heat generation must continuously be considered.

Thus it is important to identify system solutions that avoids sub optimization and gives us energy efficient buildings and an efficient energy system with a good environmental performance. IMPLICATIONS OF NEW TYPES HEAT LOAD To better understand implications of different new types of heat load (as illustrated in the right hand side of Figure 1) a life cycle assessment (LCA) has been performed regarding the use of heat instead of electricity for the three examples of house hold appliances: dish washer, washing machine and tumble drier. Basic data regarding the appliances are exemplified with those in the ―district heating villa‖ in Göteborg, Sweden. The LCA model includes energy production (electricity or/and heat) for an average use of each machine and the materials needed to produce it. Different types of energy mixes for electricity and district heat generation were studied. Details of the system boundaries and data can be found in the full report of the study [2]. The results indicate that the total energy system influences the results greatly. If we consider electricity production with large environmental impacts, to utilize district heating is a good alternative, even in cases where the district heating generation in itself is not optimally environmentally friendly. This is exemplified in Fig. 4 where we consider Swedish average district heating fuel mix (bio and residue heat, but also fossil fuels and some peat [5]) and European average electricity generation. If we for the long term development consider electricity generation that is much less fossil carbon intensive and compare it with district heating based on forest bio fuels the results are much more narrow, and it become important what environmental impact category is considered. In Fig. 5 this is exemplified with climate impact and acidification impact.

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

204

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

REFERENCES [1] Ingrid Nyström, Martin Eliasson, Torbjörn Lindholm, Morgan Fröling, Jan-Olof Dahlenbäck, Erik Ahlgren and Elsa Fahlén (2009): Energieffektiv bebyggelse och fjärrvärme i framtiden (in Swedish: Energy efficient built environment and district heating in future). Swedish District Heating Association, Stockholm, Sweden. Available as pdf from www.svenskfjarrvarme.se [2] Morgan Fröling and Ingrid Nyström (2009): Miljöpåverkan från energieffektiva hus och alternativ värme- eller elanvändning (in Swedish: Environmental impacts from energy efficient buildings and alternative heat or electricity use). Published in [2]. [3] Morgan Fröling; Charlotte Reidhav; Jan-Olof Dalenbäck and Sven Werner (2008): Is there a role for district heating in future cities with low energy buildings? 11th International Symposium on District Heating and Cooling, August 31 to September 2, 2008, Reykjavik, ICELAND [4] Göteborg Energi. Fjärrvärmehuset (published in Swedish; ―The district heating house‖). Brochure. Göteborg Energi AB.

Fig. 5 Environmental impact from using district heat for dishwasher, drier and washer. Case: bio based district heating production and Swedish av. electricity.

[5] Morgan Fröling (2004): Environmental limitations for the use of district heating when expanding distribution into areas with low heat density. 9th International Symposium on District Heating and Cooling, August 30-31, 2004, Espoo, Finland.

ACKNOWLEDGEMENT Financial support from the Knut and Alice Wallenberg foundation and the Swedish District Heating Association is gratefully acknowledged.

205

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

ADAPTIVE CONTROL OF RADIATOR SYSTEMS FOR A LOWEST POSSIBLE RETURN TEMPERATURE P. Lauenburg and J. Wollerstrand Lund University, Faculty of Engineering, Department of Energy Science, Sweden, patrick.lauenburg@energy.lth.se oversized for all other heat loads. In addition, radiator systems are generally also oversized for safety reasons, as presented in both Swedish studies [3], [12] and international ones [5], [8] and [10], thus providing further potential to reduce the return temperature.

ABSTRACT The present paper describes how the control of a radiator system connected to a district heating network via a heat exchanger can be optimised to provide the lowest possible district heating return temperature. This can be achieved for each operating point by employing an optimal combination of radiator circuit supply temperature and circulation flow rate. The control algorithm gradually creates a modified control curve for the radiator circuit, enabling it to consistently provide an optimal cooling of the district heating water. Since the heat exchanger is dimensioned for very low outdoor temperatures, it is oversized for all other heat loads. In addition, radiator systems are often oversized due to safety margins. Such facts render it possible to reduce the district heating return temperature.

Objective The objective of the study was to develop a control algorithm for determining the optimal choice of supply temperature and flow in an arbitrary radiator system for every heat load in order to minimise the primary return temperature. Limitation The present investigation has dealt with DH substations that were indirectly connected to the DH network, i.e., hydraulically separated by HEXs.

The objective of the present study was to develop a control algorithm and to test it in practice. A description is here given of the algorithm, as well as of field tests that were carried out to practically verify it. The control method could be implemented in any modern control logics for adaptive control of a radiator circuit, and the obtained results indicated that one can expect a lowering of the return temperature in line with previous theoretical calculations.

OPTIMISED HEATING SYSTEM TEMPERATURES There exist various ways to control the heat output in a heating system. Here, we have dealt with the prevailing control method used in Sweden; an outdoor temperature-compensated supply temperature, ensuring that an adequate amount of heat is supplied to the building at each outdoor temperature. The benefits with regard to the primary return temperature from adjusting the flow according to the heat load are known. The idea of using an optimal combination of flow and supply temperature was conceived by Frederiksen and Wollerstrand [2], and this theory has been further studied [13] [11]. The guidelines from Euroheat & Power [1] state that the lowest return temperature is obtained by varying the flow according to the consumption. If such a variable flow is used, it is controlled by thermostatic radiator valves (TRV) either in combination with a constant supply temperature or with an outdoor temperaturecompensated supply temperature. Langendries [4] suggests a central control of the flow rate through the pumpâ&#x20AC;&#x2DC;s rotating speed, but claims that it appears to be a rather difficult and expensive system. Petitjean [9] proposes a lowering of the pump speed at low heat loads, when the TRVs are almost fully open, but finds it problematic to determine which parameter to use for controlling the pump speed.

INTRODUCTION The present paper demonstrates how the control of a radiator system connected to a district heating (DH) network via a heat exchanger (HEX) can be optimised to provide the lowest possible DH return temperature. This is done by always choosing the optimal radiator supply temperature and flow rate. Relevance of the topic Low return temperatures are beneficial for the production as well as the distribution of DH. A specific advantage of the control method demonstrated in this paper, as opposed to, for example, conventional low flow balancing, is its robustness, enabling the lowest possible return temperatures to be consistently obtained. This is the case independently of the current outdoor temperature and heat load, even if the DH supply temperature changes, the HEX becomes fouled, or the house heating requirements change. The idea is also to utilise the fact that, since a HEX is dimensioned for an extremely low outdoor temperature, it is in fact

It should be possible to implement the control algorithm presented in this paper in any modern, state-of-the-art 206

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 often used for controlling DH substations. The control method suggests how the flow can be determined for each heat load. The flow is regulated by adjusting the pump‘s rotating speed. Speed-controlled pumps are commonly used nowadays and they provide a superior controllability [1], [10].

Table 1: A summary of flow-weighted mean primary return temperatures (bold) and resulting reduction for various temperature programmes.

Let us first study an example of an optimal control curve for a 100 % oversized system. Such a curve is presented in Fig. 1, which also shows the relative magnitude of the varying radiator flow in relation to the required flow. The blue dashed line in the diagram corresponds to the primary return temperature. For the sake of comparison, the primary return temperature for a 55/45 °C system is also shown (gray dashed line). Under the dashed line, results are shown for a system that is oversized by 100 %. The first three temperature programmes are 55/45, 60/40 and 80/30 °C, whereas the last two are optimised ones with variable flow.

100 Tp,s

Temperature [  C]

Tp,r,opt

Ts,s,opt

Ts,r,opt

The following conclusions could be drawn from the table:

Tp,r,55/45

60 50

 The oversizing of a radiator system leads, in itself, to a significant reduction of the primary return temperature, provided that some kind of compensation has been made in order for the system to work properly, i.e., that an accurate indoor temperature has been provided.

Primary return temperature reduction 40

Rel. flow [%]

30 75 ms

50 25 0

-15

-10

-5

0 5 Outdoor temperature [ C]

 By optimising the system (through the use of a variable secondary flow), the primary return temperature can be further reduced, especially if the system is oversized.

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

 By extending the radiator HEX, the return temperature can be further reduced with the temperature programmes that employ a relatively low flow.

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

 Regardless of the degree of oversizing, a combination of an optimised temperature programme and an extended HEX provides a substantially reduced primary return temperature. The values presented in the table have been calculated only for the radiator HEX. When considering the substation‘s total return temperature, it can be said to be smoothed by the DHW consumption. Calculations corresponding to those in Table 1 for a parallel and a 2stage substation for 20 flats (based on the Swedish District Heating Association‘s recommendations for sizing) result in reductions in the return temperature that are approximately 20 % lower than the values shown in the table. The difference between the parallel and the 2-stage connection is negligible when the return temperature from the radiator HEX is low or moderate, a fact that has been previously 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 that a 2-stage connection be used only in large multiresidential buildings if the primary radiator return temperature is high. However, it should not be employed if a low-flow heating system providing low return temperatures is used [1].

i.e., a longer HEX. Furthermore, with optimised control, there exists a preparedness for future changes in system temperatures in the DH network. Should the DH supply temperature be changed, an adaptive control will ensure that the lowest possible return temperature is always achieved.

The advantage of extending the HEX when the secondary flow is low actually demonstrates the optimisation problem: When the secondary flow is reduced, the secondary return temperature will decrease. In the radiator HEX, the situation is different. As the secondary flow decreases, the difference between the primary and the secondary return temperatures, increases as a result of the heat transfer coefficient in the HEX being strongly flow dependent. Fig. 2 shows how the secondary return temperature is lowered with a decreasing secondary flow while the difference between primary and secondary return temperatures increases. This results in a primary return temperature that, at first, decreases and then increases when the secondary flow is further reduced. The values in the figure have been taken from one of the test objects. For this heat load, the lowest primary return temperature was achieved for a secondary flow of approximately 30 % of the original flow.

In order to operate according to Fig. 1, the algorithm must combine a control of the radiator supply temperature with a control of the radiator flow as a function of the heat load and the DH supply temperature. In previous work [7], we have shown that it is possible to manually determine the optimal radiator supply temperature and flow. A natural continuation is to develop a method for automatic adjustment of parameter values for the optimal control algorithm.

THE TEST OBJECTS The tests have been carried out in four multi-residential buildings in the city of Karlshamn, Sweden. The houses were built in 1967-1968: three of them had three stories and a basement, and one had six stories and a basement. The number of flats varied between 20 and 30 per house. The radiators in all houses were fitted with TRVs, but these were at least ten years old. It was thus uncertain whether they functioned properly. The circulation flow was found not to vary significantly in any of the radiator circuits, which may have been an indication that many of the TRVs were not working. However, it should be noted that the presented control algorithm is independent of the use of TRVs in a system. Whatever combination of optimal supply temperature and flow that is identified for a given outdoor temperature, the heat supply will be the same. The main task for TRVs is to limit the heat supply in a room where additional heat supply (solar radiation, bodily warmth or electrical equipment) would result in an overheating of the room.

7 Tp,r,rad 6

Ts,r Grädigkeit

5 Optimum, lowest Tp,r

Grädigkeit, °C

Return temperature, °C

0 15

30 35 Flow, %

The substations were of the 2-stage type and equipped with control logics of the brand IQ Heat (Alfa Laval AB). The equipment for the building automation was manufactured by Siemens and furnished with a separate communications module that could also be used for executing minor computer programmes. There was also an internet connection, rendering it possible to communicate in a number of ways, such as via the software Saphir ScopeMeter© (Siemens), or FTP. After a reconfiguration, the pump speed could be controlled, since all pumps were equipped with communication modules.

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

Another reason for including the impact of an extended HEX in the comparison in Table 1 is the opportunity of connecting to new installations. Large parts of the housing stock in Sweden, built under strong political incentives during the 1960s and 1970s, are facing substantial renovation needs. The results of this project can be considered consistent even if fewer radiator systems be oversized in the future, whether incorporated in older, renovated, or new buildings. The smaller potential for return temperature reductions resulting from less oversized radiator systems may be compensated by the ability to install a HEX that is dimensioned for of an optimised radiator programme,

In order to monitor the circulation flow in the radiator circuits during the tests, clamp-on ultrasonic flowmeters were utilised. However, the objective was to develop a control algorithm based on modern, state-ofthe-art equipment without using additional installations. 208

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

To assure that the temperatures measured in the substation corresponded to the average temperature levels in the various risers in the radiator circuits, temperature sensors were installed in two of the houses. This enabled measurement errors or disturbances in the radiator circuit to be identified. The indoor temperature could be monitored thanks to six wireless sensors installed in each house in the area.

temperature around 0 °C, which corresponded to a load of approximately 50%. The actual flow rate was about 1.1 l/s and the temperatures corresponded to 60/40 °C, thus representing an oversizing around 100%. ADAPTIVE OPTIMISATION - METHOD In the theoretical example, the system was assumed to be 100 % oversized, while in an arbitrary system one cannot be sure of the degree of oversizing. It is also desirable to have a robust and adaptive control algorithm. The method found to function the best is described below. This approach consists in gradually modifying, by automatically performed tests, the control curve and determining the associated flow rate.

Modifications in the substations After some initial tests, the circulation pumps were found to be generally oversized to such an extent that the flow rate could not be decreased as much as desired. There exists a predetermined minimum rotational speed for this type of pump, implying that the speed could be reduced by 60–70%. Discussions with the manufacturer revealed that the lowest pump speed could not be changed in this model, for which reason the decision was made to throttle the flow with an existing shut-off valve located after the pump, which shifted the pump‘s operating range. The throttling was conducted in order for the pump to give half the flow rate at 100% rotational speed. The control curve was modified accordingly, leading to the temperature drop in the radiator circuit becoming doubled and the heat supply remaining unaltered.

Online testing By locking the control valve (CV), one can assume to have approximately the same primary flow through the radiator HEX, and since the variations in the cooling of primary water is relatively small, the heat supply is also approximately constant. If the secondary flow is reduced while the CV is maintained locked, the temperature of the secondary flow leaving the HEX will rise. When a new flow and its associated supply temperature are tested, the current level of the primary return temperature is compared to the level before the experiment. In this way, the new combination of flow and supply temperature can be either accepted or rejected. This method renders it possible to implement the adaptive algorithm in any arbitrary system, leading to the control curve becoming gradually modified. This method we suggested in [7].

We were unable to receive a comprehensive reply from the pump manufacturer with respect to the possible measures regarding the regulation of the pump. A discussion with another manufacturer implied that there were no technical limitations for how far down the pump speed could be controlled. However, such an extension of the manoeuvrable range has so far not been requested. After a simple modification of the pump‘s frequency converter, the working range could be extended from today‘s 30–100% to, in an extreme case, 2–100%.

One problem associated with this kind of optimisation is that the method is sensitive to disturbances. If the primary supply temperature, primary differential pressure or the outdoor temperature changes during the test, one cannot be sure that the heat supply is constant. In that case, a reduced return temperature could be the result of a heat supply that is too low. Such tests have to be rejected.

Existing control of the radiator circuits Although the radiator circuits within the area were designed by the same consultant, there is today a large spread in the choice of control curve and resultant temperature drop (10–30 °C). It is likely that the curves have been gradually adapted to the circuits‘ hydraulic properties and balancing, and one can assume that this is a common situation.

In order to render the tests less sensitive to disturbances, the CV is locked only briefly, in order for the HEX to stabilise. Subsequently, we return to automatic control, but instead of using the control curve, the control aims at maintaining a constant temperature drop in the radiator system. If this is successful, the heat supply is also kept constant. One can assume that the secondary flow is relatively constant: as long as tests are conducted at night, no solar radiation is present and internally generated heat is likely to be at a relatively steady level. If, for instance, the primary supply temperature or differential pressure rises during the course of a test, the CV will close somewhat causing the secondary supply temperature

When older houses are renovated and their radiator circuits are modernised, there are no guarantees that oversizing is taken into consideration. For example, the radiator HEX in a substation that was installed in 2005 in one of the houses was dimensioned for 185 kW heat output at DOT with temperatures corresponding to 80/60 °C at a flow of 2.25 l/s. However, when examining data for this substation, it turned out that the substation delivered less than 40 kW at an outdoor 209

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

to decrease, and thereby also the temperature drop and heat supply, to be detained at the same level.

Tp,s

Ts,r Tp,r,rad

A test is started by keeping the CV locked for ten minutes. This leaves enough time for the HEX to stabilise. The new level of the difference between the primary and secondary return temperatures became stable already after about two minutes in the tested objects. The CV was maintained locked for ten minutes, which should be sufficient even for very low flows and most types of HEXs. Subsequently, the control was resumed in order to ensure a constant temperature drop on the secondary side.

Tp,r,tot

Temperature [  C]

Ts,r  Ts

To To,dam p

Gr 30 20 10 0 00:30

01:00

01:30

02:00

02:30

03:00

03:30

04:00

Time

20 CV,heat

The temperature drop was controlled by verifying the current temperature drop, e.g., every five minutes, and comparing it with the desired temperature drop, i.e., the temperature that was observed when the CV was locked. If the difference exceeded a certain value, 0.2 °C has been used so far, the set-point for the supply temperature was updated according to

CV,DHW

Flow [l/s]

0 1 mp

0.75 0.5 0.25

Heat supply [kW]

Tsetpoint = Ts,r + Tsetpoint. Fig. 3 displays a performed test: At 1:00 a.m., the CV was locked and the radiator flow rate was reduced from 0.59 to 0.36 l/s with the result that the secondary supply temperature rose from 40 to 44 °C. After ten minutes, the temperature drop in the radiator circuit was automatically controlled (in this case, the temperature drop was stable and it took more than 15 minutes before the CV opening degree required adjustment). After ninety minutes, the second flow reduction was carried out, to 0.24 l/s, and the secondary supply temperature increased to about 48 °C.

60 40 20 0 00:30

01:00

01:30

02:00

02:30

03:00

03:30

04:00

Time

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

An interesting aspect of this test was that the primary supply temperature fluctuated a lot. Since the secondary temperature drop was kept constant, it had no impact on the outcome of the test. One can see that the CV generally demonstrated a lower opening degree later in the night, as opposed to before 1:00, when the primary supply temperature increased. Without the T control, the heat supply would have been too high during the last part of the test.

The total primary return temperature varied to a relatively large extent, partly because of tappings of domestic hot water (DHW), but also due to the DHW control in this substation being very unstable when no tappings were made. However, the return temperature from the radiator HEX was of interest for the tests. In this object, the difference between the primary and secondary return temperatures was very small, and even for a low radiator flow, the grädigkeit was below one degree. One can see from the figure that the return temperature had fallen from just under 32 °C to slightly over 28 °C during the test. This resulted in, for a current outdoor temperature of 8 °C, the set-point for the secondary supply temperature being changed from 40 to 48 °C while the flow should be reduced from 0.59 to 0.24 l/s.

The radiator flow was altered by changing the set-point for the pump speed, expressed as a percentage of the maximum speed. It has been found that two flow alterations of ninety minutes each are suitable per test, as this would allow the secondary return temperature to stabilise even at very low flows. The first test for any outdoor temperature, as was the case in Fig. 3, means that starting conditions include the original control curve and flow rate. It is then desirable to perform two fairly large flow reductions since, according to the theoretical calculations, one can expect to find an optimum at a relatively low flow. If, however, the flow is already on a low level, it is reasonable to attempt one slightly higher and one slightly lower flow rate. The algorithm for the adaptive control is illustrated by the flow chart in Fig. 4. 210

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

point, one could expect a stable secondary return temperature, e.g., during the last five minutes. In addition to the secondary supply temperature, also the primary supply temperature is recorded. However, the dampened outdoor temperature, i.e., the input signal to the controller, is recorded when the CV is locked for the first time. The reason for this is that the heat supply is subsequently kept constant at a level matching the outdoor temperature (and heat load) at the time before the test was started.

11:50 PM < time < 0:00 AM Get To, use modified control curve

Yes

Wait 60 min with constant pump speed

Start timer

Save current values

Wait 10 min

Test pump speed 1

Set pump speed for test

Set control valve to Auto

Test pump speed 2

Wait 5 min

< 80 min?

No, > 80 min

Pump speed 1?

Yes

Set control valve to Manual

Wait 10 min

Set T

Yes

The next step consists in using the information attained from the test to modify the control curves. Initially, the original curve was used and the pump was, in our case, controlled to give a constant differential pressure. If the result of a test is that a lower primary return temperature is obtained at a lower secondary flow rate, the control curve is updated for that outdoor temperature. A reasonable resolution is 1 °C. The original control curve, generally based on 5–8 points, was therefore initially extended to comprise values for each outdoor temperature.

No, pump speed 2

T – Tset-p > 0.2°C

Yes

Tset-p = Ts,r + Tset-p Test done

Reject test result Update curves

Not ok

Check maximum deviation for Q (e.g., 5%) and To (e.g., 2°C)

Determine Tp,r,rad,min (pump(0), pump(1) or pump(2))

Fig. 4 Flow chart describing the adaptive control algorithm.

If a modified control curve is used before a test is about to start, the control should be interrupted and the pump speed kept constant for an hour prior to the test. This way, one avoids the risk of the flow changing (due to alterations in the outdoor temperature) too close to the test, which could result in unstable radiator system temperatures.

If the experiment, as in Fig. 3 above, was performed at 8 °C, this point on the curve would be updated. Along with the new supply temperature there followed a new radiator flow, which in our case was expressed as a new set-point for the pump speed. The adaptive control continues in this manner night after night, and the control curves are continuously updated. Outside the test periods of approximately three hours each night, the modified control curves are used for controlling the heating system.

The supply and return temperatures were measured on four of the most remote risers from the substation, during the tests. A continuous matching against measurements on risers gives a good indication that the flow distribution in the system was not impaired by the optimisation. The temperature profile was closely matched to the profile at the substation. Both flow reductions resulted in increased temperature drops.

Fig. 5 shows an example of the gradual development of the modified control curve. The first graph shows a new point at 0 °C (used for 0 ± 0.5 °C). In the second (upper) graph, a point for 3 °C has been added, while the range 0 to 3 °C is complete in the third. The fourth graph shows a much more complete control curve (-5 to 10 °C). Temperature curves corresponding to constant flow systems with lower flows than the original system have been included as thinner lines. The value for 10 °C coincides with the curves of a system with a low flow, while the value of -5 °C coincides with the curves of a system with a moderately reduced flow (normal flow). The last graph clearly demonstrates that the modified curves are based on a variable flow, i.e., they coincide with various constant flow curves at different points.

Updating the control curves After the completion of a test, the obtained information needs to be evaluated. The influence of the variation of the outdoor temperature is not entirely obvious; its influence decreases with an increasing time constant for the building. Variations on the primary side normally have is compensated for since the heat supply is kept constant. As a result, it is sufficient to verify that the heat supply was maintained at a steady level during the test, avoiding any disruptions. If a test result is accepted, the primary return temperatures for each tested flow are compared in order to verify which flow resulted in the lowest return temperature. This flow also gave rise to a secondary supply temperature. It is however not obvious how to read this temperature, given that it was regulated by the controller and changed continuously. The most logical choice is to read the mean value at the end of the test period, before the pump speed changes. At this

As shown in the second graph of Fig. 5, the modified curve could emerge in sections that subsequently are combined. One way to speed up the modification of the control curves is to interpolate intermediate values rather than wait for a flow optimisation at the missing outdoor temperature. Even the return temperatures could be interpolated, since it is possible to determine 211

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

where Ts,r,n is determined in analogy with Ts,s,n, according to:

the required radiator flow for a known temperature drop (and heat supply). 70

Constant, decreased flow

Constant, further decreased flow

Original control curve

Temperature

Modified control curve

Ts ,r ,test  Ts ,r ,n1

(3)

To ensure that the heat supply is kept constant, the required flow for the new temperature drop is calculated. Since the flow is inversely proportional to the temperature drop, it can be determined from the last used flow and temperature drop, together with the new temperature drop, according to:

-10

-5

0 5 Outdoor temperature

-10

Temperature

Ts ,r ,n 

-5

0 5 Outdoor temperature

-5

0 5 Outdoor temperature

20 -10

-5

0 5 Outdoor temperature

-10

m s ,n 

Fig. 5 A stepwise modification of the control curve. The supply temperatures are drawn in solid lines while the returns are dashed.

For the first test to be carried out at a specific outdoor temperature, it is logical to let the results of this test fully replace the original points on the curve. As more tests are performed for the same outdoor temperature, one can proceed in several ways. Since the control should be adaptive and thus able to take into account changing circumstances both in the DH network and in the building, the results of new tests should be employed. However, one may expect that tests performed close to one another in time, at equivalent outdoor temperatures, still provide slightly differing results for varying reasons. A solution would therefore be to use a forgetting factor, i.e., to gradually ―forget‖ old values when the supply temperature curve is updated with new data. A possible approach for doing so consists in calculating the new supply temperature, Ts,s,n, as a mean value of the obtained, Ts,s,test, and the last used, Ts,s,n-1, supply temperature according to:

Ts ,s ,n 

Ts ,s ,test  Ts ,s ,n1 2

(4)

As mentioned earlier, the flow rate is set by changing the set-point for the pump speed. According to the affinity laws for fluid machines, the flow is proportional to the rotational speed. The process of letting the last modified supply temperature and the result of a new test form a new modified supply temperature is illustrated in Fig. 6. 60 Ts,s,test

50 Temperature

Ts,s,n Original curves

Modified curves

30 Ts,r,n

(1)

Ts,r,test

20 2

When a new test is performed at the same outdoor temperature, a new mean value is calculated, which means that older values will have less and less influence. To determine the secondary flow associated with the new supply temperature, i.e., the one providing the correct heat supply at the current outdoor temperature, the expected temperature drop is calculated as:

Ts ,n  Ts ,s ,n  Ts ,r ,n

(m  Ts ) n1 Ts ,n

4 6 Outdoor temperature

Fig. 6. An approach for modifying the control curve based on new test results.

The proposed method for updating the control curves indicates that if for instance the DH utility demonstrates a long-term change in the supply temperature in the network, the control system gradually adapts to the new temperature. However, there are always variations in the primary supply temperature. This may include both unintended and intended variations which may be the result of, for example, a charging of the network if the outdoor temperature is expected to fall. Since the primary supply temperature affects the primary return

(2)

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 also compensate for such short-term variations. One way of doing so is to develop a number of parallel control curves for various intervals of the primary supply temperature. If the temperature is greater than a certain level, an alternative control curve is employed, whereas if it is below a certain level, one utilises another. This method has yet to be tested and there is no basis for assessing how much impact one can expect from normal variations in the supply temperature or what would constitute reasonable intervals for parallel control curves in this case. Another variant could be to perform a linear adjustment for the secondary supply temperature depending on the primary supply temperature, according to:

Ts ,s  Ts ,s ,0 (1  a(Tp,s ,0  Tp,s ))

the system decreases and all risers receive a more similar differential pressure. However, one must be on the look-out for errors (e.g., short circuits) in the systems, a problem that is often emphasised in connection with low-flow systems, as these tend to be more sensitive to hydraulic imperfections [12]. Reduction of the primary return temperature To estimate a yearly mean return temperature reduction (as presented in Table 1) achieved by the adaptive control, an entire, or a major part of the, heating season needs to be evaluated. The control method presented in this paper was developed during the winter and spring of 2009, and only a limited number of tests were performed during the spring. However, Fig. 7 shows the obtained primary return temperature that was attained for the tests that were performed in one of the houses. Note that these results were ―first runs‖ for each outdoor temperature (i.e., the flow was reduced to approximately 40%), signifying that no further optimisations were undertaken. The curve displaying the original return temperatures was based on the average return temperatures from the radiator system prior to any of the modifications (i.e., for the tests or the constant flow rate change, as described in section 3.1).

(5)

where a is a constant that can be determined from tests. Regarding the measurement of temperatures and flows Regarding the temperature measurement in the substation, supply and return temperatures on both the primary and the secondary sides are required. One should keep in mind that, on the primary side, the return temperature from the radiator HEX is needed since the total return temperature is affected by the DHW system. This temperature is normally available in modern substation control equipment.

Tp,r,rad,orig

Primary return temperature

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

Tp,r,rad,opt 40

20 -10

-5

0 5 Outdoor temperature

Fig. 7. Primary return temperatures in the radiator system when the flow is reduced (dots), compared to the original return temperatures (curve).

In the test objects, indoor temperature measurements were used to verify that the adaptive control was able to give the correct indoor temperature. However, one can in fact be sure that the correct amount of energy is transferred to the system for each operating point, regardless of whether the original control curve or the optimised curve is used. A possibility is that there is an imbalance in the system. For example, the most distant riser may not receive the required flow because of a too low differential pressure when the pump speed is decreased. It is, however, more likely that a better balance in the system is achieved when the differential pressure is lowered this since the pressure losses in

CONCLUSIONS AND DISCUSSION An adaptive control algorithm was developed in order to minimise the DH return temperature. The control algorithm can be implemented in any modern control logics for building automation. Some refinement may be done by compensating for short-term temperature variations in the DH network. During the field studies, limitations in the speed control of the circulation pumps have presented a complication. A modification of the frequency converter could increase the working range. 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 modified control curves for the test objects during the present heating season. On the other hand, a control curve with an adaptive controller is never definitive; rather it increases as more operational points (different outdoor temperatures) are added and is then gradually modified if outer conditions change. In order to receive values for the primary return temperature on a yearly basis using the adaptive control algorithm, the new control curve needs to be modified for the entire temperature range. During the performed field studies, the reduction of the primary return temperature was about 3 °C. Even though the test period limited the number of tests, the temperature range was still rather wide, including temperatures from -2 to 14 °C.

[5] Liao, Z., Swainson, M., Dexter, A.L., On the control of heating systems in the UK, Building and Environment 40 (2005) 343-351. [6] Lindkvist, H., Walletun, H., Teknisk utvärdering av gamla och nya fjärrvärmecentraler i Slagsta (Technical evaluation of old and new district heating substations in Slagsta), Report 2005:120, Swedish District Heating Association, 2005. [7] Ljunggren, P., Johansson, P.-O., Wollerstrand, J., Optimised space heating system operation with the aim of lowering the primary return temperature, Proceedings from 11th International Symposium on District Heating and Cooling, Reykjavik, 2008. [8] Peeters, L., Van der Veken, J., Hens, H., Helsen, L., D‘haeseleer, W., Control of heating systems in residential buildings: Current practice, Energy and Buildings 40 (2008) 1446-1455.

It is plausible that certain circuits are more suitable for a variable flow rate, e.g., depending on hydraulic balancing. It would also be possible to map out under which circumstances other heat emitters than radiators, such as fan coil heaters, can be included in a radiator circuit where the flow varies.

[9] Petitjean, R., Total hydronic balancing, Tour & Andersson Hydronics AB, Ljung, Sweden, 1995. [10] Skagestad, B., Mildenstein, P., District Heating and Cooling Connection Handbook, published by the International Energy Agency (R & D Programme on District Heating and Cooling), 2002.

ACKNOWLEDGEMENT The Swedish District Heating Association, the Swedish 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 Differences, Report 2002:S2, International Energy Agency (R & D Programme on District Heating and Cooling) & NOVEM, Sittard, 2002.

REFERENCES [1] Euroheat & Power, Guidelines for District Heating Substations, Downloaded from: http://www.euroheat.org/documents/Guidelines%2 0District%20Heating%20Substations.pdf, 20081117.

[12] Trüschel, A., Hydronic Heating Systems – The Effect Of Design On System Sensitivity, Doctoral Thesis, Chalmers University of Technology, Gothenburg, Sweden, 2002.

[2] Frederiksen, S., Wollerstrand, J., Performance of district heating house station in altered operational modes, 23rd UNICHAL-Congress, Berlin, 1987.

[13] Volla, R., Ulseth, R., Stang, J., Frederiksen, S., Johnson, A., Besant, R., Efficient substations and installations, Report 1996:N5, International Energy Agency (R & D Programme on DHC) & NOVEM, Sittard, The Netherlands, 1996.

[3] Gummérus, P., Petersson, S., Robust Fjärrvärmecentral (Robust District Heating Substation), Report A 99-223, Dept. of Energy and Environment, Chalmers Univ. of Technology, Gothenburg, 1999. [4] Langendries, R., Low Return Temperature (LRT) in District Heating, Energy and Buildings, 12 (1988) 191-200.

214

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

POLICIES AND BARRIERS FOR DISTRICT HEATING AND COOLING OUTSIDE EU COUNTRIES 1

A. Nuorkivi and B. Kalkum

Energy-AN Consulting Energy & Utility Consulting

ABSTRACT

PRELIMINARY COUNTRY SPECIFIC SURVEYS

The policies and barriers faced by DHC in the countries outside the EU will be investigated during 2010â&#x20AC;&#x201C;2011 as a part of the Annex IX of the IEA Implementing Agreement on District Heating and Cooling (DHC), including the integration of CHP.

1. Canada 1.1. Status of DHC The old DH systems before 1985 are predominantly with steam, whereas water systems have been built since 1985. Both domestic hot water (DHW) and space heating (SH) have been included. Based on water/steam carrier, various combinations of heating and cooling are available in Canada.

The countries to be covered are China, USA, Canada, South Korea, Russia and some other selected European countries outside the EU. The work is based on both interviews of the key officers and specialists and the existing laws, regulations and policies of each selected country. The project will also provide examples of best practices useful for sustainable development of DHC as well as offer recommendations to the countries to improve the institutional set up of the DHC.

Historically, Canada has had the highest per capita energy use of the developed countries, as a result of the harsh climate and relatively low-cost, abundant energy. So the benefits of DHC would be particularly welcome to save energy. In Canada, there are records of some 120-160 DHC systems in the country, and almost a half of them located in Ontario Province alone. About 27 Mm2 of residential, industrial and institutional floor area are connected to the DHC systems. This represents about 1,3% of all floor space in Canada. The largest DHC system is in Toronto with 522 MW thermal capacity.[1]

Regarding each country, the project will review, for instance, the tariff setting, DHC related legislation, taxation rules, price regulation, customer definition and points of delivery; ownership of fixed assets; allocation of CHP costs and environmental fees; social considerations; municipal heat planning; and, heat metering and control.

Natural gas distribution has spread everywhere, which is a challenge for DHC expansion. Moreover, at relatively low electricity prices, there is a little market for CHP. No economic market for CHP exists in Canada unless the feed-in tariff is in place or the electricity is used in-house of producer. Power and gas utilities have not been co-operating so far, because there has not been any incentive to such co-operation. Because of the structure of the provincial utilities and low electricity prices, only a few CHP based DHC systems are in operation.

The project here is a twin project to EcoHeat4EU that is a thorough analysis of the barriers and opportunities of DHC as well but in the selected EU member countries. INTRODUCTION There is no reliable statistics of DHC in most of the subject countries. The countries are in different stages of DHC development, as can be read out in the paper. The market drivers and barriers are different as well. The aim of the study is to identify lessons learned from all countries, including the EU that might be useful to boost DHC development in the particular subject country. Nevertheless, the lessons learned and recommendations will be developed in fall 2010, after the Symposium, and the final and complete study will be available in May 2011. Therefore, all information presented in the paper regarding four countries, Canada, China, Ukraine and USA is based on the preliminary survey that will be finalized by October 2010.

The utilities are empowered to provide the people with gas and electricity at the lowest costs possible. Economic drivers support the selection of the proper technologies, and the provincial regulators ensure that the system availability and safety are maintained at all times. Provincial governments provide some directions to the energy industry, but limit themselves to setting overall goals only. The selection of the technologies is left to the utilities. Natural gas is widely available throughout the country, which is a challenge for other heating modes to enter the market. Serious lack of gas reserves is expected in the future, which means alternative energy sources to become increasingly 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 biomass and possibly with CHP is a superior option.

while fearing of intervening the private sector driven heating market.

For DH, two-tier tariffs are used in which energy fee is pass-through of energy costs, and the fixed fee covers the profit, the connection costs and all other cost except energy. The fixed fee can be adjusted annually/biannually with CPI (Consumer Price Index).The customer contracts are made for a long period, say 10-20 years, during which the capital cost have been discounted to the fixed fee. Municipal companies operate as non-profit but private companies with reasonable profit.

1.4. Current Activities The Integrated Community Energy Solutions (ICES) Roundtables have been established to accelerate progress toward reducing GHG emissions by bringing together senior-level stakeholders to exchange views on the best way forward from here. The Roundtables build upon ICES. The Roadmap for Action, which was released by the Canadian Council of Energy Ministers at its annual meeting in September 2009, describes the role that Canada's federal, provincial and territorial governments can play in advancing ICES and it sets out a broad strategy for action. It also includes a variety of options from which the governments can choose, according to their priorities, to advance community energy performance and complement existing energy efficiency activities in different sectors.

1.2. Market Drivers In Canada, the federal government is committed to reducing GHG emissions by 17% below 2005 levels by 2020, being the main driver of DHC. The DHC market is expanding smoothly to start creating a different infrastructure to substitute depleting resources of natural gas.

The ongoing collaboration of key energy actors and enablers across Canada from the private and public sectors through the Quality Urban Energy Systems of Tomorrow (QUEST) collaborative also informed the Roundtable discussion. In particular, preliminary results from a QUEST-led study suggest that ICES could reduce GHG emissions at the community level by as much as 40% to 50%, resulting in reduction of 65 Mt by 2020, which is about 20% of Canada's official 2020 target reductions. These results are very promising and highlight how ICES could contribute significantly to improving Canadaâ&#x20AC;&#x2DC;s energy and GHG performance.

As mental drivers, there is strong interest in municipalities to consider DHC introduction and further expansion very much based on European practise. Many municipalities have set voluntarily targets to the reduced GHG emissions. DHC systems are widely recognized as a potential measure to achieve the targets. The DHC is considered a tool for the urban planners but not an energy issue per se. As an example of investment support, Ontario Power Authority (OPA) subsidizes investments in electricity savings by paying up to $800/kW of the saved electric capacity. The subsidy used to be 400/kW, but was doubled at the end of 2009. Customers can use that money as the partial payment of the connection costs of DHC, thus DHC companies indirectly benefitting from the subsidy system as well.

2. P.R. China 2.1. Status of DHC In China, the DH development has been very strong, more than 10% annually during the past decade on average. By the end of 2005, DH supply (including steam and hot water) was over 2 100 PJ; of which CHP accounted for 47% and boilers accounted for 51%.In the supply of steam and hot water, steam supply is 715 PJ, of which CHP accounts for 81% and boilers account for 17%; the total hot water heating supply is 1395 PJ, of which CHP accounts for 29% and boilers account for 69%. The heating supplied by CHP units and boilers are respectively 992 PJ and 1086 PJ.

1.3. Main Barriers There is no formal DHC strategy or policy supporting DHC and CHP development in Canada. The Government does neither have the tradition nor the willingness to take strong position in DHC development. The private sector that could bring investments and entrepreneurship cannot be much interested, because starting the DHC is risky: long payback times ranging beyond 10 years, limited access to municipal property, challenging contracting of residential, municipal and federal buildings, overall billing and collection of different types of customers. Nevertheless, the municipalities are rather weak, because the municipal taxation only covers property and tourism taxes but no corporate or income taxes. Moreover, municipalities have no mandate on energy. The federal government hesitates to take a strong role

Apart from Europe, only SH is supplied by the DH systems, and the DHW by individual systems: solar collectors, propane, electricity, etc. [2,3] During the few years to come, China will become the largest DH country in the world.

216

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

charges and power grid balancing that need to be addressed. At present, the State Power Grid Group is responsible for the power grid operation. As such, more communication and coordination activities could be conducted between the DHC industries and the State Power Grid Group.

2.2. Market Drivers The rational of strong DH development in China is based on eliminating the small and polluting coal fired boilers in the northern, western and central provinces and to provide feasible living conditions to the population massively moving in to the cities. 

DH has been encouraged by the Chinese government for several decades. China's DH heating area has increased from over 276 Mm2 in 1991 to over 1100 Mm2 in 2000, and exceeded 2500 Mm2 in 2005, with an annual growth rate of 17%. The growth in DH mainly came from the northern and the northeast regions. In China, residential buildings account for about 70% of the total DH area and commercial buildings the balance of about 30%.

POLICY BARRIERS There also exist barriers in the area of economic support and administrative policies related to CHP/DHC, including:

The urban communities are very densely built, which effectively supports centralized heating and cooling solutions. The new buildings comprise about half of the DH connections, whereas the balance for existing buildings, the latter previously having had been heated by small coal boilers.



There is a lack of monitoring and enforcement of the government‟s policies related to the efficient operation of CHP projects. Currently, it appears that some newly- built CHP projects are operating only in thermal generation mode after they have been approved, thereby reducing their energy efficiency.



There is a lack of targeted policy for smaller CHP units. In order to fulfill the energy conservation target, China is attempting to increase the number of more efficient large power generation plants and to close down smaller, older units. While it is important that the smaller, more inefficient units be closed down, some small CHP units with high efficiency are also being targeted for phase-out. Based on the goal of increasing energy supply efficiency, a different policy should be adopted. For example, in regions with low heating loads, small CHP units could provide most of their energy needs at a fraction of the cost of larger units.

2.3. Main Barriers The DHC sector is expanding fast but there are still some barriers regarding economy, policy, financing and technology as summarized below. ECONOMIC AND PRICING BARRIERS In order to become cost-effective and an attractive investment, power and heating reform policies will need to be undertaken. Some of the key issues include: 



Centralized DHW would benefit CHP. Missing DHW load hampers economic development of CHP schemes. Without DHW, the CHP plants can operate all year round only if there is industrial steam load existing nearby.

Energy price policy reform is a priority. At present, in China, the coal price is based on the market, which has grown rapidly in recent years. However, electricity and heating prices are still controlled by the government, and have only slightly increased. While the government has provided limited subsidies to DH companies, most CHP enterprises and DH companies are currently not making a profit as a result of the lack of energy price reform.

FINANCING BARRIERS There are promising energy conservation projects – particularly in the DH sector – that could be realized if there were sufficient funds or other means available to address the gap in investment capital. In particular:

In addition, heating reform needs to be further developed. Currently, in most cases, heat tariffs are based on the building area, rather than on the actual heat consumption, which has a negative influence on improving the energy efficiency in district heat facilities and buildings.



Power sector reform is also needed. At present, the electricity produced by most DHC (and some CHP) projects cannot interconnect with the power grid, which has strongly reduced development. The technical issues of grid connection can likely be addressed. However, there are also administrative interconnection issues, such as added-capacity 217

Some planned CHP/DHC projects are not operated efficiently because they lack sufficient resources to invest in expanded heat pipeline infrastructure. Further, at many existing DHC projects, the heat loss in pipelines is high, reducing the overall efficiency of the heating system. Additional financing is needed to invest in cost-effective heat pipeline retrofit projects, which will generate sizeable energy efficiency benefits and GHG reductions.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia 

While energy service companies are expanding in the commercial building energy conservation arena, they have not yet entered the CHP/DHC area. There is some room for these types of thirdparty players to come up with innovative means to finance projects.

3. Ukraine 3.1. Status of DHC Ukraine is one of the largest DH countries in Europe. Currently almost 80% of urban housing is supplied with DH through extensive grids of hot water pipes. The DH sector is rather saturated, but in some eastern (Donbas) cities the DH systems are deteriorating fast, and customers are either adopting apartment level gas boilers or even remain without heating, thus enjoying on the heat losses penetrating to them through walls free of charge from their heated neighbours. Even the municipalities are offering investment subsidies to the apartment owners to purchase apartment level gas boilers while disconnecting the DH services.

TECHNICAL BARRIERS While CHP/DHC are proven, existing technologies that do not require major research and development, there are some advanced technologies that could be introduced from IEA Member Countries to improve efficiency and operational benefits. In addition, there is currently some debate about the relative merits of DC technology. China-specific research studies could be conducted to confirm the primary energy conservation performance of these technologies.

Such practices have led to extremely poor quality of DH services: low water and room temperatures a well as periodical heating are used to minimize fuel costs. There are coal (and anthracite) mines in Ukraine, but little used for providing fuel for DH: Most DH is based on natural gas imported from Russia. The costs of gas comprise 50–70% of the DH, which explains why the DH is vulnerable to gas price changes.

ORGANIZATIONAL BARRIERS There are some organizational barriers for optimal development as well. 



Scattered organizations with several heat suppliers and distributors prevail in one city. In the same DH system, the heat supplier is responsible for operation and maintenance until the group substations that serve several buildings through the secondary network, and the distributors being responsible from the substations to the indoor heating elements. Therefore, the holistic optimization can be often compromised by partial optimizations. The DHC companies are operation and maintenance companies only, whereas investment decisions and financing depends on the municipal and provincial budgets. This is one more reason for that there is little business minded atmosphere in the extensively staffed DHC companies.

Ukrainian heat generating facilities are ineffective for many reasons. The most important reasons are as follows:    

technology used for heat generation is outdated and inefficient; key assets are heavily deteriorated; equipment is being used in a switching mode on unspecified fuel; delays and failures to carry out regular repairs.

According to the Ministry of Fuel and Energy, more than 90% of energy units have worked out their projected service life (100 000 hours), more than 60% have been in service longer than 200 000 hours. Heat tariff for final consumers is defined as a sum of tariffs for production, transportation and supply.

2.4. Current Activities The DH systems are expanding fast in China, simultaneously restricting coal consumption and reducing overall GHG emissions of the heating services.

Tariffs for heat that is produced by CHPs, cogeneration or alternative/renewable energy sources are set by the National Energy Regulatory Commission (NERC) but they should not be higher than heat produced by other sources.

The Ministry of Construction has issued the Housing and Building Reform on Energy Efficiency (HRBEE), which requires more efficient buildings to be built as well as introduction of heat metering and consumption based billing. The first consumption based billing pilot was initiated in Tianjin a few years ago with a two-tier heating tariff. Such billing systems are slowly expanding to other regions.

Tariffs for heat production, transportation and supply other than CHPs, co-generation or alternative /renewable energy sources are approved by local governments. Due to that the tariffs differ much across the territory of Ukraine. According to the Law of Ukraine ―On Heat Supply‖, heat tariffs should cover all the economically sound expenses for heat production, transportation and supply. Tariffs should include full costs of heat 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 that is not lower than the level defined by the Cabinet of Ministers on the base of calculations by the central body of executive power in heat supply.

directed to energy saving in heat, water supply and sewerage. But according to monitoring results, the funds are allocated to other purposes. Only four regions used the funds for energy saving. Other regions used from 7% to 40% instead of the required 75% to energy saving.

If heat tariffs do not cover the cost of heat and marginal profitability level, the body that has set the tariff should provide for the compensation according to effective legislation. That is, if the tariffs for heat from thermal power station and boilers that are approved by the local government on the basis of heat producer calculation, and they are lower than economically sound cost including marginal profitability level, the local governments must compensate the losses from the local budgets.

Other measures of energy saving that would be appropriate include:    

Meanwhile, the Ministry of Economy elaborated the draft that specifies binding of the household services tariffs to energy prices. First of all, it means heat, hot water and gas supply to households. The current system of tariff setting reduces the competitiveness of Ukrainian industry, since industry is forced to compensate for low households tariffs.

Another stimulus for companies to introduce energy saving technologies is outlined in the Law of Ukraine ―On Heat Supply‖. According to the Article 8, ―in case heat supply or heat transportation companies introduce energy saving measures that result in saving of heat losses, the body of executive power, that is entitled to regulate heat tariffs according to the Law, may leave the tariffs unchanged for the three consecutive years‖.[4]

The procedure to raise the heat tariffs is rather complicated and time consuming, as follows: 1) The district heat supply company receives official notification from NERC on gas price increase. Only after that the company may start developing the proposal on the heat tariffs increase.

3.3. Main Barriers In general, there are a number of decent laws and regulations that would support DHC development, but they are not implemented properly, as mentioned above already.

2) The new heat tariffs have to be approved by the following authorities: Commissions of the Municipal Council (mis‘krada) and regional council (oblrada). The tariff proposal has to be reviewed by several instances as listed below:     

replacement or reconstruction of steam and gas boilers with efficiency that is lower than 89%; improvement of heat pipes insulation to decrease losses in transmission pipelines; installation of heat meters; and, Installation of co-generation equipment.

Therefore, there is little if any incentives to business oriented development of the heating services, but the systems are run at minimum investments and reduced technical performance. The DH companies are solely operation organizations, mainly departments of the municipality. The municipalities take care of billing and collecting based on subsidized lump sum tariffs, and on investment decisions.

Trade unions Antimonopoly Committee Department for Price Administration Department for the Protection of Consumer Rights Public hearings

There are many privileged customer categories that enjoy reduced costs of DH services. In Odessa, for instance, 25% of the customers in year 2006 enjoyed such privileged heating prices. Their billings were decreased by 20, 30, 50, 75 or even 100%, which effectively destroys the business opportunities of DH.

3) Municipal Executive Committee (misk‘vykonkom) has to approve the new heat tariffs as well. 4) The tariff changes shall be publicized via official mass media of Municipal or Regional Council. If during a month there are no official protests from the Office of Public Prosecutor, the company is entitled to apply the new tariffs.

Individual and autonomous heating in every apartment seems the most favourable option for consumers. In such a case they do not pay for heat and hot water but only for gas and cold water. In addition, they can regulate temperature in their apartments and do not suffer from overheating in spring and insufficient heating in winter. But sometimes it is impossible to install autonomous boilers in every apartment, because there is not enough space for heating equipment and the vertical ventilation ducts are not designed for flue gases. Therefore, it would be appropriate to install one

The above steps clearly show how cumbersome any tariff increase can be in practise. 3.2. Market Drivers Article 54 of the state budget of Ukraine for 2006 and the Cabinet of Minister‘s Decree No.207 of 9 March, 2006 stipulate for subsidies from the state budget to local budgets. No less than 75% of the subsidy must be 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 several buildings. Another problem for individual and autonomous heating is that in case of gas supply interruption there is no reserve fuel resources to continue heating. Reserve fuel can be provided only for centralized DH.

DHC/CHP system in New York City is the World‘s largest steam system with 1850+ customers. DHC (primarily DH currently) delivers about 3,5 % of the total final energy demand in the industrial, residential, public, and commercial sectors. In the past two decades, some 47 Mm2 has been connected to the DHC systems, but the total customer base volume number is not available.

Frequent failures in the heating systems as a result of outdated equipment and poor funding are still common throughout the country. Some service breaks in coldest winter times have caused serious impacts on human life already.

The DHC systems are predominantly (80%) with steam, the consumption being a mixture of steam heating, cooling and DHW depending on the particular case. There is little residential heat load but the majority is public: offices, malls, universities and military bases.

Legally, local authorities that establish tariffs for population lower than the cost coverage level have to compensate the difference to energy‐generating companies. In practice the compensation is not always paid in full which leads to arrears accumulation and aggravates financial state of heat‐generators. The procedure of heat tariffs increase is rather complicated, as well as time consuming.

Countrywide, the DH and DC markets are expanding at 3-4%/a and up to 10%/a, respectively, but almost solely on campuses, hospitals, military bases and in the downtown commercial and public buildings.[5] In general, however, DHC together with CHP has been tragically underutilized as a tool to combat climate change, to reduce life-cycle costs of energy supply and to defend energy independence in U.S.A.

According to the Law of Ukraine adopted in April 2006, heat producers such as CHPs and renewable sources power plants are not allowed to cross‐subsidy heat production to cover losses from heat production at the cost of electricity production or other activity.

4.2. Market Drivers The U.S. Congress has acknowledged the benefits DHC/CHP by stating that:

Nevertheless, official sources say that due to low heat tariffs for CHPs heat production is subsidized by the cost of electricity production. But the unofficial sources assert that CHPs may charge heat tariffs that are even higher than heat production cost to cover losses from electricity production, because electricity tariffs are set only by NERC while heat tariffs are set by heat production companies with the approval of local bodies of power.



 

3.4. Current Activities The DH strategy is under preparation in Ukraine as a multi-ministerial approach and it should be ready in fall 2010. CHP development is in the focus of the strategy. There has also been comprehensive framework support initiated by USAID, EBRD and EU to reformulate the national energy policy, including DHC and CHP. It is uncertain now how much the political election of April 2010 will influence availability of such foreign technical assistance in the years to come.



4. U.S.A. 4.1. Status of DHC The total DHC industry base comprises approximately 2 500 systems, in which the number of customer buildings served by a typical DHC system may range from as few as 3 or 4 in the early stages of new system development to the largest system served by Consolidated Edison in Manhattan. The downtown



approximately 30% of the total quantity of energy consumed in the United States is used to provide thermal energy – heating and cooling building space, DHW and industrial processes; thermal energy is an essential, but often overlooked segment of the national energy mix; DHC systems provide sustainable thermal energy infrastructure by producing and distributing thermal energy from CHP, sources of industrial or municipal surplus heat and from renewable sources such as biomass, geothermal, and solar; DHC systems provide advantages that support secure, affordable, renewable, and sustainable energy for the U.S., including use of local fuels or waste heat sources that keep jobs and energy dollars in local economies, stable, predictable energy costs for businesses and industry, reduction in reliance on fossil fuels, reduction in emissions of GHG, and flexibility to modify fuel sources in response to future changes in fuel availabilities and prices and development of new technologies; DHC helps cut peak power demand and reduce power transmission and distribution system constraints; and,

CHP systems increase energy efficiency of power plants by capturing thermal energy and using the thermal energy to provide heating and cooling, more 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 plants.

The definition of CHP is rather complicated. The 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. electric power to 20% by 2030 would avoid 60% of the projected increase in U.S. carbon dioxide emissions (equivalent to taking half of all U.S. passenger vehicles off the road); and, generate $234 billion in new investments.



DHC would be a critical component of this CHP growth. The local electric distribution companies (LDCs) are interested in DHC as a means to reduce the summer peak and to release transmission and distribution capacity to other electric applications that have more even consumption during the year.



at least 20 % of its total useful energy in the form of thermal energy which is not used to produce electrical or mechanical power (or combination thereof), and at least 20% of its total useful energy in the form of electrical or mechanical power (or combination thereof), and the energy efficiency percentage of which exceeds 60%.

The Thermal Energy Efficiency Act of 2009 establishes the Thermal Energy Efficiency Fund that would award grants for DHC, CHP, and recoverable waste energy projects. It includes biomass facilities. Under a federal GHG emissions regulation program, 2% of emission allowances established for each calendar year from 2012–2050 would be allocated to the Fund.

The developers of the building sector are interested in DHC as well, because it would leave more room space in the building for sale. At the municipal level, the market driver for DHC is the reduction of the GHG emissions. Many municipalities have set voluntarily targets to the reduced GHG emissions.

This legislation would dedicate 2% of revenues from climate change legislation to fund CHP, waste energy recovery, and DHC projects. Based on various estimates, this could mean roughly between $1 billion and $1,5 billion per year for clean energy infrastructure. The Thermal Energy Efficiency Act would provide 40% of its funding for institutional entities (defined as public or non-profit hospitals, local and state governments, school districts and higher education facilities, tribal governments, municipal utilities, or their designees), 40% for commercial and industrial entities, and 20% to be used in the discretion of the Secretary of Energy to fund institutional entity projects, commercial and industrial projects, or federal facility projects. A match is required of all non-federal applicants, starting at 25% from 2012-2017, and rising to 50% from 2018 to 2050. The breakdown of how the money would be used is 75% for construction of infrastructure, 15% for planning, engineering, and feasibility studies, and the remaining 10% to be used at the discretion of the Secretary for either infrastructure or planning, depending on the need.

4.3. Main Barriers In general, the barriers are very much the same as already discussed in Canada. Private sector as investor cannot be much interested, because starting the DHC is risky: long pay-back times ranging beyond 10 years, limited access to municipal property, challenging contracting of residential, municipal and federal buildings, overall billing and collection of different types of customers. Only little expansion on residential sector is recognized, and that is because there is voting needed among the condominium owners. The centralized energy systems, that the condo owners are not familiar with and perhaps difficult for them to understand the benefits, have not been adopted on the residential sector in a considerable scale so far. 4.4. Current Activities

In competition with grid power plants receiving generous allowances in ACES, CHP systems could be shut down. Unless allowances are allocated to the DHC CHP system, it will have to purchase allowances for all gas consumed in the facility, resulting in an additional cost equal to 15% of the average 2007 wholesale power price ($57 per MWh) at the $16 per metric ton allowance price projected by Environmental Protection Agency (EPA) for the year 2020. In contrast, the merchant coal plant will only have a GHG allowance cost of only 5% of the average 2007 wholesale power price, because allowances will be allocated for nearly all (83%) of its emissions.

There are several laws and regulations that are expected to support DHC development in the U.S.A.[6,7] Rising interest on development and extension of renewable energy sources as well as improving overall energy efficiency is to be converted to legislation at the moment. Unfortunately, DHC has not been successful in the legislation process so far, but both the Department of Energy as well as the DHC and CHP associations such as IDEA and USCHPA are working on it.

221

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Faced with this huge competitive disadvantage in the marginal cost of power generation, some existing CHP facilities will shut down and construction of new CHP plants will be choked off.

CONCLUSION The survey work is still underway, and therefore, the lessons learned and recommendations will be issued in the final report in spring 2011.

In the ACES, DHC systems are not directly covered entities unless they qualify as ‗electricity sources‘. District systems would be covered indirectly through the costs of allowances built into the prices of purchased fuel oil, or natural gas if purchased from the gas LDC. However, gas purchased on the wholesale market or coal users not qualifying as an electricity source would not be required to submit allowances.

ACKNOWLEDGEMENT The authors express their gratitude to the interviewed specialists, Mr. B. Gilmour (Canadian Urban Institute), Mr. K. Church (Natural Resources Canada), Mr. M. Wiggin (Public Works and Government Services Canada), Mr. R. Thornton (International District Energy Association – IDEA), Mr. D. Kaempf and Ms. P. Garland (U.S. Department of Energy), Mr. B. Hedman (ICF International) and Messrs. G. Draugelis and P. Salminen in the World Bank.

This is a fundamentally good framework with the exception of the concerns about CHP systems to be covered or not. However, if the upcoming legislative process results in modifications that make many DHC systems covered entities, it is critical that changes in allowance allocations be made as discussed below. For example, if the final climate change bill regulates all sources with emissions greater than 25 000 metric tons CO2e (the threshold generally used in the ACES as well as a number of past bills), over 70% of DHC systems and over 95% of DHC output would be capped. In such a way, more efficient systems will have competitive advantage, because the quantity of allowances needed per unit of energy will be lower.

REERENCES [1] National DHC Survey, Canadian DHC Association (CDEA), 2009. [2] Ministry of Construction, China City Construction Statistic Annual. The DH data does not include industrial steam and hot water. [3] T. Kerr, IEA Collateral, Sustainable Energy in China: The Role of CHP and District Heating/Cooling, 2008.

American Clean Energy Leadership Act (ACELA) in June 2009 and Federal Renewable/Energy Efficiency Standard establishes a Renewable Electricity Standard which includes provision for energy efficiency credits as well as renewable energy credits that can benefit DHC as well.

[4] A. Tsarenko, Overview of Heating Sector, CASE Ukraine, 2007. [5] IDEA Report, The DHC Industry, 2005. [6] DHC Services, Commercial Data Analysis for EIA‘s National Energy Modeling System, Energy and Environmental Analysis, Inc. and International DHC Association, 2007.

Renewable Electricity and Energy Efficiency Standard established by ACELA is applicable with the electric utilities selling >4 TWh a year. The utilities are required to supply 20% of demand from combination of renewable sources and increased energy efficiency meaning 15% renewable together with 5% efficiency increase. If the state determines that it cannot meet the renewable requirement, then the portion of renewable sources may fall lower to 12% but with efficiency increase equal or higher than 8%. These requirements provide important leverage for DHC/CHP development.

[7] M. Spurr, Climate Change Legislation in Dollars and Cents, presentation in IEA DHC in Gustavelund, Finland, in Aug 2009.

Energy and Water Development Appropriations Act of 2010 will provide $15 M for DHC feasibility studies.

222

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

BARRIERS TO DISTRICT HEATING DEVELOPMENT IN SOME EUROPEAN COUNTRIES 1

Dag Henning and Olle Mårdsjö 1

Optensys Energianalys, Örng 8c, SE-582 39 Linköping, Sweden, phone +46 70 536 59 22, e-mail dag.henning@optensys.se, www.optensys.se 2 Manergy, P.O. Box 271, SE-581 02 Linköping, www.manergy.se are difficult to use for individual buildings, such as unrefined biomass fuels, heat from waste incineration and industrial surplus heat. The latter may, for example, be a by-product from production of automotive biofuel. District heating can provide cheap energy to consumers by using low-cost energy sources, such as wood, waste and surplus heat. Many of these resources can be of local origin and promote local business and industry.

ABSTRACT District heating (DH) offers low primary energy demand, high security of supply and small CO2 emissions. Barriers to DH in the UK, Ireland, France, Romania and the Czech Republic have been compiled through publications and interviews. DH systems require large investments, have negative initial cash flow and long payback time, which obstructs financing. One actor should control DH from source to consumption. If the value chain is fragmented, contracts are required between the links. It increases risks and financing costs, like in the UK and Ireland, where DH is not established. There are few multi-family houses with central heating and it is expensive to build DH networks in built areas.

The main advantages with district heating are high security of supply through utilisation of domestic renewable energy resources, if available, low primary energy demand due to high conversion efficiency, as well as small CO2 emissions thanks to low fossil fuel use and the high energy efficiency. Incineration of waste with heat recovery to district heating may be used at very low cost. District heating also gives opportunity for cogeneration of power and heat with very high efficiency. District heating enables profitable heat supply with outstanding environmental performance but there are in many places various barriers to a prosperous DH development.

Most French DH systems are operated according to long-term concessions by companies that sell electricity and gas. No strong actor provides unbiased DH support. In the Czech Republic, gas offers DH severe competition. Much DH is produced at the expense of electricity that is considered more valuable, and waste incineration is not popular. In Romania, DH consumption was reduced by one-half. Distribution losses are enormous. New less polluting plants are needed.

Barriers to district heating in the United Kingdom (UK), Ireland, France, Romania and the Czech Republic, as well as barriers to export of Swedish district heating knowledge and products to these countries have been compiled from publications and through personal communication with people in public and private energy bodies and companies in Sweden and abroad [1].

Consortia from established DH countries could offer DH systems from fuel to customer if local policies facilitate DH development.

In the studied countries, there are large potentials for district-heating development and for Swedish sales of DH related goods and services. But for district heating and export to succeed, there are several barriers to overcome in Sweden as well as in the other countries. It should be emphasised that this paper focuses barriers and does not give the full picture of the conditions for district heating, which also includes many possibilities.

INTRODUCTION This paper describes barriers to district heating (DH) in various parts of Europe and to Swedish involvement in district-heating business abroad. The paper is based on a report called ―District Heating in Europe: Barriers to overcome for Swedish export‖ [1], which was prepared for The Swedish District Heating Association. The losses by energy conversion in Europe are of the same magnitude as the European heat demand and consist mainly of heat that is wasted by electricity generation [2]. District heating is a means to utilise such surplus heat to cover heat demand.

BARRIERS IN WELL-DEVELOPED DH COUNTRIES In many countries with well-developed district-heating industry, such as Sweden, much DH competence resides in municipally owned energy companies. They have system knowledge, which could be applicable in other countries. District-heating companies owned by

District heating can utilise the heat from electricity generation in combined heat and power (CHP) plants. District heating can also use other heat sources that 223

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 their business abroad to sales of services, and to a very limited extent goods. For municipal district heating companies, domestic judicial restrictions are the first barriers to overcome before operations in other countries can commence.

Another general barrier to district heating is the EU emission trading scheme, which favours individual heating because individual CO2 emissions do not need allowances.

Only certain components for production and distribution of district heating are manufactured in a single country, which calls for international cooperation. The Swedish Government provides certain but limited support to promotion of district heating business abroad. For example, Swedish district heating consultants work abroad but it is seldom followed by goods export.

In the countries analysed in this project, the barriers are of very diverse nature. The obstacles are dominated by difficulties for district heating itself rather than for foreign companiesâ&#x20AC;&#x2DC; operations in the countries. In the British Isles, it is largely a question of establishing district heating as a natural element in society. In France, it is about large domestic companies that may offer superior competition to foreign firms. In the Czech Republic, French and other companies from abroad dominate the DH business but the technical design of district-heating production may hamper DH development. In Romania, there are several problems with facilities in bad shape and public bodies that have not addressed the issues properly.

TYPES OF DH BARRIERS

FINANCING THE DH VALUE CHAIN Financing is a large barrier to district heating development. DH systems require large investments and may have long payback times. The cash flow is negative for a long time during the establishment of a new DH system. Time horizons are distant, which stresses financers in our present situation of rapidly changing conditions. Private companies often focus on short-term profit and public involvement may be necessary for the deployment, modernisation and longterm development of district heating systems.

Table I is an attempt to assess how large the various barriers are in the studied countries. The table starts with some general conditions. Ownership and organisation considers if district-heating companies are owned, or DH operations are organised, in ways that make it more difficult for Swedish companies to do business. Corruption may be a problem through, for example, indirect bribes by procurement. National and local control encompasses national laws and policy instruments that are disadvantageous for district heating, DH price regulations, as well as municipalities not facilitating district heating by planning of new developments. But rules complicating combined heat and power production are included in the CHP line in Table I.

District heating is a comprehensive concept for heat from source to consumption. Its strength lies in maintaining the value chain (Fig. 1). This may fit badly in an exaggerated market context where every little link of the value chain is organised separately with an interface of costs and revenues to other links. A fragmented value chain increases interface costs and total risk. EU regulations have a tendency to promote such fragmentation. Between the links of a fragmented supply value chain, many complicated agreements are required, which all include risks. It means a larger total financing risk, which raises interest rates and shortens amortisation periods for loans. This implies a mismatch with the depreciation in the balance sheet due to the long economical lifetime of district heating versus the short amortisation time. TWO GENERAL DH BARRIERS

Financing is one of the largest barriers to district heating, primarily because DH schemes give a low rate of return. A fragmented value chain cause contract risks at several instances. Entrance barriers for foreign companies in Table I consider additional difficulties for foreign firms besides the other parameters and the general disadvantage of not being familiar with the domestic business culture.

Two general district-heating barriers are related to CO2 emissions and the attempts to reduce these through, for example, reduced energy use. Global warming and better insulated houses reduce heating demand and, hence, the advantages of district heating because investment costs must be carried by less supplied heat.

Some parameters in Table I are related to districtheating sales. DH competitiveness includes the availability and price of other forms of heating, primarily natural gas. Customer relations concern customer attitudes toward district heating, customersâ&#x20AC;&#x2DC; and suppliersâ&#x20AC;&#x2DC; 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 disconnections have occurred or may occur. Built environment relates to how common multi-family buildings are and if these have a central heating system for the whole house. Table I ends with districtheating production and distribution issues. Biomass considers domestic biomass supplies and infrastructure for biomass fuel supply. Waste includes current waste management and attitudes toward waste incineration. CHP concerns regulations hampering CHP production as well as problems in existing plants. Finally, district heating distribution in

Table I encompasses difficulties with building networks and deficiencies in existing distribution. The assessments in Table I were primarily made within each country and secondly countries were compared but mostly the ranking of countries for a parameter is appropriate. However, every grade has a certain â&#x20AC;&#x2013;widthâ&#x20AC;&#x2013; and two countries with the same digit may differ. As an example, district heating is assessed to be somewhat less competitive in Romania than in the Czech Republic. It follows a description of barriers in the individual countries emphasising the largest barriers.

Table I. â&#x20AC;&#x201C; Height of DH barriers in analysed countries [1] BARRIER

IRELAND

FRANCE

CZECH REPUBLIC

ROMANIA

Ownership and organisation

Corruption

National and local control

Financing

Fragmented value chain

Entrance barrier for foreign companies

DH competitiveness

Customer relations

Built environment

Biomass

Waste

CHP

DH distribution

is rated as a rather large barrier in Table I because supplies are limited in the British Isles and fuel supply systems are less developed.

THE BRITISH ISLES In the United Kingdom (UK), and even more in Ireland, district heating is not really an established phenomenon. Figure 2 shows that residences mostly are heated with gas in the UK, often through a gas boiler for the individual household. Oil is the most common fuel in Irish homes but gas is expanding.

UK Government and municipalities have hitherto not facilitated district-heating development sufficiently and strong incentives for deploying district heating systems are lacking. Heating is generally not regarded as a public concern, but as a concern for each individual. National and local control is therefore indicated as a rather large barrier in Table I. In Ireland, the situation seems to be slightly better but in both countries certain regulations, designed with electricity and gas in mind, are disadvantageous for district heating. CHP suffers especially from rules on how produced heat and power may be supplied.

The largest problem is district heating distribution (Table I). It is expensive and complicated to build DH networks in already built areas and, at least in the UK, it is not straightforward to obtain a licence for putting district heating pipes into streets. The financing difficulties in the British Isles are primarily due to a fragmented value chain with many contract issues that need to be solved before a larger district heating scheme can be deployed. British thinking is based on competition and individual choices. A collective large scale solution, such as district heating, may conflict with principles and tradition. Another large barrier is the built environment. Few people live in multi-family houses in the UK and even fewer in Ireland [3], and even these buildings often lack central heating, but individual heating of apartments is common. Biomass

Customer relations are complicated because district heating is a rather unknown energy form and there is a certain resistance against collective solutions [3]. There is a lack of standardised terms of contract for connection to and delivery of district heating. Potential heat suppliers and customers feel insecure concerning how many users that will connect to a DH grid, for how long they will stay and if heat supply may be 225

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

interrupted. The competitiveness of district heating compared to gas concerning availability and price is

considered as a medium severe barrier in the British Isles (Table I).

Legend: Grade 4: Large barrier, Grade 3: , Grade 2: , Grade 1: Small barrier, Grade 0: Assessed not to be a barrier, No grade: No assessment.

100% 100% 90% 90% 80% 80% 70% 70% 60% 60% 50% 50% 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% UK UK District heating District heating

Ireland Ireland Gas

France France

Biomass BiomassPeat Gas

Czech CzechRepublic Republic

Romania Romania

PeatElectricity Electricity Oil

Oil Coal Coal

Fig. 2. Heating of residences [1], [4]â&#x20AC;&#x201C;[6]

FRANCE Table I shows that one of the largest barriers in France concerns the organisation of district-heating operations. Most DH systems are managed by private French companies according to long-term concessions [7]. The companies have successfully applied this DH management model in several other countries. By such arrangements, it is important that operators have incentives to make investments even if these have payback times longer than the concession period [8]. It is unclear if the French DH management model is disadvantageous for district heating development but it should anyway be a large barrier for foreign companies wanting to enter the French market. In general, domestic solutions are preferred. There is no strong actor who provides unbiased support for district heating. The dominating DH operators also sell electricity and gas, which both cover a large fraction of the heat demand (Fig. 2) and offer district heating severe competition. Only ten percent of the apartments and four percent of all residences have district heating today, and DH expansion is slow [6].

Renewables

Miscellaneous Natural gas CHP

Coal Oil

Natural gas heat

Fig. 3. District heating production in France [9]

Financing is considered to be a smaller problem in France. The market domination by a few actors may present an indirect financial barrier. Quite a few people live in apartments but most multi-family houses lack central heating. The large French nuclear power production is one reason for worse CHP conditions, which is assessed as a medium-grade barrier (Table I). THE CZECH REPUBLIC

Fig. 3 shows that one-half of the district heating in France is produced with natural gas, mostly in CHP plants. The main part of the renewable energy used for district heating production is waste, which is used to a slowly growing extent [7]. But French waste incineration plants are mostly built far away from towns, which makes it difficult to utilise the heat [6].

Fig. 2 shows that district heating covers a substantial part of residential heating in the Czech Republic, but electricity is used to the same extent and gas is the most common heat source. District heating covers onehalf of the apartments and 60% of urban heating [7]. A large barrier in the Czech Republic is, according to Table I, the competitiveness of district heating. 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 heating to compete with individual gas heating [7]. There are some disconnections from DH systems.

Miscellaneous

use. Many district heating users switched to gas due to low gas prices and heavy, government-regulated DH price increases [7], whereas households and district heating plants had the same gas price.

Natural gas Hard coal

Coal

Lignite

Natural gas Oil

Fig. 5. District heating production in Romania in 2005 [7]

Fig. 4. District heating production in the Czech Republic [10]

Fig. 5 shows that Romanian district heating production is completely based on fossil fuels. One-half of the heat is produced in, normally coal-fired, CHP plants. Large investments are required in the Romanian district heating systems. CHP plants and heat-only boilers must be replaced for environmental reasons. Distribution losses are enormous [7].

Domestic coal dominates Czech district heating production (Fig. 4). Most of the district heating is produced in CHP plants. The problem concerning CHP (Table I) is that a large share of Czech district heating comes from coal-fired power plants with extraction turbines where the heat is produced at the expense of electricity [7], which is considered more valuable. The benefit of this CHP production is not allocated to the heat [8]. Some biomass is used to produce district heating, but biomass use is complicated due to deficient fuel supply systems [7] and government scepticism toward renewable energy. There is also much resistance to waste incineration from the public as well as from politicians.

Organisation is a rather large obstacle for district heating in Romania (Table I). The municipalities are now mostly in charge of the district heating systems [7] but much lobbying is required to achieve improvements and it takes time to reach an investment decision. Corruption is common. Some politicians and employees try to make their own profit on DH business. Financing difficulties largely concern insecurity whether customers will remain because many have disconnected from district heating. National and local control is a certain barrier because DH companies partly get heat production costs covered by central Government and City Councils [7]. Besides the mentioned problems, the entrance barrier for foreign companies should be rather low. Waste collection and sorting are now deficient but, on the other hand, new possibilities should emerge when Romania wants to introduce waste incineration, and waste is therefore considered to be a medium-size barrier in Table I.

Financing may be a rather large barrier, partly due to a certain district heating disconnection tendency. The many private foreign district-heating companies in the Czech Republic [7] may be a difficult target for Swedish and other district heating companies from abroad that are not established in the country. There may also be some reluctance toward foreign enterprises. A certain barrier is the common corruption by public procurement (Table I). The value chain is sometimes fragmented into production and distribution run by different actors. ROMANIA In Romania, biomass covers the largest fraction of residential heat demand among the countries under study (Fig. 2). Individual boilers and stoves for wood and gas cover more than one-half of the heat use in households. Gas is the most widely used heating source for residences and it is expanding at the expense of district heating [7].

HOW TO OVERCOME BARRIERS

Table I shows that district heating has large problems with competitiveness and customer relations. Today, the DH consumption is just one-half of the previous

To have a chance to overcome the outlined barriers to any significant extent, powerful initiatives are required from countries with established district-heating

This paper focuses barriers and omits more positive circumstances for district heating. It may be depressing but the message is not that district heating has no prospects. The report should rather be understood as a realistic guide to DH development in the studied countries.

227

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

industries, such as Sweden. Initiatives should comprise many different players, for example, district-heating companies, equipment manufacturers, consultants and governmental bodies. Such consortia could offer district-heating systems from fuel supply, via heat production plants and DH networks to customer contracts. Now, many foreign groups visit municipal district-heating systems in Sweden but these opportunities are seldom utilised to sell a comprehensive DH solution.

ACKNOWLEDGEMENT The Swedish District Heating Association and The Swedish Energy Agency are gratefully acknowledged for financing this study through the Fjärrsyn programme. We would also like to thank everybody who has contributed to the study with facts and viewpoints. REFERENCES [1] D. Henning and O. Mårdsjö, Fjärrvärme i Europa: Hinder att övervinna för svensk export, Rapport 2009:3, Fjärrsyn, Svensk Fjärrvärme, Stockholm (2009) http://www.svenskfjarrvarme.se/index.php3?use=biblo& cmd=detailed&id=1440

Municipally owned district heating companies have system knowledge that can be applicable in other countries. A competence transfer may be realised through deeper involvement that might include ownership of plants in other countries. Business models should be developed, which allow utilisation of municipal knowledge abroad and give municipalities reasonable returns.

[2] S. Werner, Ecoheatcool work package 4: Possibilities with more district heating in Europe, Euroheat, Brussels (2006) www.euroheat.org/ecoheatcool

For a successful transfer of district-heating solutions from established to emerging markets, private and public companies must focus marketing on the countries, places, projects and forms of involvement that have the greatest expectations to succeed. At the same time, national and local policies should reduce and remove described barriers and facilitate district heating development as a means for increased efficiency of energy utilisation, higher security of supply and decreased environmental impact.

[3] WS Atkins Consultants Ltd, Assessment of the Barriers and Opportunities Facing the Deployment of District Heating in Ireland, Sustainable Energy Ireland, Dublin (2002) www.sei.ie/uploadedfiles/InfoCentre/DistrictHeatingRep ortatk.pdf [4] S. Werner, Ecoheatcool work package 1: The European heat market, Euroheat, Brussels (2006) www.euroheat.org/ecoheatcool [5] SEI, Energy in Ireland: Key Statistics, Sustainable Energy Ireland, Dublin (2008) www.sei.ie/Publications/Statistics_Publications/EPSSU _Publications/Energy_in_Ireland_Key_Statistics/Energ y_in_Ireland_Key_Statistics_Final.pdf

CONCLUSION There are several barriers to district heating development in the countries under study. In the UK, there are not many district heating systems. There are few multi-family buildings with central heating in Ireland. The long-term operating concessions of French district heating systems might hamper their development. In the Czech Republic, much district heating is produced in extraction turbines at the expense of more valuable electricity. Romanian district heating use was reduced by one-half by cheap gas.

[6] P. Cousinat, District Heating: A Tool for Rational Heat Management, Master thesis 2006:21, Department of Civil and Environmental Engineering, Chalmers, Gothenburg (2006). [7] Euroheat, District heating and Cooling country by country 2007 survey, Euroheat, Brussels (2007). [8] J. Zeman and S. Werner, District Heating System Ownership Guide, DHCAN project, BRE, Watford (2004) http://projects.bre.co.uk/DHCAN/guides.html

In general, it should be advantageous that one actor controls the whole district-heating value chain from source to consumption in order to utilise synergies and to avoid economic risks with contracts between the separate entities of a fragmented value chain. Like for other long-term large-scale infrastructure investments, public involvement may be necessary for district heating development.

[9] SNCU, Les réseaux de chaleur et de froid: l‘énergie citoyenne, SNCU, Paris (2004). www.fg3e.fr/public/federation/syndicats/plaquettes.php ?root_page=6 [10] T. Zenaty, CHP/DH sector in the Czech Republic: situation / problems / wishes, Energy Policy EHP meeting, Budapest, 11 September 2008, www.lsta.lt/files/seminarai/080911_Budapestas /CZ.pdf

Through cooperation among various well-established players in the district heating industry, knowledge, products and services can be transferred to evolving district heating markets, which promotes industrial prosperity for all parties and helps building sustainable energy systems in Europe. 228

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

IMPACT OF THE PRICE OF CO2 CERTIFICATES ON CHP AND DISTRICT HEAT IN THE EU27 Markus Blesl 1

Institute of Energy Economics and the Rational Use of Energy (IER) University of Stuttgart modelling language of GAMS due to reasons of being better transferable. TIMES is a multi-periodic linear optimization model based on a technical approach at which single plants are aggregated. The purpose is the evaluation of the economically optimal energy supply structure at a given need of end use energy and energy services and also at given energy and climate policy requirements. For this, the discounted system costs are minimized, whereas the single players (industry, supply, households) could have different economic considerations. The main objective of the model development of TIMES is the flexible structure to ensure a simple mathematic adjustment to the respective problem.

INTRODUCTION In the current energy and climate policy debate, one of the key points is the discussion about emission reduction targets and how they are spread among different world regions or countries and also among different sectors. To find a cost optimal burden sharing of an emission reduction target, the different reduction potentials of the particular sectors or technologies have to be known. To reach a reduction target, emission certificates in a country or region (like EU-27) are allocated among the different sectors or between different types of heat and power generation technologies. This allocation (for example, auctioning) of emission certificates is an important issue to negotiate since the costs of buying certificates could be an important factor in technology choices for investment.

The pan European TIMES energy system model (abbreviated as TIMES PanEU) is a model of 30 regions which contains all the countries of EU-27 as well as Switzerland, Norway and Iceland. The objective function of the model is a minimization of the total discounted system costs over the time horizon from 2000 to 2050. A perfect competition among different technologies and paths of energy conversion is assumed in the model. The TIMES PanEU model covers on a country level all sectors connected to energy supply and demand such as the supply of resources, the public and industrial generation of electricity and heat and the industrial, commercial, household and transport sectors. Both greenhouse gas emissions (CO2, CH4, N2O) and pollutant emissions (CO, NOx, SO2, NMVOC, PM10, PM2.5) are covered by TIMES PanEU.

The significant advantage of this approach is that the analysis of the different competing pathways to achieve emission reductions also assesses how they influence each other. In the context of efficiency improvement in industrial CHP and district heating and cooling, the use of waste heat becomes an interest field. Efficiency improvements in the residential or commercial sector is examined in the topic of energy saving. Without analysing the entire energy system the possible advantages of CHP and district heating and cooling couldn‘t be taken into account. This shows the difference to a standard cost potential curve approach, which has a fixed order of measures depending on their avoidance cost.

The transport sector is disaggregated into four areas: road transport, rail traffic, inland shipping and.aviation. The road traffic includes five demand categories for passenger transportation (car short distance, car long distance, bus, coach, motor bikes) and one for freight service (trucks). The rail traffic includes three categories: rail passenger transportation (divided into short and long distance) and rail freight transportation. The transport modes of inland shipping and aviation are represented by a non-specified general process where the development of the transport demand is embodied by the final energy demand.

This analysis will evaluate the reduction potential of CHP plants or in general the production of district heating and cooling in the EU-27 using the energy system model, TIMES PanEU /Blesl et al 2008; Blesl 2008; Blesl et al 2008b, Kuder Blesl 2009; Blesl 2009/. TIMES PAN-EU MODEL The energy system model, TIMES (The Integrated Markal Efom System), is a further development of the two model generators, MARKAL and EFOM-ENV, written in GAMS. TIMES was developed in recent years within the „Energy Technology Systems Analysis Programme―(ETSAP) from the IEA with contribution from the IER. It is classified in one category with the models MARKAL, EFOM or MESSAGE. The model generator, TIMES, was developed in the general

The household sector contains eleven demand categories (space heating, cooling, hot water, cooking, refrigeration, lighting, washing machines, laundry dryer, dishwasher, other electrics, other energy use), whereof 229

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 family houses in urban and rural areas and multi-family houses each described as existing stock and new build). The commercial sector is represented by a similar reference energy system (RES) and consists of nine demand categories (space heating, cooling, hot water, cooking, refrigeration, lighting, public street lighting, other electrics, other energy use). The first three of them are subdivided according to different building types (large/small).

Industrial heat demand by temperature and subsector in the EU27 The particular sub-sectors of the industrial sector use different chemical and physical conversion processes. Therefore, they need heat on different temperature levels (Figure 1). Processes with a need for very high temperatures (> 1400 Â°C) are e.g. blast furnaces (iron/steel industry) or kilns (cement or lime industry). Processes with lower temperature levels occur in the food/tobacco (sugar production, dairy) industry, other industries or in general for the supply of space heating and hot water. Also, the pulp/paper industry has a high need for heat at a lower temperature level (< 100 Â°C).

The agricultural sector is described by a general process with a mix of several energy carriers as input and an aggregated demand of end use energy as output.

Most of the heat is produced by the combustion of fuels. Other heat is generated by the use of electricity. Key processes using electricity for high temperature heat are chlorine electrolysis, aluminium electrolysis, electric arc processes (iron/steel) and copper electrolysis.

The industrial sector is subdivided into several branches (for example, iron and steel, cement, limeâ&#x20AC;Ś) and into energy intensive and non-intensive branches. While the intensive ones are modelled by a process orientated approach, the other industries have a similar structure but with five energy services (process heat, steam, machinery drive, electrochemical, others)..

Others

2,500

Food/Tabacco Pulp/Paper

Final energy consumption [PJ]

The generation of electricity and heat in power plants, CHPs and heating plants is differentiated into public and industrial production. The model contains three different voltage levels of electricity (high voltage, medium voltage, low voltage) and two independent heat grids (district heat, local heat).

2,000 Oth. non-metallic minerals Glass flat

1,500

Glass hollow Lime Cement

1,000

Oth. chemicals Chlorine

500 Ammonia

In the supply sector, all primary energy resources (crude oil, natural gas, hard coal, lignite) are modelled by supply curves with several cost steps. Three categories can be differentiated: discovered reserves (or developed sources), growth of reserves (or secondary and tertiary extraction) and new discoveries. Additionally, seven bio energy carriers are differentiated: matured forest, bio gas, household waste, industrial waste, as well as energy plants containing sugary, starchy and lignocelluloses..

Oth. non-ferrous metals

> 1500

1400-1500

1300-1400

1200-1300

1100-1200

1000-1100

800-900

900-1000

700-800

600-700

540-600

480-540

420-480

360-420

300-360

240-300

180-240

120-180

100-120

0-60

60-100

Copper Aluminium Iron/Steel

Figure 1: Final energy consumption for industrial heat production by temperature and sub-sector in the EU-27 in 2005

On the country level, the role of the different member states concerning a particular temperature level depends on the structure of the industrial sector in that country. In general, the final energy consumption for heat production at a specific temperature level is dominated by the bigger member states and members of the EU-15 like Germany, Italy, UK, France and Spain. However, new member states like Poland, Czech Republic or Romania also play an significant role. Some countries only play a key role at single subsectors and thus only for some temperature levels.

Due to its regional resolution, TIMES PanEU allows the consideration of country specific features, for example different structures of the stock of power plants, different extension potentials for renewables as well as potentials for storing CO2. An interregional electricity trade is implemented in the model, so that exports and imports of electricity according to the existing border capacities could be calculated endogenously in the model.

The lower temperature levels are dominated by the industrial sub-sectors pulp/paper, food/tobacco and others. Due to high activities in those areas, the heating demand is clearly influenced by France (strong for food/tobacco), Sweden and Finland (strong for pulp/paper) next to other big countries like Germany, Italy and UK. Italy and Spain play a large role, especially at very high temperatures, due to their high amount of cement production. In the Netherlands, the

The role of CHP and district heating will be influenced in the future by the heating demand for the heat, space heating and cooling processes. The following chapters describe the status and the assumed development for Europe.

230

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

occuring at this price level are analysed according to the role of the different reduction possibilities.

chemical and food & tobacco industries are the most important ones. Each country is clearly specialised in differing industrial sub-sectors.

The foundation for the CO2 price variation is set based on the CO2 price outcomes from two scenario runs with a reduction target of 15% [scenario: 15% reduction (2020)] and 40% [scenario: 40% reduction (2020)] in 2020 compared to the Kyoto base year (Table 1). In the long run (2050), both of these restricting scenarios have the same target which equals a 450ppm goal (-71% in 2050 compared to 1990).

Space heating and cooling demand in Europe today and in future The demand for space heating and cooling differs among the countries in Europe due to the differences in climatic conditions and in living standards (e.g. square meters per capita) and building standards. This is especially applicable to the assessment of current and near future energy demand.

2020

14000

Heat demand in [PJ]

2030

-45% -20%

2035

-50% -25%

2040

-55% -37%

2045

-61% -48%

-66% -60%

2050

-71% -71%

Table 1: CO2 reduction pathways for the two restricting scenarios

12000 10000

The resulting CO2 prices of these two restriction scenarios build the framework for the price variations. Within the range of the resulting CO2 prices, the carbon price varies between 10 €/tCO2 and 110 €/tCO2 in 2020 in increments of 10 €. In 2030, the price varies between 27 €/ tCO2 and 123 €/ tCO2. The price increases until it reaches the level of a 450 ppm scenario in 2050 (Figure 3). The emission reductions are evaluated using the results from the different scenarios in comparison to the case of the lowest CO2 prices (10 €/t in 2020, 27 €/t in 2030). First, the total reductions over all sectors are presented and afterwards the focus will be on the industrial sector. The drivers of the reduction are shown separately. Looking at the industrial sector, the reasons for the emission reductions could be split up into more efficient production processes, more efficient heat supply, fuel switch in heat generating units or CCS technologies in production processes and energy supply.

8000 6000 4000 2000 0 2005

2010

2015

Commercial

2020

2025

Urban MFH

2030

Urban SFH

2035

Rual

2040

2045

2050

New Houses

1200 Residential SFH

1000 Cooiling demand in PJ

2025

-40% -15%

40% reduction (2020) 15% reduction (2020)

Residential MFH Commercial

800 600 400 200 0 2000 2005 2010 2015 2020 2025 2030 2040 2050 year

450

Figure 2: Demand for space heating/hot water and for cooling in the EU27

15% reduction (2020) 40% reduction (2020) CO2_10

400

350 CO2_20

In the 2000, the useful demand for cooling was less than 5% lower than the useful demand for space heating and hot water. In the long term, the cooling demand will be dominated by the commercial sector. The increase of cooling demand in the EU27 up to 2050 will reach approx. 1120 PJ in the residential and commercial sectors.

CO 2 price [€ 2000 ]

300

CO2_30 CO2_40

250

CO2_50 200 CO2_60 150

CO2_70 CO2_80

100

CO2_90 50

CO2_100 CO2_110

0 2020

2025

2030

2035

2040

2045

2050

Figure 3: CO2 prices of the different scenarios

SCENARIO DEFINITION

OVERVIEW OF THE DYNAMIC DEVELOPMENT OF THE ENERGY SYSTEM OVER TIME

A parameter variation is used to evaluate the reduction potential and the role of CHP and district heat in the energy system of the EU27. By varying the CO2 price, the reduction potential curves are constructed. Therefore, different scenarios with different CO2 prices (one common price for ETS and Non-ETS sectors) are calculated with TIMES PanEU and the reductions

In the following analysis, the two scenarios with the lowest (10 €/t CO2 in 2020, scenario CO2_010) and the highest (110 €/t CO2 in 2020, scenario CO2_110) prices are displayed to show the range in which the results of the price variation occur. Therefore, the development over the whole modelling horizon (2000–2050) is presented to rank the more detailed 231

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The electricity generation from CHP plants in the EU27 increases by 79% from about 380 TWh in the year 2000 to 640 TWh by the year 2020 (see Figure 4). The extension of the electricity generation from CHP plants is essentially supported by gas-fired and biomass based CHP plants. Additionally, existing public CHP plants with an extraction condensing turbine are substituted by CHP plants with a higher power-to-heat ratio and there is also an extension of industrial CHP plants, which are often used in cooperation with communal facilities. The intermediate growth of CHP plants in the commercial sector between the years 2015 and 2035 are based on efficiency advantages of CHP plants with a medium sized internal combustion gas engine. In the long term, the limited possibilities of using CO2 free fuels in commercial CHPs will result in these phasing out in the commercial sector. Until the year 2050 the electricity production by CHP plants in the scenarios further increases up to a level of 1055 to 1100 TWh. CHP plants based on biomass as well as CCS CHP are an important option in the year 2050.

results from a point of time within these more general results over a period of time. Since the CO2 reduction target of the two bounding scenarios [scenario 15% reduction (2020) and scenario 40% reduction (2020)] clearly differ in the mid-term periods of 2020 and 2030 (Table 1) and the corresponding prices are more different in these periods (Figure 3), the energy system shows the most variations during this time. To show the development over the modelled time period, first of all the net electricity generation of EU-27 is displayed (Figure 4). The overall electricity generation remains almost constant at 2010 levels (about 3 200 TWh) until 2030. In later periods, there is a clear increase in electricity generation up to 4 255 TWh (2050, scenario CO2_110). The increase in the later periods is driven by stronger emission reduction targets. To fulfil the restrictions, more electricity with low specific emissions and high end use efficiency in the demand sectors is used. According to the given CO2 prices of the two scenarios (CO2_010 and CO2_110), the main differences occur in the mid term periods. While the total electricity demand in 2020 is lower in the scenario with higher emission certificate prices (-22 TWh in 2020 between CO2_110 and CO2_010), the demand is higher by 86 TWh in 2030. The increase is due to the use of more efficient technologies in the end use sectors resulting in lower electricity demand in 2020, while by 2030 the switch to electricity based technologies to fulfil the emission restrictions has already taken place.

Net electricity generation [TWh]

2000

CO2_010 CO2_110

Statistic

2020

2025

2030

2035

2040

2045

2050

CO2_010

2000 2010 2015

2010

2020

2025

2030

2035

2040

CO2_010 CO2_110

CO2_010 CO2_110 2015

2045 2050

Net electricity generation CHP in [TWh]

In addition to the net electricity generation, the primary (Figure 6) and final energy (Figure 7) consumption of the EU-27 are also analysed over the whole time period. Overall, the primary energy consumption (PEC) does not show clear changes and remains at a level of about 75 000 PJ. The lowest total PEC occurs in the mid-term periods. The total consumption is influenced by an increasing efficiency till 2030 and later on by a higher share of renewables and also CCS which both lead to a higher consumption due to the lower thermal efficiency in the combustion processes.

Oil

2005.0

Figure 5: Net electricity generation CHP by sector in the EU-27

Natural gas

500

200.00

2000

Nuclear

1000

400.00

0.00

Hydro

1500

Public Comercial Industry

600.00

CO2_010 CO2_110

Wind

3500

800.00

CO2_010 CO2_110

2500

4000

1000.00

CO2_010 CO2_110

3000

Others / Waste non-ren. Other Renewables Biomass / Waste ren. Solar

4500

1200.00

Lignite Coal

Figure 4: Net electricity generation in the EU-27

Aside from the changes in the total electricity demand, there is also a change in the structure of the electricity generation. At higher CO2 prices, less coal (-120 TWh from coal fired power plants in 2030) and more gas (+44 TWh) and nuclear (+30 TWh) are used and more electricity from renewable energy sources (+35 TWh from wind, +56 TWh from biomass and renewable waste) is generated. Furthermore, CCS is used more widely under the conditions of the CO2_110 scenario in 2030 compared to CO2_010.

Looking at the impact of the single energy carriers, there is a distinct change between the two scenarios than in the total sum of the PEC. In 2030 at a higher CO2 price, less coal (-1 675 PJ) and petroleum products (-881 PJ) and more Hydro, wind, solar (+338 PJ) and other renewables +4856 PJ) (mainly biomass) are used.

232

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia 60000

Others (Methanol, Hydrogen)

50000

70000

Waste (non renewable)

Waste

60000

Other renewables Hydro, wind, solar Nuclear

30000 Natural gas

20000 10000

Oil

Electricity

10000

Gas

CO2_010 CO2_100

Coal

Figure 7: Final energy consumption in the EU-27

Figure 6: Primary energy consumption in the EU-27

In contrast to the year 2000, the distribution of local and district heat to the household, commercial and industrial sectors changes by the year 2050 with an additional approx. 1000 PJ district consumed in the year 2050 (see Figure 8).

The final energy consumption (FEC) shows comparable results (Figure 7). The use of petroleum products declines over time in both scenarios (-9 052 PJ in scenario CO2_010 between 2000 and 2050). The use of gas increases at lower CO2 prices in the mid-term periods (up to more than 13 500 PJ in 2020 at scenario CO2_010), but declines in both scenarios at the very end. This shows that one early and cost-effective measure for emission reduction is the fossil fuel switch from petroleum products and coal to gas in the end use sectors.

Final energy consumption heat in [PJ]

3500

As already shown with electricity generation, the use of electricity also increases in the end use sectors. Especially in the long run at higher carbon prices, there is a clear rise. The use of renewable energy sources also increases constantly in both scenarios. In 2020 and 2030, clearly more renewables are used in the CO2_110 scenario due to the higher CO2 prices (+3900 PJ in 2030).

Supply Residential Industry Comercial Agriculture

3000 2500 2000 1500 1000 500

2000

2005

2010

2015

2020

2025

2030

2035

CO2_010 CO2_100

2040

2045

2050

Figure 8: Final energy consumption district heat in the EU27

In contrast to the PEC, the total FEC decreases slightly in the long run. The reason for this different development is that the higher conversion losses arising from a higher electricity demand and the extended use of renewables and CCS at the public electricity generation are balanced at PEC and do not influence the FEC.

In the long term, the CO2 contents of the heat supply for the end use sectors will be reduced from 130 kg CO2/MWh to 122 kg CO2/MWh in 2020 and from 113 kg CO2/MWh to 36 kg CO2/MWh in the year 2050, which is one explanation for achieving the CO2 reduction targets in this area. On the other hand, the possibility to use renewable energy or to install CCS, increasingly influences the penetration of CHP. By 2050, fossil heat plants will also be substituted with large heat pumps and solar thermal heat plants in combination with storages, biomass heat plants fuelled with wood or woody crops and biogas.

Even though more renewables (mainly biomass) are used, due to the higher use of electricity with its high end use efficiency and other efficiency improvements, the total FEC declines to 49 482 PJ (in 2050 at scenario CO2_110). This efficiency improvement occurs in the industrial sector mainly at industrial production processes, but is also clearly driven by efficiency improvements ain the residential and transport sectors.9

For a detailed discussion of the effects in the different end use sectors and its impact on the total final energy consumption see /Blesl et al. (2010)/

Petroleum products

2000 2010 2015 2020 2025 2030 2035 2040 2045 2050

CO2_010 CO2_100

CO2_010 CO2_110

Coal

CO2_010 CO2_100

CO2_010 CO2_110

2050

CO2_010 CO2_100

CO2_010 CO2_110

2045

CO2_010 CO2_100

CO2_010 CO2_110

2040

CO2_010 CO2_100

CO2_010 CO2_110

2035

CO2_010 CO2_100

CO2_010 CO2_110

2030

CO2_010 CO2_100

CO2_010 CO2_110

2025

Statistic

CO2_010 CO2_110

Statistic

2020

CO2_010

2000 2010 2015

Lignite

CO2_010

Heat

20000

CO2_010 CO2_100

40000

Renewables

30000

CO2_010 CO2_100

50000

40000

CO2_010 CO2_100

Primary energy consumption [PJ]

80000

Total final energy consumption [PJ]

Electricity import

90000

The overall emissions decrease is based on the emission reductions of the single sectors leading to different CO2 abatement costs (Figure 9). The total emissions correspond to the emission pathway of the two restricting scenarios (scenario ―15% reduction (2020)‖ and scenario ―40% reduction (2020)‖, see Table 1. The earliest and strongest reductions take place in the conversion/production sector. The industrial sector and the residential/commercial sector also show clear reductions. The transport sector tends 233

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

only to reduce its emissions with very strict reduction targets connected to high carbon prices.

900 Transport

Additional emission reduction [Mt]

800

4500

Emissions of CO 2 [Mt]

4000

Transport

3500 3000

2000

Households, commercial, AGR

1500

Industry

2500

Commercial

700

Residential

600

Industry

500

Conversion/ Production

400 300 200 100

1000

0 36

500

CO2_010 CO2_110

2000 2010 2015

2020

2025

2030

2035

2040

2045

2050

CO2_010

CO2_010 CO2_110

Statistic

104

114

123

Carbon price [€2000/tCO2]

Conversion, production

Figure 10: Additional CO2 reduction in the EU-27 in 2030 by sector compared to the scenario with the lowest CO2 price of 27 €/t

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 final energy demand from district heat (see figure 11). The overall increase of the district heat demand influenced by the different CO2 prices is 14%. The biggest growth can be seen in the commercial sector, where the total district heat demand for district heating grows by over 30% between the min the minimum and maximum CO2 certificate price.

ANALYSIS AT A SPECIFIC POINT OF TIME WITH FOCUS ON 2030 After the general effects are described and the scenarios with the lowest and highest CO2 prices are analysed over the whole period of time, a more detailed analysis shows the effects in the industrial sector during the mid-term periods with a particular focus on 2030.

2000

Firstly, the reduction potential of the different sectors should be analysed (Figure 10). Both conversion /production and the other end use sectors are taken into account. As in the results of the emission reduction from 2000 to 2050 (Figure 9), the industrial reduction potential plays the key role next to the conversion/production sector. Looking at the year 2030 and comparing the additional CO2 reductions when the CO2 price is increased from 27 €/t to 123 €/t, the strongest additional reduction occurs at the conversion sector (+351 Mt at a price of 123 €/t compared to 27 €/t). An additional 301 Mt of CO2 are reduced by the industrial sector.

Final energy demand heat in [PJ]

1800 1600 1400 Residential

1200

Industry Comercial

1000

Agriculture

800 600 400 200 0 27

104

114

123

Carbon price [€/T CO2]

Figure 11: Final energy demand of district heat in the EU-27 in 2030 by sector

Especially at higher prices above 94 €/t, the reduction potential of the industrial sector becomes more and more important. Its share of the total additional reduction increases from 33% (36 €/t compared to 27 €/t) to 37% (123 €/t to 27 €/t). The lowest reduction occurs in the transport sector. Till a price of 85 €/t, only an additional 6.3 Mt are reduced, while at a price of 123 €/t an additional 18.9 Mt are reduced. In the residential and commercial sector, some reduction possibilities are cost-effective even without a price on CO2. The energy savings outweigh the additional investment costs. Those reduction measures are especially connected to the building/heating sector.

However, the generation of district heat from the use of renewable sources and CCS will be one reason for the growth of the reduction potential in the conversion sector (Figure 12). The share of the use of renewables, especially biomass, will rise from 29% to 60%. More than 1300 PJ of additional biomass will be needed. Due to this increase, the average heat to power ratio of all CHPs will fall from 0.9 to 0.66. In the cases where CO2 prices exceed 56 €/tCO2, the district heat generation in CCS CHP plants grow more rapidly. The specific emissions of the district heat generation decrease from approx. 380 kg / MWh to 84 kg / MWh.

234

4200

1800

4180

1600

4160

30% SUM 25%

20%

1400 1200

Heat Plant RES Heat Plant

1000

CHP RES CHP FC

800

CHP CCS

4140 15%

4120 4100

10%

4080

CHP

600

4060 0%

400 4040

200

Change of Electricity supply by technology compared to CO 2 price of 27 €/t

2000

Use of Electricity [PJ]

District heat generation in [PJ]

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

public generation Condensing industrial CHP industrial

-5%

4020

0 27

104

114

4000

123

-10% 27

carbon price in [€/t CO2]

104

114

123

Carbon price [€2000/tCO2]

Figure 12: District heat generation in the EU-27 in 2030 by technology group

Figure 13: Use of electricity in the EU-27 in 2030 by technology

In the industrial sector, the share of CHP will grow. The additional emission reductions by the industrial sector of 301 Mt in the year 2030 could be split into industrial supply and industrial production processes. The supply side covers the industrial generation of energy commodities or energy services. These are electricity from industrial condensing power plants and CHPs, heat and steam from CHPs and boilers, space heating and heat for hot water as well as cooling. The supply activities play an important role in the industrial subsectors with a high share of space heating (such as food & tobacco or other industries) or low temperature process heat (such as pulp & paper or food & tobacco).

The other part of the industrial supply processes is the industrial heat generation. The drivers for the emission reduction in industrial heat production are a switch to biomass (from coal and clearly from gas) and the use of CCS in industrial CHPs (Figure 14). Between a CO2 price of 36 and 56 €/t of CO2 in 2030, there is a clear increase in the use of renewables in boilers. The share of renewables in the total fuel use in industrial boilers increases from 33% to 51%. As a result, the thermal efficiency of boilers has an overall decrease. In industrial CHPs, there is also a slight increase in the use of s. This switch takes place between CO2 prices of 27 €/t to 65 €/t. However, the main change concerning CHPs is the increasing use of CCS. At a CO2 price above 94 €/t, there is a clear rise in the use of this technology. These CCS CHPs are mainly gas fired10. This is why the share of renewables used in industrial CHPs declines at a price over 75 €/t again.

In total, from the additional reduced emissions, 147 Mt are reduced by industrial supply processes and 154 Mt by production processes in 2030. While at lower a CO2 price more emissions are reduced on the supply side (66% of the additional reduction based on supply processes at 46 €/t), at higher prices more and more reductions take place on the production side (49 % based on supply processes at 123 €/t).

Like biomass, the extended CCS use also leads to lower efficiencies resulting in both the efficiency of boilers and CHPs to decline over time. Accordingly, the key driver is not efficiency improvements, but the use of renewables and CCS. The effects of renewables and CCS compensate the trend to lower energy intensity within one technology. Gas boilers become more efficient and as do biomass boilers. However, the more efficient biomass boilers still use more fuel than the gas boilers.

The additional electricity needed at high CO2 prices is mainly generated by industrial autoproducers. Within this industrial production, the additional electricity mainly comes from CHP power plants. The use of electricity in the industrial sector from public generation remains relatively constant even when the CO2 price increases. Accordingly, one key way to reduce the emissions on the supply side is through the extended use of CHP plants for industrial power generation. This higher amount of electricity from industrial autoproducers (Figure 13) leads to higher conversion losses in total when the fuel use is considered. As described above, that is one reason for the difference between final energy consumption and fuel consumption. Another reason is the lower efficiency of electricity generation due to the higher use of CCS.

Looking at the heat output by technology, there is also a shift (Figure 14). At lower emission prices, the heat output from industrial boilers stays almost constant. Within this range, the share of renewables used increases (as illustrated in Figure 13). Afterwards, at a price above 65 €/t, boilers are substituted with heat from CHPs and district heat. Both heat commodities

235

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)/

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

that the climate conditions within Europe differ substantially.

are generated in combination with an increasing share of renewables, a higher CO2 price and from CCS. 90% CHP industrial η (total)

80%

CHP industrial share RES

70%

Efficiency and share [%]

Boiler industrial η

60% Boiler industrial share RES

50% CCS CHP industrial (share Fuel input)

40% 30% 20% 10% 0% 27

104

114

123

Carbon price [€2000/tCO2]

Figure 14: Efficiency of heat supply technologies and share of CCS at industrial CHP in the EU-27 in 2030 150 District Heat CHP industrial

change in heat output [PJ]

100

Within the energy system of the EU-27, there are different emission reduction pathways. The emissions could be reduced by a fuel switch in more efficient (or better, less carbon intensive) energy supply or by a change in production processes. Key drivers concerning the emission reduction in production processes and in heat demand side are efficiency improvements due to new technologies and technological improvements. The key driver concerning the supply side of electricity and heat generation is the increased use of renewables, mainly biomass, for heat generation. The CCS technology also plays an important role in the reduction of emissions. Due to the increased use of renewables in CHP and heat plants and the use of CCS, the efficiency in the supply processes decreases at higher CO2 prices. In the long run to a CO2-free world, the possibility to generate district heat with renewable energy and the use of CCS make the decarbonisation of the energy consumption in the end use sectors possible.

Boiler 50

In general, the progression of district heat depends crucially on the possibility of generating CO2 emission free district heat and electricity.

-50

-100

-150 36

104

114

123

REFERENCES

Carbon price [€2000/tCO2]

Figure 15: Heat supply by technology in the industrial sector in the EU-27 in 2030 compared to the scenario with the lowest CO2 price of 27 €/t

[1] Blesl, M.; Kober, T.; Bruchof, D.; Kuder, R.: Effects of climate and energy policy related measures and targets on the future structure of the European energy system in 2020 and beyond, Energy Policy, 2010 (forthcoming)

In total, all these described effects concerning the industrial supply processes lead to the additional emission reduction in 2030 of 147 Mt at a price of between 123 €/t and 27 €/t. In general, more emissions are reduced in boilers than in CHPs. The reasons are the fuel switch from coal and mainly gas to renewables at lower CO2 prices and later on the substitution of boilers with CHPs (less boilers are used and therewith produce less emissions).

[2] Blesl, M.; Kober, T.; Bruchof, D.; Kuder, R.: Beitrag von technologischen und strukturellen Veränderungen im Energiesystem der EU 27 zur Erreichung ambitionierter Klimaschutzziele, Zeitschrift für Energiewirtschaft 04/2008 [3] Blesl, M.: CHP and district heat in the Europe under an emission reduction regime, in: 11th International Symposium on District Heating and Cooling in Reykjavik, Island

Due to a higher use of CHPs, there is no clear increase in emissions during the mid-term ranges. When the output of heat stays constant and a higher share of CCS is used, then clear emission reductions from CHPs (additional 48.7 Mt in 2030 at 123 €/t compared to 27 €/t) occur.

[4] Blesl, M., Cosmi, C. ,Kypreos, S. , Salvia, M.: Technical paper n° Technical Report n° T3.18 – RS 2a ―Summary report of Pan European model results – BAU scenario‖ EU Integrated Project NEEDS ―New Energy Externalities Developments for Sustainability‖ October, 2008

CONCLUSION AND OUTLOOK District heating generation offers an economic potential for expansion in the future. Depending on the regions or countries, the development will be different because the starting point is economic growth and the existing national laws or cross-subsidies for competitor‘s energy carriers. In addition, it is necessary to take into account

[5] DEHSt (2010): Deutsche Emissionshandelsstelle, Kohlendioxidemissionen der emissionshandelspflichtigen Anlagen im Jahr 2009 in Deutschland, Mai 2010

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, European Union emission trading scheme (ETS) data viewer, 2010

Probabilistic Scenariosâ&#x20AC;&#x2013;, project.net/planets

2010,

www.feem-

[9] Kuder, Blesl (2009): Kuder, R.; Blesl, M.: Effects of a white certificate trading scheme on the energy system of the EU-27, Fullpaper 10th IAEE European Conference in Vienna, Austria, 2009

[7] Kober, Blesl (2010a): Analysis of potentials and costs of storage of CO2 in the Utsira aquifer in the North Sea; report work package 4: Regional analysis at North Sea level, 2010, www.fencoera.net

[10] UNFCCC (2009): GHG inventory reports for the single member states of the EU-27, submission 2009 situation / problems / wishes, Energy Policy EHP meeting, Budapest, 11 September 2008,

[8] Kober, Blesl (2010b): Perspectives of CCS in Europe considering technical and economic power plant uncertainties; in PLANETS work package 6 deliverable No. 15 â&#x20AC;&#x2022;Report on

www.lsta.lt/files/seminarai/080911_Budapestas/CZ.pdf

237

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

CONSIDERATIONS AND CALCULATIONS ON SYSTEM EFFICIENCIES OF HEATING SYSTEMS IN BUILDINGS CONNECTED TO DISTRICT HEATING 1

Maria Justo Alonso , Rolf Ulseth and Jacob Stang 1

SINTEF Energy Research, Department of Energy Processes 2 NTNU, Faculty of Engineering Science and Technology, Department of Energy and Process Engineering of EN-standards. The main goal of the Directive is to promote the improvement of the energy performance of buildings within the Community, taken into account outdoor climatic and local conditions as well as indoor climate requirements and cost-effectiveness. The main focus is on reducing the primary energy use and the associated CO2 emission of buildings.

ABSTRACT In order to harmonize the implementation of the EC Directive on the energy performance of buildings (EPBD) [1], and to provide guidelines and common calculation tools, several technical standards have been worked out by CEN in accordance with a mandate from the EC. This paper focuses on calculating system efficiencies of hydronic heating systems by using the standards EN 15316-x-x [4], [5], [6].

Figure 1 shows how the Primary energy use is calculated based on all the steps where the energy is changing its nature from the source to the end use. In the current case, the energy calculations are performed for the systems within the building to be able to calculate the delivered energy to the building. This means that the building substation with the heat exchangers and tap water storage are included.

The paper has been written in order to ease and diminish the time consuming process of interpreting details in the standards such as the numbered EN 15316-x-x, and with the goal to enlighten main parts of these standards.

In the current scenario, all the losses before the heat is delivered to the building are included in the primary energy factor (PEF) for the delivered heat. In case of considering the complete scenario, the boundaries for the energy performance indicators are the whole energy chain from the source to the end use. In this case, if a CHP plant is represented, the ―power bonus method‖ (EN 15316-4-5) should be used. This method is giving the produced district heat a bonus for the electricity produced assuming that this electricity replaces electricity production with a high PEF-value.

To exemplify some results, an apartment building of 1000 m2 floor area located in a climate like Oslo is chosen. In the base case, the design distribution temperatures in the building are 80/60. The different efficiency figures applying for this case are calculated efficiency values for the production of the heat, for its distribution through the building and its emission in the room. The room efficiency is the one that has the bigger influence on the total system efficiency. INTRODUCTION

According to the implementation of EPBD, it is crucial that the system borders are clearly defined so that the delivered energy is doubtlessly defined.

The Directive on the energy performance of buildings is carried out in order to be used together with a number

Calculation of Primary Energy use according to EPBD and mandated EN-standards Primary energy use = DEdh • PEFdh (f (x,y,z)) + DEel • PEFel (f (x,y,z)) = (Weighted delivered energy indicator (kWh/m2)) • AC Delivered Energy (DE) dh +el

(PEF might be PEFR or PEFT depending on purpose)

Net energy demand

Distribution and Transmission (el) Electricity Hot tap water Air and room + heating system DH substation

End use demand !

( AC = conditioned floor Area )

Primary energy use calculated by PEF(x,y,z)

• • •

Energy carrier (z) Energy carrier (y)

Delivered energy (el) Delivered energy (dh)

Heating systems

Distribution and Transmission (dh)

CHP-plant Heat boilers Storage Generation Transformation

Energy carrier (x)

Waste coll. Logging Extraction Processing Storage

Transportation 2010 / 04 / RU

Calculating end use and losses by EN standards worked out according to mandate from the EU Commission Heating systems efficiency

Calculating direction

System border for the energy performance indicators is the whole energy chain from the source to the end use

Figure 1.- Sketch of the calculation of Primary Energy use according to EPBD and mandate EN-standards

Figure 1 Sketch of the calculation of Primary Energy Use according to EPBD and mandate EN-standards 238

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Finally; once distributed, the heat is emitted according to the demand. For the present case, the heat is delivered either by radiators (80/60 ˚C) in the base case, floor heating (35/28 ˚C), or domestic hot water at

METHOD Definition of building and system build up In this exemplified case, the main chosen building is an apartment building. This building category seems to be the most representative concerning heat use among the building categories defined in the EPBD [1].

60 ˚C. Figure 2 gives a further visual explanation. For the present paper, the supply of heat is just done by a hydronic heating system. The possible heat loss from the distributed air is neglected since the temperature of the air is assumed to be slightly lower than the temperatures in the rooms.

In the shown example, the size of the building is chosen to be 1000 m² floor area since this size should be rather representative and be a good compromise between the previous and the proposed new recast of the EPBD. [2]

Categories of building The presented analysis shows results for five kinds of buildings described in the EPBD which are: single family house and apartment block, office buildings, hotel and restaurants, educational buildings and hospital buildings. When it comes to heat consumption for these buildings, the measurements performed in Linda Pedersen‘s PhD thesis [3] show that the consumption of the apartment buildings is about

A building of these features corresponds to a three storeys squared building with four flats of about 80 m² per storey. In this setting, the total heating system efficiency in the building is built up based on the differentiation between the three main parts of the system. It must be defined where the substation is located in the building, i.e. where the heat is exchanged from the distribution network – DH stage in Figure 2. The heat supply to the heating system within the building from the district heating system is assumed to be provided by two heat exchangers and hot water storage defined as the building substation part of the system.

116 kWh/m², while hospitals use 150 and office buildings use 100 kWh/m². These measured values include the domestic hot water (DHW) and the space heating (SH) consumption. The calculated efficiency for the system will depend on the size of the building as well. The present apartment building shows a higher efficiency value than a single family house with the same consumption. This is due to higher relative losses in the substation. Climate influence The calculations in the present paper are based on a climate like in Oslo, Norway. This climate is defined to have approximately 5100 degree days with 20 ˚C as the internal reference temperature and an external design temperature of -20 ˚C [8].

2010/05/RU

● Ventilation air

+ Hot tap water distrib.

•

+ Room heating distrib.

Substation system border

Figure 2 Sketch of the system elements for production, distribution and conditioning of the rooms

From the substation, the hot water is distributed either for air and space heating or as domestic hot water. Both uses are provided by their own heat exchanger and the necessary pipelines will now be referred to as distribution pipelines. 239

In practice, the outdoor climate can vary widely from place to place. Owing to this, the outdoor climate affects not only the heat consumption but also the relative losses. In general the relative losses are increasing with an increased ratio between the degree days and ΔT between the dimensioning internal and external temperature. The average outside temperature affects the heat consumption and the temperature variations affect the regulation of the heat emitters. This means that during cold periods, the temperature of the supply water tends to be increased imposing an increase in the losses related to the transport of water with higher temperatures. The design temperature for the radiators in the base case in this paper is 80/60, and in warmer periods, this temperature is decreased in order to reduce losses and adapt the supply temperature to the outside temperature. This affects the efficiencies in a positive way.

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 heating season. Usually, the lower the average outside temperature, the longer the heating season. This however, does not affect the DHW since this is more or less steady all the year along.

temperature is here constantly at the designed point of 60 oC. When dealing with distribution of SH, the losses are considered dependent on the kind of insulation material and the ambient and the mean water temperature in the supply and return pipes.

Positioning of substation

The heat emission to the room from the DHW draw-off tap discharge cocks is considered to be negligible in comparison to the total heat consumption.

Figure 2 shows the positioning of the substation. The heat is delivered from the district heating pipelines through two separate heat exchangers, one for heating the water in the storage tank for DHW by a circulating loop, and the other for the air heating and space heating system. The main reason for having two heat exchangers is due to the different needs of temperature levels. In the calculations dealing with the production, the used heat demand used is the total demand, while for the distribution the heat is divided into heat distribution for SH and for DHW.

Dealing with space heating a distinction is done with respect to the kind of emission. Two major groups are considered: the emission by floor heating and by radiators. The first has a low temperature distribution of 35/28 ˚C. As for the radiator system, the analyzed base case is 80/60 ˚C for supply/return design values. Besides the temperature level, the placing in the room affects the stratification efficiency and the loss through the outside wall. Furthermore, another point related to the temperature is the regulation of the room temperature, which in our case, is assumed to be a PI-regulator, even if in a lot of apartments this regulation is quite often done by on/off regulation.

In order to calculate losses related to the storage tank, it is assumed that the tank is of a common type with a common value for the stand by heat loss. The system design consists of the coupling in series of 289 litres storage tanks. There are considered two tanks for the apartment building of 1000 m2 but one more tank if the case is dealing with hospitals, educational and hotel buildings. For other sizes the number of tanks is adjusted according to the demand.

RESULTS Production efficiency The production efficiency is shown in Figure 3. By using the losses on the heat demand and the temperature difference as basis for calculation, the values in Table 1 are obtained.

Efficiencies to be studied For the present paper, as written previously, the system is divided in three smaller system parts which are independent. For every component, the efficiency is calculated following different standards:

Production efficiencies for the different types of buildings according to the EN 15316-4-5 :2007 (Oslo climate) with a distribution temperature 80/60

Production; according to EN 15316-4-5:2007[4]

0,990



Distribution; according to EN 15316-2-3:2007 and EN 15316-3-2:2007[5]

0,980



Room emission; according to EN 15316-2-1:2007[6]

System efficiency



The efficiency of the production includes the losses depending on the thickness of the insulation material, the insulation material itself, the storage tank, the complete local piping system of the substation system and the temperature difference between the two media and the ambient. It takes into account the thermal loss of the total substation. For this case the substation is considered to be in an unheated part and therefore the losses are considered as unrecoverable.

Apartment block

0,970

Office building Hotel and restaurant building Educational building

0,960

Hospital building Single family 0,950 0

250

500

750

1000

1250

1500

1750

2000

2250

Building floor area [m2]

Figure 3 Production distribution of the 80/60 ˚C district heating for different buildings

In case of the distribution, the efficiency depends on the use of the heated water. In case of being a part of a DHW system; the energy used for heating the water which is not drawn-off and which slowly gets cold in the pipelines, has to be considered as loss. Moreover, heat is used to heat up the pipes and fittings. Since the building is large enough to need a circulation loop this loop is considered to be a source of loss? The water 240

As Figure 3 shows, the bigger the building, the higher the efficiency. This effect is due to the reduction of the relative losses when the size of the substation (kW) increases. The curve profile is decreased slightly from 2000 m2 and downwards, and then decreasing rapidly from about 1000 m2 down to 500 m2.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

loop ranging up to 35 % of the total losses from the total distribution system. This loss is related to the fact that the water that remains in the distribution pipelines represents 5% of the total losses per flat.

In addition, it can also be observed that among all types of buildings, apartment houses represent somehow the highest efficiencies which justify the main focus in this study. The displayed case applies for the values where the design temperature level is 80/60 ˚C. It can be concluded from other calculations that the higher the design distribution temperature level, the lower the production efficiency. This conclusion is what could be expected considering the difference between the average temperature and the ambient temperature; the larger this difference, the larger the losses.

For this calculation it is assumed that the pipelines have insulation which a loss of 0.3 W/m∙K (the pipelines are considered to be according to the category ―installed after 1995‖ in [5]). In this calculation the losses due refilling the pipes with hot water are included. This heat could be considered as recoverable loss for space heating during the heating season but in lack of a special national annex all the losses related to the distribution of DHW should be considered as ―non recoverable‖. These losses are not related to the demand for heat and will consequently be lost or result in increased room temperatures.

As shown in Table 1, the efficiency varies only between 0.9784 and 0.9673. It can be concluded, compared to the distribution loss values that the production efficiency is not changing significantly even if the temperature level is changed. As a conclusion it can be said that the losses in the production are relatively low for bigger houses but increasing quite rapidly for smaller buildings.

When it comes to SH, the losses are related to the temperature difference in the non-heated areas where the water goes through. These losses are relatively low compared with tap water since most of these losses are considered to be recoverable. The values used are tabulated in the EN standard [5].

Table 1 Efficiency of DH production system for different design temperature levels. Kind of building

80/60

70/55

Apartment blocks

0,9776

0,9778

0,9780

0,9784

Office building

0,9729

0,9732

0,9735

0,9740

55/45

The percentage of recoverable losses is the cause of the higher efficiency for distribution of space heating which ranges 0.99, whilst the efficiency for distribution of DHW is in thee range of 0.60.

35/28

Hotel and restaurant building

0,9701

0,9703

0,9706

0,9709

Educational building

0,9676

0,9678

0,9681

0,9685

Hospital building

0,9773

0,9775

0,9777

0,9780

Emission efficiency in the rooms In this case, domestic hot water is not considered to contribute to the room heating since the losses from the discharge cocks are considered negligible. In case of space heating a difference has to be made between floor heating and radiator heating when it comes to the efficiency calculations. Floor heating is by its nature emitted at lower temperature, which has an effect on the stratification efficiency since the lower the temperature level, the higher this ηstr. By definition in [6] the stratification efficiency of floor heating is 1 whilst this parameter for radiators goes down to 0.91 on the 80/60 distribution system. This value is combined with the efficiency value of 95 % due to the positioning of the radiators on a normal external wall. Together these values make a total room efficiency of 0.93.

The quality of the insulation of the storage tank will also influence the production efficiency. Manufactures should follow the standard pr EN5044:2005 [7] in order to calculate these losses. Losses from storage tanks should be considered closely in practice, and tanks with relative high losses should be considered for replacement or to be replaced by direct heat exchangers for DHW.

However, a regulation with PI controllers for the radiators delivers an efficiency of 0.97 while the same controller remains at 0.95 for floor heating.

Distribution efficiency First the system for the distribution of tap water is analyzed. In this case, the building includes a circulation loop (in small dots Figure 2) which goes from the storage tank, to the third floor and the distribution branches (in bigger dots in Figure 2) which deliver DHW from the central loop to the consumer. The water temperature in the circulation loop is assumed to be at 60 ˚C throughout the whole year. The biggest share of the losses come from the circulation

In case of the embedded floor heating efficiency the efficiency is 0.93. Since it is considered to be normal insulation layer according to EN 1264, it results in a ηemb of 0.95, the combination results in ηemb of 0.94. Due to these three parameters, floor heating all in all has a room efficiency of 0.90 and radiators of 0.88. 241

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia Table 2 Total efficiencies of the systems for space heating with floor heating and hot tap water system

Total system efficiency The total system efficiency is in this paper calculated as the product of the efficiencies of the different pasts of the heating system in the building.

Total system efficiency

Figure 4 shows the room efficiency and the total system efficiency related to the temperature level supply/return. The delivering water temperature to the radiators is the parameter which affects the efficiency the most.

System efficiency for an hydronic heating system in an apartment building of 1000m2 floor area

Efficiency

0.92 0.9 Room Efficiency Total system efficiency

0.82 Δθ=80/60

Δθ=70/50

0.59

The introduction of an energy performance certificate for buildings according to EPBD requires a transparent calculation model according to the standards in the EN 15316 series. This paper gives a picture of system efficiencies for hydronic heating systems and also an idea of the time consuming process that has to be performed in order to calculate the efficiency of a system in detail. Therefore, it is concluded that some user-friendly guiding material should be desirable in order to enlighten and facilitate the calculation process.

0.94

0.84

Hot tap water

The design temperature level for the system is the most important factor when referring to the efficiency of a hydronic heating system in buildings supplied by district heating. . Therefore the possibility of lowering the design temperature level of the heating system should be considered closely. This increases the emission efficiency in the room and reduces the losses from the distribution pipelines. It saves energy and increase the cooling of the district heating water through the substation. Changing the positioning of the radiator from the external wall to the internal wall actually decreases the room emission efficiency.

Another conclusion from Figure 4 is that the efficiency of the complete system varies significantly with the design temperature level of the heating system. This is due to the temperature difference between the heating system components and the ambient. With a lower distribution temperature the losses will be smaller.

0.86

0.87

CONCLUSIONS

The total system efficiency follows the pattern of the room efficiency since this parameter has far the largest influence.

0.88

Floor heating

Δθ=55/45

Temperature distribution

In the present paper the potential heat losses from the ventilation system are neglected due to the fact that it is assumed that the air temperature is distributed at temperatures slightly below the room temperatures.

Figure 4 Room efficiency and total system efficiency for the different design distribution temperatures according to EN 15136-2-1:2007 (Oslo climate). Radiators with thermostatic valves mounted on normal external walls and with heat supply from district heating

ACKNOWLEDGEMENT This work has been supported by SINTEF, NTNU and has been related to the project ―Systemvirkninsgrader‖ (System efficiencies) which was initiated by Standard Norge and paid by the Norwegian Water Resources and Energy Directorate. It has also been supported by the Primary Energy Efficiency project which is paid by Nordic Energy Research and companies in the heating field in Norway.

Calculated efficiencies for a floor heating system and the hot tap water system are presented in Table 2. Space heating with floor heating has a slightly lower efficiency than radiators due to the lower efficiency for the emission of the heat in the room. Tap water systems have a lower efficiency since the system is by its nature losing a considerable amount of heat when leaving the hot water in the pipes between the tapping cycles. This water is cooled down inside the pipelines and is then being tapped without being useful. In the present case the distance from the substation to the furthermost apartment forces an installation of a circulation loop in order to reduce the waiting time for hot tap water at the tapping cocks. This is a stand by source of loss. These two factors cause the rather low efficiency of the hot tap water system.

REFERENCES [1] European Parliament and Council on energy efficiency of buildings, ―Directive 2002/91/EC on the energy performance of buildings‖ (EPBD) [2] Proposal on a recast of Directive 2002/91/EC on the energy performance of buildings, 2009-11-25 242

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 Buildings in Mixed Energy Distribution Systems‖, Department of Energy and Process Engineering, NTNU,(Norwegian University of Science and Technology), Trondheim

[6] EN 15316 ―Heating systems in buildings - Method for calculation of system energy requirements and system efficiencies – Part 2-1: Space heating emission systems.”, 2007

[4] EN 15316 ―Heating systems in buildings – Method for calculation of system energy requirements and system efficiencies – Part 4-5: Space heating generation systems, the performance and quality of district heating and large volumes”, 2007

[8] “VVS-tekniske klimadata for Norge”, byggforskningsinstitutt, Håndbok 33

[7] CEN: “Efficiency of domestic electrical storage water-heater – German version pr EN 50440”,2005

[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

Norges

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

HEAT LOAD REDUCTIONS AND THEIR EFFECT ON ENERGY CONSUMPTION 1

Christian Johansson and Fredrik Wernstedt

Blekinge Institute of Technology, PO Box 520, SE-372 25, Ronneby, Sweden, chj@bth.se NODA Intelligent Systems AB, Drottninggatan 5, SE-374 35, Karlshamn, Sweden, fw@noda.se balance. This definition is based on the fact that the heat load reduction will continue to exert an influence on the buildings thermal buffer for some time even after the heat load reduction in itself is ended. The length of this interval is specific to each building and is related to the thermal inertia of the building in question.

ABSTRACT In this paper we investigate the consequences of using temporary heat load reductions on consumer substations, from the perspective of the individual consumer as well as the district heating company. The reason for using such reductions are normally to save energy at the consumer side, but the ability to control the heat load also lie at the core of more complex control processes such as Demand Side Management (DMS) and Load Control (LC) within district heating systems. The purpose of this paper is to study the way different types of heat load reductions impact on the energy usage as well as on the indoor climate in the individual buildings. We have performed a series of experiments in which we have equipped multiapartment buildings with wireless indoor temperature sensors and a novel type of load control equipment, which gives us the ability to perform remotely supervised and coordinated heat load reductions among these buildings. The results show that a substantial lowering of the heat load and energy usage during periods of reductions is possible without jeopardizing the indoor climate, although we show that there are differences in the implications when considering different types of heat load reductions.

In this paper we study the consequences of using different types of heat load reductions, and try to analyse the way the thermal buffer of the building is affected along with the actual heat load and energy usage from both a local and a global perspective. We study the performance of both long low-intensity heat load reductions (e.g. night time set-back) as well as short high-intensity reductions (e.g. those frequently used in DMS schemes). The use of night time set-back has received some attention in previous works, e.g [1], and the possibilities to use the building as a heat buffer has been evaluated [11], but heat load reductions such as those used in DSM and LC have to the knowledge of the authors not been thoroughly investigated. Night Time Set-back Night time set-back means to lower the wanted indoor temperature during night time, with the purpose of saving energy through reduced heat losses due to decreased difference between indoor and outdoor temperature. This is the most common way to perform temporary heat load reductions, and many commercial control systems support this feature. This is normally done by a parallel displacement of the heat control curve during night hours. During night time set-back the wanted indoor temperature will be set to one, or a few, degrees lower than during normal operations. There is, however, an ongoing debate on whether night time setback actually gives an energy saving or not [4], and most practical implementations of night time set-back suffer from morning peak loads when the control system returns to the original operational level. Still, almost all control equipment companies sell equipment that facilitates the use of night time set-back, and the use of this technique is widespread.

INTRODUCTION The main purpose of this paper is to investigate the consequences of using temporary heat load reductions on consumer substations within a district heating network. The most common way to perform temporary heat load reductions is to use night time set-back, i.e. to lower the wanted indoor temperature during night time while social activity is expected to be low. Emerging technologies like Demand Side Management (DMS) and Load Control (LC) also use temporary heat load reductions in order to accomplish system wide control strategies, although the characteristic of these head load reductions differ significantly from night time set-back. In the context of this study we regard a heat load reduction to be the whole process from the initial change of heat load, through the return to normal heat load, and until no evidence of the heat load reduction can be noticed in the dynamics of the building energy

Demand Side Management and Load Control

244

While night time set-backs are a solely local energy saving technique, DMS and LC are usually performed with a system wide perspective in mind. A building

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 energy consumption, while the district heating company is more interested in being able to optimize the whole production and distribution process. Optimizing the production normally translates to avoiding expensive and, more often than not, environmentally unsound peak load boilers or trying to move heat load demand in time in order to maximize utility during combined heat and power generation. Basically, from the perspective of the district heating company it is a question of finding a balance between lowering expensive heat load demand while still selling as much energy as possible. Implementing this on a system wide scale requires complex coordination control strategies that dynamically adapt to the state of the district heating system [2]. On the local building level this is implemented by performing temporary heat load reductions. On a local level these reductions are normally very short, i.e. one or a few hours, but they can be of high intensity, even sometimes completely shutting of the heat load during shorter periods of time. This behaviour requires the control system to be highly adaptive in relation to the dynamics of the buildings thermal inertia in order to avoid jeopardizing the indoor climate. By coordinating such local heat load reductions among a large group of buildings it is possible to achieve system wide DMS and LC.

Most of the previous work done on the subject is based on simulated results. This is expected since the dynamic thermal processes within a building are extremely complex and it is not surprising that comparisons between measurements and calculations sometimes show large discrepancies. It is noted that most calculations are dependent on variables that cannot be measured and verified, and that the building time constant is really not a constant [6]. EXPERIMENTAL METHOD In order to study the effects of temporary heat load reductions we equipped a building with several wireless temperature sensors in order to measure the fluctuations in indoor temperature. The building in questions is an office building with semi-light thermal characteristics (light construct with concrete slab) and a time constant of about 150 hours [7]. The indoor temperature sensors were placed on different locations within the building in order to get a good overview of the thermal behaviour of the indoor climate. In addition to the existing outdoor temperature sensor an extra wireless sensor was also placed on the outside of the building. Unlike the existing outdoor temperature sensor the wireless one was placed in a position were it was fully exposed to any possible sunshine. This gave us an extra indication of the impact of free heating through window areas, even though we did not have any ability to measure the actual solar irradiance.

Previous work Most previous work regarding temporary heat load reductions deals with night time set-back. This is a technique that has been around for a long time, and is based on the general idea that if you decrease the difference between the outdoor and indoor temperature in a building you will save energy. One of the first large-scale evaluations of night time set-back was performed in 1983 when buildings in Sweden, USA, Belgium and Denmark were evaluated. This experiment concluded that night time set-back did not save as much energy as was expected, at most a few percent for multi-apartment buildings [3]. In hindsight it is possible to see that these meagre results were a consequence of several interacting factors. First of all the control systems of the time were not capable of properly handling the transition from night time setback to the original operation mode, which causes a considerable over-compensation of heat load when the systems tries to find the new control level. This extra boost in heat load during the mornings counteracts large portions of the energy saving done during the night. The theoretical part of the experiment also had a few draw-backs, e.g. assuming optimally adjusted radiator systems and linear relations between indoor temperature and energy savings. Other articles show that there is indeed a substantial level of energy saving to be found by controlling the local heat load [5]. 245

In order to control the district heating consumer station we connected a load control platform for system wide LC and DSM [8]. This platform is based on a novel form of hardware and software which enables us to manage the heat load of the substation without any major alterations or any damage on the existing hardware. The software system is based on the open source Linux operating system and is equipped with an application programming interface (API) for I/O. This makes it easy to apply additional sensors, e.g. for measuring the forward and return temperatures of the radiator system. The platform also features connections to a database system which enables realtime logging and analyse of sensor data. The actual heat load reductions are implemented by supplying the existing control system with adjusted outdoor temperatures, which gives us the ability to manage the behaviour of the heat load without exchanging any existing hardware. This adjusted outdoor temperature can be managed with a resolution of at most 60 seconds. The computer platform uses either Ethernet or GPRS modems to communicate with the database. In our case we used the existing Internet access in the building. In addition to this primary experimental building we also collected and analysed data from previously installed buildings using the same basic computer platform.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

system performs a controlled heat load recovery in order to avoid unwanted heat load peaks after the reduction.

Energy and heat load usage was primarily evaluated by studying the dynamic differences between the forward and return temperature of the radiator system in relation to the flow. These readings were then verified by specifications from the district heating provider regarding energy consumption and momentary heat load usage.

The same values are shown for a long heat load reduction in Figure 2. The heat load reduction starts slightly before the 600 minute mark and continues for several hours until about the 900 minute mark. After that the control system performs a controlled recovery in order to return to the original operational state.

Using this set-up we scheduled different types of temporary heat load reductions and studied their effects on the measured data. During this study we studied three primary types of temporary heat load reductions: 

Long – Four to eight hours of continuous heat load reduction with different intensity



Short – Up to one hour long heat load reductions with different intensity



Recurring – Several short subsequent heat load reductions with short pauses in between

When we studied the different types of heat load reductions we took care in allowing the buildings thermal process to return to its original state between each reduction so that the reductions would not influence each other. This was done in between each reduction except in those cases when then purpose was to explicitly study the interaction between subsequent heat load reductions.

Figure 2: dT in radiator circuit with long heat load reduction

Figure 3 shows the same values for a series of recurring heat loads.

EXPERIMENTAL METHOD Figure 1 shows the temperature difference between the forward and return temperature in the radiator circuit during a short heat load reduction.

Figure 3: dT in radiator circuit with recurring heat load reduction Figure 1: dT in radiator circuit with short heat load reduction

Each of the heat load reductions in Figure 3 is one hour long intersected by one hour long recovery periods. The first reduction starts at the 60 minute mark and continues until the 120 minute mark.

The heat load reduction starts at about 60 minutes and continues until the 120 minute mark. Between the 120 minute mark and about the 160 mark the control 246

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Figure 6 shows recurring heat load reductions instead of single long ones. It is clear that the building is able to respond to the control scheme in this example also. The largest heat load reduction during the recurring scheme is about 25%.

Figure 4 shows the energy consumption in relation to the outdoor temperature during week long periods with and without heat load reductions implemented as LC. The squares are from periods without LC and the triangles are from periods with LC. LC in this regard means that temporary heat load reductions are being performed in recurring sets throughout the week as long as the thermal inertia of the building allows it, i.e without jeopardizing the indoor climate. In this example the energy usage is about 8.2% lower during periods of heat load reductions.

Figure 6: Heat load reductions shown 24 hours without reductions (black), 24 hours with reductions (dark grey) and control scheme for reductions (light grey)

Figure 7 shows a range of indoor temperature readings during periods with heat load reduction (triangles) and during periods without (squares). The average deviation during heat load reduction is about 0.29 while the average deviation during periods without reductions is about 0.19.

Figure 4: Energy usage in relation to outdoor temperature. The squares are values during periods without LC, and triangles show periods with LC

Figure 5 shows the heat load (kW) during 24 hours when using reductions compared to not using reductions. The control scheme is also added to the figure in order to show when the reduction was performed.

Figure 7: Indoor temperature during periods with heat load reductions (squares) and during periods without heat load reductions (hourglass) Figure 5: Heat load showing 24 hours without reductions (black), 24 hours with reductions (dark grey) and control scheme for reductions (light grey)

Figure 5 clearly shows that the reduction in heat load closely follows the control scheme. The largest heat load reduction is about 30% in this example. 247

Figure 8 shows readings from two different outdoor temperature sensors during a time period of two days. The graph shows the outdoor temperature sensor which is connected to the actual consumer sub-station in the building (black line). Normally these sensors are placed somewhat in the shadow to avoid large fluctuations due to solar radiation. We added another temperature sensor (grey line) in order to estimate the

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

impact of this solar radiation. Hence this sensor was placed in full view of the sun. The first day was sunny during most of the morning until midday, while the second day was cloudier.

will lower the need of additional heating from the radiator system, by coordinating the thermal inertia of the building with freely available heat, e.g. heat from sunlight or electrical appliances, to balance the heating need. This notion is supported by our results as we have shown that the thermal inertia of even a small or medium sized multi-apartment building is considerable. How people perceive the indoor climate is dependant not only on the actual indoor temperature itself but also on other factors like air quality, individual metabolism and behaviour, radiation temperature and air movement. In relation to this it can be noted that previous work have shown that about five percent of any group of people will always be unsatisfied by the indoor climate [9], and that it is not possible to create a perfect climate that will make everyone happy. CONCLUSIONS There is an ongoing debate whether night time setbacks lead to an energy reduction or not. Results from this study clearly show an energy saving in relation to heat load reductions, although this assumes that the control system is able to smoothly handle the transition from reduction to normal operation. The results showing energy saving is evaluated in relation to the total energy usage which also includes tap-water usage. Normally this is estimated to about 30% of the total energy use in a multi-apartment building.

Figure 8: Outdoor temperature sensors placed in the shade (black line) and in full view of the sun (grey line)

DISCUSSION When dealing with temporary heat load reductions it is important to include the whole process of the reduction. This also includes what happens after the actual heat load reduction has been performed. For example, when just restoring the wanted control level after a long reduction, e.g night time set-back, the forward flow temperature in the radiator system will rise much faster than the return flow temperature. This causes a substantial, although temporary, heat load increase in the radiator system which negates large portions of the energy saving done during the actual reduction. Apart from decreasing the local net energy saving this behaviour is also less than desired from a system wide perspective, since it causes massive heat load peaks if done in many buildings simultaneously, e.g. contributing to morning peak loads. In order to avoid this it is important to factor in the whole process of the reduction, and make sure that the control system properly handles the transition from the reduction level to the original level. The inability among most commercially available control systems to properly handle this over-compensation is most likely contributing a great deal to the lingering controversy whether night time set-back actually gives an energy saving or not. It is important to realize that the definition of an acceptable indoor temperature is not about having the indoor temperature at a certain precise level at all time, but rather to have it within a certain, socially acceptable, temperature interval at all time. This has been discussed at great length in previous work [6]. The general idea is that a greater temperature interval

248

In prior studies of temporary heat load reductions the focus has been on the fluctuations in the indoor temperature as a way of evaluating the energy saving [3]. This idea is based on the widespread notion that any energy saving is linearly proportional to the temperature difference between the indoor and outdoor temperature. This model might be true in a steady state simulation where the temperature difference is assumed to have had time to permeate the air mass as well as the entire building structure, but it is obviously inadequate in a dynamic situation. We have instead focused on the heat load and energy usage directly, i.e. the difference between forward and return temperature in relation to the flow within the radiator circuit. In most of the buildings evaluated there has been a considerable reduction of energy consumption without any noticeable change in indoor temperature. The reason that there does not need to be a measurable change of the indoor temperature is due to the dynamics of the thermal inertia of the building, e.g. the time constant of a building is not a constant [6]. This aspect comes into play when using very short heat load reductions, at most one or a few hours long. During this first part of the reduction it is mainly the actual air mass that is influencing the indoor temperature drop since this body has a low resistance to change, i.e. the short time constant [10]. If the heat load reduction is prolonged, like during a night time set-back, the

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 thus stabilizing the continuing temperature drop, i.e. the long time constant [10].

REFERENCES [1] N. Björsell, Control strategies for heating systems, University-College of Gävle-Sandviken.

The influence of external and internal free heat is large enough that when these heat sources interact with other parts of the thermal process it hides shorter heat load reductions in the ambient temperature. This can be seen in Figure 7 where it is shown that although the average indoor temperature is not noticeably affected there is still a somewhat larger deviation in the indoor temperature which implies that there is indeed a higher level of temperature flux within the air mass and that this is triggered by the heat load reductions. The control policies used during this work obviously set a high bar for the control system to handle, but as the average hardware develops it should be possible to implement such techniques on a larger scale.

[2] F. Wernstedt, P. Davdisson and C. Johansson, ―Demand Side Management in District Heating Systems‖, in Proc. Of Sixth International Conference on Autonomous Agents and Multiagent Systems, Honolulu, Hawaii, USA, 2007. [3] L. Jensen. ―Nattsänkning av temperatur I flerbostadshus‖, R64:1983, Byggforskningsrådet, 1983 (In Swedish). [4] H. Lindkvist and H. Wallentun. ―Utvärdering av nio fjärrvärmecentraler i Slagsta‖ Report ZW 04/05, ZW Energiteknik, 2004 (In Swedish) [5] F.B. Morris, J.E. Braun and S.J. Treado ―Experimental and simulated performance of optimal control of building thermal storage‖, ASHRAE Transactions, Vol. 100, No. 1, 1994

Figure 8 gives another clear indication of just how substantial such sources of free energy can be. This extra heating due to solar radiation through the windows directly interacts with the mass of air inside the building, thus raising the temperature.

[6] E. Isfält and G. Bröms. ―Effekt- och energibesparing genom förenklad styrning och drift av installationssystem I byggnader‖, ISRN KTH/IT/M-22--E. Institutionen för Installationsteknik. Kungliga Tekniska Högskolan, 1992. (In Swedish)

In addition to being able to help save energy usage in a building temporary heat load reductions also form the backbone of DSM and LC, in which the goal is to manage the heat load (kW) rather than the energy usage (kWh).

[7] S. Ruud. ―Energimyndighetens program för passivhus och lågenergihus‖ Remissversion 200903-10. Forum för Energieffektiva byggnader, 2009. (In Swedish)

FUTURE WORK In the future we plan to further develop models in order to dynamically estimate the temperature flux within buildings and develop theoretical and practical interfaces for incorporating this data dynamically into the control systems.

[8] F. Wernstedt and C. Johansson. ― Demonstrationsprojekt inom effekt och laststyrning‖. ISBN 978-917381-041-8, The Swedish District Heating Association, 2009. (In Swedish)

ACKNOWLEDGEMENT

[10] C. Norberg. ―Direktverkande elradiatorers reglering och konstruktion‖ Vattenfall Utveckling AB, Rapport nr F-90:5, Älvkarleby, 1990. (In Swedish)

[9] J. Skoog, ―PM avseende komfort‖, ÅF-Infrastruktur AB, 2005. (In Swedish)

This work has been financed by Blekinge Institute of Technology and NODA Intelligent Systems AB.

[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

VERIFICATION OF HEAT LOSS MEASUREMENTS 1

J.T. van Wijnkoop , E. van der Ven 1

Liandon B.V, Thermaflex International Holding B.V. conditions, as defined by the European standard [1]. The ability to conduct equally based heat loss measurement result in an objective comparison of different types of (semi) flexible piping systems, providing the opportunity to highlight strengths and weaknesses of (competitive) piping systems. Furthermore, in contradiction to most heat loss tests, the test time in the Thermaflex test-rig is only a few hours so the test can be performed during production. This provides the opportunity to optimize the production process real-time and measure the heat loss of the product several times during a production run. This guarantees the quality of the produced batch.

ABSTRACT Heat loss tests are performed on different samples of the Thermaflex Flexalen 600 series and one ST-PUR-PE sample at the Thermaflex heatloss equipment and two German test facilities. At these facilities two different testing methods are used. These methods are both described in the European standard [1] but show significant differences in the results. In this paper the different methods of testing are described. Furthermore the Thermaflex heat loss equipment is verified with the test institute that uses the same testing method.

In addition the handling of the equipment is made easy, so no specially trained staff is needed for testing, making it possibly for operators to carry out the tests.

INTRODUCTION Last year Liandon developed a test-rig for Thermaflex to measure heat loss of insulated plastic piping systems. With this test-rig it is possible for Thermaflex to test the in house produced pre-insulated, semi flexible pipes in various diameters.

EUROPEAN STANDARD METHOD DESCRIPTION The European standard EN 15632 [1] allows two different methods of heat loss or thermal conduction testing. These methods both state the same on internal heating of the service pipe but vary on the method of compensation for heat loss in axial direction.

To verify the test results, the results of the Thermaflex heat loss equipment are compared with the test results of two acknowledged institutions. For this paper two German institutions are chosen, since they both measure in compliance with the European standard EN 15632 [1], however with different methods described in this paper. In order to give an appropriate comparison, knowledge of the testing methods of both systems is required. In this paper the testing methods of all three systems is covered, together with the comparison of the test-results. Since the testing facilities use two different methods described in the standard, the comparison refers to the test methods and the test results.

The first method, the guarded end method, states no axial heat transfer is permitted. This should be accomplished by the use of end guards, an extra pair of heating elements at both ends of the service pipe as shown in Fig 1. By heating the ends separately to the same temperature as the middle test section no heat transfer will take place to the ends of the service pipe. In this case a theoretical compensation is not required since the test section only has losses in radial direction. This method is used in the Thermaflex heat loss equipment and at one of the institutes.

The objective of this paper is to compare the test methods and test results of the two different test institutes with the Thermaflex heat loss equipment and verify the outcome. As in “Heat loss of flexible plastic pipe systems analysis and optimization” (E. Van der Ven et Al.) [4] and “Performance of pre insulated pipes” (I. Smits et Al.) [6] these results are used to compare different sizes of the Flexalen 600 series and competitive products.

Fig 1, Longitudinal section guarded end heating probe

NOVELTY AND MAIN CONTRIBUTION

The second method described for compensating for axial heat loss is the calibrated or calculated end method. The calibrated end method will not be covered in this paper since it is not used in our comparisons.

The Thermaflex test-rig is newly developed for the research of heat loss of pre-insulated pipes. The novelty of this system is its ability to measure the overall heat loss of different samples under similar 250

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

X, L: distance to next measuring point from the middle, sample length

The calculated end method states the ends of the service pipe shall be insulated with a known thermal 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 overall heat loss. These samples consist of 2 or 3 m of the pre-insulated piping system. More information about the Flexalen 600 system can be found in “Heat loss of flexible plastic pipe systems analysis and optimization” (E. van der Ven et Al.) [4]. Furthermore, method comparison tests are performed on competitive pre-insulated piping systems, a comparison of the products themselves is given in “Performance of pre insulated pipes” (I. Smits et Al.) [6].

Fig 2, Configuration calculated end cap.

The service pipe is heated, using a heating element with only one section. During the tests a thermal profile is made of the outer casing of the sample, showing lower values at the ends. After testing the heat loss is compensated for the end loss with the van Rinsum or Nukiyama theory. For this investigation only the van Rinsum theory is used and therefore described. According to the van Rinsum theory, the axial heat loss causes a decrease in temperature not only towards the ends of the service pipe, but in the test section as well. With the use of the equations (1), (2), (3) this temperature decrease in the test section can be calculated and added to the measured value, compensating the end loss. This corrected temperature is used in equation (4) to calculate the overall thermal conductivity. This method is used by one of the German institutes.

 D2    D0 

The tests on the Flexalen 600 products are performed by Thermaflex and by one of the acknowledged institutes, using the different methods. To ensure the effect of ageing in the Flexalen 600 system is the same during all tests, the Flexalen 600 samples are tested simultaneously. To exclude effects of the production process both tested samples are half of a 6 meter stick. An alternative method is used for Flexalen 50A25 and competitive products. Here the same sample is tested at the different test facilities.

  ln

 calc  2   L T



0m  T2

The comparison of the results is based on the outcome of heat loss per meter, calculated as described in the European standard [1]. This loss per meter is only conclusive on a very small part of the entire system. Therefore the complete Flexalen 600 system will be covered in paper “Heat loss system optimisation” (J. Korsman et Al.) [3] and ‗‟New economical connection solutions for flexible piping systems” (C. Engel et Al.) [5].

(1)

2    calc

c 

D 

A1  1  A2  2 ln D2   0

T 0m 

(2)

T0m  T0X

cosh X  c

 D2    D0 

In this report the following diameters of the Flexalen 600 piping systems are used for comparison of the measurements:  Flexalen 600:  50A25, two guarded end tests* and calculated end test.  160A90, one guarded end test* and calculated end test.  200A110, one guarded end test* and calculated end test. Competitive products:  Sample 1 two guarded end tests  Sample 2 two guarded end tests* *At the time of writing the second test results were not yet available.

(3)

  ln  





2   L T0m  T 0m  T2

(4)

λcalc: approximate value of thermal conductivity D2/D0: outer/inner diameters of casing and service pipe A1, A2: areas of the heating probe, inner service pipe λ1, λ2, λ: thermal conductivity of heating probe, thermal conductivity of medium in the service pipe, thermal conductivity total sample. 251

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The physical part of the Thermaflex heat loss equipment consists of three segments.

section. With this method it is possible to measure the heat loss by measuring the power needed to maintain a constant temperature of the test sample. In contradiction to the measurements at the test institutes, the Thermaflex heating probes temperature is regulated by PID controlled power supplies. In the test results Graph 2 the power consumption versus test time is shown. This variable power supply makes it possible to pre-heat the probes in a short period of time, shortening waiting times considerably. Furthermore the use of the actual pipe material as a heating probe increases the accuracy. Moreover it eliminates all additional heat loss by convection that will be present with the use of smaller, not inner service pipe connecting heating probes.

The first is the water cooled compartment in which all tests are performed. This compartment is kept at a constant temperature, (23 째C), during each measurement.

For testing competitive products with different diameters these advantages are lost. However by the use of thermal compartments in the service pipe the test results can be guaranteed.

THERMAFLEX HEAT LOSS EQUIPMENT 600 The Thermaflex heat loss equipment is specially designed for the Thermaflex Flexalen 600 series. One of the major design goals was to develop a fast and easy to use test rig with the precision of a laboratorial test. These goals have resulted in a test rig that is able to measure heat loss in a few hours, allowing direct optimization during the production process, and is operable by the production staff without the loss of accuracy Physical test facility

The second is a heat source, for which heating probes are used. These heating probes are custom made by equipping a two meter Thermaflex piping segment, of all available diameters, with three heating coils.

Thermaflex method of testing For testing, the heating probe with the appropriate diameter is inserted in the insulation covered with outer casing, and inserted in the cooled test section. After connecting the probe to the control unit the measurement can be started. Different testing conditions can be entered at this point such as the inner pipe temperature, representing the internal medium. When the test is started the heating coils heat the inner side of the probe until the desired temperature is reached. When the inner temperature is considered constant and uniform throughout the three heating coils, the actual measurement is started. To ensure a constant temperature in the probe, a waiting time is built in the software that will reset the measurement if temperature exceeds preset temperature values.

The third part of the heat loss equipment is the control unit. Here the heating probe is powered and all thermal readings are done. By applying custom made software all desired readings can be done. The final output is the actual heat loss in W/m through the entire pre-insulated Flexalen pipe, consisting of the service pipe, insulation and outer casing.

The heat loss measurement is done by measuring the energy required to keep the probe at a constant temperature, by measuring the current at constant voltage in the heating coils, and calculating the power consumption. Since the middle/testing coil is exactly one meter in length the required energy represents the exact heat loss through one meter of piping and insulation in W/m. Since the actual piping material is used during the measurement, there are no other losses, nor advantages, than there will be in practise, ensuring an objective measurement. Furthermore a realistic fit of the insulation material is guaranteed. As stated in the foregoing paragraph these advantages are lost for divergent diameters. However during this investigation the probes have proven suitable for testing, as both testing institutes also use smaller heating probes.

Fig 3, Thermaflex heat loss equipment 600.

Measurement principle Thermaflex The Thermaflex test rig is designed in compliance with the European standard [1] and also the tests are carried out according to ISO 8497 and EN 15632. In the design of the heating probes the most realistic method, the guarded end method, is used. According to this method the heating probes are equipped with three heating coils with separate power supply. As shown in Figure 2, two 400 mm heating coils located at each end of the 1000 mm test section. These two sections provide a thermal insulation at both ends of the test section since all three are kept under uniform temperature, eliminating axial heat loss of the test 252

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Test results Thermaflex heat loss equipment In this paragraph the test results are presented for the tests carried out with the Thermaflex heat loss equipment. For this study four different types of the Flexalen 600 series were tested. The tests for the Flexalen 600 series took place at three different temperatures, 60, 70 and 80 °C. The values at lower temperatures are calculated using the linearization method described in the European standard EN 15632 [1]. In the following tables and graphs the test results of the Thermaflex heat loss equipment are presented. The power usage during the testing cycle is shown in Graph 2. In this graph the first 40 minutes represent the heating and stabilisation time for the heating probe and insulation, whereas the last 30 minutes is the actual test time. Since, as the figure shows, the temperature is constant, the power usage equals the heat loss through the piping system in radial direction during the last 30 minutes. The results, as given in Table 1, are calculated by using the mean of the power consumption during the last 30 minutes of the heat loss test. The results in Table 1, are also displayed in Graph 1 for the three tested samples.

Graph 2, Power and temperature of the Thermaflex heating probe.

Outcome Competitive products for comparison of testing method: For the comparison with test institute two, two samples of competitive products are tested. As these samples are ST-PUR-PE system, a correction has been made for using the PB heating probe using the Wallentén [2] method. First the thermal conductivity of the insulation is determent by the use of equation (5), hereafter the heat loss is recalculated without the heating probe, using the temperature of the inner service pipe in equation (6). The results are presented in Table 2 and Graph 3.

Table 1, Results heat loss equipment for the Flexalen 600 products Heat loss of the Flexalen 600 series in W/m tested on the Thermaflex heat loss equipment Product

40 °C

50 °C

60 °C

70 °C

80 °C

50A25

3.6

6.4

9.3

12.0

15.0

160A90

6.2

10.1

14.0

17.9

21.8

200A110

6.5

12.0

17.5

23.0

28.5

i  2 

Heat loss results thermaflex heat loss equipment

 Tp  Tc  probe

 corrected 

Heat loss [W/m]

 d3  W  d2    m K (5) d 1  d2  1  6  1  d4   ln   ln   ln st  d1  p  d  c  d3       5 ln





2  T st  T c

1 st



 d2  1  d3  1  d4  m (6)   ln   ln  d1  i  d2  c  d3       

 ln

Where: Tp,Tc, Tst=Probe, Casing and Steel pipe temperature

0 40

d1 to d6 = inner/outer diameters of service pipe, casing and heatingprobe 50

λst, λi, λc, λp = heat coefficient of service pipe, insulation, casing and probe

Temp erature inner service pipe [°C]

Φprobe, Φcorrected =probe power and corrected heat loss.

Flexalen 50A25 Flexalen 160A90 Flexalen 200A110

Graph 1, Results heat loss equipment Flexalen 600 products.

253

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

made of metal. Furthermore no heat guards are used. This means the outer ends of the piping system are insulated and the heat loss is corrected with a calculated value. In the paragraph ―European standard method description‖ a more detailed description is given. As can be seen in Fig 4 the heat distribution in this case is not uniform along the test specimen, proving the need for the van Rinsum correction.

Table 2, Results heat loss equipment for the competitive products Heat loss of competitive products in W/m tested on the Thermaflex heat loss equipment Sample

40 °C

50 °C

60 °C

70 °C

80 °C

Sample 1

2.5

4.5

6.4

8.4

10.3

Sample 2

6.5

8.8

11.2

13.5

15.9

Sample 3

Heat loss results thermaflex heat loss equipment

Heat loss [W/m]

Fig 4, Thermal image of the sample at institute one

0 40

In contradiction to the Thermaflex test rig, no integrated computer controlled power supply system is used. The power for the heating probe is first theoretically calculated and manually set to this value. For the temperature measurement thermocouples and a data logger with computer link are used.

Temp erature inner service pipe [°C] Competitive samp le 1 Competitive samp le 2

Graph 3, Results heat loss equipment for the competitive products

Method of testing

Physical test facility

The heating probe is positioned in the centre of the test pipe with positioning foam in three sections of the pipe. On these foam blocks four thermocouples are placed in 0, 90, 180 and 270 degrees on the inner surface of the service pipe. For the outcome of the pipe inner temperature the mean of the four values is used. To measure the temperature on the outside casing of the insulation, five groups of four thermocouples are used in the same configuration as the inner pipe. The difference being that the thermocouples are placed both on and in between the corrugations of the casing. The test sample, with the heating probe, is placed in the conditioned container thereafter the test can be started. The power supply of the heater is turned on by setting the voltage and current of the power unit to a fixed value so the electrical power equals the calculated heat loss.

The physical part of the test facility is similar to the Thermaflex test rig and also consists out of the three elements: A temperature controlled compartment where the tests are carried out at a constant temperature of 23 °C. FIW also uses heating probes as a heat source but, since it is not specially designed for the Flexalen 600 system, they are made to fit all systems. To ensure the fit of the probes in all different systems the diameters are smaller, and for durability

Depending on the diameter of the test sample and the test temperature the waiting time for the heating of the sample is five to eight hours due to the low, fixed power input. After a constant temperature of the outer casing is achieved the actual test cycles start. Each test cycle consists of a measurement of 30 min in which the outer casing temperature is to be constant. If not the cycle has to be restarted. In total ten cycles will be performed on each sample. After the test the values are corrected

TEST INSTITUTE ONE This institute is specialized in measuring heat loss in different types of insulation. The test facility used for the Flexalen 600 system is specially designed for measuring the heat loss of (pre-) insulated piping systems. This means the facility is designed to measure all different types and diameters. Measurement principle institute one The measurements are all based on the calculated end apparatus, using the van Rinsum theory as correction, as described in the paragraph European standard [1] method description of this paper.

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 resistance and overall heat loss are calculated.

An update to this paper will be made as soon as the Flexalen 600 results will become available.

Test results institute one

Testing method institute two

As the actual measurement data are not available due to the correction factor, only the calculated values can be discussed in this paragraph. As soon as the actual measurements become available this section will be updated. Furthermore the results are not given for exactly 60, 70 and 80 °C due to the fixed power supply with no temperature set point, the displayed results are calculated heat loss values at the set temperatures to make the data more interpretive. For this calculation the linearization method described in the European standard [1] is used. In Graph 4 the data from Table 3 is presented as a graph.

The method used by this institute is generally the same as the method used by Thermaflex; however the test facility itself is different. Physical test facility The testing facility at institute two consists of a temperature controlled room, kept at the prescribed 23 °C. As a heat source a heating probe, consisting of a 2 m test section and two 50 cm end guards is used. At the time of writing no further information on the test facility was available. This paragraph will be updated when this information becomes available.

Table 3, Results test institute one for the Flexalen 600 products

Method of testing Prior to testing, the sample is prepared by placing thermocouples in various locations on the inner service pipe and outer casing. Subsequently the sample is placed in the temperature controlled room and the heating probe is inserted. By setting the power supply to a calculated value for all three heating coils the heating process of the sample is started. Because of the low fixed value of the power supply, this heating will take approximately 5 to 8 hours. After the desired temperature is reached at the test section as well as at the guarded ends, the actual test is performed. The test consists of a power reading during a 30 min cycle where het temperature of the test section and guarded ends may not exceed the limit of an yet unknown bandwidth.

Heat loss of the Flexalen 600 series in W/m tested at test institute one Product

40 °C

50 °C

60 °C

70 °C

80 °C

50A25

5.6

8,8

11.9

15.1

18.3

160A90

9.1

15.1

21.1

27.1

33.0

200A110

9.8

15.1

20.5

25.8

31.2

Heat loss results test institute one

Heat loss [W/m]

Test results test institute two

The test results of institute two are given in Table 4 and Graph 5. As not all data was available during writing there tables and graphs will be updated.

0 40

Table 4, Results test institute two for the competitive products

Temp erature inner service pip e [°C] Flexalen 50A25 Flexalen 160A90 Flexalen 200A110 Graph 4, Results test institute one for the Flexalen 600 products

Heat loss of competitive products in W/m tested at test institute two

TEST INSTITUTE TWO For the second test institute in this research, an institute using the same guarded end method is chosen. This makes it possible to provide a correct comparison between the test results and not only the testing method. The tests carried out by test institute two at the time of writing are of competitive products only as the facility was already running on full capacity. 255

Sample

40 °C

50 °C

60 °C

70 °C

80 °C

90DN25

1.95

4.00

6.02

8.06

10.09

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

competitive products show consistency with test institute two as shown in Graph 7. The heat loss equipment values are just a little higher, which can be explained by the need to cut the sample in order to place the heating probe with thermocouples in the right position. The difference between the outcome of the test on sample 1 are 0.39 and 0.22 W/m at an inner service pipe temperature of 60 and 80 °C respectively. These values are within the combined accuracy range of both facilities. This comparison, although only based on one test, proves the worthiness of the Thermaflex heat loss equipment and will be updated as more data comes available.

Heat loss results test institute tw o

Heat loss [W/m]

0 40

Temp erature inner service pip e [°C] Heat loss c omparison Thermaflex and test institute two

Competitive sample 1 Competitive sample 2

Heat loss [W/m]

Graph 5, Results test institute two for the competitive products

COMPARISON OF THE TEST RESULTS

Comparison of the Thermaflex flexalen 600 series: Although both methods, guarded end and calculated end, are approved and described in the European standard [1], the difference between the results is substantial as displayed in Graph 6. Moreover all results vary more as the temperature difference increases. This can be explained by the use of the calculated end caps that conduct more energy at higher temperature differences. As these end cap losses increase, the corrected thermal conduction for the sample also increases, resulting in a higher calculated heat-loss.

0 40

CONCLUSION During this research it has become clear that the European standard [1] tolerates differences in heat loss values by allowing different testing methods. The outcome of the tests indicate that the result of the guarded end cap method varies from the result of the calculated end cap method, however no assumptions can be made based on only one comparing measurement. Further study that is being conducted at this moment will provide more comparison data. This will be updated with this data as soon as becomes available. This new data could point out that the van Rinsum theory is not suitable for accurate heat loss measurement of plastic piping systems.

Heat loss [W/m]

25 20 15 10 5

Graph 7, Comparison results of the heat loss equipment and test institute two

Temp erature inner service pip e [°C]

Competitive samp le 1 Competitive samp le 1 Competitive samp le 2 Competitive samp le 2

Heat loss c omparison Thermaflex and test institute one

0 40

Temp erature inner service pip e [°C] Flexalen 50A25 Thermaflex result Flexalen 50A25 Institute one result Flexalen 160A90 Thermaflex result Flexalen 160A90 Institute one result Flexalen 160A90 Thermaflex result Flexalen 160A90 Institute one result

The comparison of the guarded end method results from test institute two and the Thermaflex heat loss equipment conclude that the results of the heat loss equipment are correct and comply with the European standard [1]. This validation makes the results of the Thermaflex heat loss equipment valid for not only in house testing but also for publication as done in “Heat loss of flexible plastic pipe systems analysis and optimization” (E. van der Ven et Al.) [4] and “Performance of pre insulated pipes” (I. Smits et Al.) [6].

Graph 6, Comparison results of the heat loss equipment and test institute one

Comparison of competitive products: Although the Thermaflex heat loss equipment was designed for Flexalen series, test results on 256

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

for their devotion on all the heat loss measurements they performed during this research in a short amount of time.

FURTHER INFORMATION Questions concerning the paper can be addressed to: Thermaflex International Holding B.V. Veerweg 1 5145NS Waalwijk The Netherlands

REFERENCES [1] NEN-EN 15632 and NEN-EN-ISO 8497

Liandon B.V. Dijkgraaf 4 6920AB Duiven The Netherlands

[2] P. Wallentén, ―steady-state heat loss from insulated pipes‖, Lund Institute of Technology, Sweden, 1991 [3] J. Korsman and G. Baars, ―Heat loss system optimization‖, 12th ISDHC 2010

ACKNOWLEDGEMENT Acknowledgments go to both the test institutes for their open and honest explanation of their testing methods and facilities and for even showing the entire facility and methods.

[4] E. van der Ven and R. van Arendonk, ―Heat loss analysis and optimization‖, 12th ISDHC 2010 [5] C. Engel and G. Baars, ―New economical connection solution for flexible piping systems‖, 12th ISDHC 2010.

Furthermore acknowledgements go to all involved employees of Thermaflex Isolatie B.V. and Liandon B.V. who made this research possible. Special acknowledgements go to P. Blom and P. van Rijswijk

[6] I. Smits and E van der Ven, ―Performance of pre insulated pipes‖, 12th ISDHC 2010.

257

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

DISTRICT HEATING AND COOLING WITH LARGE CENTRIFUGAL CHILLER-HEAT PUMPS Ulrich Pietrucha Friotherm AG, Switzerland ABSTRACT

Number of units

With prices for primary energy resources soaring, the recovery of "waste energy" was getting into the focus of attention within the last years. Also the global climate change reminded us to limit the use of primary energy resources to a minimum, thus exploiting "waste energy" potentials wherever feasible. The process of upgrading low grade waste heat is especially interesting where large amounts of such energy are available at one point, e.g. next to sewage water treatment plants, alongside main sewers, in power plants or close to ground water sources.

Type

UNITOP® 50 FY/34FY

Refrigerant

R134a

Heat source medium

Raw waste water

Raw sewage water inlet

10.0 °C ... ~ 15 °C

Raw sewage flow water flow

3800 m3/h

Heating water temp. in/out

60 / 90 °C

Heating water flow

824 m3/h

Power at terminal

9‘750 kW

Heat capacity

27‘600 kW

Coefficient of performance

2.83 up to >3.0

Even if the "waste energy" potential is abundant and easily exploitable, the aspect of overall thermal efficiency is considered crucial for the final decision to invest in large heat recovery installations. INTRODUCTION Described are five applications of large centrifugal heat pumps-chillers for the use in large district heating/cooling systems. Application 1: Heat recovery from raw sewage water and hot water production at 90 °C.

One of the Skoyen heat pumps

Application 2: Combined heating and cooling with a raw sewage water heat pump/chiller installation. This plant is operated successful since 1989.

2. THE SANDVIKA PLANT IN OSLO: COMBINED HEATING AND COOLING FROM A RAW SEWAGE WATER HEAT PUMP

Application 3: Combined heating and cooling: a combination of cooling with simultaneously heat production in summer and heat recovery from cleaned sewage water in winter.

This is the oldest combined chiller/ heat pump installation in the world, producing simultaneously cooling, taking out heat from raw sewage water and producing heating capacity for the district heating system.

Application 4: Heat recovery from wet flue gas cleaning process Application 5: Combined heating and cooling in Stockholm

The heat pumps are in successful operation since 1989 and each one has an additional heat exchanger, which is used either as raw sewage evaporator or as raw sewage water condenser.

1. SKOYEN VEST PLANT IN OSLO: HOT WATER PRODUCTION AT 90 °C This is the world's largest heat pump plant using raw waste water as heat source. It is installed in a cavern alongside one of the main waste water channels in Oslo. With 2 heat pumps a heating capacity of 27'600 kW is generated by recovering heat from raw waste water.

Each heat pump has an overall operating time of about 160'000 hours, means the heat pumps were operated since 1989 each year for more than 8'400 hours. A 3rd larger heat pump was taken into operation in 2008.

258

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

points with lower district heating temperatures where a COP of up to 6.5 can be achieved.

2 Friotherm heat pumps Type UNITOP 28CX-71210U

Heat source capacity 15.5MW 24.3°C

From districtheating-network

Heat sink capacity 19MW

50°C

Flue-gas Flue- gascondensing

Combined chiller / heat pump at Sandvika plant

Flue-gascleaning

3. THE KATRI VALA PLANT IN HELSINKI: COMBINED HEATING AND COOLING

Power consumption 3.5MW

34.2°C

Steam turbin e

This is the largest combined chiller heat pump installation in the world producing simultaneously 60MWth cooling and 90MWth heating, i.e. total produced thermal energy is 150MW. The required electrical input is 30MW i.e. a superb COP of 5 can be achieved (150MW / 30MW).

59.2°C

Generator

Boiler

Waste-to-Energy plant SYSAV Malmö Sweden

During Winter season the required cooling is done by sea water, while heat is produced by using cleaned waste water as heat source.

5. NIMROD STOCKHOLM: COMBINED HEATING AND COOLING Due to the fact that with every cooling process there is also waste heat generated, Friotherm AG, which has worked since many years on chillers with heat recovery, has worked out a concept which allows various operating modes in order to operate the chiller / heat pumps more efficient over a longer period and, making therefore the investment more attractive: There are 4 chiller / heat pumps installed in the Nimrod plant. The centrifugal compressors are switched in parallel for Summer cooling production of 48MW. However during this period heat recovery is not required as there is sufficient capacity available from the existing heat pumps.

Typical Unitop 50FY heat pump (Qheat 15 to 23MW)

The same units are producing during Spring, Autumn and Winter a cooling capacity 24MW with a full heat recovery of 35.6MW at a temperature level of 78 °C. For heat recovery operation mode the centrifugal compressors are switched in series.

4. SYSAV MALMÖ: HEAT RECOVERY FROM WET FLUE GAS CLEANING PROCESS SYSAV Malmö in Sweden has built a new waste-toenergy plant. An important part in this plant was the installation of a 19MW heat pump using the flue gas condensation as heat source. The heat pump is supplying hot water with a temperature of up to 70 °C to the district heating system of the community of Malmö.

Each chiller / heat pump consists of two centrifugal compressors Type Uniturbo 33CX and 28CX and is able to operate at the following modes, described below:

The two heat pumps are connected in series on the heat source side and on the heat sink side; this improves considerably the COP. There are operating

5.1 Cooling only: During Summer with high cooling demand, the waste heat from the condenser is removed with sea water of 259

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

max. 22 °C, therefore the condenser and sub cooler are equipped with Titanium tubes.

Nimrod single stage operation 2 compressors in parallel

The two compressors Uniturbo 33CX and 28CX are then working in parallel, in a single stage mode, with a single stage expansion, producing a cooling capacity of up to 7MW plus 5MW = 12MW i.e. with 4 units a total of 48MW.

Sea water

Depending on the cooling demand, one or the other, or both compressors can be put in operation.

2 8 8

If needed, the part load of each chiller / heat pump can be controlled down to 10% of its nominal capacity, with a reasonable high efficiency, with the use of inlet guide vanes. The chilled water temperature outlet is kept constant to 5 °C 5.2 Combination of cooling and heating: During Spring, Autumn and Winter, with moderate cooling demand of up to 24MW, but simultaneous need of heating, the waste heat from the condenser is supplied to the district heating network at a temperature outlet of 78 °C and a maximum heat capacity of 35.6MW.

Cooling capacity 12 MW

Nimrod two stage operation 2 compressors in series

The two compressors Uniturbo 33CX and 28CX are then working in series in two stage compression mode, with two stage expansion using an economiser after the first stage expansion.

Heating water

The compressor Type Uniturbo 33CX with the larger volume flow is working as 1st stage and the Type Uniturbo 28CX with the smaller volume flow as 2nd stage compressor.

The control system is controlling the required cooling capacity; the surplus heat is supplied fully to the district heating network at a temperature level of up to 78 °C. I.e. this operation mode delivers heat which can be sold in addition to the cooling, with a total COP of above 5. The cooling only mode and the combination of heating and cooling mode are explained in the below P&I‘s:

Heating capacity 9 MW Cooling capacity 6 MW

CONCLUSION Today developments are the extension of the heat pump operation range in temperature and capacity to exploit new heat sources and to extent the field of applications.

Reliability of technology, future developments and challenges About more than 140 heat pumps, with temperatures above 70 °C, wide since 1980. The heat pump this article are only showing a nowadays available applications.

producing hot water are installed world plants described in small part of the

The adaptation of the centrifugal heat pumps to new refrigerants with GWP close to zero is already on the way.

Almost all of the installed heat pumps plants starting from the early 1980's are today still in operation, which is showing the high reliability of this technology.

REFERENCES Text and pictures from Friotherm AG / Switzerland.

260

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

NEW ECONOMICAL CONNECTION SOLUTION FOR FLEXIBLE PIPING SYSTEMS Christian Engel, Gerrit-Jan Baars Thermaflex International Holding B.V. Together with the University of Leoben, the long term durability of two types of plastic medium pipes, made of PB and PE-X were investigated. The research made by Dipl.Ing. E.Kramer and Univ.Prof.Dr.J. Koppelmann [2] was based on OIT (oxygen induction time), tear strength, elongation at break and internal pressure tests to determine the lifetime of plastic pipes at 80, 95 and 110 °C. The final results were in favour for pipes made of PB. The calculation of the lifetime for PB pipes was based on a typical temperature profile used in secondary district heating networks of STEWEAG. The lifetime expectancy was stated with 36 years.

ABSTRACT Most Energy Companies are facing the same problem: Connection costs per house shall be cheaper and faster to install to reach more customers. At the same time high level of durability and a system free of maintenance must be guaranteed. This paper shall give an insight of the practical experience with new solutions showing the economic and ecological advantages in projects with several Energy companies.

The decision of STEWEAG was made for PB pipes due to their more homogenous structure, superior flexibility and allowance for welded joints. More than 250 km of this system have been installed since 1981 in secondary networks operated by STEWEAG. See also Univ. Prof. Dr. E. Hönninger, STEWEAG [6].

INTRODUCTION District Heating & Cooling networks are a major cost factor for Energy Providers and subject to permanent search for cost improvements. Flexible plastic pipe systems have been a major step for cost reduction in low temperature networks. With the new EN 15632 [1] the necessary basis for certification of these systems has been laid. This is a milestone in terms of acknowledgement for flexible systems as a proven part of future network developments.

WHAT CAN BE SOLVED WITH FLEXIBLE PLASTIC SYSTEMS Apart from the high and long term investment costs, the following main problems had to be solved as well: 

As flow temperatures and pressures are reduced, the field of application for flexible plastic systems is increasing. Until recently only a small percentage of District Heating companies have started to use plastic pipes in their networks. These were kind of pioneers who co-created systems together with the industry.

 

Corrosion problems in conventional Systems made of Steel/PUR/PE or Cu/PUR/PE or Cu/Mineral wool/PE Heat loss due to wet and aged insulation System shut downs for maintenance and repair

The first co-development of such a system was started already in 1980 by the Austrian Electricity company STEWEAG. They were looking for a pre-insulated piping system as easy to install as an electric cable.

Photo no.: 2 corroded steel pipe connection

THE NEW FLEXIBLE PIPE GENERATION In 2001 the Dutch Energy Provider NUON started a development co-operation with Thermaflex to create even more flexible and moisture resistance piping systems. Target was again to reduce the connection costs for new district heating projects. The new system

Photo no.:1 first Flexalen installation 1981 261

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

up to d63 – there trenches can be compared to cable ducts. Another fact is that connections are most of the time only necessary at branches, for sticks trenches need to be suitable for execution of the welding and the insulation process.

developed is called FLEXALEN 600, the improved version of the system used by STEWEAG. The system consists of a PB (Polybutene) medium pipe and PO (Polyolefin) insulation foam welded to a HDPE (high density Polyethylene) outer casing. With a new inline production process it was possible to weld the moisture resistant insulation to the outer casing. The targets of a corrosion proof and moisture resistant insulation were met.

Photo no.: 4 Steel compared to Flexalen

Although material cost for plastic pipes are higher especially for larger dimensions, the total installed system costs are lower, especially when using double pipe systems wherever possible (see Photo nr. 5).

Photo no.:3 Flexalen 600 longitudinal cut

FLEXALEN has been the first system to pass a certification and 3rd party control by KIWA, which is similar to the new EN 15632.

BRANCH SOLUTIONS Until now 2 types of branch solutions have been used. On site welded solution with Half-shells plus insulation to cover (see Photo no.:4). This technology has been used for smaller networks. Due to homogenous welding techniques, either with polyfusion or with electrofuction fittings, the branches are corrosion free and offer the same inner diameter as the pipes.

DECREASING INSTALLATION COSTS The most obvious advantage is the chance to reduce the installation time with flexible systems supplied in coil lengths of 100m and more. Compared to rigid systems the following relative costs have been realized in actual projects Table 1 Pre-insulated steel pipes

FLEXALEN

Material costs

100%

90–150%

Installation time

100%

20–25%

Trenching

100%

50–70%

Total

100%

60–85%

Material costs are depending on the dimensioning of the system in the first place. In case of optimization of pipe sizes and lengths according to the advantages of PB pipes and connection systems, as described later on, the material costs can be reduced for Flexalen.

Photo no.: 5 Polyfusion welded Flexalen branch

Pre-insulated Tees are another way to secure more reliable network quality due the reduced number of joints to be made on site (see Photo no.: 6). The straight connections are insulated with special kits with a robust slide over HDPE tube, which is sealed to the outer casing with heat shrinks.

Installation costs are proven in practical experiences since almost 30 years. Flexalen systems can be installed 5 times faster than rigid systems. Lower costs in trenching is related to the fact that Flexalen systems are supplied in double line systems 262

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Photo no.: 7 Narrow space for district heating lines under buildings in NL

The result was the ―Flexalink (Flexalen T-Link)‖ solution a very small, flexible and 100% watertight system, prefabricated and pressure tested by Thermaflex NL.

Photo no.: 6 Pre-insulated branch

The pressure for even more economic solutions for connections has led to further innovations in close cooperation between the Dutch Energy Provider Eneco and Thermaflex.

Photo no.: 8 Flexalen T-Link

PRE-FABRICATED NETWORKS

This solution combines the following advantages:

Compared to the branch solutions described before, these new solutions take full advantage of flexible welded systems, in order to further reduce the number of joints on sites.

 

A new type of pre-fabricated network has been developed. High flexibility and a minimum of connections was the goal.

 

The first application was the district heating network Capelle a/d Ijssel in the surroundings of Rotterdam for renovation in difficult circumstances under houses (high ground water level) to replace corroded heating and sanitary distribution systems.

 

The space under the houses is so small that neither welding nor mechanical connections can be carried out in a safe way. Steel welding is even forbidden under these conditions.



Factory made welding and branch insulation – all watertight and pressure tested. Connections are made under clean manufacturing circumstances. No weather influences, no failure costs. Customer made connections according to the real situation. Light weight and flexible for easy sliding into the trench or under the house Fast Installing time (first project experience 10 houses/day) Reducing connection costs in new building projects. Less system parts on the building area.

This development covered all the wishes and requirements of the Dutch Energy Provider ENECO. They have already ordered this system for 800 house connections in 2010. 263

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Photo no.: 9 Installation of Flexalen T-Link

Photo no.:12 Welding of T-Link to the next section

All experiences so far have been very positive and have led to further applications already. RENOVATION IN PURMERENT The situation of the current district heating network in Purmerend (Energy supplier Stadsverwarming Purmerend) is very critical. Due to higher ground water levels than expected the current metal pipe systems have corroded and need to be replaced.

Photo no.: 10 Installation of Flexalen T-Link

These 2 photos (no 8 and 9) demonstrate how easy it was to slide the connection into the duct under the house.

As the network has been installed under the basement of the attached housing schemes, the space for the installer is very tight and it is not allowed to use any steel welding process in these circumstances. Flexalen T-Link has been identified to be the ideal solution.

Also further connections between the pre-fabricated sections were made before sliding the entire system under the house. Only the last connection had to be made in the duct. See also photos no.:11 & 12.

Purmerend has ordered this system for 300 house connections for 2010 already. For the renovation market in the Netherlands, this solution has shown big advantages. This solution is now available for Energy Provider worldwide not just for renovation, but also for new networks. PRE-FABRICATED NETWORKS FOR NEW PROJECTS The conditions in new building situations are much easier and this solution can help to reduce the connection costs.

Photo no.: 11 pipes under the building

Both Dutch Energy Provider, Eneco and Nuon are investigating this new solution for new building projects. 264

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Fig. 1 network scheme suitable for T-Link

Especially for networks with short distances between the branches and the connection to the houses, a high degree of pre-fabrication can be offered.

PB systems can be operated with much higher flow speed; hence smaller dimensions can be used for the same load requirement. See also J. Korsman, I.M. Smits, E. van der Ven [4].

One solution is the T-Link as described before. Another one can be a main line up to 100 m with factory welded and insulated Tees. This reduces the work on site to just 1 welding for the house connection line. This solution is interesting for longer distance house connections.

With relation to the topic of this paper, the following additional savings can be made during the network design: Looking for a new building area, mostly streets with block of houses there are two possibilities: Installing under the floor or Installing in the streets.  For every house connection under the floor only two welds and two insulation sets are necessary. Reduction of the installing time/costs by 50–60%.  For every house connection in the street, a pre-fabrication e.g. for 8–10 house-connections built in into one 100m coil can safe installing time/costs totally including excavating the trenches of 70% Taking all these possibilities for savings and optimization into account, the next most important topic for Energy Provider, the efficiency of the network in operation, can be tackled as well. Due to the possible reduction in network length and pipe diameter, the overall heat loss can be reduced as well. See also results from the work of I.M. Smits, J. Korsman, J.T. van Wijnkoop and E.J.H.M. van der Ven [5] and J.T. van Wijnkoop, E.J.H.M. van der Ven [3]. 

THE IMPORTANCE OF THE NETWORK DESIGN Flexible PB piping systems offer important advantages compared to other plastic and steel systems in terms of layout and design. Compared to steel pipes flexible PB systems can be laid more direct as the system is flexible and fully selfcompensating. Expansion loops and elbows can be saved. The saving in pipe length can be calculated with 7–10%. PB systems offer low friction loss and show no calcification or incrustation during the lifetime. The polyfusion welded fittings have at least the same inner diameter as the pipe and offer the same high abrasion resistance. Taking this into account, some extra security factors used in pipe dimensioning can be eliminated. 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 requirements of Energy Companies in terms of lower investment costs, faster installation time and durability, have been co-created with leading Energy Suppliers in Austria and The Netherlands.

[1] EN 15632 District heating pipes, Pre-insulated flexible pipe systems, Requirements and test methods [2] Dipl. Ing. E. Kramer, Univ.Prof. Dr. J. Koppelmann, „Untersuchung zur Dimensionierung einer flexiblen Fernwärmeleitung aus Kunststoff―, University Leoben, Austria; 1984.

These solutions are based on flexible and weld-able plastic systems and have been used successfully up to 29 years in secondary networks with maximum operation temperatures of 95 °C (peak temperature) and maximum pressure of 8 bars.

[3] J.T. van Wijnkoop, E. van der Ven, ―Verification of heat loss measurements‖, 12th ISDHC 2010. [4] J. Korsman, I.M. Smits, E.J.H.M. van der Ven ―Heat loss analysis and optimization of a flexible piping system‖, 12th ISDHC 2010.

The latest development is going into the direction of a higher degree of pre-fabrication, by including the entire connection line to the houses as well as parts of the main line into one piece, made up and fully pressure tested in the factory.

[5] I.M. Smits, J. Korsman, J.T. van Wijnkoop and E.J.H.M. van der Ven, ―Comparison of competitive (semi)flexible piping systems by means of heat loss measurement‖, 12th ISDHC 2010.

The experiences in recent projects are showing installation times 5–10 times faster compared to conventional pre-insulated steel. The number of connections to be made on site is significantly reduced.

[6] Univ. Prof. Dr. E. Hönninger, „Sekundärnetze fördern die Fernwärmeanwendung―, STEWEAG, Fernwärme International 14/85.

Successful projects with Energy Suppliers in The Netherlands are confirming the advantages of this new connection solution.

[7] C. Engel, „Polybutene – The alternative material for 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

COMPETITIVENESS OF COMBINED HEAT AND POWER PLANT TECHNOLOGIES IN ESTONIAN CONDITIONS 1

E. Latõšov and A. Siirde 1

Department of Thermal Engineering Tallinn University of Technology, Tallinn, ESTONIA energy consumption effective methods, where CHP production from the renewable fuels is preferable. [1]

ABSTRACT The goal of this paper is to evaluate competitiveness of market ready combined heat and power (hereafter called as ―CHP‖) technologies for CHP expansion potential locations in Estonian. The main criteria to indicate preference of CHP technology is a heat price by which the internal rate of return is equal to investors‘ expectations.

During the last 2 years a few CHP plants working on woodchips and peat were build in Estonia. A few of biomass CHP plants are under active development. All of them are planed or constructed in major Estonian cities and are based on backpressure steam turbine technology. At the same time feed-in tariffs as well as possibilities to get grants for expanding of CHP and usage of renewable fuels makes CHP expansion more attractive for locations with a lower heat demands.

Calculation results shows, that in spite of the advantages of gas engines (relatively low investment costs and high electrical efficiency) the calculated heat prices are the highest. Heat price for expected 7% IRR is 53–61 EURO/MWhheat depending on heat demand. It is mainly because of relatively high natural gas price. Under 5 MWel ORC is competitive to steam turbine/engine technology. Heat prices are lower for 1–4 EURO/MWhfuel, depending on heat demands.

Steam turbine technology is a classic for CHP plants. But in relatively small-scale boilers and district heating systems use of steam turbines is connected to economically less efficient operation (commonly higher specific investment costs, O&M costs and lower electrical efficiency) where use of other alternative CHP technologies could be preferable.

Heat prices for places with annual heat demand under 20 000 MWh are mainly above 45 EURO/MWhfuel (average heat prices for biomass boiler houses in Estonia is between 40–45 EURO/MWh). Developing of CHP plants in such areas is feasible in the case of grant payments for investments. CHP plant development based on wood chips or peat could be feasible without grant payments in the places where heat demand exceed 30 000–40 000 MWh annual. Carefully selected CHP technology and capacity can afford higher IRR when keeping competitive heat prices.

The goal of this paper is to evaluate competitiveness of market ready CHP technologies for CHP expansion potential locations in Estonian. The main criteria to indicate preference of CHP technology is a heat price by which the internal rate of return (hereafter called as ―IRR‖) is equal to investors‘ expectations. The paper is structured as follows. After an overview of places where construction of CHP plants can be reasonable the paper provides principles for evaluation of CHP technologies competitiveness. Next sections provide an overview of the CHP technologies which can be used in CHP plants and descriptions of main fuel sources for energy production in Estonia. The last section provides heat price calculation examples based on proposed principles for evaluation of CHP technologies competitiveness.

The most feasible places for CHP expansion in Estonia are Maardu, Viljandi, Rakvere, Valga, Haapsalu, Võru, Paide and Põlva. INTRODUCTION This paper draws on ongoing project ‗Analysis on the technical and economic consequences of renewable energy based CHP systems in new areas with the lowered useful heat demand or after implementation of energy conservation measures in the areas with older buildings‘ within the project ‗Primary Energy Efficiency‘ partly financed by NER, which contributes to the effort of enhancing the primary energy efficiency (PEE) and reducing CO2 emissions in the energy sector.

LOCATIONS OF POTENTIAL BIOFUELED CHP PLANTS IN ESTONIA Fig. 1 shows major Estonian cities and municipalities which are distributed by the annual heat demands. Places where CHP plants are already constructed or under construction, as well as in a state of active development are marked separately. Fig. 1 reflects well known principles, where the consumers with higher annual heat consumptions are more preferable.

Present-day world energy policy is based on two main directions: energy efficiency and environmental protection. Efficient CHP production is one of the 267

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 TECHNOLOGIES COMPETITIVENESS

Income from the heat sells depends on amount of sold heat (computable value) and heat price.

The revenues of a CHP company are generated from the heat and electricity sales. Theoretically they must cover the operation and maintenance (hereafter called as ―O&M‖) costs of the CHP plant completely and provide an expected IRR. Main CHP plant related costs and incomes are shown in Fig. 2.

Knowing investment costs (specified in section CHP technologies), and other above mentioned costs and incomes the power plant operation annual net cash flows can be calculated and IRR defined. The principle for evaluation of CHP technologies competitiveness is based on finding such heat price which will cause an expected (proposed) IRR, where calculation/estimation rules for the other cash flows components are clearly defined. CHP TECHNOLOGIES There are numerous CHP technologies that can be theoretically used for small scale CHP systems, but not all of them are economically and technically feasible. The list of main CHP technologies ordered by market readiness and common heat outputs are shown in Fig. 3.

Fig. 2 CHP plant incomes and costs

Fuel costs, pollution charges and ash handling costs are mainly depend on used fuel properties and are estimated in section Fuel sources for energy production. CHP technology related fixed operation and maintenance costs depend on selected CHP technology and are defined in % from the investment costs annual. They are estimated in section CHP technologies. Electricity sells depends on amount of produced electricity (computable value) and fuel prices. Fuel prices are estimated by taken into account feed-in tariffs described in Electricity Market Act [2].

It is important to consider the market ready solutions first of all, such as a steam turbine (hereafter called as ―ST‖), steam engine (hereafter called as ―SE‖), ORC technology (hereafter called as ―ORC‖) and gas engine (hereafter called as ―GE‖). Hereafter SE and ST are considered jointly, where capacities less than 1 MWel correspond to SE by default. For CHP plant economical calculations it is important to know such CHP plant parameters as efficiencies, price and O&M costs. Above mentioned parameters are obtained and systemized on the basis of information regarding CHP plants collected from different information sources such as [3, 4, 5, 6, 7, 8].

268

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Fig. 3 Main prime mover CHP technologies

Values for investments as well as nominal electrical capacities used to calculate fuel prices depend on CHP nominal electrical capacities and CHP technology, as shown in Table 1. Table 1. Values for investments and nominal electrical capacities for selected CHP technologies Capacity

Specific investment costs

Electrical nominal efficiency

MEURO/MW el

ST/SE

10,3

Technology

MW el 0,1 0,1

1,6

ST/SE

5,1

ORC

5,8

1,0

ST/SE

3,2

ORC

4,5

0,8

ST/SE

2,9

ORC

4,2

0,8

heat load is 35% for steam engine/turbine, 80% for gas engine and 85% for ORC from the nominal electrical efficiency. It is assumed, that CHP technology related fixed O&M costs for SE/ST, ORC and GE are relatively 2.5%, 2% and 3.5% from the investment costs annual. FUEL SOURCES FOR ENERGY PRODUCTION Main fuel sources for under 10 MWel CHP plants in Estonia are natural gas, peat and wood chips. Fuel prices The fuel prices taken as basis for heat price calculations are as follows:

In this paper investment means all costs before CHP plant commissioning.



Peat price – 11.7 EUR/MWhfuel. Proposed price is based on average peat price levels obtained from Tootsi Turvas AS, the biggest peat milling and exporting enterprise in Estonia.



Wood chips price – 12.8 EUR/MWhfuel. Proposed price is based on latest data, published by Estonian Institute of Economic research in their web based price information system [9].



Natural gas – 35 EUR/MWhfuel. Proposed price is an average price for the latest data published by Statistics Department of Estoni [10].

For the evaluation of CHP competitiveness the efficiency drop working at partial load is taken into account. It is assumed, that minimal CHP heat load for all technologies is 25% from the nominal heat load. It is assumed, that electrical efficiency working at minimal 269

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Ash handling costs Ash handling costs assumptions, that:



calculations

are

based

Peat ash content is 5%. Average calorific value is 3.3 MWh/t [11];

Taking into account above mentioned information the ash handling costs per MWh of fuel energy content for the peat and wood chips are ~0.19 and 0.72 EURO respectively. Pollution charges



Wood chips ash content is 1%, calorific value 2.4 MWh/t;



Natural gas combustion does not emit any ash;

Pollution charges and levels are calculated base on the Environmental Charges Act [12], Regulation No 99/2004 [13] and No 94/2004 of Estonian Minister of Environment [14].



Regarding to information obtained from different landfill owners, an average for year 2012 expected ash removal costs (ash transportation to landfill, and storing) are 45 EUR/t.

The method described in [13] takes into account different combustion technologies, flue gas cleaning technologies, control devices as well as capacities to define emission factors of pollutants.



The combustion plant is equipped with dry ash removing system.

Table 2. Summarised results of the heat price calculations for different CHP expansion scenarios District heating area Heat demand MWh 5000 5000 5000 10000 10000 10000 10000 10000 20000 20000 20000 20000 20000 40000 40000 40000 40000 40000 80000 80000 80000 80000 80000

CHP plant capacity

Maximum heat capacity Technology

Investment

ST/SE ST/SE Gas engine ST/SE ST/SE ORC ORC Gas engine ST/SE ST/SE ORC ORC Gas engine ST/SE ST/SE ORC ORC Gas engine ST/SE ST/SE ORC ORC Gas engine

Heat

Electrical

MW h

MW e

Peat Woodchips Natural gas Peat Woodchips Peat Woodchips Natural gas Peat Woodchips Peat Woodchips Natural gas Peat Woodchips Peat Woodchips Natural gas Peat Woodchips Peat Woodchips Natural gas

0,83 0,83 0,83 1,65 1,65 1,65 1,65 1,65 3,30 3,30 3,30 3,30 3,30 6,60 6,60 6,60 6,60 6,60 13,20 13,20 13,20 13,20 13,20

To avoid complexity of the analysis to be issued from different combinations of capacities, combustion technologies, fuel gas cleaning and control equipment it is assumed that:

 



IRR 7%

IRR 12%

Fuel

MW 1,5 1,5 1,5 3 3 3 3 3 6 6 6 6 6 12 12 12 12 12 24 24 24 24 24

Heat price, EURO/MWh

MEURO

0,13 0,13 0,63 0,27 0,27 0,34 0,34 1,36 0,60 0,60 0,70 0,70 2,94 1,44 1,44 1,44 1,44 6,40 3,90 3,90 2,96 2,96 14,00



1,31 1,31 0,77 2,54 2,54 2,93 2,93 1,29 4,46 4,46 4,94 4,94 2,65 7,08 7,08 8,10 8,10 5,34 14,57 14,57 15,20 15,20 11,66

Without grant 55 53 61 54 51 53 49 56 49 46 46 42 55 41 38 40 36 53 39 35 38 34 53

With Without With grant grant grant --67 --39 65 44 --66 ----65 --37 62 43 --66 --32 61 38 --60 ----59 --34 56 39 --57 --29 53 34 --59 ----49 --29 45 32 --49 --25 44 29 --57 ----47 ----42 ----46 ----42 ----57 ---

Combustion plant is equipped with the most effective flue gas treatment technology mentioned in [13].

Calculated levels for pollution charges for year 2013 are:

Thermal capacity of combustion plants is below 50MW; Selected combustion technology provides lowest emission level than the others in [13] mentioned combustion technologies; Combustion plant is equipped with the most effective control systems mentioned in [13];

  

270

~0.07 EUR/MWhfuel for wood chips; ~0.95 EUR/MWhfuel for peat; ~0.43 EUR/MWhfuel for natural gas.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Heat prices for places with annual heat demand under 20 000 MWh are mainly above 45 EURO/MWhfuel where an average heat prices for biomass boiler houses are between 40–45 EURO/MWh [16]. Developing of CHP plants in such heat demand areas is feasible in the case of receiving of grant payments for investments.

HEAT PRICE CALCULATIONS Calculations of heat prices are provided in correspondence with principles described in section Principles for evaluation of CHP technologies competitiveness. Heat prices are evaluated for different scenarios. Scenarios include described heat demands, considered fuels and technologies.

CHP plant development based on wood chips or peat could be feasible without grant payments in the places where heat demand exceed 3000-40000 MWh annual. Carefully selected CHP technology and capacity can afford higher IRR when keeping competitive heat prices.

Heat prices are calculated for 7% and 12% IRR. Heat price for CHP plant developing scenarios which satisfy the requirements described in regulation [X], which define conditions for grant payments to expand renewable energy production and construction of CHP plants in Estonia, are calculated separately.

The most feasible places for CHP expansion in Estonia are Maardu, Viljandi, Rakvere, Valga, Haapsalu, Võru, Paide and Põlva.

For calculating heat prices in addition to information from previous paper sections, some other figures have to be specified:

    

Calculation results are valid for assumed cases only. Other particular cases should be calculated individually.

Cash flows are calculated for 20 years; CHP starts energy production in the beginning of 2013; Expected rate of inflation is 1.5%; Heat loses in district heating network are 15%; Heat load profile is estimated based on heat load model described in [15] taking as a basis the heat load duration curve shape of Tallinn.

REFERENCES [1] C. Dötsch and A. Jentsch, ―District heating (DH) in areas with low heat demand density (HDD): A chance for the integration of renewable energy sources (RES)‖, 10th International Symposium on District Heating and Cooling, 3–5, September 2006, p. 2 www: http://www.lsta.lt/files/events/20_doetsch.pdf [20.01.2010]

The results matrix of heat price calculations is shown in Table 2.

[2] Electricity Market Act www: https://www.riigiteataja.ee/ert/act.jsp?id=13279771 [14.05.2010]

CONCLUSION The technologies for smaller CHP applications are more expensive (specific price) and less efficient than those for larger CHP plants.

[3] Schwaiger, H., Jungmeier, G, (2007) Overview of CHP plants in Europe and Life Cycle Assessment (LCA) of GHG emissions for Biomass and Fossil Fuel CHP Systems CIBE Conference „Cogénération biomasse dans l'industrie et sur les réseaux de chaleur opportunités – retours d'expérience-perspectives―

At present peat is considered as a good alternative for wood chips. Lower fuel price (11.7 EUR/MWh) smooth over higher than for wood chips ash handling costs and pollution charges. At the same time wood chips are more preferable because of higher feed-in tariffs for produced electricity.

[4] Obernberger, I., Thek, G, Techno-economic evoluation of selected decentralised CHP appications based on biomass combustion in IEA partner countries Graz (2010)

The advantages of gas engine CHP plants are relatively low investment costs and high electrical efficiency. But because of high natural gas price (MWhfuel price is 2.5–3 times higher than for wood chips and peat) and relatively high fixed O&M costs the calculated heat prices are the highest. Heat price for expected 7% IRR is between 53 and 61 EURO/MWhheat depending on heat demand.

[5] Bryson, T., Major, W., Darrow, Ken. Assessment of On-Site Power. Opportunities in the Industrial Sector, Carlsbad (2001) www: http://www.uschpa.org/files/public/Assessment%20 of%20Onsite%20Power%2001.pdf [14.05.2009]

Under 5 MWel ORC is competitive to SE/ST technology. Calculated heat prices are lower for 1–4 EURO/MWhfuel, where higher fuel price difference corresponds to places with lower heat demands.

[6] Kirjavainen, M., Sipilä, K., Savola, T. Small-scale biomass CHP technologies. Situation in Finland, 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

www: http://www.opetchp.net/download/wp2/small_scale_biomass_chp_t echnologies.pdf [14.05.2010]

[12] Environmental Charges Act, [14.05.2010] www: http://www.riigiteataja.ee/ert/act.jsp?id=13316043 [14.05.2010] [13] Procedure and Methods for Determining Emissions of Pollutants from Combustion Plants into Ambient Air www: http://www.riigiteataja.ee/ert/act.jsp?id=789462 [14.05.2010]

[7] Institute for Thermal Turbomachinery and Machine Dynamics, Cogeneration (CHP) Technology Portrait, Vienna (2002) www: http://www.energytech.at/pdf/techportrait_kwk_en.pdf

[14.05.2010]

[14] Välisõhu eralduva süsinikdioksiidi heitkoguse määramismeetod www: http://www.riigiteataja.ee/ert/act.jsp?id=127572 15 [14.05.2010]

[8] U. S. Environmental Protection Agency Combined Heat and Power Partnership, Biomass Combined Heat and Power Catalog of Technologies, (2007) www: http://www.epa.gov/chp/documents/biomass_chp_ catalog.pdf [14.05.2010]

[15] Latõšov, E., Siirde, A. (2010). Heat load model for small-scale CHP planning. In: Proceedings of International Conference on Renewable Energies and Power Quality: International Conference on Renewable Energies and Power Quality (ICREPQ‘10), Granada (Spain), 23-25th March, 2010., 2010.

[9] Estonian Institute of Economic research www: http://www.ki.ee [14.05.2010] [10] Statistics Estonia www: www.stat.ee [14.05.2010] [11] Paappanen, T., Leinonen,A. Fuel peat industry in EU, 2005, p. 134 www: http://turbaliit.ee/index.php?picfile=21 [14.05.2010]

[16] Estonian Competition Authorities approved district heat maximum prices (without VAT) to end-users www: http://www.konkurentsiamet.ee/file.php?15416 [14.05.2010]

272

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

DISTRIBUTION OF HEAT USE IN SWEDEN Margaretha BorgstrĂśm, Sven Werner 1

School of Business and Engineering Halmstad University, PO box 823, S-301 18 Halmstad Sweden corresponding to 310 million square metres in multifamily buildings and service sector premises. The survey sample thus constituted a sizable portion of the entire building stock.

ABSTRACT The current heat use refers normally to the average heat use in a country or a sector during the course of a year. But it is also important to be aware of the distribution of high to low use when estimating the potential for reducing total heat use.

This energy statistical data, published in the annual reports from Statistics Sweden, have been supplemented with a deeper analysis of the distribution of the heat use and the systematic causes regarding high heat use. Independent variables for explanation of variations were number of degree-days, construction year, ventilation system, energy efficiency measure, and co-use of heat supply. High and low users were also analysed by location, construction year, heat supply method, ownership, and building size [3]. In this short paper, the specific heat use will be presented by its distribution, construction year, degree days and energy efficiency measures.

Energy statistical data published in the annual report from Statistics Sweden have been supplemented by a deeper analysis of distribution of heat use and systematic causes regarding high heat use. The aim of this paper is to explain the variation in heat use with respect to construction year, degree days and energy efficiency measures. In the Swedish energy efficiency debate, many voices refer to systematic causes for high heat use. However, the results from this study do not support this opinion, since the use distribution mostly comes from individual causes. The most important implication of the study results is that systematic policy measures will have a low impact on the total national energy efficiency.

1. Distribution of heat use The total distribution of specific heat use as a function of the percentage of the building area of all multi-family buildings and service sector premises in Sweden is shown in Fig. 1.

INTRODUCTION Multi-family residential buildings and service sector premises constitute 80% of the customer stock in the Swedish district heating systems. The level of future heat use in these buildings will thus have a strong influence on the future district heating economy and the corresponding investment demand. It is therefore of interest to collect information and make analyses of the costumer heat use and how the heat use will develop in the future.

Heat use kWh/m2 400 Multi-family buildings

350

Premises

300 250 200 150 100 50

Specific heat use in multi-family buildings and service sector premises has decreased considerably since the 1970â&#x20AC;&#x2DC;s. In 2006, the specific heat use in multi-family buildings has decreased by 38% compared to the heat use in 1972. The lower heat use is due to increasing energy prices and more energy efficient buildings.

0 0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Share of all national building space

Fig. 1 Heat use distribution during 2006 as a function of the share of all national building space. The diagram is an estimation for all multi-family and service buildings in Sweden.

An extensive study of the current heat use for buildings in Sweden has been performed. The input information for this study was constituted by the anonymous responses to the annual survey of energy use in multifamily buildings and service sector premises performed for 2006 by Statistics Sweden, [1] & [2]. The responses provided input data from 11253 buildings having a total area of 77.6 million square metres. By using scaling factors, estimates could be made for the entire country,

The area under each curve is the total heat used in multi-family buildings and service sector premises during 2006. The figure shows that 13% of the area in multi-family buildings had a specific heat use of more than 200 kWh/m2, and 12% of the area in service sector premises had a specific heat use of more than 200 kWh/m2. This result shows that there are no major differences between the percentages of the building 273

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The relationship between construction year and high heat use in buildings has been analysed. The definition of high heat use is 200 kWh/m2 or more. Fig. 4 shows the results for multi-family buildings. There were a total of 179.3 millions square metres in multi-family buildings in 2006 and 13% of the heated area had heat use of at least 200 kWh/m2.

area with high heat use in multi-family buildings and service sector premises. The results in Fig. 1 also show that 11% of the building areas in multi-family buildings, and 31% of the building area in service sector premises have a specific heat use lower than 100 kWh/m2. Buildings with heat use 200 kWh/m2 or more have been further analysed and the results are presented in the following section considering construction year.

Of special interest are buildings built during the period 1965–74, when a large part of the existing buildings in Sweden were built. During this period there were no requirements for low energy use in buildings.

2. Construction year

In multi-family buildings built during the period 1965–74, 30% of the total area had heat use of at least 200 kWh/m2 and for buildings built in the period 1941–60. 42% of the total building area had heat use of 200 kWh/m2 or more.

Fig. 2 shows specific heat use in multi-family buildings as a function of construction year. The figure also includes the average value each year, together giving the total average specific heat use of 152 kWh/m2. There are no major differences in heat use in buildings constructed before 1980. After 1980, the heat use was approximated 15% lower than the average heat use for all buildings in Sweden. 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 by construction year.

50 0 1930

1940

1950

1960

1970

1980

1990

2000

2010

Construction year

Fig. 2 Specific heat use as a function of construction year for 4285 multi-family buildings

Heat use kWh/m2 400 350 300 250 200 150 100

Fig. 5 Total square metres where heat use is higher or equal to 200 kWh/m2 in service sector premises categorised by construction year.

50 0 1930

1940

1950

1960

1970

1980

1990

2000

2010

Construction year

Fig. 3 Specific heat use as a function of construction year for 4061 service sector buildings.

The relationship between high heat use in service sector premises and construction year is shown in Fig. 5. During 2006, 15.6 million square metres had heat use of at least 200 kWh/m2. Service sector premises built between 1965 and 1974 had high heat use in 3.5 million square metres.

The heat use in service sector premises is shown in Fig. 3. Also in these buildings, the average heat use after 1980 is lower (about 10%) than the average heat use in all service sector premises in Sweden. 274

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

in the theoretical analysis of the optimal wall insulation as a function of degree-days.

The results show that the period 1965–1974 did not have a dramatically higher heat use in the construction year analysis.

The results show that the average difference between Northern and Southern Sweden was small, implying a small climatic impact on heat use. The main conclusion from this analysis is that the individual variation in each climate area is much higher than the local impact of climate. This astonishing conclusion can have several different explanations:

3. Degree days The climate in Sweden varies with a much colder climate in the northern part compared to the southern part. Since the statistical data consist of buildings from different parts of Sweden, the influence of the local climate on the heat use in buildings can be analysed. This has been done by analysing the correlation between the number of degree days for the location of a building and the corresponding specific heat use. The number of degree days, according to the Swedish definition, varies from approximately 3000 in the south up to 7000 in the north of Sweden. Each building in the analysis was connected to one of 14 climate areas.



Higher awareness and consequences of low building heat resistances in Northern Sweden



Lower regional GDP in Northern Sweden giving higher incentive to reduce heat costs



More frequent snow cover in Northern Sweden giving extra heat resistance during the winter.

4. Energy efficiency measures. The statistical data shows the energy efficiency measures during the period 1995–2005. The energyefficiency measures were:

Heat use kWh/m2 400 350

Supplementary insulation

More energy efficient windows

Balancing heating- and ventilation systems

100

Electrical efficiency measures

Heat recovery in ventilation systems

300 250 200 0,28

150

0 2000

y = 15,63x

2500

3000

3500

4000

4500

In multi-family buildings, one or several energy efficiency measures were implemented for an estimated floor area of 57.6 million square metres during the period 1995–2005. No energy efficiency measures had been performed for an estimated floor area of 92.2 million square metres during the same period.

5000

Degree-days

Fig. 6 Specific heat use for 5111 multi-family buildings as a function of the number degree days in each climate area.

Heat use kWh/m2

In service sector premises, with an estimated floor area of 37.2 million square metres, one or several measures had been taken during the period 1995-2005. During the same period, no measures had been taken for an estimated floor area of 70.3 million square metres.

400 350 300 250 200 150 100

The most common measures in multi-family buildings and service sector premises were balancing of heatingand ventilation systems.

y = 10,37x0,30

50 0 2000

2500

3000

3500

4000

4500

5000

In many buildings, a combination of two or several energy efficiency measures had been taken in the same building. In some buildings, up to five measures have been taken in the same building.

Number of degree-days

Fig. 7 Specific heat use for 6041 service buildings as a function of degree days in each climate area.

The average heat use in multi-family buildings and service sector premises in relation to measures taken is shown by bars in Fig. 8 and Fig. 9. The horizontal lines show the average heat use in buildings, in which no energy efficiency measure was performed.

Fig. 6 and Fig. 7 show the specific heat use as a function of degree days for multi-family buildings and service sector premises. The figures also show the average curve and its equation for specific heat use as a function of degree-days. You should also note that the exponent in the fitted equations has only the magnitude of 0.3 instead of the 0.5 exponent obtained 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:

Fig. 8 Average heat use in multi-family buildings in relation to the measures performed. The measure figures correspond to the measures defined in the text.

Fig. 9 Average heat use in service sector premises in relation to the measures performed. The measure figures correspond to the measures defined in the text.



Individual variations dominate compared to systematic causes regarding the specific heat use in multi-family and service sector buildings.



The district heating companies can help their customers by identifying them as high, medium or low users of heat.



On the short term, a significant potential exists for lower heat use in the Swedish multi-family and service sector buildings.



More efficient heat use in buildings will probably represent the most important competitor to district heating supply in the future.



In the Swedish energy efficiency debate, many voices refer to systematic causes for high heat use. However, the results from this study do not support this opinion, since the distribution of heat use mostly comes from individual causes. The most important implication of the study results is then that systematic policy measures will have a low impact on total national energy efficiency.

REFERENCES

As shown in the figures 8 and 9, there were no substantial differences in heat use between buildings where energy-saving measures had been taken and those where they had not. The conclusion from this analysis is that the measures taken during these 10 years were taken by late-comers rather than by early adopters, since heat use after measures were taken generally corresponds to the average level for all buildings.

276

[1]

Statistics Sweden, Energistatistik för flerbostadshus 2006 (Energy statistics for multifamily houses during 2006). Statistiska Meddelanden EN16SM0702.

[2]

Statistics Sweden, Energistatistik för lokaler 2006 (Energy statistics for premises during 2006). Statistiska Meddelanden EN16SM0703.

[3]

Andreasson M, Borgström M, Werner S, Värmeanvändning i flerbostadshus och lokaler (Heat use in multi-family buildings and premises 2006) Fjärrsyn report 2009:4, Stockholm 2009. Available at www.svenskfjarrvarme.se

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

DAMAGES OF THE TALLINN DISTRICT HEATING NETWORKS AND INDICATIVE PARAMETERS FOR AN ESTIMATION OF THE NETWORKS GENERAL CONDITION 1

Aleksandr Hlebnikov , Anna Volkova , Olga Džuba , Arvi Poobus , Ülo Kask

Department of Thermal Engineering, Faculty of Mechanical Engineering, Tallinn University of Technology, Kopli 116, 11712 Tallinn, Estonia 2 Tallinna Küte, Punane 36, 13619 Tallinn, Estonia ahleb@staff.ttu.ee, anna.volkova@ttu.ee a different heat supply alternative. Often the decentralized heating is not an effective solution for regional heat supply strategy and it decreases potential of combined heat and power production.[3].

ABSTRACT District heating networks in Estonia are mostly old and in bad condition. The state of the district heating networks of Tallinn is typical for the rest of Estonian DH networks. The paper includes analysis of the Tallinn district heating networks. Valid data about damages in district heating systems received for the last 12 years were used for an analysis of the networks damages.

Nowadays DH systems operate both in big cities and in small towns, which means, that there is enough heat load for the installation of new cogeneration equipment. But before new energy sources installation it is important to define and analyse the situation with DH networks.

Different types of network damages are analysed: external corrosion, internal corrosion, defect of installation, factory defects, defect of construction and other reasons. The number of damages for the different elements of networks is compared in the paper: armature, compensator, construction and pipes. Main factors, which influence damages in district heating networks, are the age of networks, the quality of construction works and the network operation conditions.

The purpose of this paper is to define the valid condition of typical old networks in Estonia, to define the reasons of damage occurrence on the basis of operational data and to make forecasts for operation of a DH network for the next 20 years. The paper includes analysis of Tallinn district heating networks. The valid data about damages in district heating systems collected during past 12 years was used for analysis of networks damages.

The damage quantity dependence on the age of networks is also defined and analysed in the paper. The number of damages can be diminished by reducing the average age of networks. This is possible by replacing old pipelines and other network system elements. Pipes average age changes for 20 years period are simulated according different intensities of renovation works.

THE PRESENT CONDITION OF TALLINN DISTRICT HEATING SYSTEM District heating networks in Estonia are mostly old and in bad condition. The state of the district heating networks of Tallinn is typical for the rest of Estonian DH networks. In Tallinn the heat is transmitted to the consumers through a 406-kilometres long heating network including the 93 km of pre-insulated pipes (23%). District heating systems of Tallinn were constructed mostly during the 1960-1980 period and their average age is 22 years.

INTRODUCTION District heating (DH) allows centralized heat production for an area and hot water transportation to the buildings through a network of pipes. District heating systems offer the potential to use energy-efficient and renewable heat generation technologies, such as cogeneration technologies which implement both fossil fuels, as long as biomass and waste [1]. District heating system is traditional in Estonia. It has formed approximately 70 per cent of all heating in the country. The share of heat produced by combined heat and power production stations is approximately one third. At the same time, the technical situation of the district heating networks (and production equipment) is poor. [2] Unsatisfactory condition of DH networks and unreliable heat supply can doubt on future of district heating and the consumers can make a choice towards

The AS Tallinna Küte enterprise makes operation of the bigger part of district heating networks and boilerhouses of Tallinn. District heating systems of Tallinn consist of five districts of the central heat supply: Kesklinna district (total length ~92 km, length on the balance of AS Tallinna Küte ~76 km), Lääne district (total length ~162 km, length on the balance of AS Tallinna Küte ~141 km), Lääne district local networks (total length ~12 km, length on the balance of AS Tallinna Küte ~11 km), Lasnamäe district (total length ~114 km, length on the balance of AS Tallinna Küte ~106 km), 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 balance of AS Tallinna Küte ~14 km). [4]

(natural gas, 232 MWth); the CHP Väo (wood chips, 25 MWel, 65 MWth); the boiler house Mustamäe (natural gas, 390 МWth); the boiler house Kadaka (natural gas, 290 MWth).

District heating systems of the areas Kesklinna and Lasnamäe are connected through the pump station Laagna. The total length of heating networks is 406 km from which on the balance of AS Tallinna Küte there are 348 km, or 85,7%.

Besides the abovementioned there are some smallscale boiler houses. In Fig. 1 is displayed the basic scheme of Tallinn heat supply.

The following CHP stations and boiler-houses supply heat to the districts of Tallinn: the CHP Iru (natural gas, 190 MWel, 748 MWth), the boiler house Ülemiste

Mustamäe boiler-house 390 MW 390 MW

Kadaka boiler-house 290 MW

District heating systems of Tallinn were constructed mostly during the 1960–1980 period and their average age is 22 years.

Iru CHP 748 MW (190 MW)

Ülemiste boiler-house 232 MW (in reserve)

Laagna pump station 200 MW

Mustamäe network 325 MW

Lasnamäe network 268 MW

Kesklinna network 180 MW

Maardu network

Väo CHP 65 MW

Fig. 1 The basic scheme of Tallinn district heating system

The state of DH networks varies for the different districts of Tallinn.

(districts Mustamäe and Õismäe). Initially there had been two separate networks which were merged later on as a result of growth. In the area Lääne the construction of district heating systems began in 1960. The length of the Lääne area network is ~141 km. The diameters of the main pipelines are less than those in the Lasnamäe area.

In Lasnamäe the construction of district heating systems began in 1970, and the network length is ~106 km at present time. Assuming the actual load the heating systems of Lasnamäe district are the most overloaded in town.

The length of the main pipelines with diameter DN400–900 is ~27,8 km. The heat losses of the network in 2008 were 16% from the total produced heat.

The length of main pipelines DN1000–1200 is ~19 km, the length of pipes DN400-800 is ~4,4 km. The share of the main networks is quite big and it is ~22% of total network length in Tallinn. Thermal isolation is made of glass wool according to old soviet building norms and it is the reason of big heat losses in the network. The heat losses in Lasnamäe network in 2008 were 21% from the total produced heat.

The speciality about the heating system of the area Lääne is that in past there was an open system of hot water supply. The water added to the system had no time to purify sufficiently and oxygen and water hardness led to an intensive internal corrosion of pipes.

The interconnected district heating systems of boilerhouses Mustamäe, Kadaka and Karjamaa (not in operation at present time) are related to the Lääne area

In Kesklinn area the network construction began in 1959. Initially the heat supply was carried out by the 278

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

30 years and by today they are already worn out. The probability of failures sharply increases. By today the 84% of all compensators should be replaced. Some parts of the old locking armature also have to be replaced. The service life of armature has exceeded 25 years. Armature and compensators are partly renovated; however some pieces of it are old and also require replacement. [5]

Tallinna Soojuselektrijaam heat and power station and later on by the boiler-house Ülemiste. The district heating system of Kesklinn area is the oldest in Tallinn. The average age of the Kesklinn area network is 25 years, the total length is ~76 km. The length of the main pipelines with diameter DN400900 is ~13,8 km. The share of main pipelines in Kesklinn area network is ~18,1%. Relative heat losses of Kesklinn network are within the limits of 15...18%. In comparison with other areas the relative heat losses are less. The reasons for this are: the bigger network loading, the not oversized pipes and the significant share of preinsulated pipes.[5]

300 250 200

armature compensator

150

THE ASSESSMENT OF DAMAGES The analysis of networks damage statistics for Tallinn is made on the basis of valid data collected during the past 20 years.

100

The distribution of damages of Tallinn district heating network is shown in Fig. 2 according the periods of construction. It is obvious that the most critical situation is with the sites constructed during the 1980–1985 period. It can be explained by the poor quality of both construction works and materials used in construction. During that period the networks were being constructed in a hurry and with lack of proper supervision.

construction pipes

1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Fig. 3 Places of damage in Tallinn district heating networks elements

In Fig. 4 the nature of damages is summarized. There are no data about the character of damage for all the areas of Tallinn within past 10 years, that‘s why the damage allocation by character of damage is shown for a five year period.

damages during 1998-2009

400

In Tallinn network the significant part of damages is caused by external corrosion of pipes. Main reasons of external corrosion are the bad waterproofing of underground channels and chambers and the collapsed drainage. Amongst other reasons are the defects of pipe supports and the destruction of concrete channels.

350 300 250 200 150 100

0 …1965

1965 1970

1970 1975

1975 1980

1980 1985

1985 1990

1990 1995

1995 2000

2000 2005

2005 2008

external corrosion

years of construction

internal corrosion

Fig. 2 Damages of Tallinn district heating networks according the periods of construction

deffect of construction

deffect of installation

In Fig. 3 the places of damage in the network elements are shown: armature, compensators of thermal lengthening, construction and pipes. The major part of all damages was the pipes.

wrong service

other reasons

10 0

During the 1997–2003 period there were many problems with armature and compensators; after 2003 the quantity of damages to these elements had considerably decreased. The oldest thermal lengthening compensators work since 1959. The resource of axial compensators is no more than

2004

2005

2006

2007

2008

Fig. 4 Nature of damages in Tallinn district heating networks

The second main cause of damages is the internal corrosion. In 2004 many pipes damaged by internal 279

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

corrosion were revealed. Internal corrosion is the most serious problem in Lääne network where an open system of hot water supply earlier has been used. Besides the damages caused by defects of installation, defects of construction, factory defects and improper maintenance, the other reasons have also been registered.

D  0,0096  A2  1.8985  A  1.0496

(1),

where D – Number of damages/100 km per year A – Age of networks Before using this regression for further calculations, we should check if this equation is appropriate. One of the main parameters for estimation of regression equation is the correlation coefficient. It is considered, that the correlation is good in case when R>0.8. In the case of damage dependence on pipes age, R is 0.802. R2=0.643, which means that the equation characterizes the 64,3% of damage number changes, but the 35,7% of changes are characterized by another factors. There is still an influence of other factors, which can not be changed, such as construction and installation problems in the past.

The main factors, which have an affect on the damages in district heating networks, are the age of networks, the quality of construction works and the network operation conditions. The two latter can be regulated by control authorities and proper legislation, however, the influence of these factors has been reduced in comparison with the 1970–1990 period. Then quality of construction works was very low, drainage systems were installed incorrectly or were not installed at all and isolation materials were not qualitative. As regards district heating operation conditions, the aforementioned open vented hot water supply system used in some networks has led to intensive internal corrosion of pipes.

Data about damages allocation by the group and approximation of these data is shown in Fig. 5. The regression equation can be used for the damage forecasts in future.

One important reason for damages reduction is that in recent years the networks have significantly reduced pressure. The network works in a stable temperature mode, the reliability of heat sources is improved and the quantity of equipment emergency stops forced by sharp fluctuations of the heat-carrier temperature has decreased.

damages/100 km per year 70 60 50 40

Other operation condition factor which influenced number of district heating system damages was higher water temperatures in networks (up to 130 t °C) than nowadays (up to 110 t °C). Finally we can conclude that such factors as quality of construction works and quality of network operation are close to their optimum at present time in comparison with previous years.

30 20 10 0 0

Damage quantity also depends on the age of networks. The number of damages can be reduced by reducing the average age of the networks. This is possible by replacing the old pipelines and other networks systems elements.

15 20 Age of networks

Fig. 5 Damage number dependence on the age of pipes in district heating systems for the 2005–2007 period

As it has been mentioned before, the age of networks depends on the intensity of renovation works. In Fig. 6 the length of all repaired sites is shown split by years.

Reconstruction and replacement works are made in Tallinn, but the intensity of replacement is rather low and not enough for a stable system operation. It is important to define, how intensive the network reconstruction should be.

Since 1980 the serial repair of Tallinn district heating system is being carried out. Basically the investments have been directed towards the increase of reliability and the reduction of quantity and duration of faults in heat supply. It has been invested a lot in the locking armature.

Data for the three past years were used for defining the damage dependence (number of damages/km/year) on the age of networks. Data about damages were collected for 7 age groups (0–5 years, 5–10 years, 10–15 years, 15–20 years, 20–25 years, 25–30 years, 30–35 years).

For the past 10 years ~35 km of district heating pipelines have been replaced, which is 10% of total length of the district heating systems in Tallinn area.

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 average age of pipes for each year was calculated according equation (2)

The annual replacement of pipes is in average about 3,06 km per year, which is less than 1 percent from the length of Tallinn DH system pipelines. Length, km

Aav j 

8 7

i a

 li  l j ; i=c, if

i=b, if

ia

i a

lj b

 li ( j  i)   li ( j  i)  ( j  c)  (l j   li ) c

l i a

 lj

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

J – current year;

1994

0 1993

I – year of construction; 1992

1 1991

li is length of pipes, constructed in i year

1990

1989

Aav is average age of pipes in j year

1988

1987

where

1986

1985

(2)

a – year of construction of the oldest pipes, operating in the current year.

Fig. 6 Length of replaced pipelines by years in Tallinn district heating network

As a result of simulations, seven forecasts for pipes average age were calculated according different intensity of renovation works: for current intensity of renovation (3,06 km/year) and for intensities when 1%, 1,5%, 2%, 2,5%, 3% and 4% of total DH system length would be annually renovated. The forecasts were simulated for the 20 year long period.

THE FORECASTS FOR DISTRICT HEATING SYSTEM AGE One of the tasks was to assess, how big the renovation works should be in order to stop increasing the average age of pipes. A simulation model, which uses both real data and also some assumptions, was created for such estimation.

The results of simulation are shown in Fig. 8. age, years 40

Assuming that the length of pipes (360,67 km) will not change during the forecast period and that the annual scope of renovation works will remain the same during whole of the period means that the length of renovated pipes also will not change. Besides it‘s was assumed that every year just the oldest pipes would be renovated; however in reality the renovation works are based on the pipes actual state estimation.

35 30 25

1% 2% 3% 4% 1,50% 2,50% current

20 15

Allocation of pipes ages for starting point (2008) is shown on Fig. 7 [5].

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040

length, km

Fig. 8 Pipe age forecasts for different intensity of network renovation works

As it can be seen from Fig. 8 in case the renovation stays on the same level, the average age of pipes will grow till reaching 39 years in 2040. In case the length of annually changed pipes is 1% or 1,5% higher, the average age will still rise, but in a less steep way. When the 2% of DH system length is annually renovated there will be the minimal changes in age during first 5 years, after that the age will start rising and only after 15 years it will begin to decrease.

0 age, years

If renovation intensity is 2,5% of the length or higher, the average age will not rise at all or will decrease. For reducing the damages occurrence probability influenced by the networks age, the amount of repaired

Fig. 7 Length of DH networks by pipes age (in 2008)

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 process of ageing will slow down and also the average age will stabilize on a certain mark. One of the possible solutions is to replace the pipes with higher intensity of 3–4% until reaching the 17–20 years average age and then reduce the length of renovated pipes per year to the 2–2,5% of the whole length of DH network.

Seven forecasts for pipes average age according different intensity of renovation works were simulated: for current intensity of renovation (3,06 km/year) and for intensities when 1%, 1,5%, 2%, 2,5%, 3% and 4% of total DH system length would be annually renovated. It was concluded, that for maintaining the networks average age at least at former level, the rate of old pipelines replacement should exceed the 2,5% of the whole length of DH system.

CONCLUSIONS District heating networks in Estonia are mostly old and in bad condition. The state of the district heating networks of Tallinn is typical for the rest of Estonian DH networks. That‘s why the result of damage analysis made for the DH network of Tallinn can be used for the other networks in Estonia.

AKNOWLEDGMENT This work has been partly supported by the European Social Fund within the researcher mobility programme MOBILITAS (2008–2015), 01140B/2009 REFERENCES

The AS Tallinna Küte enterprise makes operation of 85% from the length of district heating networks in Tallinn. Tallinna Küte data about the damages were used for assessment.

[1] Cogeneration and district energy sustainable energy technologies for today…and tomorrow, International Energy Agency, 2009.

Places of damages in the DH system are following: armature, compensator, pipes and construction. Most of the damages happened in the pipes.

[2] Long-term Public Fuel and Energy Sector Development Plan until 2015, Riigi. Teataja, RT I, 23.12.2004, 88, 601

As regards the character of damages, the typical damages are caused by external corrosion, internal corrosion, defect of construction, defect of installation and wrong service. The major part of damages is caused by external corrosion of pipes.

[3] Hlebnikov, A.; Siirde, A. The major characteristic parameters of the estonian district heating networks, their problems and development. // The 11th International Symposium on District Heating and Cooling: University of Iceland, 2008, 141–148.

The age of networks, the quality of construction works and the network operation conditions are the most important factors, which influence the damages in district heating networks. The number of damages can be reduced by reducing the average age of the networks. This is possible by replacing the old pipelines and other networks systems elements. The intensity of replacement works during last 25 years was less than one percent from the whole length of pipes.

[4] Tallinna küte webpage, www.soojus.ee [5] A. Hlebnikov "The analysis of efficiency and optimization of district heating networks in Estonia", Doctoral Thesis, Tallinn University of Technologies, 2010.

282

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

EFFICIENCY OF DISTRICT HEATING WATER PUMPING IN FINLAND 1

Antti Hakulinen , Jarkko Lampinen and Janne Lavanti 1

Pöyry Finland Oy

ABSTRACT

The investment costs of a pump

The objective of this study was to determine the savings potential in district heating pumping in Finland. A measurement method was also developed to quickly estimate the efficiency of district heating pumping.

90 000 € 80 000 € 70 000 € 60 000 €

Total

50 000 €

Pump

The work was based on the data gathered from district heating statistics. The work is divided into two parts.

40 000 €

Frequency converter

The district heating statistics reveal a number of district heating networks whose consumption of energy needed for pumping is exceptionally high. These companies should clarify the reasons for that.

10 000 €

Motor

30 000 € 20 000 €

0€ 0

100

200

300

400

500

pow er kW

Fig. 1. The investment costs of a pump.

In addition, companies with an exceptionally low consumption of pumping energy should check their measurements and data gathering routines.

1.2 Total costs of pumping Total costs of pumping include capital, maintenance and energy used in pumping. The pump lifetime costs are mainly energy costs as we can see from figure 2 on the next page. The lifetime costs are calculated with the following assumptions: energy price € 60/MW/h, operating lifetime 15 years, utilization period of maximum load 5000 h/a, interest rate 5 per cent and the O&M 1.2 per cent of the investment.

On average the electricity needed for district heating pumping should not be over 0.5 per cent of the total energy supply (=sold+losses). If the density (supply/length of the network) of the district heating network is less than 3 GWh/km, the energy needed for pumping may rise. In any case the proportional pumping energy should be lower than 1 per cent of total energy supply.

Pum p I, pow er: 16 kW

The Finnish potential for saving in district heating pumping is estimated to be 20 per cent of the current pumping energy i.e. 30 GWh/a. This is equivalent to a yearly saving of approximately € 2 million.

14 % 2% Capital O&M Energy

84 %

PART 1. INTRODUCTION Pum p II, pow er: 131 kW

In the Finnish district heating systems no typical pumping arrangements have been used at heat production plants or at booster pump stations. The ways of dimensioning and connecting pumps have varied a lot. This has led to incorrect dimensioning and connections of pumps, which in turn has caused higher investment costs and greater pumping energy usage than expected, operational problems and in the worst case many interruptions in the use of the network. The Finnish district heating system is based on the variable flow operation (consumer driven scheme)

Capital O&M Energy

92 %

Pum p III, pow er: 283 kW

Capital O&M

1. COSTS

Energy

1.1 The investment costs of a pump (including motor and controls)

94 %

The calculated investment costs of a pump including motor, control, and pump are shown Fig. 1. 283

Fig. 2. An example of lifetime costs of three different sized pumps.

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

As we can see the pump‘s efficiency plays a huge role because the lifetime costs mainly consist of the operational energy (84–94 per cent). For that reason a lot of attention should be paid to the efficiency when making the investment. Pumps with a low efficiency may eat into the savings of the investment many fold.

cent or 2*70 per cent in parallel connection with individual rotating speed controls. In that way the pumping of maximum heat load can be managed and there is a room for possible expansion of the district heating network. The other pump will act as a ―summer pump‖ so that the efficiency of pumping remains high also when the heat load is low.

1.3 Booster pump station and costs

By dividing the pumping capacity between many pumps it is possible to save pumping energy even if pumping is handled from one point or from the heat production plant and the booster pump station. The possibilities to divide the pumping must be examined case by case by taking into account every single thing that might have an effect on the costs.

A booster pump station should be considered when the primary pumps of an energy station do not have enough capacity to ensure the pressure difference at the last customer. Typical reasons for the building of a booster pump station can be: long transmission lines, expansion of network, optimization of pumping energy and controlling of pressure level. The investments of a booster pump station including pump, motor, frequency converter, building, automation systems, etc. are shown in the following figures 3 and 4.

3. MAXIMUM WATER FLOW The actual cooling of the district heating system in operational conditions of pumping should be taken into account when determining the calculated maximum water flow. It is worthwhile to specify the water flow according to slightly worse cooling than the actual conditions require so that there is some design margin for unusual conditions.

The investment costs of a booster pump station 500 000 € 480 000 € 460 000 € 440 000 €

4. OPERATION POINT

420 000 € 400 000 €

To change the rotating speed of a pump with a frequency converter is a good way regarding energy efficiency because the pump‘s efficiency often remains on high level within the whole adjusting area but the need for power reduces strongly when the rotating speed goes down.

380 000 € 360 000 € 340 000 € 320 000 € 300 000 € 0

100

150

200

250

300

350

400

450

500

pow er [kW]

Fig. 3. The investment costs of a booster pump station (only 1 pump).

The investment cost of a booster pump station (flow + return) 760 000 € 710 000 € 660 000 € 610 000 € 560 000 € 510 000 € 460 000 € 410 000 € 360 000 € 0

100

150

200

250

300

350

400

450

500

Fig. 5. An example functional diagram of a pump.

pow er [kW]

An example functional diagram is shown in Fig. 5. When the rotating speed changes, the efficiency remains good regardless of the changing rotating speed. The pumping of a district heating network follows this theoretical situation very well. However when choosing a district heating pump it is important to

Fig. 4. The investment costs of a booster pump station (with 2 pumps).

2. PUMPING ARRANGEMENTS At the primary station it is usually sensible to divide the pumping between a few pumps, for example 2*60 per 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 the minimum from 50 to 60 per cent of the nominal rotating speed.

Specific pumping energy vs. heat supply Supply 0 - 2 500 GWh/a Specific pumping energy (electrical power / heat supply)

1.6 %

PART 2. INTRODUCTION The goal of the second part was to motivate the district heating companies to analyse their pumping methods and, hopefully, to lower their pumping costs. Total savings potential in district heating pumping in Finland was also estimated.

0.8 % 0.6 % 0.4 % 0.2 %

500

1000

1500

2000

2500

Heat supply, GWh/a

Fig. 6a. Example Electricity used for pumping in relation to the size of a district heating company, heat supply 0–2 500 GWh/a. Specific pumping energy vs. heat supply Supply 0 - 300 GWh/a

Specific pumping energy (electrical power / heat supply)

1.6 %

STATISTICAL FINDINGS The used pumping energy in different companies was analyzed by comparing the pumping energy to the following parameters: Heat supply (sold heat + losses) Length of the district heating network Heat density (supplied heat energy divided by the length of the DH network)

1.4 % 1.2 % 1.0 % 0.8 % 0.6 % 0.4 % 0.2 % 0.0 % 0

100

150

200

250

300

Heat supply, GWh/a

Fig. 6b. Electricity used for pumping in relation to the size of a district heating company, heat supply 0–300 GWh/a.

The following parameters were also examined but no clear correlation was to be seen, and the results are therefore not reported in this paper:

The average pumping energy is 0.6 per cent of the heat supply. The bigger the company, the smaller the proportional pumping energy.

Heating output density (daily maximum heating output divided by the length of the DH network) CHP production Share of small (< 30 kW) consumers Peak load utilization time Losses of DH network

Companies with exceptionally high pumping energy are marked with a circle. 2. Length of the district heating network Specific pumping energy vs. length of the DH net Length 0 - 500 km

Specific pumping energy (electrical power / heat supply)

   

1.0 %

The biggest companies have reported the pumping energy, thus, the pumping figure is available to companies which supply almost 70 per cent of all district heat in Finland.



1.2 %

0.0 %

This part is based on the Finnish district heating statistics of the year 2007 [1]. The statistics cover the data of nearly 200 district heating companies but only about 60 of them have reported the electric power used in district heating pumping.

  

1.4 %

1. Heat supply Heat supply is the same as sold heat + losses. The assumption used was that the bigger the company the smaller the proportional pumping energy. The situation is presented in figures 6a and 6b, where only the companies which supply less than 2500 GWh/a (Helsinki not included) are shown.

1.7 % 1.5 % 1.3 % 1.1 % 0.9 % 0.7 % 0.5 % 0.3 % 0.1 % 0

100

200

300

400

Length of the DH net, km

Fig. 7a. Electricity used for pumping in relation to the length of the district heating network, 0–1 300 km 285

500

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia Specific pumping energy vs. heat density

Specific pumping energy vs. length of the DH net Length 0 - 70 km

Specific pumping energy (electrical power / heat supply)

1.7 % 1.7 % 1.5 % 1.3 % 1.1 % 0.9 % 0.7 % 0.5 % 0.3 % 0.1 % 0

1.5 % 1.3 % 1.1 % 0.9 % 0.7 % 0.5 % 0.3 % 0.1 %

Length of the DH net, km

Heat density, GWh/km

Fig. 8. Electricity used for pumping in relation to the heat density.

Fig. 7b. Electricity used for pumping in relation to the length of the district heating network, 0–500 km.

It is natural that in a DH network with not too many pipes in proportion to sold heat the need for pumping of DH water is lower.

Specific pumping energy (electrical power / heat supply)

Specific pumping energy vs. length of the DH net 1.6 %

FURTHER INFORMATION:

1.4 %

Pöyry Finland Oy PL 93 (Tekniikantie 4 A) FI-02151 Espoo Finland antti.hakulinen@poyry.com

1.2 % 1.0 % 0.8 % 0.6 % 0.4 %

CONCLUSION

0.2 %

District heating networks enlarge and change continuously and therefore the conditions of pumping will also change. For that reason, it is important to check every now and then if the actual operating point of the pump is as designed and what the efficiency of the present operating point is. The pumping could still work technically well but the pumps could be operating with low efficiency.

0.0 % 0

200

400

600

800

1000

1200

1400

Length of the DH net, km

Fig. 7c. Electricity used for pumping in relation to the length of the district heating network, 0–70 km.

For big companies the proportional pumping energy is almost constant 0.5 per cent of heat supply.

The most important issues in designing and operating of district heating pumping are:

The longer the DH network, the smaller the proportional pumping energy. The result is partly the same as in the previous chapter: the bigger companies have smaller proportional pumping energies.

 

If a company seems to have a high proportional pumping energy in figures 7a–7c it may be due to poor heating density (lots of pipes in areas with not so much consumers).



3. Heat density



Heat density is the heat supply divided by the length of the district heating net.

A sufficient but not too big pressure difference must be guaranteed for customers. There must be enough pressure in all parts of the network at all circumstances and at the same time the maximum pressure level must not be exceeded. When designing pumping it is important to study all possible pumping cases. Good operating point should be verified when designing and operating pumps.

Pumping energy is dependent on certain parameters. The best parameter is considered to be heat density on which pumping energy is clearly dependent. And this is a quantity every district heating company measures. 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 red line to help the reader estimate the pumping energy of his own plant. If the pumping energy is above the red line some measures ought to be taken.

The following figure illustrates an example case in which the heat density is over 2.5 GWh/km. The figure can be utilized when estimating the losses in real money if the proportional pumping energy is over the average of 0.5 percent.

1.7 %

1.5 %

1.3 %

1.1 %

0.9 %

0.7 %

0.5 %

0.3 %

Value of "excess" pumping energy Heat density > 2.5 GWh/km, Value of power 60 EUR/MWh 400

350

Proportional share of pumping energy 0.8 % Proportional share of pumping energy 0.7 %

0.1 %

Value of excess pumping energy, 1 000 EUR/a

Specific pumping energy (electrical power / heat supply)

Specific pumping energy vs. heat density

300

Proportional share of pumping energy 0.6 %

250

200

150

100

0 100

600

1100

1600

2100

Heat supply, GWh/a

Heat density, GWh/km

For example, if the heat supply of the company is 1.1 TWh/a and the proportional pumping energy is 0.7 percent, the losses of unnecessarily high pumping energy is â&#x201A;Ź 130 000 per year.

Fig. 9. Electricity used for pumping in relation to the heat density of the district heating network + trend line

Figure 9 shows that on average the electricity needed for district heating pumping should not be over 0.5 per cent of the total energy supply (=sold+losses). If the density (supply/length of the network) of the district heating network is less than 3 GWh/km, the energy needed for pumping may rise. In any case the proportional pumping energy should be lower than 1 percent.

Some of the pumping energy is converted to heat. This decreases the value of the losses. In total, the potential savings in all Finnish district heating companies are approximately 20 percent of the current pumping energy, i.e. 30 GWh/a. This is equivalent to a yearly saving of approximately â&#x201A;Ź 2 million. 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

MODELLING DISTRICT HEATING COOPERATIONS IN STOCKHOLM – AN INTERDISCIPLINARY STUDY OF A REGIONAL ENERGY SYSTEM 1

D. Magnusson , D. Djuric Ilic

Department of Thematic Studies – Technology and Social Change, Linköping University, SE-581 83 Linköping, Sweden 2 Department of Mechanical Engineering, Division of Energy Systems, Linköping University, SE-581 83 Linköping, Sweden in Karlstad in 1948 and during the following decades the largest cities built their own systems, as was the case in Stockholm [2]. Because of the large amount of energy in the systems, the fuel used in the plants has a major impact on greenhouse gas (GHG) emissions, and there is also a large potential for using combined heat and power (CHP) technology in the systems. CHP technology is becoming more important as a part of creating sustainable energy systems, which for example can be seen in the EU directive for promotion of cogeneration [3]. In Sweden, as well as in Stockholm, large investments are made in building new CHP plants, in large part thanks to the electricity certificate system [1]. Another important potential with CHP generation is through the Electricity Directive of 1996, in which the EU prescribed common rules for creation of an open and competitive electricity market [4]. With a fully integrated electricity market, the Swedish prices of electricity can be expected to increase. However, as long as they are lower than Europe‘s there is a large potential for exporting electricity. From a marginal power production perspective, which will be discussed further in the paper, there is a potential for decreasing global emissions of CO2, if the exported electricity comes from non-fossil fuels.

ABSTRACT In this paper, a combination of methods from social science (interviews) and technical science (modelling) have been used to analyse the potential for cooperation in the present and future district heating system in Stockholm. The aim of the paper is to explore barriers and driving forces for energy cooperation in the Stockholm district heating system and to analyse the potential for combined heat and power generation in the system. In the study it was found that with better connectivity in existing systems, the annual system cost would decrease by approximately 10 million €, and with new CHP plants a similar potential exists. There is also a large potential for decreasing the local and global emissions of CO2 with CHP plants. The results from the interviews showed that the existing cooperation has a long history and is working well today. The advantages are higher supply security and economic benefits, while disadvantages are a need for more administration and control because of a more complex system. That the barriers to cooperation are seldom technical is another conclusion. With the combination of methods, we have gained a better understanding of the actual potential for the development of the system.

A large enough system is an important prerequisite for investment in CHP plants, in order to take advantage of the economy of scale of district heating and CHP generation. In Stockholm, the largest urban region in Sweden, there are already well-developed district heating systems. The systems started as smaller units that gradually have been interconnected and today consist of three large networks. However, since there are eight different energy companies in the city region, a working cooperation between the energy companies is important. With this in mind we will analyze how the actors perceive existing and future cooperation. The study is conducted with an interdisciplinary approach where interviews have been combined with modelling the systems' performance with present and possible future interconnections, present plants and future CHP plants, and finally with a hypothetical introduction of natural gas. The aim of the paper is to explore barriers and driving forces for energy cooperation in the

NOMENCLATURE CO2 – carbon dioxide; LECO2 – local emissions of CO2; GECO2 – global emissions of CO2; CHP – combined heat and power; BCHP – CHP plants fuelled by solid biomass; NGCHP – CHP plants fuelled by natural gas; TGC – tradable green certificates; GHG – greenhousegas. 1. INTRODUCTION Swedish district heating has a long history and is today one of the dominant heating forms with approximately 55% of market share, and an annual energy production of approximately 55 TWh[1]. The first system was built 288

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

them, although with some differences depending on the company. Because a semi-structured interview is a qualitative method, the possibility of using open questions is an advantage, and since we are interested in a specific situation, the interviewees have the chance to give their opinion. It also gives the opportunity to analyze the answers in different ways, to understand the opinions expressed [7].

Stockholm district heating system and to analyse the potential for CHP generation in the system. 2. CASE STUDY There are three large district heating networks in Stockholm that deliver more than 12 TWh of heat annually, produced in some 70 heating plants [5]. Table I shows the heat production, types of base production and installed heat and electricity capacity in those networks. Six of the plants in the system are CHP plants with total installed electricity capacity of about 600 MW, which gives a possibility for production of over 2 TWh of electricity annually [5].

3.2 Modelling Stockholm’s district heating system Based on the data from ―Open district heating network in greater Stockholm‖ [5] a model of Stockholm‘s district heating system has been constructed. Purchases and sales prices of electricity, taxes and tradable green certificates (TGC) are included in the model (Table II) as well as the operating and maintenance costs for all plants and fuel prices. However, due to agreements with the contact persons from the district heating companies, the prices for fuel are not presented in the paper.

Table I. – Major district heating networks in Stockholm. [5] Southcentral

Northwest

Southeast

Heat production in the year 2005 [TWh]

9.4

2.2

0.53

Installed heat capacity [MW]

4000

700

300

Installed electricity capacity [MW]

493

105

Current price of electricity [€/MWh]

CHP waste, CHP coal

BCHP

NGCHP

Purchase

Sale

Sale with TGC included

70.10

35.46

67.56

Base production

Table II – The average annual purchases and sales prices of electricity, including all taxes and TGC. [8], [9], [1]

European price of electricity [€/MWh] Purchase

Sale

Sale with TGC included

83.30

48.65

80.77

3. METHODS A combination of methods from social science and technical science has been used; modelling with MODEST and semi-structured interviews with representatives from the largest energy companies. MODEST is an energy-system optimisation model with time-dependent components that was developed at Linköping University in Sweden. MODEST uses linear programming to calculate the most profitable combination of existing and potential new facilities and shows which investment options are financially viable [5].

Carbon dioxide emissions used in this paper are shown in Table III [10]. However, since the greenhouse effect is a global problem, carbon dioxide (CO2) emissions are not simply analysed from a local perspective but also in regard to a global perspective. The global emissions of CO2 (GECO2) of the system are calculated with the assumption that electricity produced in the plants is going to replace marginal power production in the integrated European electricity market. Since coalfired condensing power plants have the highest variable cost compared with other sources of electricity in the EU, they work as the marginal power production [11]. When assuming that the coal-fired condensing power plants have an electricity efficiency of 33%, each megawatt-hour of electricity generated in such a plant releases approximately one tonne of CO2. According to that, any increase in electricity production in Stockholm‘s district heating system can lead to reduced production in the marginal coal condensing power plants, and consequently to a reduction of one tonne of CO2 emissions. However, it is necessary to mention that considering the EU Emissions Trading Scheme (EU ETS), the decrease of CO2 emissions in

3.1 The structure of the interviews The interviews were conducted during the spring of 2009, with representatives from the five largest energy companies producing and/or distributing district heating in the Stockholm region. These are Fortum, Norrenergi, Söderenergi, E.ON and Vattenfall. The representatives were chosen by the companies themselves, since they could better decide who would be most appropriate to answer questions regarding interconnections, cooperation and future strategies. We decided to let the respondents remain anonymous, since one of the interviewees wanted this. The interviews were semistructured, as we had similar questions for most of 289

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Since electricity generation will probably be the primary production in all district heating companies in the future, when the Swedish electricity price becomes as high as the typical European price, in scenarios 6-9 our research focuses on the cogeneration potential in Stockholm's district heating system. Scenario 1 has been used as a reference scenario for scenario 6. Scenario 3, where the influences of a higher electricity price on the system with the existing plants have been analysed, has been used as a reference scenario for scenarios 7–9. Scenarios 6-9 are analysed as possible future cases that may exist more than 10 years from today. Because of that, all plants in the scenarios are new so the investment costs for all plants are considered. While in the scenarios 6, 7 and 8 the system consists of 31 CHP plants fuelled by solid biomass (BCHP), there are a total of 46 CHP plants fuelled by natural gas (NGCHP) in scenario 9. In scenario 9 it is assumed that the natural gas network exists along the Swedish east cost.

electricity production sector does not necessarily to lead to reduction of GECO2 [12]. But the marginal electricity concept still has significance for future measurement of and planning for future limitations of CO2 emissions and the future trading system. Table III. – Net emissions of CO2 [10]. Fuel

Emissions kg/MWhfuel

Oil

280

Coal

330

Waste

100

Biomass

Electricity

950

Natural gas

230

3.3 Description of chosen scenarios Nine different scenarios have been analysed considering the possible future cases (Table IV), with special attention to economic and environmental aspects.

The characteristics of the CHP plants that have been integrated in the model of the district heating system in scenarios 4–9 are presented in Table V [13].

The existing district heating system (scenario 1) and the system with three new CHP plants that are planned to bee built according to the interviews and documents (scenario 4) have been analysed. Since the base productions in the networks differ, the differences between the production‘s costs in different parts of the system are notable. Because of that, in both cases (scenarios 1 and 4) the influences of a better connectivity between networks have been studied (scenarios 2 and 5).

Table V. – The characteristics of the new integrated CHP plants in scenarios 4–9 [13]. Technical characteristics Fuel Sc.

Table IV. – List of the chosen scenarios.

Sc.

1 2 3 4

Plants in the district heating system existing existing

6 7 8

existing + new CHPP + new CHPP BCHP BCHP BCHP

NGCHP

Connectivity

existing one 1 network existing existing

TGC

Nordpool Nordpool

exist exist

EU Nordpool

exist exist

one network

Nordpool

exist

one network one network one network

Nordpool EU EU

one network

exist exist do not exist -

Fuel efficiency

MWe

Α*

biomass

110

0.45

waste

0.32

biomass

110

0.46

6–8

biomass

113

0.51

natural gas

150

1.41

4–5

Electricity price

Electrical output

Economic characteristics Process plant cost €/KWe 4–5 6–8 9

Operating and maintenance % of PPC

€/MWh fuel

2 745

1.5

2.45

5 440

9.31

2 110

1.5

2.45

2 110

1.5

2.45

715

2.5

0.9

* electrical/thermal output 3.4 Previous studies

1) Interconnections between the south-central and the north-west networks have been introduced as well as interconnections between south-central and south-east networks. Capacities for existing pipes have been increased.

Two studies regarding Stockholm´s district heating system were done in the years 2005 [10] and 2006 [14], and the results showed that benefits for better connectivity between some parts of the system existed. 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 replaced with BCHP plants, with an electricity-to-heat output ratio 0.46, up to 10TWh electricity can be produced and the potential for decrease of GECO2 of the system would be 3 tons CO2 annually. If all plants in the system are replaced with NGCHP plants, with an electricity-to-heat output ratio 1.2, the electricity generation in the system can increase to 11TWh and the potential for decrease of the GECO2 of the system would be about 5 tons CO2 annually. However, since these two studies were done, a new connectivity between networks has been built and the total installed electricity capacity in the system has increased by 20% [5]. Furthermore, new CHP technologies are constantly being developed, which enable greater electricity efficiency and consequently greater benefits from economic, energy and environmental viewpoints.

4. RESULTS OF THE SCENARIOS The results from the scenarios are presented in Table VI and Table VII. According to the optimisation results, if better connectivity is introduced, some economic benefits exist. In both cases the case with only existing plants in the system and the case where the new plants are introduced in the model (scenarios 2 and 5) the decrease in system costs would be about 10 million € annually. The potential for decrease of the environmental impact of the system is more notable. If better connectivity were introduced in the system today, the biomass share in total fuel use would be 8% higher and consequently both the local emissions of CO2 (LECO2) and GECO2 of the system would be about 0.25 million tons lower annually. The potential for decrease of GECO2 of the system if better connectivity is introduced after the building of new CHP plants (scenarios 4 and 5) is 0.4 million tons annually.

Regarding interconnection and cooperation of DH systems, some studies have been conducted in a Swedish context. However, none of them have focused on cooperation between energy companies. They have instead focused on cooperation between energy companies and industry.

Table VI. – Results for the scenarios – economic aspects.

Thollander et al. [15] found that technical aspects are seldom barriers to cooperation. The barriers are rather risk, different aspects of information during negotiations, and other social factors such as inertia among personnel. Driving forces have been economic factors such as an aim for lower costs and means of control, as well as environmental values. In a study with a similar aim, Fors [16] found the same results, that technical aspects are seldom barriers. Information during negotiations, stable contracts and the importance of involving the personnel at the plants in the process are important factors. It is also important that the cooperation benefits both parties. Grönkvist et al.[17] reached a similar conclusion in a study that emphasises the importance of the willingness of people on both sides to cooperate. The main advantages of the cooperation are lower costs and benefits for the environment, while the main disadvantages are less flexibility as both parties work under contracts.

Sc.

Annual system costs

million €

Historically, interconnection of technical systems has been seen in the theory of Large Technical Systems as one way for systems to grow. Systems start in a local context, but when the technology is transferred to other geographic areas, the systems grow and can then be interconnected as they often have grown into each other. Interconnection of systems can also be explained through the fact that larger systems have a higher load factor and better economic mix [18], [19], [20].

CHP heat production share

Electricity

Annual production

The income from electricity

TWh

Million €

258

2.30

122

245

2.31

125

243

2.35

150

204

2.96

164

192

3.15

176

403

344

546

504

424 100

6.39

482 281

100

17.66

777

As the electricity price increases, the system would earn extra income from the electricity sold, and thus the heat production cost would decrease (scenarios 1, 3). This gives an even bigger advantage to CHP generation compared with pure heat production.

291

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

also notable that in scenarios 6, 7 and 9 the annual income from electricity is higher than the annual system costs. However, since all plants in those scenarios are new, the total investments are high. Because of that, if the analysed time period is just 10 years, the annual system costs are much higher then today.

Table VII. – Results for the scenarios – environmental aspects.

Biomass share in the system

LECO2

GECO2 of

[million tons/year]

2.50

0.32

2.25

0.06

2.46

0.23

2.12

– 0.69

1.91

– 1.08

100

– 6.07

7.80

– 8.98

Sc.

the system

The lowest GECO2 of the system are in the scenarios where all plants in the system are BCHP (scenarios 68) and NGCHP (scenario 9) plants. In those two cases GECO2 in Sweden, which is about 60 million tons annually [21] would be reduced by approximately 9% and 15% respectively, with the assumption that the electricity produced would replace the marginal electricity. LECO2 in the system is highest in the scenario where all plants are NGCHP but at the same time GECO2 of the system is lower because of the high electricity production.

6 7 8 9

5. RESULTS FROM THE INTERVIEWS In the following section the results from the interviews will be presented. The interconnections between the systems make it possible to cooperate regarding heat production and distribution.

The income from the electricity sold in scenario 3 is about 30 million € higher then the income in scenario 1, and because of that the system cost is 6% lower. The difference between the electricity production in scenarios 1 and 3 is not significant, but in spite of that, the decrease of GECO2 of the system in scenario 3 is almost 100%. The reason is higher biomass share in the total fuel used in the system in scenario 3, and consequently lower LECO2 in the system.

5.1 The system today The interviews show that the interconnections have a historical background. Most of them were made during a period when a regional energy company called STOSEB (Greater Stockholm Energy Company) existed, where the municipalities, which to a large extent owned the systems then, were represented. The main reason for the interconnections then was supply security. When the systems were interconnected, the companies could help each other during stops, and this is still the case. All representatives say this, and the representative from Söderenergi expresses it this way:

The introduction of three new plants in the system (scenario 4) would lead to a significant reduction of the heat production cost compared with the system today. The income from the electricity sold would be 35% higher and, as a result, the annual system costs would be 20% lower. This confirms that heat production in CHP plants has a major influence on the economic efficiency of the district heating system. With the assumption that the electricity produced would replace the marginal electricity in the European electricity market, reduction of GECO2 of the system would be almost 1 million tons annually.

…At the same time it is a common good. It is good that the systems are interconnected. It is an extra security if one plant should stop for some reason [22]. The advantages historically and foremost today are also economic. The emissions trading makes it advantageous, since the companies can use the production better by making ―capacity trades‖ and even out the production cost between the companies:

If all plants in the system are BCHP (scenarios 6–8) or NGCHP (scenario 9) plants, the annual electricity production would be as high as 4.5% and 12% of the total electricity production in Sweden, which was about 145 TWh in the year 2008 [1]. The annual income from the electricity sold in those scenarios is much higher then the income from the electricity sold in the other scenarios. In the scenarios with typical European electricity price, (scenarios 7–9), the income from the electricity sold is 220%, 90% and even 420% higher then in scenario 3, where the system with the existing plants is analysed with the higher electricity price. It is

We see that we can use existing production more effectively. Most of the trades are a trade to mid-price so to speak. You can say that we split the profit. Capacity trading (effektköp) is also common. Like we have here with Söderenergi, we have partly a production cooperation and partly we buy capacity. They have more capacity than they need today [23].

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 peak load, which often is oil-based, which is costly both for the fuel price but also because of the emissions. Another point is that, as the Fortum representative said, it is possible to even out effect between systems. One system may have cheaper base load than the other, and for tax reasons it may be cheaper to buy from the other than to use peak load.

modelling, the interconnected:

systems

are

also

already

well

Yes, the principal structure is already established. (…) It is this connection, between the central and the northwest system, it is the only one. That is not solved yet [23]. This particular connection would interconnect the two main systems, and has been discussed in some investigations [26], [27]. However, it is yet to be done. This connection is most important for Fortum, as for example E.ON thought that it made little difference to them.

One factor that is pointed out for a successful cooperation is that both parties can benefit from it. As in all business, it is important that the cooperation be correct from a business standpoint and that both parts are satisfied [24].

The other main connection still missing is connection between the south system and Vattenfall's system in the southeast. Vattenfall thinks that the question has been raised on occasion, although never realized. They give no specific reason for this; they state that all cooperation is important and that different investigations have shown the advantages, although it is difficult to quantify what it means practically [28]. Stockholms Energi (now Fortum) previously owned one of the plants, and there were plans to interconnect the systems then. Fortum gives no explanation for why the interconnection has not been done earlier or now.

The extent of cooperation varies between the companies. Some have more extensive cooperation with daily trades, like Fortum and Söderenergi, while others, for example Fortum and E.ON, do not trade every day. In the latter case, they normally do not trade as much during winter, although sometimes when peak load is needed it is decided quickly [25]. Another advantage with the interconnections is that the companies can cooperate regarding revisions of the plants. While one company has revision during summer, the other can produce for the other company. The factors that are seen as barriers are seldom technical. The companies think that the technical problems often can be solved while making the interconnection and at that point there is a need to negotiate certain aspects. For example, who provides the electrical energy for the pumps and takes responsibility for the regulation of the water pressure in the culverts and repairing the system in a joint part of the system? However, this is often solved:

Although no direct comments regarding the lack of interconnection were made, one of the interviewees who previously worked at Vattenfall said that there was an opinion at Vattenfall that they prefer to keep to themselves, without interconnections, and should not work towards cooperation. Comments without a specific direction were also expressed in interviews that there was a lack of will to cooperate from some companies. There is also a history of rivalry between Vattenfall and the former Stockholms Energi [29]. It is possible that this rivalry stills exists. Fortum also expressed opinions about the fact that other companies are building their own CHP plants instead of trying to find regional solutions.

Yes, the other things we can handle while building the technical parts. At that point we hopefully have identified all technical barriers so that they can be taken into account. They should not appear during production. Settlement of account and such things, they are not a big problem although complicated. However, it is nothing that makes you pass on a profitable cooperation [23].

5.3 Building CHP in the system As seen in the scenarios, in the near future in Sweden many CHP plants are planned and will start to be built. In Stockholm most of the companies have plans for CHP, and two of them have already built in the last years, for example Igelsta (Söderenergi) and Jordbro (Vattenfall). Other companies are making plans, such as Norrenergi, EON and Fortum. The reasons for building CHP are varied, but the most clear is that they see economic advantages in selling electricity, and our stagnating heat load ahead. By selling electricity there is a possibility to keep profits high, even with a stagnating heat load. The system is also relatively old and well established; the potential for further connections are getting smaller as saturation in the

In the above quote, we see one of the disadvantages with today's cooperation, on which all the companies agree, and that is the settlement of accounts. It is complicated to control the systems and the trades, and it requires staff to do so. 5.2 Barriers towards more co-operations In the interviews the companies expressed satisfaction with the present cooperation. Few actual barriers as such were expressed, except the ones that today‘s situation creates. For example, it is almost geographically impossible to expand the systems to 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 difficult to expand:

actors say that they can optimise the system's performance, and our scenarios have shown that more cooperation could benefit them even more economically. Even though the gain is not extremely high, since the lower system cost would be approximately 5%, there is potential. However, since there seems to be reluctance to cooperate between some actors, it is difficult to fulfil the potential.

A rough rule of thumb has been that the expansion with new customers that have been, (...), has been eaten up by the efficiency we could achieve together with the customers in their buildings. So basically, the heat load has been static in our area for quite some time. (…) …[The reason for building CHP] is the electricity. We, as the producing company, have the problem that we can not expand. We have our two customers and district heating is not a new thing in the municipalities so the chance of getting new customers is limited [22].

Advantages with the cooperation are said to be a possibility to even out the production in the system and thus avoid peak load. The disadvantages with the cooperation are the need for more administrative work to control the system and the trades; the control of the system becomes more complex. This study also confirms previous studies that have pointed out that technical aspects are seldom barriers to cooperation. Most things are solved while the systems are being interconnected, and the will of the persons involved to cooperate is important.

The other representatives are of a similar opinion, that a stagnating load can be expected, and CHP is a way to keep profits high. The Swedish certificate system also makes it advantageous to build new bio-fuelled CHP-plants. Another reason, arguably of a more rhetorical character, is that building CHP is more economically and environmentally correct since the fuel efficiency is higher with CHPs. As scenarios 4–8 show, there is major potential for reducing local and global CO2 emissions.

There is a large potential in building new CHP plants, both from an economic and an environmental perspective. If all the plants in the system were replaced by BCHP or NGCHP, the electricity produced could make up to 4.5 or 12% of total Swedish electricity production, based on the fact that total production in Sweden in 2008 was 146 TWh [1]. The reason that the difference between the electricity productions in those two cases is so large is a big difference between the electrical/thermal outputs (see Table V). The introduction of NGCHP is a less likely future since it can be considered only with the assumption that the natural gas network already exists along the Swedish east cost. On the other hand, introducing more BCHP in the district heating system would increase the system‘s dependence on biomass availability and the heat production cost would become highly sensitive to the solid biomass cost. The actors are highly aware of the potential for CHPs. Since they are expecting a stagnating heat load, the sale of electricity is a way to keep profits high. However, none of them think that natural gas will become a reality in the near future, and even if it did, the introduction is expected to be somewhat problematic, since the fuel can be considered fossil fuel and substantial infrastructure is needed.

In the interviews we also asked questions about the possibility of an introduction of natural gas in the region. Investigations have been made earlier by the above mentioned STOSEB; however, the plans never came to reality. Generally the representatives did not think that an introduction would happen. Since most of them also have strategies to be climate neutral, natural gas probably is not an option. The large investments in infrastructure are another barrier: These are such large infrastructure investments and natural gas is not especially cheap either. It is difficult to come in with natural gas in this energy system. It is rather stable [23]. What the representative here points at is also the inertia in the system. In LTS terms it is called momentum: as the system is stable, it is difficult to change the structure [18], [19]. 6. CONCLUDING DISCUSSION The study has shown the advantages of an interdisciplinary approach. Advantages with interconnections and CHP have been shown in the modelling; however, as there are many different actors involved, there is a need for a will to cooperate. The interconnections have a historical background, with an aim for higher supply security, and today most of them continue to cooperate, despite the fact that the structure and ownership of the companies in some cases have changed since the deregulation of the electricity market in 1996. As previous studies have shown, the main advantages with cooperation have been economic, as is also the case in this system. The

The study has shown a potential for decreased LECO2 and GECO2. The largest potential from a local perspective is from BCHP; so, since the LECO2 would be low and with high electricity production, the potential for lower GECO2 would exist. The high electricity-toheat output ratio in NGCHP has a high potential for decreasing GECO2 of the system. If all plants in the system would be replaced with NGCHP the GECO2 of the system would be -9 million tons annually. However, in that case LECO2 would be much higher than today. 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 does interviewing the actors involved. When combining the methods there is a possibility of getting a better and deeper understanding of the actual potential for cooperation. The historical and social aspects cannot be neglected; they can in many cases explain why potentially beneficial cooperation is or is not done, while modelling can show the actual potential.

[9] M. Melkerson and S-O. Söderberg, Dynamiska elpriser – elprissättning på en integrerad europeisk elmarknad (Dynamic electricity prices – pricing in an integrated European electricity market – in Swedish), Sweden: Institute of Technology, Dept of Mech Eng, Linköping University, Linköping (2004) [10] N. Levinson and R. Freiman, Optimal kraftvärme och nätinvestering I Stockholms fjärrvärmesystem (Optimal CHP and district heating network investments in Stockholm), Linköping University, Linköping (2005)

7. ACKNOWLEDGEMENTS This article has been carried out in two PhD projects in the Energy System Program, financed by the Swedish Energy Agency. The authors would also like to thank Jenny Palm (Linköping University) and Louise Trygg (Linköping University) for valuable comments on the paper.

[11] Statens energimyndighet (Swedish Energy Agency). ‗Marginal elproduktion och CO2-utsläpp i Sverige‘ (Marginal electricity production and CO2emissions in Sweden, in Swedish), Swedish National Energy Administration, ER 14:2002, Eskilstuna, Sweden. (2002)

8. REFERENCES

[12] E. Dotzauer, Greenhouse gas emissions from power generation and consumption in a nordic perspective, Energy Policy 2010, Vol. 38(2), pp. 701-704.

[1] Statens energimyndighet (Swedish Energy Agency), Energy in Sweden 2009. Swedish National Energy Administration, ET 2009:30, Eskilstuna, Sweden (2009)

[13] H. Hansson, S-E. Larsson, O. Nyström, F. Olsson and B. Ridell, El från nya anläggningar (Electric Power from New Plants), Elforsk, Stockholm (2007)

[2] S. Werner, Fjärrvärmens utbredning och utveckling (The development and expansion of district heating), Värmeföreningen, Stockholm (1989)

[14] M. Danestig, A. Gebremehdin and B. Karlsson, Stockholms CHP potential – An opportunity for CO2 reductions? Energy Policy 2007, Vol. 35(9), pp. 4650-4660.

[3] European Union, Directive 2004/8/EC, ‗Directive on the promotion of cogeneration based on a useful heat demand in the internal energy market‘, (2010), homepage: http://www.managenergy.net/products/R81.htm, 2010-04-28

[15] P. Thollander and I-L. Svensson, ‖Vägen till framgångsrika värmesamarbeten – en fallstudie‖ (Road to succesful heating co-operations – a case study), In: L. Trygg, L. et al (2009) optimala fjärrvärmesystemi symbios med industri och samhälle, Rapport 2009:13, Svensk fjärrvärme (Swedish district heating association) (2009)

[4] European Union, , Directive 96/92/EC ‗Second report to the Council and the European Parliament on harmonisation requirements concerning common rules for the internal market in electricity‘, (2010), homepage: http://eur-lex.europa.eu/LexUriServ/LexUriServ.do? uri=CELEX:32003L0054:EN:NOT , 2010-04-28

[16] J. Fors, ‖Spillvärme från industri till fjärrvärmenät – sammanfattning av intervjuer på 5 orter‖, (Excess heat from industry to district heating systems) Rapport 2004:5, Svensk fjärrvärme (Swedish district heating association) (2004)

[5] B. Dahlroth, Öppnade fjärrvärmenät i Storstockholm (Open district heating network in greater Stockholm), Stockholm, Sweden: Fastighetsägarna Stockholm, (2009) [6] D. Henning, Optimisation of Local and National Energy Systems. Development and use of the MODEST model. Dissertation No. 559. Linköpings universitet, Linköping (1999)

[17] S. Grönkvist and S. Sandberg, ―Driving forces and obstacles with regard to co-operation betweenmunicipal energy companies and process industries in Sweden‖, Energy Policy 2006, Vol. 34, pp. 1508–1519.

[7] S. Kvale and S. Brinkmann, Den kvalitativa forskningsintervjun (The qualitative research interview), Studentlitteratur, Lund (2009)

[18] T.P. Hughes, Networks of Power: Electrification in Western Society 1880 – 1930, John Hopkins University Press, Baltimore (1983)

[8] Nordpool. Nordpool electricity spot market, (2009), homepage: http://www. nordpool.com. 2009-12-18.

[19] B. Joerges, Large Technical Systems: Concepts 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

Development of Large Technical Campus Verlag, Frankfurt (1988)

Systems,

[26] STOSEB, STOSEB 92 – Energiframtider för Stockholms län (STOSEB 92 – Energy futures for Stockholms County), STOSEB, Stockholm (1992)

[20] J. Summerton, Changing Large Technical Systems, Westview Press, Boulder, CO (1994)

[27] Fortum & Stadsbyggnadskontoret, Möjlighetsstudie: Nätintegration Storstockholm, (Possibility study: Netintegration in greater Stockholm) Fortum & Stadsbyggnadskontoret, Stockholm (2005)

[21] SCB (Statistics Sweden). Utsläpp av växhusgaser, (2010), homepage: http://www. scb.se. 2010-02-17.

[28] Vattenfall, Head of business development and Senior advisor, 090320

[22] Söderenergi, Production manager, 090318 [23] Fortum, Site manager and Senior advisor, 090325

[29] STOSEB, 25 Energiska år – Om Stor-Stockholms Energi AB (25 Energic years – about greater Stockholms Energy AB), STOSEB: Stockholm, (2003)

[24] Norrenergi, Production manager, 030304 [25] E.ON., Group manager production, 090323

296

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

CUTTING COSTS OF DISTRICT HEATING SYSTEMS BY USING OPTIMIZED LAYING TECHNIQUES 1

Alexander Goebel , Dr. Stefan Holler 1

MVV Energie AG, Mannheim, Germany

ABSTRACT

MATERIALS AND METHODS

The soil covered plastic jacket pipe is the common state of the art laying technique in the district heating sector: A preferable shallow trench is dug out and backfilled with cable sand after the installation of the two pipes. Alternative possibilities concerning the digging of the trench, the backfill and the piping itself are evaluated in this paper. Results show, that an optimized laying technique can save construction or running costs under the right boundary conditions: Backfill materials with insulation properties can reduce the heat losses by about 25 %. Using glass-reinforced plastic pipes (GRP) instead of steel pipes leads to pump energy savings of about 40 %.

The cost saving potentials of the alternatives, concerning the digging of the trench and the backfill, are mainly evaluated by outlining the results of research reports. Calculations are used in order to estimate the insulation properties of special backfill material. Also the cost saving potentials of pipes with low friction losses are evaluated with simple equations. RESULTS Reuse of the excavated material [1], [4] Earlier research activities have proven, that plastic jacket pipes could be used with backfill material showing a greater grit size than cable sand. Special protection material is available not only for the muffles but also for the pipes. Field tests have shown, that there are promising money saving potentials because of the significant reduction of transport and disposal costs. A consideration of reusing the excavated soil is also reasonable from an environmental point of view.

INTRODUCTION In the first place, excavation costs could be cut by digging smaller and shallower trenches. However, this is only possible if the location of the construction site is appropriate. In an urban area the situation is completely different from a rural area concerning space and regulations. The paper describes the boundary conditions and compares different methods from the technical as well as the economical perspective using the example of the district heating system in Mannheim, Germany.

Following points are important, when it comes to an evaluation of this possibility at an individual construction site:       

The second approach which will be presented in the paper is the potential to reuse the excavated material and to use self-compacting material when refilling the trench. Furthermore, it is also possible to use new materials with better insulation properties in order to cut down heat losses. In the paper the different properties of the new materials will be compared and evaluated.

   

A third possibility to reduce costs is the use of specialized piping systems wherever possible. Nowadays a wide range of products is available on the market. In many cases a specialized system fits some applications better than a standard system does. Not only insulation properties but also compensation, ductility and friction losses are important characteristics of modern piping systems. In the paper it will be shown, how costs could be reduced by using less or no compensation measures (cold laying, flexible pipes, fibre pipes), by avoiding welding measures (flexible pipes for house connections, fibre pipes) or by reducing friction losses (fibre pipes).

 

the grading of the excavated material sandy or cohesive ground compacting properties the friction between the ground and the jacket pipe protection measures for muffles and the pipe underground construction regulations a place for the storage of the excavated material (beside the trench, container or any place near the construction site) formation of dust (especially in the summer) contamination pH-value an improvement of the excavated material with lime (especially with cohesive ground) a removal of the coarse material a separation of the material, if the ground is split up in different layers

The use of self-compacting material [1], [4], [5] It is important to distinguish between the following two types of self-compacting material: 

297

stabilised sand mix

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

An economical justification is only achievable, if the transport costs are low and the heat price is high. From a technical point of view, the compaction behaviour has to meet the requirements and regulations. The jacket temperature must not exceed the maximum of 50 °C and the friction between pipe and the material should be in the common range.

excavated material mixed with water and special additives in order to get a self-compacting behaviour The use of self-compacting materials offers a wide range of advantages and applications: 



 



it is possible to dig out a narrower trench, because no machines are needed for the critical compaction around the pipes the backfill process is significant faster – after the trench is filled up, it takes normally only one day until the material is hard enough to walk on self-compaction is more reliable within difficult conditions (many crossing pipes etc) without the use of compaction machines, buildings nearby the construction site are stressed less (no vibrations) there is less inconvenience for residents living nearby the construction site, because of the noise reduction in combination with the pipeline laying technique, the sheeting can be omitted, because nobody needs to work in the trench

A calculation method for heat losses of plastic jacket pipes is described in EN 13941 ANNEX D [2]. Figure 1 shows the influence of the thermal conductivity of soil λs on the heat losses. Normally the value of λs lies in between 1,0 and 2,0 W/m*K [2]. The curve becomes very non-linear below a value of 1,0 W/m*K. This indicates that it is necessary to customise the calculation method in order to get realistic results. The heat losses are cut down by 30%, if the λs is reduced from 1,5 to 0,35 W/m*K. 75

heat losses of the flow and return pipe Φ f + Φ r [W/m]



A common problem is the local availability of the technology. The price is also an issue, if the reason of the application is the approach to save money. Another problem concerning the dimensioning of the compensation measures is the bad predictability of the friction between the jacket pipe and the selfcompacting material. Depending on whether the pipes are taken into service during or after the hardening time, which is about a month long, a more or less crucial tunnel effect is observed [4].

70 65 60 55 50 45 40 0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

Thermal conductivity of the soil λ S [W/m*K]

Fig. 1 Heat losses of a district heating line as a function of λs (DN 250, 120/50 °C, Z = 0,6 m, C = 0,55 m, λi = 0,03 W/m*K)

The reuse of the excavated soil as base material is more elegant, than the stabilised sand mix, because of the recycling aspect. Research projects have even shown that sharp particles are less problematic, because they are enclosed in the self-compacting mass. An advantage of the stabilised sand mix is the easier application.

The insulation material should solely be integrated in the calculation as an additional thermal resistivity (Rλ,embedment), since the soil around the pipes is not made completely out of it. The heat dependency of the insulation foam‘s thermal resistivity should also be taken into account. Figure 2 illustrates, what is meant with ―additional insulation layer‖.

If the district heating line does not run under a street, compaction measures around the pipes can be avoided simply by watering the cable sand, which is filled in layers into the trench. Cost saving potentials of backfill material with insulation properties If a reduction of the heat losses comes into consideration, the change to a higher insulation series is evaluated. Calculations show that in most cases an economical justification is not given for this measure. The idea of filling the trench around the pipes with material that provides an additional insulation seems to be promising. Like in the case of the self-compacting material, the local availability is the greatest problem.

Fig. 2 The different layers of the heat conductivity problem 298

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The modified calculation procedure is made up of the following equations:

1   Di  

R ,steel

Ri 

Rs 

Uflow 

Ureturn 

(3)

 DC  1   ln 2    C  Dinsulation 

 4  (Z  R0  s ) 1  ln 2    s  DC  2  sembedment

  

Rs  Ri ,return  Rh  R  R ,steel  R , jacket  R ,embedment

i  i ,50C  0,0001 Ti ,average  50K 

   (5)

 Tflow  Treturn   Tsoil  2  

f  r  U flow  U return   

(10)

(11)

Ti ,average  Tfluid  Q Ri  R  R,steel  R, jacket  (12) 

(6)

180

without insulation material

30%

with insulation material

160

heat losses of the district heating line [W/m]

(8)

(9)

(4)

 D  2  sembedment 1  ln C 2    embedment DC 

(7)

1 Rs  Ri ,flow  Rh  R  R ,steel  R , jacket  R ,embedment

(2)

D  1  ln insulation  2    i  Do 

R ,embedment 

4    embedment

(1)

D  1   ln o  2    steel  Di 

R , jacket 

Rh 

  2  (Z  R   )  2   0 s    ln1     DC  s embedment    

reduction in %

140

25%

120 100 20% 80 60

heat loss reduction

R 

15%

40 20 0 0

200

400

600

800

10% 1000

nominal diameter [DN]

Fig. 3 Reduction of the heat losses for DN 15 to DN 1000 temperature was at 120 °C and the return temperature was at 50 °C.

The average temperature of the insulation was calculated with following equation and put back into (10). A VBA script was used to iterate five times.

Fig. 3 shows, that savings are significant lower with small diameters. Also the specific thickness of the PUR insulation, which differs because of standardised jacket pipe diameters, has an impact.

Fig. 3 shows the results of the calculation. The insulation material was taken into account with a value of 0,33 W/m*K (λembedment). Around and in between the flow and the return pipe a space of 0,2 m for each pipe size was chosen (sembedment). The depth of cover had a value of 1 m (Z). Like in the previous example, the flow

The heat losses of a DN 250 pipe are reduced from 67 to 51 W/m (24%). This means, that the heat loss reduction is 6% less compared to Fig.1. 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 with huge diameters, a DN 700 pipe was chosen for an example scenario. An annual average for the flow and return temperature was taken into account. For the calculation of the required backfill volume in the embedment, 0.2 m space in every direction of the pipes was estimated. It is important for the calculation to take only the additional costs into account. That means the price difference between cable sand and the insulation material including the transportation costs.

A common value of 6% was chosen for the required rate of return (i). The net present value C0 was calculated with the following equation: T

C0  I   (Rt )  (1  i ) t

(13)

t 1

The internal rate of return shown in Fig. 4 was calculated with the IRR- function in Excel.

Table 1 Scenario for insulation material

Nominal diameter: Average flow temperature: Average return temperature: Annual hours of operation: Heat price (at the time of the invest): Required volume of insulation material:

5000

DN 700 95 50 8760 15 1.85

°C °C h

16 - 17

Required volume for the whole line:

9250

Heat loss with use of the material:

69.6

W/m

Heat loss without use of the material:

90.7

W/m

Energy savings:

23.3

Annual savings of the whole line:

924

MWhth

148,000 – 157,000

Time of cash flow:

Annual growth rate of the heat price:

0 – 3.5

8.0%

40,000 30,000

16 €/m

6.0%

20,000 10,000

17 €/m 1.0%

2.0%

NPV

4.0%

internal rate of return

2.0%

0.0%

3.0%

The growth rate of the heat price is difficult to predict, but has an important influence within the given period of 20 years. Presumably the heat price is mainly influenced by emission trading, governmental subsidies and the development of the fossil fuel price. Other scenarios may estimate higher growth rates, but in order to get realistic results, the rate was varied from 0% to 3,5%.

€

6.0

10.0%

The results in Fig. 4 show, that the additional specific costs should be below 17 €/m3 in order to get a positive value spread, assumed that the required rate of return is 6 %. A reduction of specific costs of 5% (16 €/m3) results in an increase of the value spread by 1%. The net present value after 20 years rises about 10,000 €.

Required rate of return:

Fig. 4 NPV and IRR of the scenario defined in Table 1 depending on the additional specific costs of the insulation material.

€/m

17 €/m

growth rate of the heat price

€/MWhth m

16 €/m

50,000

0 0.0%

Additional specific costs:

Additional investment (I):

60,000

internal rate of return

Length of the district heating line:

Unit

Value

net present value (20 years) [€]

Parameter

Table 2 gives an example of materials with low heat conductivity that could be interesting to use as backfill. It is obvious to look for natural products, because of the price and environmental regulations.

300

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

Flexible systems with a corrugated service pipe have significant higher friction losses, which has to be taken into consideration (dimensioning).

Table 2 Heat conductivity of different materials W/m*K

When it comes to money saving potentials, the most important properties of flexible pipes are the following:

hard plaster [9] 600 kg/m

0.18

900 kg/m

0.30

1200 kg/m

0.43

1500 kg/m

0.56

    

less welding measures self-compensating less insulation work less work concerning the monitoring system less head access holes, because of the reduced welding measures less risk of leaks, because of less weld joins faster laying of the pipes

light sediment natural stone [9]

0.85

 

porous rock, e.g. lava [9]

0.55

Pipes with low friction losses

natural pumice [9]

0.12

Service pipes made of glass-reinforced plastic (GRP) have significant less friction losses than steel service pipes. Because of their chemical resistance, GRP pipes are used mainly in the chemical industry. It is important do distinguish between filament-wound pipes and centrifugally cast pipes. Because of the Poisson‘s effect almost no compensations measures are needed, if filament-wound pipes are used. Centrifugally cast pipes need to be compensated, but have an even smoother inner surface, which means the lowest possible friction losses. Also the temperature resistance is a little bit higher. The greatest problem of GRP pipes is the fact that the service life is cut down by high temperatures in combination with high pressures (derating factor).

bitumen [10] 2100 kg/m

0.70

as matter, 1050 kg/m

membrane, 1100 kg/m

0.17 3

0.23

expanded volcanic rock (perlite) [11] 3

loose perlite, 50 - 130 kg/m

perlite compressed with filaments, 170 - 200 kg/m

Thermosand [8]

0.07 0.06

Fig. 5 shows the possible savings, if a GRP pipe with a surface roughness of k = 0,01 mm is compared to a steel pipe with a roughness of k = 0,2 mm.

0.33

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 service pipe:    

Re 

plastic (e.g. PE-Xa, Polybuten) copper mild steel corrugated stainless steel

w d



(14)

The value of the kinematic viscosity () was taken with 2.941*10-7 m2/s, the density (ρ) with 958.77 kg/m3 (water with 100 °C and a pressure of 10 bar) [7] The pipe friction factor was calculated with the following equation [6]:

Flexible pipes have a significant higher operating pressure (16 or 25 bar) [3], if the service pipe is made of metal. Also the maximum and continuous operating temperatures differ much. Because of this fact, systems with plastic service pipes could normally not be used within huge district heating networks. In a smaller network with lower flow temperatures, which e.g. was built to distribute the heat of a small block heating station, a system with a plastic service pipe might have an application.

k d

  0,11  

68   Re 

0,25

(15)

The pressure loss was calculated with the following equation:

p 

301

 L  w 2 d



(16)

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The equation for the pump power:

CONCLUSION A consideration of an alternative laying technique is usually worth the work, because the money saving potential is often higher than expected. It depends strongly on the single project and the local boundary conditions (heat prices, rural or urban area, availability of technologies/materials etc), whether a different technique makes sense from an economical point of view.



Ppump 

V  p



(17)

Equation (17) shows, that the correlation between pump power and pressure loss is linear. The savings are expressed as a percentage. They do not depend on the diameter or length of the pipe, because only the friction factor differs.

Containing a bunch of alternatives in district laying techniques, Table 3 gives a rough overview with a simple rating. Techniques appearing in the table, which are not discussed is this paper, are listed there, because they are also belonging to the ―alternatives‖ and will be evaluated in future studies. When the word alternative is used, it means every technical aspect, which differs from the standard laying technique defined in the abstract.

reduction of pressure loss

47.5%

45.0%

42.5%

40.0%

Table 3 Overview of alternative laying techniques

37.5%

35.0%

highly recommended to take into consideration from an economical point of view

a closer look seems promising

an economical benefit can be achieved, if special boundary conditions are given

because of technological or economical reasons, the use of the technology is not recommended

the technology is not available or can not be applied reasonable under the given boundary conditions

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

flow speed [m/s]

Fig. 5 Pump energy savings of a GRP pipe

Another important aspect of GRP pipes are the joints: the pipe ends are glued together with a two component adhesive, which is heated up for the curing process. This can be an advantage, because welding measures on a construction site are often problematic (lack of space, wind). Statistics show that in most cases leaks are caused by bad weld seams [12].

302

GRP pipes

flexible pipe

pipeline laying technique

self-compacting material

reuse of excavated soil

combined laying with other supply pipes

stacked laying

insulation as a casting compound

insulating backfill

trench-less laying techniques

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

nominal diameter < DN 150

nominal diameter > DN 150

new district heating line

renovation measure

construction site in an urban area

construction site in a rural area

yet to be built housing estate

existing housing estate

NOMENCLATURE

Rλ,jacket

insulance of the jacket pipe

λs

the coefficient of thermal conductivity for the soil

Rα

insulance of the convective heat transfer inside the pipe

λi

the coefficient of thermal conductivity for the PUR insulation

surface transition insulance

the coefficient of thermal conductivity for the PUR insulation at 50 °C

Rλ,embedment

λi,50 °C

insulance of the insulating material used in the embedment

the coefficient of thermal conductivity for the jacket pipe

inner diameter of the service pipe

λC

outer diameter of the service pipe

λsteel

the coefficient of thermal conductivity for the service pipe

Dinsulation

outer diameter of the PUR insulation

outer diameter of the jacket pipe

λembedmen

tthe

Tflow

flow temperature

Treturn

return temperature

Tsoil

temperature of the soil

Ti,average

average temperature of the PUR insulation

Tfluid

flow or return temperature

Uflow

heat loss coefficient for the flow pipe

Ureturn

heat loss coefficient for the return pipe

coefficient of thermal conductivity for the insulating material used in the embedment

heat transfer coefficient

insulance of the heat exchange between flow and return pipe

insulance of the soil

insulance of the insulation material

Rλ,steel

insulance of the service pipe 303

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

distance from surface to the middle of the pipe

distance between the centre lines of the two pipes

sembedment

space between the pipe and the trench wall

Φf + Φr

heat loss per pipe pair

[3] European Committee for EN 15632, Brussels (2009)

net present value

investment

[4] Dipl.-Ing. Heinz-Werner Hoffmann, Dipl.-Ing. Torsten Göhler and Dr.-Ing. Manfred Klöpsch, Fernwärmeleitungsbau mit Recyclingmaterial, MVV Forschungsbericht, Mannheim (2006)

net cash flow (annual savings)

required rate of return

given period

the time of the cash flow

surface roughness of the service pipe

Reynolds number

flow speed

inner diameter of the service pipe



the kinematic viscosity

the pipe friction factor

Δp

pressure loss

length of the pipe

the density of the heating water

Ppump

pump power

REFERENCES [1] Alexander Goebel, Alternative Verlegesysteme, Mannheim (2010)

[2] European Committee for Standardization, EN 13941 ANNEX D, Brussels (2009)

flow rate

pump efficiency

Standardization,

[5] Dipl.-Ing. Heinz-Werner Hoffmann and Zoltan Dioszeghy-Günter, Neuartige Verlegetechniken für das Kunststoff-Verbundmantelrohr-System Band 1, MVV Forschungsbericht, Mannheim (1995) [6] Fratzscher et. al., Energiewirtschaft für Verfahrenstechniker, VEB Deutscher Verlag für Grundstoffindustrie, Leipzig (1982) [7] VDI-Wärmeatlas, Verein Deutscher Ingenieure, Heidelberg (2006), Dba 5, Dba 13 [8] KE KELIT Kunststoffwerk Gesellschaft m.b.H., Thermosand© (Broschüre), Linz (2006), p. 6 [9] VDI-Wärmeatlas, Verein Deutscher Ingenieure, Heidelberg (2006), Ded 12 [10] VDI-Wärmeatlas, Verein Deutscher Ingenieure, Heidelberg (2006), Ded 10 [11] Heinz Schmid, Excel mit VBA in der Wärmetechnik, C. F. Müller Verlag, Heidelberg (2008), p. 26 [12] Dipl.-Ing. (FH) Frank Espig, Schadensstatistik KMR 2007 des AGFW, article published in the „EuroHeat&Power― (2008), issue 10



Fernwärme-

304

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

ANALYSIS OF HEAT TRANSFER IN HEAT EXCHANGERS BY USING THE NTU METHOD AND EMPIRICAL RELATIONS 1

O. Gudmundsson, O. P. Palsson and H. Palsson

Faculty of Industrial Engineering, Mechanical Engineering and Computer Science Hjardarhagi 2-6, IS-107 Reykjavik, Iceland a steady state condition it has proven to be relatively simple, since analytical and empirical relations can be derived for different heat exchanger types and used for all necessary calculation regarding time invariant conditions, see e.g. [1]. If a dynamic operation exists it becomes more complex to monitor the condition of the heat exchanger and more complex models are used, see e.g. [2] and [3].

ABSTRACT Heat exchangers are widely used in domestic and industrial applications involving transfer of energy from one fluid to another, for example in district heating systems. The wide usage underlines the importance to have a good technique to detect if the effectiveness of an heat exchanger is diminishing. There are number of things that can cause diminishing effectiveness of an heat exchanger, for example fouling, changes in fluid properties as well as corrosion. In many cases the fouling is a particular problem, for example when geothermal water is used. Geothermal water is very mineral rich which can cause serious fouling problems. The method presented in this paper is simple and easy to use and can be used to detect a diminishing heat transfer coefficient in many types of heat exchangers, in this paper the method is used on cross flow heat exchanger. The method uses measurements of the inlet and outlet temperatures as well as the mass flows, these measurements are usually easy to gather under normal operation. The method uses the well known Number of Transfer Units (NTU) method as well as empirical relations to estimate the overall heat transfer coefficient, which is then statistically analyzed. The data used in this study was gathered from a simulated cross-flow heat exchanger where the overall heat transfer coefficient was gradually decreased to simulate diminishing effectiveness of the heat exchanger. The conclusion of this study shows that the derived detection method can detect fouling based on the data from a simulated cross-flow heat exchanger, with a good accuracy and consistency. Further analysis on real data is scheduled.

In this study, a mathematical model is used that has been developed to simulate accurately the temperature and flow transients in a cross flow heat exchanger. The model is based on the finite volume method (FVM) where a mathematical representation of a general cross flow heat exchanger is solved numerically. One possible application of such a model is to generate data that can be used to compare and tune more simple dynamic models based on either black box methods or state space modelling. An important application in this context involves methods to detect fouling in heat exchangers under dynamic operation. Description of the model can be seen in [4]. Fouling in heat exchanger can be categorized in the following categories, precipitation fouling, chemical reaction fouling, corrosion fouling, particulate fouling, biological fouling and freezing fouling. Usually fouling is a combination of the categories. The fouling process in heat exchanger can be described as a process where the separating metal inside the heat exchanger accumulates deposits from the fluids. This is very common and poses problems and results in reduced efficiency of the heat exchangers. There are numerous methods available to address the effect of fouling, see [5â&#x20AC;&#x201C;8]. Finally, decrease in the thermal efficiency of a heat exchanger due to property changes in a working fluid will have similar effect on the heat exchanger as fouling.

INTRODUCTION Heat exchangers are widely used in domestic and industrial applications involving transfer of energy from one fluid to another. General classification of heat exchangers are parallel flow, counter flow and cross flow. Their size and complexity can also vary greatly. Their operating conditions can be classified into two main classes, steady operation where mass flow and temperatures are relatively constant and dynamic operation where mass flow and temperatures can vary greatly with time.

There are number of ways to detect fouling but according to [9], classical methods involve a) examination of the heat transfer coefficient, b) simultaneous observations of pressure drops and mass flow rates, c) temperature measurements, d) ultrasonic or electrical measurements and e) weighing of the heat exchanger plates. Methods aâ&#x20AC;&#x201C;c) require the heat exchanger to be operating in steady state condition, d) can only monitor local fouling and e) requires the process to be stopped. These restrictions can be too strict or costly. Another

During operation it is important to have some knowledge of the condition of the heat exchanger. For 305

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

allowed to progress to a maximum of Rf=0.00033, which corresponds to 25% decrease in the overall heat transfer coefficient.

approach is to model the heat exchanger and look for discrepancy between model predictions and what is actually measured, see [10] and [4]. The method used in this study falls into category a). To make the method valid for dynamic operating conditions, empirical relations for the mass flow rates are furthermore used.

THE DETECTION METHOD The fouling detection is done by estimating the overall heat transfer coefficient, U, by using NTU relations and monitor the means of U for shift that can be related to diminishing efficiency either because of accumulation of fouling or property changes of the working fluid. NTU method is commonly known and a description of it can be seen in [1].

Although district heating systems usually operate in relatively steady state it can be argued that methods that work well to detect diminishing efficiency under dynamic operation should work very well under steady state condition. DATA USED

It is known that effectiveness of a heat exchanger can be calculated by

The data used in this study was the same data as was used in [4]. The data was generated by a simulator representing an unmixed cross flow heat exchanger. The advantage of using simulated data is that it is possible to control when and how much fouling will occur in addition to controlling the inlet temperatures and the mass flows. The data used had temperatures for the hot side in the interval [53, 67] 째C and the cold side [12, 27] 째C, the mass flow rates for the hot and cold side were in the interval [0.30, 1.45] kg/s. Description of the simulator can be found in [4].

(1) The minimum fluid is the fluid that has the minimum value of the production of mass flow and specific heat, . Effectiveness for a unmixed cross flow heat exchanger can also be calculated by the following relations of the effectiveness to NTU.

Fouling

(2)

During design a heat exchanger is commonly designed to operate under mild fouling by assuming a fouling factor in the interval 0.0001 to 0.0007. According to [11] and [12] there is usually an induction time before a noticeable amount of fouling has accumulated. In [13] it is shown that the fouling will grow with increased rate during the fouling period. Figure 1 shows the evolution of the fouling factor from the time the heat exchanger starts to accumulate fouling until the simulation is stopped. A dimensionless time is used to make easy comparison between different lengths of data series.

In normal use, the overall heat transfer is usually unknown and it is therefore not possible to calculate NTU directly. It is therefore necessary to estimate NTU from the relation between NTU and the effectiveness. The estimation is done by minimizing a score function with respect to NTU. The minimization was done by using the minimization routine fmincon in Matlab, see [14]. The parameter NTU is defined by (3) From Eq. (3) it is easy to derive the formula for U

(4) EMPIRICAL RELATIONS In the case of heat exchanger under dynamic operation where big variations can occur during operation, it is hard to see shift in the overall heat transfer coefficient that can be related to diminishing efficiency in the heat exchanger. In [15] it is proposed to use empirical relations of U to make a heat exchanger model valid over a wide range of operating conditions. The heat transfer coefficient can be written as

Figure 1. Evolution of the fouling factor from the time

The simulated data sets used in this study include 200 sets without fouling and 200 sets with fouling, the data sets are further divided equally between slow and fast fouling. In the fouled cases the data set was without fouling for the first 25% and then the fouling factor was 306

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The effect of the empirical relations can be seen in Figure 2. It can clearly be seen that including the empirical relations really helps to reduce the variations in the overall heat transfer coefficient.

(5) Where i is constant, ii is temperature dependent and iii is mass flow dependent. In this study only mass flow dependency was considered, since it has been shown by previously by [15] that the temperature dependency can be neglected. The relation for the heat transfer coefficient can therefore be written as (6) By assuming that Eq. (6) applies to both the hot and the cold side and neglecting the thermal resistance in the separating metal, the overall heat transfer coefficient, U, can be written as

(7)

Figure 2. The figure shows the evolution of the number of transfer units and the overall heat transfer coefficient with and without the empirical relations.

where y is the exponent of the Reynolds number. In [1] it is recommended to use y=0.8 for turbulent flow, which is expected in a heat exchanger.

To detect fouling a CuSum chart is used, see [16]. The CuSum chart was chosen since it is known to be effective to detect shift in mean values. When using CuSum charts it is necessary to define two CuSum parameters, a decision limit to prevent false detection and a reference value for deviations. Detection is made when the cumulative sum of deviations goes over the decision limit.

It can be practical to normalize U with a reference mass flow. The overall heat transfer coefficient according to the reference mass flow

and

is similarly

It can be seen in Figure 3 that the method is very consistent in detecting diminishing efficiency. Figure 4 shows the detection if no empirical relations are used.

(8) After inserting Eq. (7) and (8) into Eq. (4) to make it mass flow dependent and normalizing, the estimated overall heat transfer coefficient will become

(9) The overall heat transfer coefficient in Eq. (9) is the variable that is used to detect the fouling in the heat exchanger. RESULTS As mentioned above the method was applied to the same data set as was used in [4]. Measurement errors were added to the inlet and outlet temperatures as well as the mass flows to make the measurements more realistic. Measurement errors of 0.2 Â°C were assumed on the temperatures and 1â&#x20AC;&#x201C;2% measurement errors to the mass flows.

Figure 3. The CuSum chart quickly detects the shift in the overall heat transfer coefficient.

307

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

method seems also to be very stable in detecting the fouling. DISCUSSION AND CONCLUSION The results indicate that the method proposed can be used to detect fouling in cross flow heat exchangers operating under dynamic condition by using measurements that can be obtained under normal operation. The detection method is based on the well known method of Number of Transfer Units, with addition of empirical relations to make the method valid over wide range of mass flow rates. By monitoring the calculated overall heat transfer coefficient, it is possible to detect changes that are due to fouling or changes in the working fluid. Unlike conventional methods, this method can detect fouling in heat exchangers that are not operated in steady state conditions. The fouling detection is performed within the designed fouling factor interval.

Figure 4. The CuSum chart quickly detects the shift in the overall heat transfer coefficient.

Comparison of Figures 3 and 4 shows that it is possible to detect fouling in heat exchangers operating in dynamic condition with quite good accuracy by using the NTU method and empirical relations.

Further work will include application of the method on data from a real heat exchanger.

In Table 1 a comparison between the method in [4] and the method presented in this paper is shown. From the table it is apparent that the method presented in this paper gives better results. The fouling detection interval for the drift corresponds to fouling factors on the intervals [0.00002, 0.00004] and [0.00001, 0.00003] respectively for the fast and slow fouling. The method is therefore giving considerable better results than the method described in [4].

ACKNOWLEDGEMENT This work has been supported by the Environmental and Energy Research Fund of Orkuveita ReykjavĂkur, National Energy Fund and Energy Research Fund of Landsvirkjun. REFERENCES

Table 1: Comparison of detection time between the two methods, where method 1 is from [4]

[1] J. P. Holman Heat Transfer. Ninth edition, McGraw Hill, 2002.

Method 1

[2] M. Mishra, P. K. Das and S. Sarangi. "Effect of temperature and flow non-uniformity on transient behaviour of crossflow heat exchanger". International Journal of Heat and Mass Transfer, 2008, p. 2583-2592.

Percentiles

Method 2 Fast

2.5%

0.59

0.26

50%

0.83

0.35

97.5%

0.98

0.40

Percentiles

[3] H. Kou and P. Yuan. "Thermal performance of crossflow heat exchanger with nonuniform inlet temperatures". International Communications in Heat and Mass Transfer, 1997; 51(9-10):357-370.

Slow

2.5%

0.63

0.23

50%

0.81

0.30

97.5%

0.93

0.35

[4] O. Gudmundsson, H. Palsson and O. P. Palsson. "Simulation of fouling in cross-flow heat exchanger and a fouling detection based on physical modeling". In: Proceeding of The 50th Conference on Simulation and Modelling, Fredericia, Denmark, 7-8th of October, 2009.

Typical fouling factors are, as stated above, on the interval [0.0001, 0.0007]. The results therefore indicate that the method can be used to detect fouling in cross flow heat exchangers that are operating in non-steady state condition prior to the time a typical fouling factor heat exchangers are designed for is reached. The

[5] W. L. Pope, H. S. Pines, R. L. Fulton and P. A. Doyle. "Heat exchanger design "why guess a fouling factor when it can be optimized?". Energy Technology Conference and Exhibition. Huston, Texas, 1978. 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. MüllerSteinhagen. "Ion implantation of stainless steel heater alloys for anti-fouling applications". In: Proceedings of the International Conference on Ion Implantation Technology, 1999;2:869-872.

[11] B. Bansal and X. D. Chen. "Fouling of heat exchangers by dairy fluids – a review". In: Proceeding of Heat Exchanger Fouling and Cleaning – Challenges and Opportunities, Kloster Irsee, Germany, June 5-10, 2005.

[7] P. K. Nema and A. K. Datta. "A computer based solution to check the drop in milk outlet temperature due to fouling in a tubular heat exchanger". Journal of Food Engineering. 2005;71:133-142.

[12] F. Fahiminia, A. P. Watkinson and N. Epstein. "Calcium sulfate scaling delay times under sensible heating conditions". In: Proceeding of Heat Exchanger Fouling and Cleaning – Challenges and Opportunities, Kloster Irsee, Germany, June 5-10, 2005.

[8] S. Sanaye and B. Niroomand. "Simulation of heat exchanger network (HEN) and planning the optimum cleaning schedule". Energy Conversion and Management. 2007; 48:1450-1461.

[13] M. W. Bohnet. "Crystallization fouling on heat transfer surfaces – 25 Years research in Braunschweig". In: Proceeding of Heat Exchanger Fouling and Cleaning - Challenges and Opportunities, Kloster Irsee, Germany, 5-10th of June, 2005.

[9] G. R. Jonsson, S. Lalot, O. P. Palsson and B. Desmet. "Use of extended Kalman filtering in detecting fouling in heat exchangers". International Journal of Heat and Mass Transfer, July, 2007;50(13-14):2643-2655.

[14] MathWorks http://www.mathworks.com/. 20th of April 2010.

[10] O. Gudmundsson, O. P. Palsson, H. Palsson and S. Lalot. "Fouling detection in a cross flow heat exchanger based on physical modeling". In: Proceeding of Heat Exchanger Fouling and Cleaning, Schladming, Austria, 14-19th of June, 2009.

[15] G. R. Jonsson and O. P. Palsson. "Use of empirical relations in the parameters of heatexchanger models". Industrial and Engineering Chemistry Research, June, 1991;30(6):1193-1199. [16] NIST/SEMATECH e-Handbook of Statistical Methods, April 30, 2009, http://www.itl.nist.gov/div898/handbook/.

309

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

HEAT LOSS ANALYSIS AND OPTIMIZATION OF A FLEXIBLE PIPING SYSTEM 1

J. Korsman , I.M. Smits and E.J.H.M. van der Ven

Liandon B.V. Thermaflex International Holding B.V. (PP) or Polybutylene (PB) improves flexibility. From these, PE does not have adequate strength at higher temperatures and PP is rather stiff. This leaves PE-X and PB, of which the latter can be welded without difficulty. This is therefore the material of choice for this study. In accordance with the temperature duration profile mentioned in the BRL5609/EN15632, PB is suitable up to a maximum temperature of 95 °C.

ABSTRACT The object of this paper is to evaluate heat losses of a flexible PB-PE-PE piping system in the field, compared to a conventional rigid Steel-PUR-PE piping system. The flexible system is optimized in both insulation quantity (thickness) and quality. The heat loss for pairs of pipes in the field, with 70 °C supply and 40 °C return temperature, is based on heat loss measurements in the laboratory and has been evaluated using the multipole method.

As with other plastics, PB is prone to some diffusion of oxygen and water vapor. These effects have been investigated by Korsman et al. 2008 [7]. To prevent oxygen diffusion, an EVOH oxygen diffusion barrier may be used. Unless fully submerged for years, the diffusion of water vapor will not be much of a problem. When kept completely under water, it will take at least 30 years for all cells to fill with condensate.

Since the hydraulic properties of Polybutylene and steel medium pipes differ, hydraulic calculations of a demonstration distribution network, fitted with either system, are made. Total system heat losses for this demonstration network are calculated by summing the product of the heat loss per pair of pipe and the amount of pipe used.

It is not all that easy to compare the heat loss of a conventional Steel–PUR–PE piping system to a flexible PB–PE–PE piping system. Internal diameters differ, as does the friction coefficient, because PB is smoother than steel. A pipe for pipe comparison yields skewed results. One way around this problem is to compare complete distribution systems, as was done in Korsman et al. 2008 [2]. A demonstration (or reference) network is used to design and compare similar networks. For reference purposes, the same network is used in this study.

INTRODUCTION Flexible piping systems for district heating and cooling have several advantages when compared to rigid piping, mainly during installation, some even in use. Flexible pipe can be utilized much like cable, arrives on large reels, requires less engineering and fewer has joints. However, flexibility comes at a price. It seems harder to reach comparable levels of insulation, see Smits et al. 2010 [1].

The differing properties that complicate comparison between piping systems, can also be used to minimize distribution system heat loss. The object of this study is to reach comparable heat loss for the flexible system, by exploiting specific properties, whilst transporting the same amount of heat with comparable pressure losses.

The reason for this lies in the specific properties of the material most commonly used for insulation: Polyurethane foam. PUR foam has a crystalline structure and tends to be quite rigid. It is not very suitable for flexible applications. Bending may lead to a breakdown of the crystalline structure and may also compromise the bonding between foam and medium pipe, thus creating a channel. This channel may accelerate the exchange of foaming agent and air with the environment, thereby speeding up the ageing process. Flexible variants of PUR are available, but do not seem quite as good. Insulation foam made of polyolefins show ample flexibility and quite good insulation properties for small diameters. Furthermore, aging typically is a faster process than in rigid systems.

1. HEAT LOSS IN THE GROUND Heat losses have been measured on test rigs as described by van Wijnkoop et al. 2010 [3] and have been evaluated by van der Ven et al. 2010 [4]. With the results of these tests, the in-ground heat losses are calculated using the multipole method by Johan Cleasson and Camilla Persson in 2005 [5]. Note that the mentioned heat losses are calculated for a pair of pipes, run at 70 °C supply and 40 °C return temperature.

As for the medium pipe, metals may be flexible enough for the smaller diameters, but are too rigid for the bigger pipes. Again, using polyolefins like Polyethylene (PE), cross linked Polyethylene (PE-X), Polypropylene

For rigid piping, some room between the pipes is required for welding, see Fig. 1a. 310

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The installation of a supply and return closely together in the ground has a small, but positive effect on the heat loss of the pair, as is shown in Fig. 3. PB std insulation thickness

PB close

50 45

Heat Loss [W/m]

Fig. 1a, Supply and return pipe

For flexible systems, there is no space requirement between the pipes for installation purposes, and pipes are best installed right next to each other, see Fig. 1b.

35 30 25 20 15 10 5 0 16

110

Nominal diameter [mm]

Fig. 3, Heat loss per pair, standard and close together

For the rest of this paper, flexible pipes are supposed to be installed closely together. 3. PIPE PER PIPE COMPARISON Even though internal diameters and friction coefficients are different between Steel-PUR-PE and PB-PE-PE and therefore will lead to a different selection of diameters in the engineering process, an approximate comparison can be made, see Fig. 4.

Fig. 1b, Supply and return closely together

When supply and return are installed closely together the temperature profile in the ground is altered in benefit of the in-ground heat loss. Dependant on the refill, it may be difficult to achieve defined compaction when supply and return are installed too closely together. However, similar temperature effects can be reached by installing likewise in vertical orientation.

St std

PB close

50 45 40 Heat Loss [W/m]

As mentioned before, the in-ground heat losses are calculated using the multipole method by Johan Cleasson and Camilla Persson in 2005 [5]. See Fig. 2 for a calculated temperature profile.

35 30 25 20 15 10 5 0 16

110

Nominal diameter [mm]

Fig. 4, Heat loss per pair, flexible PB versus Steel

On the left of the graph, two diameters are included for which no steel counterpart has been incorporated. The reason for this is that PB allows for higher fluid velocities. However, in the current range, using the smaller diameters does not create a heat loss reduction.

Fig. 2, Temperature profile supply and return closely together in the ground 311

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The graph in Fig. 6 shows that the modified range (in blue) is much closer to the reference (Steel-PUR-PE, in purple). Heat loss is (almost) proportional to the diameter, which seems about right. However, the modified range of insulation thickness in Fig. 5 (purple) still shows a somewhat erratic distribution, which suggests that the current range of customary dimensions for the outer casing does not lead to an optimal distribution of insulation thickness. It may prove worthwhile to develop a new range of outer dimensions, adapted to the diameter of the medium pipe.

4. SELECTION OF INSULATION THICKNESS The reason for the relatively high heat losses for the two smallest diameters is explained by Fig. 5. PB std insulation thickness

PB extended insulation thickness

Thickness insulation [mm]

40 35 30 25 20

5. IMPROVEMENT OF INSULATION QUALITY

Due to the new testing facilities described by van Wijnkoop et al. in [3], the process of product improvement has been speeded up considerably. During the course of the investigations, resulting in this paper, it is becoming clear that further improvement of the insulation quality is feasible. The measurement principle used for the determination of heat losses does not allow for direct measurement of the insulation properties of the foam; however, some sort of ―equivalent lambda‖ can be derived from the data by calculation. As explained by van der Ven et al. in [4], insulation quality differs for different diameters. For production reasons, it is not expected that insulation quality will reach the same level over the entire product range. Typically, the higher values will be reached in the smaller dimensions. Still, an educated guess can be made as to which levels are feasible from a technical viewpoint, see Table 1.

10 5 0 16

110

Nominal diameter [mm]

Fig. 5, Insulation thickness, standard and increased

In red, this graph contains the insulation thickness of the current PB-PE-PE product range. The somewhat erratic distribution of insulation thickness over the range is caused by the use of customary dimensions for the outer casing. It can be seen that for the two smallest diameters the insulation is rather thin, which explains the relatively high heat losses. In purple, the graph in Fig. 5 shows a modified range, with increased insulation thickness for some of the smaller diameters, as in general it is easier to achieve a good insulation quality for the smaller dimensions. The heat losses of the modified range were calculated and are presented in Fig. 6.

Table 1, Improved insulation quality, ―equivalent‖ or ―synthetic‖ lambda at 50 °C mean temperature Type 50A25

Area 1074

Lambda fresh 0.0283

63A32

1701

0.0287

0.0330

75A40

2364

0.0291

0.0335

90A50

2993

0.0295

0.0340

90A40

3670

0.0300

0.0345

90A32

3886

0.0301

0.0346

125A63

6204

0.0316

0.0363

160A90

10790

0.0345

0.0397

160A75

12611

0.0357

0.0411

200A110

16879

0.0385

0.0442

St std PB std insulation thickness PB extended insulation thickness

Heat Loss [W/m]

40 35

Please note: The lambda values in Table 1 are not the measured lambdas of samples of the insulation foam, and may not be interpreted as actual physical properties of the insulation material. The values were calculated on the basis of heat loss measurements of sections of pipe according to EN15632, and therefore are some sort of ―synthetic system lambdas‖.

5 0 16

32 40 50 63 75 Nominal diameter [mm]

Lambda Degassed 0.0326

110

Fig. 6, Heat loss per pair, including increased insulation thickness 312

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The red graph in Fig. 7 represents the predicted heat loss values for the combined effect of both increased insulation thickness and insulation quality improvement. For most diameters, these are on par with or slightly better than the reference in Steel-PURPE. These data are valid only for the recently produced or ―fresh‖ product. As there is no experimental data available on the rate of degassing and therefore the rate of ageing, it is difficult to predict heat loss over the life time of the product.

The values presented do of course largely depend on the actual physical lambdas (in W/m.K) of the insulation material, but the underlying measurement data suggest that other factors come into play as well, such as the geometry of foam in combination with the temperature dependence of the ―physical‖ lambda of the foam. Therefore, the values in Table 1 are valid only for calculation / prediction purposes, in exactly the same calculation model from which they were derived (also according to EN15632). The values in Table 1 are supported by experimental data on four samples at the time of writing this paper.

However, it is possible to speed up the process of ageing artificially, until all the foaming agent has been replaced by air. The predicted values for this condition are also presented in Table 1, as lambda degassed. These are ―synthetic‖ as well, and suitable for calculation purposes only. Calculated heat loss results with these values are presented in Fig. 8.

When all parameters are known, equation 1 can be used to calculate heat loss:

 

  i 



2  Tprobe  Tcasing



(1)

 d2  1  d3  1  d4   ln    ln    ln  s  d1   i  d2   c  d3  1

Pump I, power: 16 kW St std

Where:

 

Tprobe, Tcasing represent probe (medium) and casing temperature d1 to d4 represent inner/outer diameters of service pipe and casing λs, λi, λc = heat coefficient of service pipe, insulation and casing

15 10 5 0 90

110

Of course, ageing is also applicable to the reference product, but not included here for two reasons. First, the ageing process for rigid systems is expected to be significantly slower than for flexible systems, and second, the reference samples were not fresh, as could be judged by the gas content. Therefore, it is not likely that the values presented for the reference system will deteriorate much further during lifetime.

In Fig. 8, the purple graph represents the reference, Steel-PUR-PE as measured, see Smits et al. 2010 [1]. The red graph represents the prediction of improved, fresh PB-PE-PE and green the prediction of fully degassed PB-PE-PE. The values vary a bit, but are generally in the same range. During the lifetime of the product, heat loss is expected to increase from the red values to the green values.

Nominal diameter [mm]

84 %

Fig. 8, Heat loss per pair, including improved insulation quality, fully degassed

St std PB std insulation thickness PB extended insulation thickness PB impr. fresh

For the Steel-PUR-PE reference, see [1], values can be determined in a similar fashion. ―Synthetic‖ λi values for PUR foam, determined from measurement of samples, were typically in the range of 0.030 to 0.032 W/m.K, with λs for steel 50 W/m.K

and λc= 0.40 W/m.K. On a test rig, T probe, Tcasing and heat loss are measured, so for specific test samples, eq. 1 can be used backwards to calculate ―synthetic‖ values for λi in Table 1.

14 %

In this case, λs and λc are known: λs = 0.19 W/m.K

Heat Loss [W/m]

PB impr degassed

Heat Loss [W/m]



PB impr fresh

110

Ageing can be slowed down considerably if measures are taken to prevent the exchange of blowing agent with the environment. If successful, these measures

Nominal diameter [mm]

Fig. 7, Heat loss per pair, including improved insulation quality 313

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The graph in Fig. 10 represents the pressure in the supply network (in m water column), as a function of the distance from the source. For standard symmetrical networks, the return network is similar, but mirrored over a horizontal axis.

would result in the â&#x20AC;&#x2022;redâ&#x20AC;&#x2013; heat loss values during lifetime. Moreover, a new generation of blowing agents is under development. These new agents aim at lower conductivity values for the gas and larger molecules. This may result in lower conductivity values for the product as well as a slower ageing process.

Using the flexible and smooth PB pipes allows for smaller diameters, mainly because PB is less prone to the transmission of hydraulic noises. This is due to the low modulus of elasticity of PB when compared to steel. In contrast, a steel pipe filled with water is quite a good conductor of sound. To prevent noise caused by high flow velocities, these are limited in the design for steel networks to 1 m/s.

6. HYDRAULIC CALCULATIONS The pipe per pipe comparison between Steel-PUR-PE and PB-PE-PE as demonstrated in Fig. 8, gives an indication of field results, but is not conclusive. Internal diameters differ, as do friction coefficients. Therefore, for the comparison between distribution systems fitted with either pipe, hydraulic calculations are needed. To this end, a reference network is introduced in Korsman et al. 2008 [2]. The same network is used here. It is installed in a housing estate near Arnhem, the Netherlands, and has been designed using Pipelab, developed by Prof. Dr. PĂ ll Valdimarsson in 1995 [6]. See www.pipelab.nl. Standard design criteria were used. A total of 247 houses are connected by 3.02 km of DH network (6.05 km of pipe), 12.2 m per house.

A network, specifically designed for PB, is shown in Fig. 11. Smaller diameters in the periphery of the network as a result of a higher permitted fluid velocity causes higher pressure drops. This has to be compensated by bigger pipes closer to the source to reach the same overall pressure drop.

Fig. 11, Design pressure drop PB network

Fig. 9, Aerial photograph of reference housing estate

In the design of district heating networks, the maximum design point is chosen considerably below the sum of the installed power in the connected buildings. It is not uncommon to have a design point of 50% of the total installed power for larger numbers of connections, depending on the experience and the courage of the designer. A design point of 50% of the total installed power was used in both designs in this paper. In practice, no problems have arisen with this design point, partly because not all installed power is used at the same time. However, this statistical effect does not apply to individual connections. Therefore, a design trick is used in the periphery of the network, to prevent problems in the service pipes connecting the buildings. The flow in these pipes is raised artificially above the design point, up to 100% load. The result of this calculation is shown in Fig. 12, which can be compared to Fig. 10.

Fig. 10 shows an output graph of Pipelab.

Fig. 10, Design pressure drop steel network 314

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia (1) ST.PUR.PE Ref

(2) PB as measured std eng.

1200

Pipe Length [m]

1000 800 600 400 200 0 16

Fig. 12, Pressure drop steel network with increased service pipe flow

90 110

Nominal diameter [mm]

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 design maximum fluid velocity is rather low for steel.

As a result of the use of smaller diameters with PB, the distribution of pipe lengths generally shifts to the left in the pipe histogram. As heat loss increases with diameter (see Fig. 8) this should have a positive effect on the total distribution system heat loss.

This may prove different for the PB network, which is designed with smaller diameters in the periphery. See Fig. 13, which can be compared to Fig. 11.

This shift to the left may be taken one step further, since the wall thickness of the smallest PB medium pipes currently is chosen a bit larger than the strength class (SDR11) requires. This is done for ease of installation. If the thickness of these pipes is chosen no larger than SDR11, there is a slight additional shift to the left, see Fig. 15.

The graph in Fig. 13 indeed shows an increased pressure drop in the service pipes connecting the houses, when the flow in those pipes is artificially increased to 100% of the installed power. However, the total pressure drop stays within the same limits as does the steel network under similar conditions (Fig. 12).

(1) ST.PUR.PE Ref (3) PB impr. Fresh all SDR11

(2) PB as measured std eng.

1200

Pipe Length [m]

1000 800 600 400 200

Fig. 13, Pressure drop PB network with increased service pipe flow

0 16

90 110

Nominal diameter [mm]

The result of both design calculations is plotted in Fig. 14, steel in red and PB in green.

Fig. 15, Pipe length histogram steel, PB and PB SDR11 networks

315

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

There is currently no experimental data on the rate of ageing of PB-PE-PE as a result of the exchange of blowing agent with air. However, it is possible to calculate a worst case situation (see fig. 18), using the predicted values plotted in fig. 8.

6. TOTAL SYSTEM HEAT LOSS To calculate the total system heat loss, pipe lengths as shown in the pipe length histograms are to be multiplied by the respective heat losses per pipe pair, as shown in the heat loss value histograms. Fig. 15 â&#x20AC;&#x2022;multipliedâ&#x20AC;&#x2013; 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 PB as measured std eng. PB increased insulation thickness PB increased insulation thickness all SDR11

70.0 60.0

Heat Loss [kW]

80.0 70.0 Heat Loss [kW]

60.0 50.0

40.0 30.0

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 to the worst case value in blue. Average heat loss during lifetime will be somewhere in-between.

The graphs in Fig. 16 show the reference heat loss for steel-PUR-PE in red and the currently measured heat loss for PB-PE-PE with non-optimized insulation thickness in orange. In green, total system heat loss is shown for current insulation quality but with optimized insulation thickness. The exclusive use of SDR11 (blue) has a rather small effect.

FUTURE RESEARCH Hydraulic calculations in combination with insulation thickness form an interesting optimization problem: what diameter to select and which insulation thickness to choose?

Improving insulation quality, as described in paragraph 5 and shown in Fig. 7, leads to slightly lower total system heat loss for freshly produced PB-PE-PE when compared to the reference, see Fig. 17. ST.PUR.PE Ref

Current design strategies for hydraulic networks, aiming at linear pressure drop with distance, seem too adventitious to be optimal. In addition, heat loss calculations using standard casing dimensions show rapidly diminishing yields with each step up in insulation thickness, suggesting the optimum is somewhere in-between.

PB impr. insulation thickness/quality

60.0 50.0

Heat Loss [kW]

50.0

First attempts have been made to use Pipelab [6] in a double optimization routine, trying to find optimal hydraulic performance in combination with optimal insulation thickness distribution over the network.

40.0 30.0

Given the specific hydraulic properties of PB (high fluid velocities permitted) and the specific insulation properties of PE foam (better at small size), this may lead to rather different design strategies when compared to conventional rigid piping systems for district heating and cooling.

20.0 10.0 0.0 system [-]

Fig. 17, Total system heat loss, reference and prediction for improved insulation quality 316

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

FURTHER INFORMATION

REFERENCES

Questions concerning the paper can be addressed to:

[1] I.M. Smits, J.T. van Wijnkoop, E.J.H.M. van der Ven, ―Comparison of competitive (semi) flexible piping systems by means of heat loss measurement‖, in Proc. of the 12th International Symposium on District Heating and Cooling, Tallinn, Estonia (2010).

Liandon B.V. www.liandon.com www.pipelab.nl Dijkgraaf 4 6920 AB Duiven The Netherlands

[2] J. Korsman, I.M. Smits, S. de Boer, ―System optimization of a new plastic piping system‖, in Proc. of the 11th International Symposium on District Heating and Cooling, Reykjavik, Iceland (2008).

Thermaflex International Holding B.V. www.thermaflex.com Veerweg 1 5145 NS Waalwijk The Netherlands

[3] J.T. van Wijnkoop, E.J.H.M. van der Ven, ―Verification of heat loss measurements‖, in Proc. of the 12th International Symposium on District Heating and Cooling, Tallinn, Estonia (2010).

CONCLUSION Flexible Polybutylene piping, insulated with Polyethylene foam is a recent development, leaving ample room for product improvement. Experimental data shows rapid improvement in heat loss performance.

[4] E.J.H.M. van der Ven, R.J. van Arendonk, ―Heat loss of flexible plastic pipe systems, analysis and optimization‖, in Proc. of the 12th International Symposium on District Heating and Cooling, Tallinn, Estonia (2010).

By optimizing both quantity (thickness) and quality of flexible Polyethylene foam, and by using the specific hydraulic properties of Polybutylene piping, the heat loss performance of conventional rigid steel piping systems insulated with polyurethane foam, is within reach.

[5] J. Claesson, C. Persson, ―Steady-state thermal problem of insulated pipes solved with the multipole method‖, Chalmers University of Technology, Report 2005:3. (2005) [6] P. Valdimarsson, "Graph-theoretical calculation model for simulation of water and energy flow in district heating systems", in Proc. of the 5th International Symposium on Automation of District Heating Systems, Helsinki, Finland. (1995).

If the current rate of improvement of the PB-PE-PE flexible piping is maintained, total distribution system heat losses will be comparable to conventional rigid Steel-PUR-PE piping. The evident benefits of flexibility would become available without the current heat loss penalty.

[7] J. Korsman, S. de Boer, I.M. Smits, ―Cost benefits and long term behaviour of a new all plastic piping system‖, IEA DHC|CHP Annex VIII research report (2008)

ACKNOWLEDGEMENT Special thanks are due to Ivo Smits who did all the calculations for this paper and made all the graphs. A small change in approach may result in a lot of recalculation. Thanks also to Camilla Persson for supplying us with a MathCad implementation of the multipole method back in 2008 and Pàll Valdimarsson for the invention of Pipelab, which set me off on this road at the symposium in Helsinki, 1995.

317

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

FREE OPTIMIZATION TOOLS FOR DISTRICT HEATING SYSTEMS 1

Stefan Gnüchtel , Sebastian Groß 1

Institute of Power Engineering, Technische Universität Dresden, 01062 Dresden, Germany modify the system configuration to check how the system reacts under new conditions and how the

ABSTRACT At the Technische Universität Dresden, Institute of Power Engineering, Chair of Power Systems Engineering as part of the project ―LowEx Fernwärme – Multilevel District Heating‖ [1] supporting by the Federal Ministry of economy and technology (FKZ 0327400B), two public available and cost free software tools have been developed, which enable the user to find A)

the optimal unit commitment of district heating generators: FreeOpt

the optimal pipeline route with the optimal pipe diameter of district heating networks: STEFaN

operating costs will change. In all cases the tool gives valuable information. So FreeOpt provides help for any municipality that need a first guess on the feasibility and operating costs of a new district heating network or who need to improve the operation of an existing one. And of course it does not matter if the network is supposed to be extended or build from scratch or an existing system just to be analysed. How does the software tool work? First the simulation time is divided into time steps. For every time step several variables (i.e. generated heat and electric power, amount of fuel, energy consumption) exist inside given boundaries as well as individual costs and proceeds are stated (i.e. for fuel, CHP refund). Different kinds of generators can be chosen like heat plants, combined heat and power plants, solar thermal plants or heat pumps. Hot water storage tanks and electricity transferred in or out the grid via contracts improve the flexibility of the whole system (Fig. 1 ).

at a minimum of costs. Both software tools are very easy and intuitive to handle. Lead time required to learn how to operate the programs is short. Both tools intend to support general design decision pro district heating systems. In this paper an overview on their advantages and fields of application, as well as example calculations are presented. PROGRAM FREEOPT FreeOpt is an optimization tool to find the optimal unit commitment of district heating generators for a given time domain. Optimal decisions are found to minimize total costs related to thermal and electric loads.

System boundary

SQfuel

Local district heat networks are becoming more common, so it is important to know how to operate even small systems in terms of minimal costs and highest efficiency. When should which generator be switched on or off? How to handle the storage? Which influences have contracts for electric power?

Block heat and power plant Combined heat and power plant Heat plant/ boiler

There already is a lot of existing software for unit commitment. As commercial and generalised software for large systems it is mostly very expensive. So FreeOpt has been developed for any cases of district heating networks for most efficient operation of all heat and power generators. Mainly operators of smaller supply areas purchasing an expensive software solution would be uneconomical can reach monetary savings.

Solar thermal system

Heat pump

Power demand

Pdem

Grid connection Qdem

Heat demand

Heat storage

Fig. 1 System boundary and interaction plan

A stable version of FreeOpt is already finished. The program‘s power is demonstrated with a simple example determining cost optimal operation of a specific district heating network. Furthermore with the help of parameters and figure lines it is very easy to

The modular design of FreeOpt allows to form easily any generator system. All one has to know are the respective figure lines and parameters for all available generators and the network. After fixing the demand for 318

The 12th International Symposium on District Heating and Cooling, th th September 5 to September 7 , 2010, Tallinn, Estonia

The most important ones are the two balance equations for power (2) and heat (3).

heat and electric power as well as the parameters the optimization problem is defined automatically. FreeOpt calculates the minimal costs to satisfy the demand. For testing purpose there is no limitation of the values.

(2)

The mathematical model belongs to the mixed integer problems. So the objective function and all constraints are linear, all variables are continuous or discrete. It is written in the mathematical modelling language GMPL [2] and solved with the COIN-OR brunch-and-cut solver CBC [3]. Both GMPL and CBC are open-source software under GNU GLP license [4]. An intuitive user interface enables to enter all input data like variable boundaries, cost coefficients, starting values, demand values or figure lines in a very easy way. The internal data flow between user interface, optimization model and solver is realised with help of txt-files (Fig. 2).

stands for the generated electric power,

for

power transferred in or out the grid via contracts, for the electric power demand and

for the own

consumption of electric power. (3) stands for the generated thermal power, the heat demand, cooler and

for

the heat disposed in the auxiliary

for the heat transferred in or out of the

heat storage. Example As an example the following heat network of a local energy supply company is given in Fig. .

Electricity network

Fig. 2 FreeOpt user interface

Running the generators cause costs. Therefore the main aim is to minimize the total cost

Heat boiler plant

CHP CHP1

CHP 2

CHP 3

Heat storage storage

(1) in which proceeds):

are the following operating costs (and



costs for fuel

Fig. 3 Flow scheme of a local heat network



costs or proceeds for transferred electricity

The heat demand of the customer is provided by



costs for start-up procedures



costs for network access



costs for maintenance



costs for CO2-cerfiticates



CHP-refund



EEG-refund



proceeds for avoiding network access



costs for electricity tax



penal costs for balance violation (virtual costs)



3 CHPs (Block Heat & Power Plants) – base load



Heat plant (natural gas) – peak load



Heat storage – used for optimization

and is given for every hour of one year. The electricity demand is not directly considered because of intern clearings inside the energy supply company. The whole generated power is transferred in the grid and refunded as well as used to satisfied the own consumption. It is also possible to transfer electrical power out the grid when all CHPs are switched off.

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 heat and power generation units as well as on the cost assumptions for fuel and the refund of the energy tax.

Analysis operation mode First the real operation mode of whole year 2008 is analysed retrospectively (case I) and the optimal operation mode is determined with the help of FreeOpt (case II). Following the annual operation costs are calculated for both cases (Tab. 3).

Tab. 1 Design parameters of generation units Generation unit

CHP 1

CHP 2

CHP 3

Heat Plant

/ kW

911

774

911

/ kW

911

774

911

/ kW

1200

1020

1200

200

/ kW

1200

1020

1200

2000

0.35

0.87

(

Own consumption

)

0.76

2.0 %

(of

)

Tab. 3 Comparison annual operating costs and proceeds (case I and case II) Annual operation costs and 3 proceeds / 10 €

0.5 %

)

(of )

0.04444

0.045 74

Refund energy tax / €/kWh

0.0055

0.002 2

0.08150

1331.13

Proceeds electricity contracts

468.48

473.65

CHP-refund

58.17

58.67

Start up costs

10.01

6.62

Costs network access

2.83

2.47

812.30

807.90

System configuration changes The energy supply company considers to replace the CHP 2 through a new one. Of course the new CHP has new parameters (Tab. 4). Tab. 4 System configuration changes CHP 2 Generation unit

Tab. 2 Electricity contracts

Costs / €/kWh

1326.12

Furthermore costs for start ups and network access can be reduced. But altogether the total annual operating costs for case I and case II are nearly equal. In this particular example savings are under 1% of the operation costs. So the energy system is operated nearly in an optimal way.

The base own consumption of the whole energy system is 50 kW. In the 100 m3 heat storage it is possible to store 4000 kWh. Tab.2 gives an overview of all intern electricity contracts.

6am – 22pm

22pm – 6 am 0.05700

Optimal operation mode (case II)

Costs fuel

Total costs

One specific characteristic of the energy system is that all three CHP cannot be operated in part load, so the value of the maximum and the minimum power have to be equal.

Electricity purchase

Real operation mode (case I)

)

Costs fuel / €/kWh

Start up cost / €/start up

The calculations in Tab.3 show that the costs for fuel increase but so the proceeds through electricity sale and CHP-refund increase too.

Electricity sale

Old CHP

New CHP

/ kW

774

404

/ kW

774

404

/ kW

1020

535

/ kW

1020

535

0.35

0.36

9.5

0.04735

The CHP-refund is 0.0056 €/kWh and costs for network access are 0.0386 €/kWh.

/% Start up cost / €/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. Main difference is the smaller range of performance. The remaining energy system left unchanged except for the CHP-refund expiring the next year.

curve helps to determine the recommended operation mode. The red line marks the given heat demand which is satisfied at all time steps. The small-sized CHP 2 operates continuous. When the heat demand is higher than the output of the CHP 2 the heat plant is switched on or the heat storage is discharged mostly. Charging the heat storage takes place in low demand times and by switching on one of the others CHPs (CHP 1 or CHP 3) for up to four hours. Using the CHP is more economic than using the heat plant but start up costs and the size of the heat storage restrict the operation of a second CHP.

The optimal operation mode considering the old (case III) and the new (case IV) parameters of the CHP 2 are determined. Case III serves as a reference case. The calculated operation costs per week are shown in Fig. 4 for both cases. Operating costs per week 35

Total cost / 103 €

30 25

Heatbalance 1800

1600

heat flow / kW

5 0 J

Month Old CHP 2 (case III)

New CHP 2 (case IV)

Fig. 4 Operation costs per week (case III and case IV)

heat plant

1200 1000

heat demand

800 600 400 200

CHP production

0 0

In Fig. 4 it can be seen that the operation costs for case IV are below the operation costs for case III in every week, especial in the summer months because the smaller size of the new CHP 2 suits better to the heat demand. Altogether total operating costs amounting to about 27770 € (3.2 % of total costs) can be saved (Tab.5). Main reasons are the huge fuel savings which settle easily the decreasing proceeds through electricity sales. By given investment cost it is very simple to check if the renewal of the old CHP 2 is economic reasonable.

charge of storage by CHP

1400

discharge of storage 72

96 108 120 132 144 156 168

time / h CHP production

storage load

heat plant

storage unload

heat demand

Fig. 5 heat balance curve on one summer week

Finally it can be summarized that first experiences and calculations show, how the FreeOpt allows in an easy and quick way to check beforehand if certain system configurations are useful or contra productive. PROGRAM STEFAN

Tab. 5 Comparison annual operating costs and proceeds (case III and case IV) Annual operation costs and 3 proceeds / 10 €

Old CHP 2

New CHP 2

(case III)

(case IV)

Costs fuel

1323.87

1208.79

Proceeds electricity contracts

462.87

367.12

Start up costs

7.72

1.20

Costs network access

1.92

0.00

870.64

842.87

Total costs

Application field The network optimization is a special case of the research-main focus „optimization of the technical structure of district heating systems― of the 5th energy research program of the German Federal Government. Due to the relatively high net costs of district heating systems it is necessary (beside the application of actual piping systems) to optimize the nets concerning their design parameters, in particular the pipe diameter and the pipe routing. Therefore the software tool ―STEFaN‖ has been created for the combined pipeline routes and diameter optimization. This Windows program for the support of the application of the district heating has interfaces to geographical information systems (GIS) and is complementary with these. Its application is possible in 3 planning phases:

As example the heat balance curve on one summer 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

In the conceptual design planning phase for the cost estimate and the principle decision for the district heating (yes or no).

In the detailed planning phase the localization of the pipeline route occurs for the approval planning.

In the execution planning phase the final determination of the dimension occurs, but no mechanical calculation (stress-strain analysis) is done by the program. Still the required proofs according to e.g. EN 13941 have to be done.

If the edge is not used for the site development ( then holds.

As variables are required beside the diameter further variable than auxiliary variables to the formulation of the constraints: vector of the mass flows



vector of the pressures



binary variable to the capture of the jump at

of the vertices

First Kirchhoff‘s law: Point rule. The sum of all mass flows in a vertex is equal zero. ( - vertex matrix)

Model

(8)

The hydraulic calculations establish the technical basis which performs constraints of the optimization model.

Second Kirchhoff‘s 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 coefficient of friction

of the edges



Thus the constraints can be formulated. These are the equation (6), local and technical limitations as well as equations. (8) and (9).

In addition, the program can be used for the hydraulic calculation of existing district heating networks.

with the surface roughness

matrix):

of the pipe and the

(9)

is applied (4).

This mathematical model is simple to describe, but difficult to solve (already for medium-sized graphs). If the diameters are eliminated by the equation (6) as a

(4)

variable, the variables

In the mentioned planning phases the coefficients of drag are included blanket into the pressure loss according to (5):

and

whose impact on the

objective function is discussed in detail in [5] remain. The principal dependency of the objective function on the vector of the mass flows is displayed in Fig 6 schematically.

(5)

1 6

where

as a extra charge of length.

For a pipe of the length and the diameter for the pressure loss of a plain pipe.

(6) arises (6)

Thus the following mathematical optimization model arises: The investment costs of every new route come into the objective function (investment costs, annual costs or net present value). They are included in the form of (7) in the model.

K konkav K Baum

(7) 0

The bracket of the first summand contains the investment costs of the route per meter as a total lumpsum price and must be multiplied according to by the length of pipeline. In the second summand "obstacles" can be included as direct costs dependent from the diameter. The exponent is set =1 in the present program version for linear dependence. The parameters and are input data.

Fig. 6 Schematic dependence of the objective function on the mass flow

On the abscissa the circulatory mass flow

of a mesh

is displayed. The ordinate shows the non-convex objective function which shows jumps by the binary variables with the rhombuses (the filled rhombus is the function value). 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 requirements mentioned above a graph free of meshes, thus a tree, arises.

Project processing The help file and the user's manual of this program contain detailed instructions to its operation and for the project processing which occurs typically in six steps in the change of STEFaN (steps 2, 3, 5 and 6 a) and of a GIS (steps 1, 4 and 6 b).

This complex course of the objective function requires especially suitable methods. Development of mathematical procedures

These necessary steps 1 to 6 are demonstrated at a fictive example. As a GIS system the recommended and provided program ShapeUp (www.nilione.com) is used. It is freeware too.

Different mathematical methods are used: a) The classical non-linear optimization which is applied for the diameter optimization for fixed development ways.

Example

b) Topological optimizations to the determination of the shortest ways (shortest path problem) and the shortest networks (spanning tree problem) which are combined under use of the procedure from a) to a special iteration process. c)

Step 1: Gathering of Geographical information With the GIS layers (themes) with geo-referenced information (vertices for the source and for the customer as well as edges for consisting and possible routes) invested and in a special standardized format (MIF â&#x20AC;&#x201C; MapInfo Interchange format) exports:

Stochastic methods for the improvement of the optimization results of the algorithm of b): A special implementing of the Monte Carlo Method and a special implementing of the Evolutionary Algorithm.

a) The geo-referenced background image is imported (ÂŠ OpenStreetMap, pale colors in Fig. 7). The 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 pentagon on the top left in the Fig. 7, through violet circle marked) and the sinks (customer â&#x20AC;&#x201C; yellow flags in the Fig. 7, as house service connection stations in the house lines arranged) with the given attribute for the heat demand.

Step 3: Generating the network topography

Step 4: Verification of the generated network topography

After the import of the files invested by the GIS the network topography of the possible routes is created by the program. Gaps between the inputted routes (step 1c) are complemented by the program to a graph with entire and varied development.

The input of possible routes (thick green lines in Fig. 7) by using the mouse and the assignment of the attributes (table in the right section of Fig. 7), for quite available pipes with attribute (branch pipe from the source in DN 150 â&#x20AC;&#x201C; input value: for the laying procedure ( ground,

With this step the files generated by the program can be imported in the GIS: The generated edges (thin blue lines in Fig. 8) and vertices (blue dots in fig. 8) complete the entered network topography and can be checked.

) and for street and

Step 5: Determination of the optimal development

for cellar corridor and available

channel).

The route optimization is carried out by the program.

Step 2: Providing the non-geographical data

Step 6: Evaluation

The files with the non-geographical data (general entries to the network as for example media temperatures as well as economic data) are entered on forms.

Fig. 8 step 4 324

Output of a result report and export of the optimization results to the GIS.

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 treatment of the results by the GIS: The result can be visualized in the GIS (Fig. 9). The different colored lines in the left section show the ascertained route planning, and in the right section of Fig. 9 a part of the data base with the site-

in m,

related data is displayed: Length of pipeline the diameter

in mm, the mass flow

in kg/s in bar.

and the pressure difference

Fig 9 step 6 CONCLUSION

REFERENCES

Stable versions of FreeOpt and of STEFaN are already finished. The program‘s power was demonstrated with simple examples. They can be downloaded from [6].

[1] www.bmwi.de: LowEx Fernwärme – Multilevel District Heating, Fördergeber: Bundesministerium für Wirtschaft und Technologie, FKZ 0327400B.

FreeOpt calculates the optimal operating solution of district heating networks at a minimum of costs to estimate saving potentials. With the help of parameters and figure lines it is very easy to modify the system configuration to check how the system reacts under new conditions and how the operating costs change. In all cases the tool gives valuable information.

[2] www.gnu.org/software/glpk: GNU Linear .Pro-gramming Kit [3] www.coin-or.org: Computational Infrastructure for Operations Research. [4] www.gnu.org/licenses/gpl.html: Public License

Unfortunately, the user guide, the help file and the manual are only available in German at the moment for both programs.

GNU

General

[5] S. Gnüchtel, Ein Beitrag zur Strukturoptimierung von Fernheiznetzen, PhD thesis TU Dresden (1981) [6] http://tu-dresden.de/die_tu_dresden/ fakultaeten/fakultaet_maschinenwesen/iet/ew/ forschung_und_projekte/mldh/download_ml 325