12th International Symposium on District Heating and Cooling

<strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heatingand CoolingSeptember 5 th –September 7 th , 2010Tallinn, ESTONIAISBN: 978-9949-23-015-0

IMPROVED PRIMARY ENERGY EFFICIENCY OF DISTRICT HEATING NETWORKS BY INTEGRATION OFCOMMUNAL BIOMASS-FIRED COMBINED HEAT AND POWER PLANTS WITH BIOMASS PYROLYSIS ........... 168T. Kohl, N.A. Pambudi, T. Laukkanen and C.-J. FogelholmCHP OR POWER STATION? – QUESTION FOR LATVIA ....................................................................................... 177D. Blumberga, G. Kuplais, F. Romagnoli and E. VigantsLCA OF COMBINED HEAT AND POWER PRODUCTION AT HELLISHEIÐI GEOTHERMAL POWER PLANT WITHFOCUS ON PRIMARY ENERGY EFFICIENCY ........................................................................................................ 184Marta Ros Karlsdottir, Olafur Petur Palsson, Halldor PalssonFLEXIBILITY FROM DISTRICT HEATING TO DECREASE WIND POWER INTEGRATION COSTS .................... 193J. Kiviluoma and P. MeibomDAILY HEAT LOAD VARIATION IN SWEDISH DISTRICT HEATING SYSTEMS .................................................... 199H. Gadd and S. WernerDISTRICT HEATING AS PART OF THE ENERGY SYSTEM: AN ENVIRONMENTAL PERSPECTIVE ON‗PASSIVE HOUSES‘ AND HEAT REPLACING ELECTRICITY USE ....................................................................... 202Morgan Fröling and Ingrid NyströmADAPTIVE CONTROL OF RADIATOR SYSTEMS FOR A LOWEST POSSIBLE RETURN TEMPERATURE ........ 206P. Lauenburg and J. WollerstrandPOLICIES AND BARRIERS FOR DISTRICT HEATING AND COOLING OUTSIDE EU COUNTRIES ................... 215A. Nuorkivi and B. KalkumBARRIERS TO DISTRICT HEATING DEVELOPMENT IN SOME EUROPEAN COUNTRIES ............................... 223Dag Henning and Olle MårdsjöIMPACT OF THE PRICE OF CO2 CERTIFICATES ON CHP AND DISTRICT HEAT IN THE EU27 ...................... 229Markus BleslCONSIDERATIONS AND CALCULATIONS ON SYSTEM EFFICIENCIES OF HEATING SYSTEMS IN BUILDINGSCONNECTED TO DISTRICT HEATING .................................................................................................................... 238Maria Justo Alonso, Rolf Ulseth and Jacob StangHEAT LOAD REDUCTIONS AND THEIR EFFECT ON ENERGY CONSUMPTION ................................................ 244Christian Johansson and Fredrik WernstedtVERIFICATION OF HEAT LOSS MEASUREMENTS ............................................................................................... 250J.T. van Wijnkoop, E. van der VenDISTRICT HEATING AND COOLING WITH LARGE CENTRIFUGAL CHILLER-HEAT PUMPS ............................. 258Ulrich PietruchaNEW ECONOMICAL CONNECTION SOLUTION FOR FLEXIBLE PIPING SYSTEMS ........................................... 261Christian Engel, Gerrit-Jan BaarsCOMPETITIVENESS OF COMBINED HEAT AND POWER PLANT TECHNOLOGIESIN ESTONIAN CONDITIONS..................................................................................................................................... 267E. Latõšov and A. SiirdeDISTRIBUTION OF HEAT USE IN SWEDEN ........................................................................................................... 273Margaretha Borgström, Sven WernerDAMAGES OF THE TALLINN DISTRICT HEATING NETWORKS AND INDICATIVE PARAMETERS FOR ANESTIMATION OF THE NETWORKS GENERAL CONDITION .................................................................................. 277Aleksandr Hlebnikov, Anna Volkova, Olga Džuba, Arvi Poobus, Ülo KaskEFFICIENCY OF DISTRICT HEATING WATER PUMPING IN FINLAND ................................................................ 283Antti Hakulinen, Jarkko Lampinen and Janne LavantiMODELLING DISTRICT HEATING COOPERATIONS IN STOCKHOLM – AN INTERDISCIPLINARY STUDY OF AREGIONAL ENERGY SYSTEM ................................................................................................................................. 288D. Magnusson, D. Djuric IlicCUTTING COSTS OF DISTRICT HEATING SYSTEMS BY USING OPTIMIZED LAYING TECHNIQUES ............. 297Alexander Goebel, Dr. Stefan HollerANALYSIS OF HEAT TRANSFER IN HEAT EXCHANGERS BY USING THE NTU METHOD AND EMPIRICALRELATIONS ............................................................................................................................................................... 305O. Gudmundsson, O. P. Palsson and H. PalssonHEAT LOSS ANALYSIS AND OPTIMIZATION OF A FLEXIBLE PIPING SYSTEM ................................................. 310J. Korsman, I.M. Smits and E.J.H.M. van der VenFREE OPTIMIZATION TOOLS FOR DISTRICT HEATING SYSTEMS .................................................................... 318Stefan Gnüchtel, Sebastian Groß

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaINTEGRATION OF AN IP BASED LOW-POWER SENSOR NETWORKIN DISTRICT HEATING SUBSTATIONSJ. Gustafsson, H. Mäkitaavola, J. Delsing and J. van DeventerDiv. of EISLAB, Dept. of Computer Science and Electrical EngineeringLuleå University of Technology, 971 87 Luleå, SWEDENABSTRACTIn this study, the implementation of a wireless, lowpower,sensor network with IP capabilities in a districtheating substation was evaluated. The aim of thestudy was to show that an open standard solution istechnically feasible. Low-power wirelesscommunication was established using IPv6/6LoWPANon an IEEE 802.15.4 wireless network. Anexperimental district heating substation was equippedwith sensor platforms in vital devices located within ornear a district heating substation. As a result, allconnected devices could obtain a direct internetconnection.A system with open standards facilitates theintroduction of new energy services such as individualmeasurements and improved space heating control.In this study, we found that resource-limited batterypowereddevices possess a life expectancy of over10 years, using small batteries while participating inIPv6 compatible communication.INTRODUCTIONEmbedding low-power wireless devices in districtheating substations and surrounding equipment suchas temperature sensors could provide useful servicesto consumers and producers. Currently, manydifferent substation control systems on the market canconnect to the internet and have various wirelesssensor reading systems. However, these systemstend to be specialized and are only compatible withequipment from the same manufacturer. Moreover,internet-compatible control systems are often alsorelatively expensive, and provide bad scalability.In general, commercially available heat meters cannotcommunicate through the current infrastructure; thus,specialized communication methods such as mbus,pulse, and infrared readings must be employed.Therefore, poor communication standards limit thecurrent usage of heat meters and other equipment inthe substation. However, by sharing information withother devices in the substation, the heat meter couldprovide useful feedback and sensing information,which can be used to improve the substation controlfunctionality.Fig. 1 provides an overview of the development ofsensor networks over the last 20 years. Unfortunately,most equipment currently used in district heatingsubstations is antiquated.Fig. 1. Evolution of wireless sensor networks. Althoughthe scalability of the sensor network has increased, manyindustries still use vendor-specific cable solutions. (Thefigure was obtained from the literature [1])If heat meters, control systems, and other non-districtheating equipment could communicate, new servicesthat have impact on both economy and theenvironment could be developed.The infrastructure required to achieve wireless devicecommunication may be attained with low-powerwireless technology. Small sensor platforms withdirect internet access through standardized wirelesstechnology can provide a solid platform for newservices.A lack of standardized communication protocols iscommonly encountered when connecting electronicdevices from different vendors. In general, devicesmanufactured by different companies use differentcommunication protocols, which limits the functionalityof the substation.District heating substations can be divided intosections based on metering, space heat control, andtap water control. For a visual overview of a commonparallel connected district heating substation, seeFig. 2, this is also the substation type used in thestudy. Typically, information is not shared betweenthese sections; thus, each system can only be locallyoptimized. To achieve complete substationoptimization, information must be shared betweensections. To this end, wireless sensor-platforms wereinstalled in temperature-sensors, heat-meters,circulation pumps, and control valves, and new controlmethods and services were tested. This empowers us4

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniato develop new control methods, and implementingnew services to heat suppliers, building owners andend users.Fig. 2. A systematic overview of a parallel coupleddistrict heating substation divided into three sections:metering, heating and hot water system.SERVICESTo control or reduce their energy bill, district heatingcustomers require specific information to determinethe appropriate action. Currently, the only informationavailable to the customer is the information providedin the bill or on the heat-meter display.If information on all devices was available online,customer could easily monitor their usage and interactwith the substation. Examples of services that couldbe provided by the substation are explained in thefollowing sections.Improved substation controlCombined heat and power plants are becoming morecommon; thus, the importance of the distributionsystem ΔT is increasing. In a combined heat andpower plant with a flue-gas condensation system, ahigh ΔT is even more important to obtain satisfactoryfuel efficiency.To maintain high energy efficiency, the hot waterproduced by the plant must be delivered to customerswith a minimal heat loss. Once the hot water istransported to the customer, a maximum amount ofenergy per volume of water should be extracted andused for heating purposes, such as hot tap water andspace heating. To achieve a maximum ΔT, energytransfer between the distribution medium to the pointof consumption should be maximized, while thetemperature of the returning distribution mediumshould be minimized.Unfortunately, there are many challenges inmaintaining the efficiency of a district heating network.Problems related to the equipment that controls thetemperature of radiator water and hot tap water areoften encountered. These devices tend to becalibrated to satisfy the desires of the customer only;thus, the effects on the energy efficiency of the entiredistrict heating system are often ignored. One keyfactor in obtaining a high ΔT across a district heatingsubstation is the radiator circuit supply temperature.The radiator circuit supply temperature does not onlyaffect the indoor comfort, but also the primary returntemperature as the returning radiator circuit mediacools the primary media through the heat exchangingunit. Specifically, water returned from the radiatorcircuit cools the primary supply through the heatexchange unit. Currently, the radiator circuit supplytemperature is based on the local outdoortemperature, which produces a stable indoortemperature. However, the primary supplytemperature also affects the ideal radiator supplytemperature and the radiator circuit flow. Therelationship between outdoor temperature and primarysupply is often assumed to be linear (colder outdoorair leads to a warmer primary supply). However,significant deviations from the ideal curve arecommon. More information on the effect of primarysupply temperature and radiator control on the indoorair temperature and ΔT of the system can be foundin [2].Adaptive radiator control is another intelligent way ofcontrolling the radiator circuit and obtaining a high ΔT.More information on this method can be found inprevious studies by Lauenburg [3].Fault detectionControl valves in the district heating substation oftenpossess inappropriate dimensions, resulting inintermittent control, pressure shocks, and high returntemperatures. Due to the high thermal time constantof a building, the indoor temperature is not directlyaffected. Therefore, an error in the control valve maygo unnoticed for a considerable amount of time.Error identification can be achieved by evaluating highfrequent meter readings, which to some extent aredone today.A fouling valve that is stuck or does not move inaccordance with the control signal may also be difficultto detect. A direct comparison of the valve controlsignal with the heat meter, which measures theprimary flow through the district heating substation,can be used to identify a broken fouling valve [4].5

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaIndividual measurementsIndividual measurements are common in somecountries and are gaining interest in others. To obtainmeasurements of each apartment, tap point, orradiator, new metering devices must be installed. Themost straight forward method is to install flow metersat each tap point and/or radiator. In general, highresolution flow meters are quite expensive; thus,installing one on every tap point/radiator can be costprohibitive.An alternative method has been evaluated by Yliniemi[5]. In this method, temperature sensors wereinstalled at each tap point, and one central flow meterwas used to measure the flow through a section,which contained up to 40 tap points. The flowrecorded by the meter and the temperature measuredat the tap points were synchronized, and the integrityof each tapping point was verified by installinginexpensive temperature sensors at each site and alimited number of central flow meters throughout thebuilding.Load balancingDynamic load balancing is a method used to removeheat load peaks and divide power consumptionbetween buildings. Dynamic load balancing is basedon the presence of a large thermal time constant ofeach building. For instance, in a building with a highthermal time constant, the heating system can beturned off when the price of heat is high or duringpeak energy hours. An online automatic andindependent auction system is used to decide whichbuildings will be shut down or provided a limitedamount of thermal power. In this system, allconnected buildings are involved in the biddingprocess. Specific details on dynamic load balancingare provided in the literature [6].Visualized energy efficiencyIf a large number of district heating substations wereconnected to the internet, the performance of differentsubstations could be compared. For instance, thesupply/return temperature, ΔT, energy usage, etc. ofall substations could be plotted in a graph, table ormap. Fig. 3 displays a map of the return temperatureof a substation, which allows the consumer tocompare the performance of their house to others inthe area. Moreover, the map provides the utilitycompany with an overview of the network andimproves the detection of leaks and short circuits.Moreover, the utility company can identifydeteriorating substations or individual installations thatperform poorly.NETWORK TECHNOLOGYA common method of visualizing a networkcommunication protocol is in the form of stack. A stackconsists of layers that are separated by function; thus,a communication stack contains different layers oftasks related to data transportation. The layers can bedivided and visualized in many ways. For example,the five-layer internet model has been usedextensively in previous studies and is displayed inFig. 4 [7]. In this paper, only the layers that aresignificant to the results of this research will bediscussed. Thus, the network, link, and physical layersare considered in more detail.Fig. 3. Performance of a district heating substationvisualized on a map. The red square can represent thesupply/return temperature, energy usage, or heat flow inthe connected building.Fig. 4 A generic five-layer internet model and itsimplementation in an IEEE 802.15.4 wireless network.IP (Network Layer)The internet protocol (IP) is the most well-known andcommonly used network protocol in the world. Alltraffic 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, thelatest version.6

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaIPv4Currently, internet protocol version 4 is the mostwidely used IP, and almost all computers connectedto the internet use this version.An IPv4 address is 32 bits long and is typically writtenin 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 ofaddresses is reserved and cannot be used for onlinepurposes. The total number of usable IPv4 addressesis approximately 3.7 billion. As the number of devicesconnected to the internet increases, IPv4 addressesare beginning to run out.Technology such as network address translation(NAT) and port address translation (PAT) havepostponed the depletion of IPv4 addresses; however,the number of available IPv4 addresses decreasesevery day.IPV6IPv6 was developed to compensate for the limitednumber of IPv4 addresses. IPv6 uses a longeraddress than IPv4 and has several convenientfeatures. IPv6 uses a 128 bit address, which meansthat there are 2128 possible addresses. Thus, thenumber of address per square millimeter of the earth‘ssurface is 6.7·1017. Hopefully, the addressesobtained through the implementation of IPv6 will lastfor a long time.With the additional address space, it is possible togive every small device its own unique IP numberwithout implementing NAT. Thus, directcommunication over the internet can be achievedwithout any special gateways. However, the newaddress space increases the overhead of datapackages, which negatively impacts small, low-powerdevices because more battery energy is wasted onheader data in every wireless data transmission.However, a new adaptation layer (6LoWPAN) wasdeveloped to limit the amount of lost energy. Moreinformation on 6LoWPAN can be found in the nextsection and in [1].In addition to a wider address space, IPv6 alsoincludes stateless autoconfiguration, which is afunction that can be used to automatically configurenewly connected devices without any special servers.To obtain stateless autoconfiguration, newlyconnected devices broadcast a router solicitation (RS)message to every listening device. When a routerreceives the message, it responds with a routeradvertisement (RA) message. The device adds theIPv6 prefix from the router to the local link layeraddress, creating a complete IPv6 address. To ensurethat another device does not possess the same IPaddress, the device broadcasts a neighbor solicitationmessage to search for a duplicate address. If anotherdevice has the same IP number, the new device shutsdown.6LOWPAN6LoWPAN is an adaptation layer that separates thenetwork and data link layer of the protocol stack. Thepurpose of the layer is to compress IPv6 headers andminimize unnecessary data transmission whilemaintaining IPv6 compatibility. According to theliterature, [8] the 6LoWPAN header uses less than10% of the total energy used during packettransmission.IEEE 802.15.4 physical and data link layers are oftenused in combination with 6LoWPAN; however, otherstandards can also be applied.802.15.4 (Link and Physical Layer)The most common data link and physical layer usedwith 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 802.15.4standard specifies operation at low frequency bandssuch 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 differencesbetween low and high frequency bands are thebandwidth and communication range. The 2.4 GHzband supports a higher bandwidth but the range islimited, especially in armored concrete buildings. Thelow frequency bands have a moderate bandwidth anda considerably larger range. In a district heatingsubstation, bandwidth usage can be minimizedbecause rapid changes are uncommon (compared tomany other control/measurement situations) and lowfrequency groups are preferred. However, only 2.4GHz sensor platforms were available at the beginningof this study; thus, these platforms were used in mostof the tests.7

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTo obtain IPv6/6LoWPAN functionality in the Mulles,the lightweight operating systems Contiki [14] andTinyOS [15] have been successfully ported to the Mulleplatform. Both operating systems were specificallydesigned to be compatible with resource limitedembedded systems such as Mulle. Moreover, Contikiand TinyOS both support IPv6 and 6LoWPAN.However, TinyOS was selected for this study becausestability issues due to edge-routing problems withContiki.Sensor platform energy usageObtaining an acceptable life expectancy is one of thebiggest challenges to battery powered, wirelessdevices. In Sweden, heat meters are inspected every 5to 10 years, depending on the size of the meter. Thelife expectancy of wireless devices should beequivalent to the inspection period to avoid frequentand expensive battery replacements. All sensor nodesdo however not need to be battery powered. In thecase of available electric power in close proximity, e.g.for platforms mounted in pumps or valves there is noexplicit need for batteries since there are electricityavailable. At other sensor platforms, battery power isthe only feasible solution, for instance outdoortemperature sensors.To determine the amount of energy used by a wirelesssensing device, the current at the sensor platformassociated with IPv6/6LoWPAN communication wasmeasured. To measure the current used by the device,a 1 ohm high precision resistor was connected in seriesto the Mulle power connector. The voltage dropgenerated across the resistor was amplified 100 timeswith a MAX4372H amplifier circuit. Using an analogacquisition card, the amplified signal was measuredand stored in an ordinary PC. Due to poor precision atvery low current, complementary measurements wereperformed with a high precision ampere-meter todetermine the current usage of the Mulle, when it wasin deep sleep mode.To evaluate the energy cost of transmitting datapackets with UDP on IPv6/6LoWPAN, packets withpayload sizes between 1 and 100 bytes weretransmitted, and the expected lifetime of the sensorwas calculated. Fig. 8 displays the expected lifetime ofa sensor with a 500 mA battery and a 15 minutetransmission interval. Out of curiosity, both TinyOS andContiki were programmed to transmit UDP packets ofdifferent sizes at consecutive time intervals to observeany differences in energy usage between the two. Theresults indicated that the energy usage of 50 to 80-bytepayloads in Contiki and Tiny OS were significantlydifferent. The observed difference between operatingsystems is most likely related to the method of headercompression. Specifically, Contiki uses HC1, whileTinyOS is based on HC01. However, both methods area part of the 6LoWPAN standard. Additionally, TinyOSuses short addressing, while Contiki employs longaddressing. The type of addressing and headercompression used by the OS can be changed, but inthis particular test, default settings were used.For payload sizes greater than 60/90 bytes, the IPpacket had to be divided into two separate 802.15.4frames because the maximum frame size of IEEE805.15.4 is 127 bytes. The separation of IP packetsincreased energy usage and decreased the expectedlifetime of the sensor. Thus, software developersshould consider the maximum frame size if absolutemaximization of sensor lifetime targeted. Howeverincreased payload sizes can of course becompensated with a larger battery.As shown in Fig. 8, the fixed transmission interval wasset to 15 minutes, and the effect of transmissioninterval on the expected lifetime of the sensor wasanalyzed. Additionally, sensor lifetime was evaluated atvarious transmission frequencies and a fixed payloadof 80 bytes, as shown in Fig. 9. In accordance to theorythe results indicated that a low transmission frequencyhas a positive effect on sensor lifetime. In the case ofcontext aware sensors, which only transmit data whenrequired e.g. when a measured temperature exceeds aset threshold, sensor life expectancy will in most casesbe increased. However, the impact of thesleep/standby energy usage will make up a largerpercentage of the total energy usage, which hence willmean that the importance of keeping the sleep currentlow will be even bigger.Fig. 8. The effect of payload size on the expected lifetimeof a sensor platform at a transmission rate of4 transmissions per hour (1 to 100 bytes).9

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAll of the substation devices used in this study weremodule-based, which allows manufacturers to produce6LoWPAN module for large scale deployment.RESULTSFig. 9. The effect of transmission frequency on theexpected lifetime of a sensor platform at a payloadof 80 bytes.The predictive life expectancy calculations did not takeinto account the fact that batteries loose energy overtime, even if they are not in use. Depending on batterytype, this can significantly reduce the expected lifetimeof a sensor.SENSOR INTEGRATIONTo provide wireless accessibility to devices in thedistrict heating substation, some simple interfaceelectronics were developed to integrate Mulle withdevice hardware. As shown in Fig. 10, a heat meterwas integrated with a Mulle in the bottom modulelocation.When digital communication interfaces were available(heat meter and circulation pump), the correspondingapplication protocols were kindly provided by thevendors (Kamstrup and Grundfos). The control valve(Siemens SQS-65) was not equipped with any digitalcommunication interface; however, an analog 0–10 Vinput used to control the position of the valve and a0–10 V output used to read the position of the valvewere available.Wireless devices in a district heating substation weresuccessful integrated to support a IPv6/6LoWPANnetwork. Due to the range limitations of 2.4 GHzmodules, deployment of several platforms wasrestricted. However, new 868 MHz platforms are nowavailable and show excellent preliminary results.2.4 GHz platforms will be replaced with 868 MHzplatforms during the spring/summer of 2010.A lifetime of 10+ years can be achieved with 500 mAhbattery and an average transmission interval of15 minutes using IPv6 compatible communication;thus, the life expectancy of battery powered sensorsdid not have a negative effect on integration.CONCLUSIONIntegrating an IPv6/6LoWPAN wireless network in adistrict heating substation can significantly increase thefunctionality and scalability of the substation and supplynew services to both producers and consumers.Using an open, well documented, and tested protocolincreases the possibility of interoperability betweenproducts of different manufacturers. This studyrevealed that available technology can be used toachieve IP-based wireless communication. However, aconsiderable amount of work on smart applicationlayers must be conducted before wireless sensornetworks in district heating substations can bedeployed and used to its full potential.FUTURE WORKTo achieve complete device compatibility, theapplication layer(s) of the integrated network mustfurther developed. One interesting approach is to adaptthe service oriented architecture in web-based servicesto low-power sensors. Available service orientedarchitectures (SOA) such as DPWS 1 are developedprimarily for large enterprises and are not intended tobe used with a resource limited device that possessesa low-bandwidth link. However, the functionality of thisarchitecture would support a convenient solution fordirect sensor integration in enterprise systems.The integration of sensors and SOA such as DPWS isa challenging but intriguing task.Fig. 10. A Mulle sensor platform integrated with aKamstrup Multical 601 heat meter.Mulle is marked by a blue square, and the interfacecard is indicated by a purple square.1 Device Profile for Web Services10

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaREFERENCES[1] Z. Shelby and C. Bormann, 6LoWPAN: TheWireless Embedded Internet, November 2009.[2] J. Gustafsson, J. Delsing, and J. van Deventer,―Improved district heating substation efficiency witha 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[3] P. Lauenburg, ―Improved supply of district heat tohydronic space heating systems,‖ Ph.D.dissertation, Dept. och Energy Sciences, LundUniversity, P.O Box 118, SE-22100, Lund,December 2009.[4] K. Yliniemi, Fault detection in district heatingsubstations. 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.[5] K. Yliniemi, ―Individuell mätning av varmvattenförbrukning,‖http://www.svenskfjarrvarme.se/download/4774/Kimmo Yliniemi.pdf, 2007.[6] F. Wernstedt, P. Davidsson, and C. Johansson,―Demand side management in district heatingsystems,‖ in AAMAS ‘07: Proceedings of the 6thinternational joint conference on Autonomousagents and multiagent systems. New York, NY,USA: ACM, 2007, pp. 1–7.[7] J. Kurose and K. Ross, Computer Networking aTop-Down Approach featuring the Internet, 2nd ed.Pearson Education <strong>International</strong>, 2003.[8] G. Mulligan and 6lowPAN Working Group, ―The6lowpan architecture,‖ in Proceedings of the 4thworkshop on Embedded networked sensors, June2007.[9] ―IEEE 802.15.4-2006 standard‖,http://standards.ieee.org/getieee802/802.15.html,April 2010.[10] ―Embedded internet system technology botnia AB,‖http://www.eistec.se/, March 2010.[11] ―Crossbow technology,‖ http://www.xbow.com,March 2010.[12] ―AVR raven,‖ http://www.atmel.com, April 2010.[13] Sensinode,‖ http://www.sensinode.com, April 2010.[14] ―Contiki,‖ http://www.sics.se/contiki/, March 2010.[15] ―Tinyos,‖ http://www.tinyos.net, March 2010.11

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaABSTRACTON THE RADIAL CONTACT PRESSUREOF PARALLEL BURIED PIPES FOR DISTRICT HEATINGI. Weidlich 1 , M. Achmus 21 AGFW, German Heat and Power Association, Research & Development,Stresemannallee 28, 60596 Frankfurt am Main, i.weidlich@agfw.de2 Institute of Soil Mechanics, Foundation Engineering and Waterpower Engineering,Leibniz University of Hannover, Appelstr. 9A, 30167 Hannover, achmus@igbe.uni-hannover.deFor the design and calculation of buried district heatingpipe systems the magnitude of radial contact pressuresacting on the pipes is of importance, since thesepressures affect the friction forces which may bemobilized. For parallel buried pipes, the stressdistribution is generally expected to be different fromthe case of a single pipe. The present investigationcompares radial stresses according to current designdirectives for buried single pipes with numericallycalculated stresses for parallel buried pipes. Thecalculations show a deviation of the radial stressdistributions in particular for the springline area. Theresults are compared with former theoretical investigations,which predicted a reduction of radial contactpressures between the two pipes. This is verified forsmall-diameter pipes. With larger pipe diameters astress increase was identified between the pipes.However, with regard to the average radial pressureonly slight differences between single pipes andparallel buried pipes were found.INTRODUCTIONAs a part of the underground infrastructure of modernsettlements, district heating pipe networks are animportant medium of economic heat transportation. Hotwater is pumped in a flow pipe from the supply stationto the consumer at a high temperature and under highpressure, and the used water is pumped back to thesupply station in a return pipe.For buried district heating pipes the earth pressure onthe pipe, respectively the radial contact pressure, is animportant value for the design, since it affects thefriction forces which may be mobilized. The frictionforces determine the axial deflections of the pipe andthe distribution of normal stresses, which are inducedby the temperature loading of the pipe. According tothe European Standard EN 13941, the normal stresson the pipe coating is calculated for single pipe trenchconditions dependent on the overburden weight of thesoil, the diameter, the pipe weight and an earthpressure coefficient [1]. However, in practice flow andreturn district heating supply pipes are buried side byside in the same trench. Fig. 1 shows a typical situationfor buried district heating pipes according to Floss [2].Fig. 1. Typical trench condition for DH-pipes afterFLOSS [2]The distance between the two pipes depends on therequirements of the laying technique and procedure.For small distances between the two pipes aninteraction between the two pipes is to be expected.PREVIOUS WORKPrevious theoretical investigations were based on thecalculation method developed by Leonhardt, taking intoconsideration the deformation behaviour of pipe andsoil and their influence on each other [3]. Leonhardtintroduced the ―shear resistant beam on elasticbedding‖ theory, in which the backfill above the pipe isconsidered to be a shear resistant beam, which is ableto transfer shear loads, but no bending moments.Using this model it is possible to determine the shearforces activated by the deformation of the ―shearresistant beam‖ caused by different stiffnesses of thepipe and the surrounding soil, which leads to aredistribution of stresses in the soil with correspondingconcentration factors .For practical application in Germany regulationATV A 127 was published employing Leonhardt‘stheory for buried pipes [4]. This regulation can beapplied analogously to all kinds of buried pipes. Thespecial application of regulation ATV A 127 for buriedpreinsulated district heating pipes was first investigatedby Beilke [5].12

For the case of parallel buried pipes former analyticalcalculations by Rizkallah and Achmus usingLeonhardt‘s theory showed a reduction of the verticalstresses between the two pipes [6]. The system usedfor these calculations with the ―shear resistant beam onelastic bedding‖ for two parallel buried pipes is shownin Fig. 2.The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 4. Ratio B2/ B dependent on relative overburdenheight and pipe distancesFig. 2. Shear resistant beam for parallel buried pipesThe concentration factors B . B for the area beside andbetween the pipes were found with the followingassumptions:The influenced area for the determination ofthe concentration factor B beside the pipe( B1 ) is defined by a line with an inclination of60° shown in Fig. 3. This angle coincides withthe theoretical slope inclination of a noncohesivesoil with an internal angle of frictionof ' = 30°.Between the pipes ( B2 ) the full interspace istaken as the area of influence.A calculation method for parallel buried pipes instepped trenches was proposed by Hornung and Kittel[7]. With this calculation method the total loading onone pipe is derived from the sum of the partial loadings,which correspond to the trench shape to the right andleft of the pipe. The typical trench condition for districtheating pipes provides a non stepped trench with theflow and return pipes installed on the sameunderground level. The presented study was thereforecarried out without employing the Hornung and Kittelcalculation method.NUMERICAL INVESTIGATIONSNumerical calculations were carried out with the twodimensional finite element program PLAXIS, version8.6. Two standard situations with different outer pipediameters D (DN65, D=140 mm; DN250, D= 400 mm)of two parallel buried district heating pipes wereinvestigated. The distance between the pipes waschosen to be A=10 cm (see Fig. 1). The overburdenheight of the backfill material of the trench wasH/D=3.0. The finite element mesh used for the DN65pipe is shown in Fig. 5 as an example.Fig. 3. Method to determine the concentration factors forparallel buried pipesThe calculations by Rizkallah and Achmus showed onlysmall deviations for the concentration factor B of singlepipes and the concentration factor B1 for parallelburied pipes. However, a significant reduction of thevertical stresses (i.e. B2 < B ) was determined betweenthe pipes. As an example, the ratio of the stress factorsis shown in Fig. 4, dependent on the relativeoverburden height H/D and the relative distance of thepipes A/D.Fig. 5. Finite element mesh for the case DN65, H/D=3The installation process was simulated by a ―stagedconstruction‖ process, considering a retained trenchand the backfilling procedure with several layers. Thecompaction process was accounted for by applying astatic distributed load of p=10 kN/m² on each of thelayers. Ground water was not considered in thisinvestigation.13

Sand in a medium dense to dense state was assumedas backfill material. The mechanical behaviour of thesoil was modelled with the Mohr-Coulomb constitutivelaw, which is a linear elastic / ideal plastic materialmodel. The parameters used for the model are shownin Table I.Table I. – Soil parameters used for sand in Mohr-Coulombmaterial modelThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDefinition and UnitSizeUnit weight [kN/m³] 18Oedometric Elasticity Modulus E oed [MPa] 70.5Poisson‘s ratio 0.3Internal angle of friction ‘ [°] 40Angle of dilatancy [°] 10Interface friction R inter [1] 0.536Between pipe and soil, the Coulomb friction law with aaccording to Eq. (1).tan Rinter *tan 'i (1)Fig. 7. Horizontal effective stresses h (DN250 pipe,H/D=3)In Fig. 8 the stress concentration is shown by thedistribution of the radial contact pressure for the lefthandpipe. 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 pressureof r,avg,calc = 18.81 kN/m² the deviation is about 12.2%.In order to keep the model as simple as possible thepipes were assumed to be rigid.In the numerical calculations the initial soil stress statedue to the soil unit weight was established first. Theinstallation procedure was then simulated and theresults were evaluated.In the first model of pipes with an outer diameter ofD=140mm (DN65), no significant stress concentrationbetween the pipes was observed. The radial contactpressure obtained for both pipes is shown in Fig. 6.Fig. 8. Contact pressure on the left-hand DN250 pipe,H/D=3From the DIN EN 13941 regulation the average radialpressure on a single buried pipe can be derived for theinvestigated trench condition according to Eq. (2).r,avg,13941 D 1k * H * 2 2 (2)Fig. 6. Contact pressure on the DN65 pipes, H/D=3However, in the second numerical model of pipes withan outer diameter of D=400 mm (DN250), a stressconcentration between the pipes was evident. Thedistribution of horizontal effective stresses acting afterthe installation process is shown in Fig. 7. The stressesare significantly larger between the pipes than beneaththem.In Table II the results of the numerical investigation arecompared to the expected radial pressure from the DINEN 13941 regulation.Table II. – Average contact pressure r,avg for H/D=3.0DNSingle pipe according toDIN EN 1394165 6.15 kN/m² 7.25 kN/m²Parallel buried pipeaccording to numericalresults250 17.58 kN/m² 18.81 kN/m²Regarding the average radial contact pressure thedifference between the expected values from the DIN14

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaEN 13941 regulation for single pipes and the calculatedvalues from the numerical simulations for parallelburied pipes is rather small. The values for the parallelburied pipes are slightly larger.CONCLUSIONThe earth pressure on district heating pipes is animportant design value and should be determined asexactly as possible. In the presented work the earthpressure on parallel buried pipes was investigated.The evaluation of the radial stresses on the pipe innumerical calculations showed a stress concentrationbetween two pipes buried in the same trench for shortpipe distances and large diameters. However, formertheoretical investigations led to a reduction of radialcontact pressure between the two pipes, which wasobserved in the numerical simulations for smalldiameters and small overburden heights.Because typical trench conditions with two parallelburied pipes are not considered in current designdirectives for district heating pipes the numerical resultswere compared with the values derived from thecurrent design regulations. For the observed systemsonly small deviations regarding the average normalpressure between single pipe and parallel buried pipeswere found. Thus, as long as the exact distribution ofstresses along the pipe perimeter is not of particularrelevance, current calculation directives are alsosuitable for parallel buried pipes. Only for conditionswith large pipe diameters and small distances betweenthe pipes and also relatively large overburden heightsis a significant deviation to be expected.Furthermore, inhomogeneous backfill compaction,which is probable for small pipe distances under in situconditions, affects the contact pressure. In order totake into account the real compaction process withinthe trench, only direct measurements seem to lead tocorrect results. Further research work is necessary atthis point.REFERENCES[1] DIN EN 13941, Berechnung und Verlegung vonwerkmäßig gedämmten Verbundmantelrohren fürFernwärme, Deutsches Institut für Normung e.V.,Beuth Verlag Berlin, 2003.[2] R Floss, „Handbuch ZTVE-StB 94/1997―,Kommentar mit Kompendium Erd- und Felsbau, 3.Auflage, Kirschbaum-Verlag, 2006.[3] ]G. Leonhardt, „Belastung von starrenRohrleitungen unter Dämmen―, PhD Thesis,Institute of Soil Mechanics, FoundationEngineering and Waterpower Engineering,University of Hannover, 1973.[4] ATV A 127, Richtlinie für die statische Berechnungvon Entwässerungskanälen und -leitungen,Arbeitsblatt A 127 der AbwassertechnischenVereinigung e.V., 2000.[5] O. Beilke, „Interaktionsverhalten des Bauwerks„Fernwärmeleitung–Baugrund―, Institute of SoilMechanics, Foundation Engineering andWaterpower Engineering, University of Hannover,1993.[6] V. Rizkallah, M. Achmus, ―Zur Größe derReibungskräfte an erdverlegten Fernwärmeleitungen‖,Forschungsvorhaben WechselwirkungenFernwärmeleitung – Bettungsmaterial, Institute ofSoil Mechanics, Foundation Engineering andWaterpower Engineering, University of Hannover,1993.[7] K. Hornung, D. Kittel, „Statik erdüberdeckterRohre―, Bauverlag GmbH Wiesbaden und Berlin,ISBN 3-7625-2039-9, 1983.In order to avoid large deviations in the contactpressures, good and consistent backfill compaction anda certain minimum distance between the flows andreturn pipes is recommended.15

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaANALYSIS ON FLAT STATION CONCEPT.PREPARING DHW DECENTRALISED IN FLATSThorsen, Jan EricSenior Project Manager, M.Sc., Danfoss District Energy, DK-6440 Nordborgjet@danfoss.comABSTRACTIn some countries the flat station concept is becoming acommon way of realising heating and domestic hotwater (dhw) installation in blocks of flats. Anyhow, inother countries it is at the very beginning. Experiencefrom those countries reveal a number of questions whenunderstanding and evaluating the flat station concept.A number of parameters can be addressed to and beevaluated to disclose qualities and performance of theflat station concept in relation to traditional concepts forheating and dhw installations.This paper aims at analysing main parametersregarding quality (comfort) and performance of the flatstation concept, covering block distribution system, flatstation itself and flat installation. Parameters in focusare: riser system, instantaneous dhw principles, heatlosses, comfort of dhw, investments and energysavings, metering and hygienic issues for dhw.INTRODUCTIONAreas of district heating distribution systems, buildingheating installations and domestic hot water (dhw)installations show a high degree of conservatism andtraditions, which are reasonable due to their lifetime. Butthis also implies a number of questions when newconcepts like the flat station concept are to beintroduced. Not only questions addressed to the flatstation concept but also to existing systems, wheredetailed knowledge is faded out due to the maturity ofconcepts. This paper aims at analysing mainparameters regarding quality and performance for theflat station concept, covering block distribution system,flat station itself and flat installation.THE PARAMETERS ADRESSEDInvestments:–Distribution system–Basement sub station versus flat stations–Energy metersEnergy Savings:– eat loss in primary distribution system–dhw circulation pump consumptionComfort:–dhw temperature stability and variation–dhw recovery time after idle periodHygienic issues– considerations on Legionella related to the system‘sphysical layout.InvestmentsReference for comparing the flat system concept with aconventional concept is based on modern way ofmaking block pipe distribution systems [1]. In bothcases it is a horizontal pipe layout in flats with a verticalpipe tunnel for distribution. Pipe distribution systems areshown in fig. 1. Main differences are to be seen in thenumber of pipes installed. Since dhw is prepareddecentralized in flats, dhw pipe and dhw circulation pipeare eliminated. Centrally located dhw station in thebasement is replaced by decentralized flat stations.Balancing valves for heating as well as for dhwdistribution is saved for the flat station concept.Regarding metering then the dhw meter is eliminated,since the primary supply to the flat station covers flatheating and dhw as well. According to measurements ofmore than 2500 dwellings in Denmark, includingdetached houses as well as multi storey buildings,individual metering, say individual billing, resulted insavings of 15–30%, [2]. Therefore, this analysisassumes metering of all thermal energy deliveries toflats.Fig 1. Pipe distribution systems in blocks of flats. C:Modern reference principle. F: Flat station principle.A recent investment example comparing flat stationconcept (F) with traditional system (C) is included infig. 2. Data are based on a Danish case from Århusarea where a block, built in 4 levels and a basementlevel consisting of 24 flats, will generally be modernised.Investments compared are based on conceptspresented in fig. 1. Main conclusion is that investmentlevel is approx. break-even for the two systems, for this16

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniatypical Danish case. For other countries implying othercomponents/costs levels, level could change. Ingeneral, the experience is that flat stations are onbreak-even cost level or slightly higher. This is valid fornew buildings as well as for renovation projects.Item1Investments comparison Flatstation conceptÅrhus Case, block of 24 flatsSaved eneregy meter fordhwSaved balancing vavlesfor dhw circulationSaved balancing valvesfor heating distributionSaved dhw pipes, incl.circulationSaved dhw preparationcentrally locatedInvested in Flat Stationdemand for a 1970 Danish block building (not includingenergy for dhw).Table 1. Energy losses for traditional system C, and flatsystem F based on the Århus case.Concept Pipe T T E E Net E Energ. Energ.length pipe amb. loss loss/y loss/y price costs[m] [W/mK] [°C] [°C] [W] [kWh] [kWh] [ /kWh] [ /year]Trad. con. C Sum. flow 120 0.20 40 20 480 4205 2102 0.05 105Trad. con. C Sum. return 120 0.20 25 20 120 1051 526 0.05 26Flat st. con.F Sum. flow 120 0.20 55 20 840 7358 3679 0.05 184Flat st. con.F Sum. return 120 0.20 30 20 240 2102 1051 0.05 53Trad. con. C Unit heat loss 1 pcs. 300 W/unit 2628 1314 0.05 66Flat st. con.F Unit heat loss 24 pcs. 25 W/unit 3816 1908 0.05 95dhw circ. C Summer 240 0.20 53 20 1584 13876 6938 0.05 347dhw circ. elec. Sum. + win. - - - - 30 260 - 0.25 65Trad. c. C Total 10880 544Flat st. c. F Total 6638 332Diff. C-F Total (ex. electrical consumption) 4242 212-20000-15000-10000-5000Euro05000Fig. 2. Investment balance for traditional system C and flatsystem F. Block of 24 flats.Energy savingsMain contribution to energy saving is originated frominstalled hot distribution pipes. To begin with, it isassumed that half the yearly distribution energy loss isnet loss (summer time), meaning not contributing toheating up the building. Wintertime temperatures areassumed to be identical for the two concepts, becausefor this period the heating system defines temperaturelevels. To quantify losses a room temperature of 20 °Cis assumed. Danish Technical Insulation Standard [3]requires minimum allowable heat loss constants (W/m),depending on temperatures, annual operation time andpipe diameter. These constants turn out to be quitesimilar to all pipes in question. To simplify preconditionsa heat loss coefficient of 0.20 W/mK has been chosenfor all hot pipes. Table 1 shows a comparison of pipetemperatures, heat loss and electrical dhw circulationpump.Flats in this first case are provided with floor heating inbathrooms; therefore, heating is active all year. Due tofloor heating, temperatures for the traditional conceptare lower during summer season compared to the flatstation concept, since floor heating typically operates atlower temperatures. For the flat station concept a dhwtemperature at 45 °C is assumed, demanding a primarytemperature of 55 °C.Comparing the two systems regarding heat loss, thenfavour is towards the flat station concept. For the Århuscase it means approximately 4200 kWh/year savingscorresponding to 210 Euro/year (ex. pump. costs). Thismeans a saving of approx. 2 kWh/m2/year. Thisrepresents a saving of approx. 2% of the yearly heat10000Secondly, a situation is analysed where heat loss is notutilised in the building distribution system at all. Winterenergy losses for the flat station is assumed to beusable and no floor heating is active during summer.Table 2. Energy losses for traditional system C, and flatsystem F based on the Århus case.Concept Pipe λ T T E E Net E Energ. Energ.length pipe amb. loss loss/y loss/y price costs[m] [W/mK] [°C] [°C] [W] [kWh] [kWh] [€/kWh] [€/year]Trad. con. C Sum. flow 120 0.20 20 20 0 0 0 0.05 0Trad. con. C Sum. return 120 0.20 20 20 0 0 0 0.05 0Flat st. con. F Sum. flow 120 0.20 55 20 840 7358 3679 0.05 184Flat st. con. F Sum. return 120 0.20 30 20 240 2102 1051 0.05 53Trad. con. C Winter flow 120 0.20 70 20 1200 10512 5256 0.05 263Trad. con. C Winter return 120 0.20 30 20 240 2102 1051 0.05 53Flat st. con. F Winter flow 120 0.20 70 20 1200 10512 5256 0.05 263Flat st. con. F Winter return 120 0.20 30 20 240 2102 1051 0.05 53Trad. con. C Unit heat loss 1 pcs. 300 W/unit 2628 2628 0.05 131Flat st. con. F Unit heat loss24 pcs. 25 W/unit 3816 1908 0.05 95dhw circ. C Sum. + win. 240 0.20 53 20 1584 13876 13876 0.05 694dhw circ. elec. Sum. + win. - - - - 30 260 - 0.25 65Trad. c. C Total 22811 1141Flat st. c. F Total 12946 647Diff. C-F Total (ex. electrical consumption) 9865 493Comparing the two systems regarding heat loss, thenfavour is again towards the flat station concept. For theÅrhus case it means approximately 9900 kWh/yearsavings 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 yearlyheat demand for a 1970 Danish block building.Additionally, as for the flat station concept there is noneed for dhw circulation pump, thus no need for theelectric energy of 260 kwh/year. A part of this saving isanyhow spent for the flat station concept due toadditional circulation of primary water. It is assumed thatthis is approx. half the electric energy for dhw circulationpump of 130 kwh/year.17

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaWhen looking at annual energy consumption savings inpercent, figures might appear rather low and of minorimpact. In this respect it has to be remembered thatenergy saving relates to a typical 1970 building.Present building codes require energy savings in theorder of 50% reduction for 2010 established buildingsand another 50% for 2015 established buildings. Thismeans savings in relative numbers for the flat stationconcept will triple towards 2015 compared to 1970building standards. Range of relative savings goes from2-4% to 8-16% towards 2015.ComfortComparing the two ways of preparing dhw, i.e. bystorage tank and by heat exchanger [4]/[5], it is obviousthat dynamics of control tasks is quite different. Atcontinuous tapping from full charged storage tanktemperature will be constant and also independent ontapping flow changes until colder layers (cold water)have ―refilled‖ the storage tank. At this point comfortdrops drastically. If tappings are made periodically andin shorter duration then temperature will be constantwithin each tapping, but will vary between tappings dueto mixing of temperature layers. A typical questionregarding instantaneous prepared dhw is how stable aretemperatures when applying dynamics. Regardingdynamic control performance an example is included infig. 3:Fig. 3. Dynamic control performance (step test) forthermostatic and pressure controlled heat exchanger fordhw production [6]Fig. 3 shows that stability, temperature peaks at loadchange and total dhw temperature (T22) variation islimited to 3–4 °C. Regarding oscillations at low tappingflow it should be noted that T22 is measured at heatexchanger outlet. As example a 5 m ø 22 mm pex pipereduces peaks and amplitudes additionally, dependenton frequencies, but typically 50%. This example is forvery high primary supply conditions. Oscillations appearat tap flow of 100 l/h or below. This level shall be seenin relation to the fact that a typical tapping flow for onetap is 200–400 l/h.Another relevant question is how fast dhw temperatureis on desired level if supply is in idle condition. Heredynamics are heavily influenced by idle bypassthermostat setting. Also pump dynamics are influencing,meaning how fast is the primary circulation pumpreacting on rapid changes of hydraulic conditions, sayopening of primary valve.Fig. 4. Dynamic control performance (idle recovery) forthermostatic and pressure controlled heat exchanger fordhw production. Heat exchanger is cold during idle. [6]Fig. 4 shows a flat system with cold heat exchanger.Bypass temperature setting corresponds to primarysupply temperature (Tf.dh) of 40 °C and primary returntemperature (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 0.25bar at the beginning of the tapping. In this casetemperature in circulation (Tsupply) is approx. 67 °C.Primary branch pipe from supply to the flat station is4m, ø 20 mm.Measurements show that primary supply has a delay ofapprox. 7 sec. to reach a level of 55 °C. Additional delayis then caused by heating up the heat exchanger anddhw water, this delay is additional approx. 3 sec. toreach a minimum demanded level of 45 °C. After5 meter of pex pipe of ø 22 mm additional delay isapprox. 7 sec. By this the total delay from tapping thestart to reach 45° at the tap is approx. 17 sec. In thisexample a very long idle branch pipe length is used,more realistic would be 0–2 m, resulting in a ―primaryside‖ delay of not more than a few seconds. Alsodiameter of secondary dhw pipe is rather big andrepresents a typical shared pipe dimension,representing one pipe for several taps.Anyhow, this delay is only relevant for the first tapping,since thermal capacities combined with efficientinsulation is maintaining temperature, typically with timeconstants of 1–2 hr.18

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaComfort level is increased by applying a higher bypassthermostat setting and/or a ―hot‖ heat exchanger duringidle. Fig. 5 shows an example of flat station with ―hot‖heat exchanger and thermostatic controlled heatexchanger [7]. Idle temperature is approx. 50 °Ccorresponding to dhw tapping temperature.For simulations a branch pipe flow (Q1) of 800 l/h isassumed. This represents a situation where thethermostat is fully open until the desired set temperatureis reached. Further a step vice flow change from zero toQ1 or zero to Q2 is assumed. Tapping flow is assumedto be on a high level flow for one tap, which is typicallyapplied when opening the dhw. Q2=400 l/h for allsimulations.161412Delay until reaching 45°CL2=5m & 10m (internal ø10mm) - Heat Exchanger hot & cold at idlehot - dt at T2 - L2=5mhot - dt at T4 - L2=5mcold - dt at T3 - L2=5mcold - dt at T4 - L2=10mhot - dt at T3 - L2=5mhot - dt at T4 - L2=10mcold - dt at T4 - L2=5mdT [sec]10864Fig. 5. Dynamic control performance (idle recovery) forthermostatic controlled heat exchanger for dhw production.Heat exchanger is warm during idle. [7]Fig. 5 shows a flat system with ―hot‖ heat exchanger atidle. Bypass temperature setting corresponds to aprimary supply temperature (T11) of 58 °C and primaryreturn temperature (T12) of 44 °C. This setting is thehigh end, meaning in ―high‖ end regarding comfort. Forthis system there are no primary delays, and dhwtapping temperature at the flat station is available afterapprox. 2 sec. Additional delay due to dhw pipingtowards tap would be similar to previous example.In many practical matters a compromise between thetwo examples regarding idle temperature setting fulfilsdemands for good comfort with reasonable energyconsumption.In the following a general trade off is included betweenbranch pipe length, dhw pipe length, idle condition forheat exchanger and temperature delay on dhw, basedon dynamic simulations. Pipes are simplified by simpledelay models with no heat loss. Heat exchanger isbased on a lumped capacity model described in [5].200 1 2 3 4 5 6L1 [m] (internal ø20mm)Fig. 7. Dynamic simulation for hot and cold heat exchangerduring idle. Delay (dt) for dhw temp. of 45 °C.Heat exchanger simulated is Danfoss XB06H-40 [6]. Itcan be seen from figure 7, that influence on hot or coldheat exchanger is in the range of 2 sec. delay. Branchpipe length (L1) has minor impact on time delay. This isdue to the fact that temperature is maintained with atemperature gradient along pipe during idle, reflectingT1 to T2. Basically water in branch pipe is heated to acertain level already before tapping. Anyhow, due toenergy loss and return temperature, idle bypasstemperature is lower than dhw tapping temperature inthis case.Main influence on time delay is dhw pipe diameter andlength (L2). Connection in flats shall be of ―starcoupling‖ principle where every tap has its own supplypipe with a small inner diameter. Temperature in dhwpipe water is assumed to be room temperature prior totapping. In general, additional delays of typically 3 to 6seconds shall be expected due to thermal interactionwith thermal capacities along the way to tap andhydraulic dynamics on branch pipe side and hydraulicdynamics on dhw side.Simulated waiting time for a dhw temperature of 40 °Cis included in figure below:Fig. 6. Basic application for flat station, including boundaryconditions for dynamic simulations.19

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniadT [sec]1614121086420Delay until reaching 40°CL2=5m & 10m (internal ø10mm) - Heat Exchanger hot & cold at idlehot - dt at T4 - l"=5mcold - dt at T4 - l2=5mcold - dt at T4 - L2=10mhot - dt at T4 - L2=10mcold - dt at T3 - L2=5m0 1 2 3 4 5 6L1 [m] (internal ø20mm)Fig. 7. Dynamic simulation results for hot and cold heatexchanger during idle. Delay (dt) for dhw temp. of40 °C.First of all it can be seen that time delay for reaching40 °C at tap is only a bit shorter than reaching 45 °C.This is due to the fact that the T4 temperature profilehas an almost step vice nature, i.e. if temperature goesup after the dhw pipe is flushed through, it goes almostlike a step.Different dhw controllers have different performanceregarding time delay. In case of pure proportionalcontrol for dhw system, time delay is longer at part load.This is because primary flow is proportional tosecondary flow, and the lower the flow the longer thewaiting time. Looking at the example for Q1=800l/h,Q2=400 l/h, L1=4 m, L2=5 m then time delay (dt) at T4is 6.9 sec to reach 45 °C dhw temperature. In case ofproportional controller with parameters Q1=400 l/h,Q2=400 l/h, L1=4m, L2=5 m then dt=11.0 sec to reach45 °C. This has considerable effect on time delay as L1gets longer. In case of a thermostatically controlled dhwsystem or a combination of a thermostatically andproportionally controlled dhw system, time delay isshorter because no matter how small tapping is, as longas the desired set point temperature is not reached, theprimary valve will be fully or almost fully open resultingin high primary flow. Regarding delay to reach a dhwtemperature of 40 °C this is only related to dhw pipedimension since 40 °C is the bypass temperature if heatexchanger is hot during idle. In case the heat exchangeris cold during idle, then this introduces an additionaltime delay as described above. In all cases, time delayis dependent on dhw flow, resulting in delay in the dhwpipe.Hygienic considerationsLegionella is a well-known bacterial risk for dhwsystems. Normally it is not the question whetherLegionella is present in the dhw system or not, butrather Legionella bacteria concentration in the dhw.Facts influencing on potential for Legionellaconcentration growth are dhw temperature, exchangerate of hot water in distribution pipes, and volume ofdhw water in the entire hot system. Also other factorsare influencing, e.g. systematic cleaning of showeroutlets, but this will be not addressed to here, since theeffect is similar for concepts compared.Comparing volumes of dhw in pipes for concepts, theflats station solution has significantly lower volumecompared to the traditional system. Furthermore dhwpipes should be ―star‖ connected, meaning one small(diameter) pipe from the flat station to each individualhot tap. This eliminates problematic dead end or lowflow areas.Typically volume of heat exchanger is 0.25 to 0.50 litre.Typical dhw pipe volume is 0.10 l/m, equal to 1.0 litre for10 m pipe. In total this is a volume of 1.5 to 2 litrepr. flat. The comparable centrally placed dhw systemwith dhw distribution will have a volume of 5–7 litre pr.flat. By installing a dhw storage tank this will increasesignificantly. The German DVGW regulations states thatheating dhw up to 60 °C, due to e.g. Legionella, is notrequired if volumes of heat exchanger or volume of dhwpipes is less than 3 litres [8]. Based on those physicalconcept differences Legionella bacteria risk is reducedfor the flat station concept.Future energy supply/demand perspectiveOne important challenge for DH is to convert to 4thgeneration DH systems. Intention is to realise efficientDH systems for urban areas where heat demands willdecrease due to modernisation and new building energysaving codes. In this context one way to go is to reducetemperatures in DH networks [9]/[10]. This allows forcost effective geothermal sources as well as otherrenewable low temperature sources. For dhw, normaltemperature level is 45 to 60°, where highertemperatures typically are based on considerationstowards Legionella. A way to reduce temperature levelsin DH networks is to set dhw temperature at 45 °C. Bythis a primary temperature at sub station of 50 to 55 °Cwill be sufficient. A precondition for this is to use heatexchangers for dhw production, like the flat stationconcept.CONCLUSIONThe two pipe flat station concepts, consisting ofdecentralised instantaneous dhw production, open thepossibility of reducing general DH net work temperature,which for the future will be even more relevant due todecreasing building heat demand and increasedavailability of renewable energy. For building owners,the investigated case shows that the flat station conceptis on brake-even investment level compared to20

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniatraditional systems. The flat station concept has a netenergy saving due to less installed hot pipes. Energysavings are in the range of 2 to 4 kWh/m^2/y for theinvestigated cases. Comfort level has beeninvestigated, revealing well acceptable dynamic controlperformance. Dhw temperature recovery after an idleperiod for the instantaneous preparation of dhw is,however, a trade-off between comfort and energysaving. Related to Legionella, then risk can be reducedwhen installing flat stations as presented in this paper.REFERENCES[1] Kristjansson, H. Comparing Distributions Systemsin Blocks of Flats, SDDE 2009, Slovenia[2] Gullev, L., Poulsen, M. ―The Installation of MetersLeads to Permanent Changes in ConsumerBehaviour‖, the magazine ―News from DBDH‖,#3/2006.[3] DS 452, Code of practise for thermal insulation oftechnical service and supply systems in buildings,2. Revision, Dansk Standard, 1999[4] Thorsen, J.E. Cost considerations on Storage Tankversus Heat exchanger for htw preparation, The10th <strong>International</strong> <strong>Symposium</strong> on District Heatingand Cooling 2006.[5] Thorsen, J. E. Control Concepts for DH CompactStations Investigated by Simulations, The 9th<strong>International</strong> <strong>Symposium</strong> on District Heating andCooling 2004.[6] http://www.danfoss.com/Products/Categories/List/HE/Temperature-Controllers/Temperature-controllers/IHPT-and-XB-06/b1c8a73c-59f1-4fef-8b52-f49c97b6019b.html[7] http://www.danfoss.com/Products/Categories/Group/HE/District-Heating-Substations/Substations-Direct-Heating/Flat-Stations/8f81605b-bab9-4644-961b-51a3f0503f05.html[8] DVGW regulations, Germany, Arbeitsblatt W551,April 2004[9] Olsen, P.K., Lambertsen, H., Hummelshøj, R.,Bøhm, B., Christiansen, C.H., Svendsen, S.,Larsen, C.T., Worm, J. A new Low-TemperatureDistrict Heating System for Low-Energy Buildings,The 11th <strong>International</strong> <strong>Symposium</strong> on DistrictHeating and Cooling 2008.[10] Paulsen, O., Jianhua, F., Furbo, S., Thorsen, J. E.Consumer Unit for Low Energy District Heating NetWorks. The 11th <strong>International</strong> <strong>Symposium</strong> onDistrict Heating and Cooling 2008.21

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaIMPROVED TEMPERATURE PERFORMANCE OF RADIATOR HEATING SYSTEMCONNECTED TO DISTRICT HEATING BY USING ADD-ON-FAN BLOWERSPer-Olof Johansson 1 , Janusz Wollerstrand 21 & 2Lund University, Department of Energy Sciences, Division for Efficient Energy SystemsCorresponding 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 17ABSTRACTDistrict heating (DH), which is the most common heatsource in multifamily houses and commercial buildingsin Sweden, can be produced in several different type ofproduction units.In order to gain thermal efficiency in a DH system it isimportant that DH supply and return temperatures arekept low. The temperature demand in the DH system is,during the heating season, dependent on thetemperature level in the heating system of the DHconnected buildings. Many production units benefit froma lowered DH return temperature, while others are moreaffected by a reduced supply temperature. In a CHPstationthe heat to power ratio will increase when theDH supply temperature is decreasing. In order to reducethe temperature demand, low temperature heatingsystems are of interest, as well as systems resulting in alow DH return temperature.To increase the heat output in an existing radiatorheating system, the radiators can be complementedwith small electric fans resulting in an increased shareof forced convection in the heating system. Field studieshave shown that the heat output, with constant supplytemperature and mass flow through the radiator, canincrease with more than 50%.INTRODUCTIONFor many years, return temperatures in DH networkshave been an important issue for DH research. A lowDH return temperature is in many cases favorable forthe DH production units. However, if also the supplytemperature could be kept at a low level the share ofelectricity produced in a CHP station could increase.This would lead the way towards an increased share ofelectricity produced by non fossil fuels. In Sweden morethan 30 % of the DH is produced in CHP stations [4].In many reports the gain from a reduced temperaturelevel in the DH network has been discussed andquantified in economic terms, see e.g. [12], [13].The DH supply temperature level in the DH network is,during heating season, dependent of the temperaturedemand in the DH-connected buildings heating system.In modern buildings low temperature heating systemsare common, which may allow reduced DH temperaturelevel. In order to reduce the temperature demand inexisting buildings the idea of using small add-on-fanblowers placed under the radiator to increase the heatoutput due to an increased share of forced convectioncame up.ObjectiveThe field study presented in this paper investigates thepossibility to reduce the space heating temperatureprogram and estimates the impact on the DH supplyand return temperature. Possible reduction of the DHflowrate is also calculated.This paper is focusing on buildings indirectly connectedto the DH network through a substation with heatexchangers (HEX).DESCRIPTION OF ADD-ON-FAN BLOWERThe add-on-fan blower that is tested in this studyconsists 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 radiatorswere tested, a panel and a column radiator.T ssSpace heatingradiatorAdd-on-fanT srIncreased air flowFloorm sOuter wallFig. 1 The add-on-fan blower mounted on a panel radiator.The add-on-fan blowers in this study are provided by aSwedish company: A-energi AB (the product is called―fläktelement‖ in Swedish). The company describes thefeatures of the add-on-fan blower as a possibility toreduce the temperature program without replacing theradiators, with the aim to reduce the electricity demandfor buildings supplied with heat from heat pumps [5].22

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTHEORYIn this section a theoretic analysis of the impact offorced airflow on heat output from radiator with a lengthof (L) 1 m and height of (h) 0.59 m is described. Theradiator is in this study approximated by a flat verticalplate. The indoor temperature is assumed to beconstant at 21°C and equal to T inf .Heat outputThe heat output from the radiator to the room arisesfrom convection and radiation. The heat transferprocess from heating water to the room through aradiator is summarized in equation 1 [7], [8].Q m c ( T T) ( k A)(1)spsssrk is the heat transfer (convection) from the water to thesurrounding metal, conduction through the metal andconvection from the outer surface of the radiator to theroom according to equation 2.1 1metal 1(2)k watermetalmetalconv The dominating parameters in this equation are theconvection and radiation between the radiator and theroom (α conv and α rad ), while the other terms, in this case,can be neglected. This results in a new equation forenergy output, see equation 3.Q Q Q A A (3)radconvconvconvThe temperature Δθ is the logarithmic meantemperature difference according to equation 4.Tss TsrTss TilnT Tsriradradrad(4)3Q C (7)radT mConvectionThe convection that arises due to the temperaturedifference between the radiator surface and thesurrounding air is a function of the Nusselt number (Nu),see equation 8. Nu h(8)conv/The heat output due to convection is divided into threesections, natural, mixed and forced convection.For natural convection, the Nu number is dependent onthe Rayleigh number (Ra), which is a product of thePrandtl number (Pr) and the Grashof number (Gr). Forair, Pr can be considered constant, Pr=0.71, wile3hGr g (9) 21/T inf 1/ T iwhere g is the gravity force, ν is kinematic viscosity andβ is the coefficient of expansion.Several empirical relations describing Nu are available.In this study a relation described by Churchill has beenused [9], see equation 10 and 11.0.250.67 Ra9Nu 0.68 Ra 10 (10)9 /16 4 / 9[1 (0.492/ Pr) ]Nu0.50.387 Ra 0.825 [1 (0.492 / Pr)1/ 6]9 /16 8 / 279Ra 10 (11)For forced convection the Nu number is calculated byequations described by Holman [10], see equations 12and 13.Nu 0.664 Re0.5 Pr 1/35Re 510 (12)RadiationAccording to Trüschel [8] the heat output from radiationcan be estimated according to equation 5.Qrad 4 radrad AradAA radradradiator (1 radT)3m Arad Where the temperature, T m , is the mean temperature ofthe radiator surface and the surfaces in the rooms, seeequation 6. For a panel radiator the A rad /A radiation =1 [8].Tm Tradiator Troom,surface( Tsf T ) / 2 TSince the A rad and emissivity, ε rad , are constant for aspecific radiator, the relation can be simplified toequation 7.sr2i(5)(6)1/30.857Nu Pr(0.037 Re 871)5 10 Re 10(13)and the Reynolds number, Re, is described as:u LRe (14)The product of Gr/Re 2 describes the dominating type ofconvection. If Gr/Re 2 >10, natural convection isdominating, if Gr/Re 2 ≈1, both natural and forcedconvection is of importance and if Gr/Re 2

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaImpact of air speed on space heating temperatureResults from the theoretical analysis, using theequations above, are shown in Fig. 2 to Fig. 4. The heatoutput for a radiator designed for the temperatureprogram 60/45 °C is illustrated as a function of the airspeed in Fig. 2. The supply temperature and the massflow through the radiator are kept constant. Two caseshave been derived, one with heat output only fromconvection, and one with heat output from both radiationand convection. With ε=0.9, the share of heat outputfrom radiation will be 65% at DOT.C and % (W/m 2 K)15010050T sf(C)T sr,conv(C)% additional Q, convT sr,rad&conv(C)% additional Q, rad&conv00 2 4 6 8 10 12 14 16air speed (m/s)30201000 2 4 6 8 10 12 14 16air speed (m/s)Fig. 2 Calculated heat output improvements at T ss=60 °C,T sr0=45 °C with increasing airspeed. m s is kept constant.As seen, the additional heat output from the radiator isincreasing rapidly when the air flow is increased. Withradiation taken into account, the increase is somewhatlower since the mean temperature, T m , is decreased,see equation 7.In Fig. 2 the heat output is increasing. In Fig. 3 andFig. 4 the supply temperature to the radiator is reducedinstead to keep the heat output constant. New T ss andT sr can now be calculated under the assumption that thetotal heat output and the mass flow (m s ) through theradiator are constant. The impact of the air flow isdescribed for three different heat loads (Q rel =100%,50%, 25%) with standard 60/45 °C temperatureprogram as a reference. See Fig. 3.Temperature ( C)6055504540353025T ss0= 60C , Q rel= 100 %T sr0= 45C , Q rel= 100 %T ss0= 43.1C , Q rel= 50 %T sr0= 35.6C , Q rel= 50 %T ss0= 33.6C , Q rel= 25 %T sr0= 29.9C , Q rel= 25 %200 5 10 15 20 25 30air velocity (m/s)Fig. 3 Possible T ss and T sr to for three heat load situationsat different air speeds. Q rad=65% at DOT.New temperature programs have been derived for somemoderate air speeds, see Fig. 4. As seen the impact ofan increased air flow, expressed in °C, is larger at highrelative heat load.Temperature ( C)65605550454035302520NEW TEMPERATURE PROGRAMT sf0U= 0.0m/sT sr0U= 0.0m/sU= 0.5m/sU= 1.0m/sU= 2.0m/sU= 3.0m/s150 20 40 60 80 100relative heatoutput (%)Fig. 4 New space heating temperature programs atdifferent air speeds. Red lines: T ss, Blue lines: T sr.Q rad=65% at DOT.In the calculations performed, the radiator is assumed tohave the same heat output from both sides of theradiator. The air flow is assumed to be uniformlydistributed through the length and height of the panelradiator. This is not the case in the real add-on-fanblower applications, however, one can expect resultsfollowing the same pattern.EXPERIMENTAL STUDYTo investigate the performance of the add-on-fanblower, two radiators of different type were supplied withsuch device during the heating season 2009/2010. Thepower supply to the fans was scheduled to switch onand off while the mass-flow (m s ) through the radiatorwas kept at a constant level.Field study objectThe radiators are situated in two offices at LundUniversity. The original temperature program for the24

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaradiators in the building is running at 60/45ºC at DOT(represented by narrow black lines in Fig. 8 troughFig. 13).The radiator types tested were: Panel radiator, see Fig. 5 Column radiator, see Fig. 6.* Calculated for new radiators of the same dimensionsmanufactured by Lenhovda radiator factory [3]The radiators are located in traditional officeenvironment in a building built in 1960.Data acquisitionMeasured parameters in the test were: secondarysupply and return temperature (T ss and T sr ), indoortemperature (T i ) and outdoor temperature (T out ).The impact on return temperature for a given supplytemperature was then calculated.In Fig. 7, a screenshot from the logger software isshown. The return temperature is decreased by 5°Cwhen the fan is switched on. This results in anincreased heat output by more than 60 %.fanT sr,0 =39 T sr,Fan =34Fig. 5 Add-on-fan blower mounted on a panel radiator.T ss (ºC)T sr (ºC)T i (ºC)T out (ºC)U fan (V)Fig. 7 Log file from field study.New reduced temperature program will be derived innext section.MODIFYING SPACE HEATING TEMPERATUREPROGRAMFig. 6 Add-on-fan blower mounted on a column radiator.For each radiator the fans have been run at twodifferent rotation speeds. The net electric powerconsumption (P fan ) has been measured. See Table 1 forP fan and the design heat energy output at DOT. Notethat the electric power to the add-on-fan blower isconstant and not dependent on the relative heat output.Table 1 Radiator and add-on-fan blower design.RadiatortypePanelColumnP fan(el)2.7 W1.9 W3.0 W2.2 WQ @ DOT, 60/45ºC(Heat)360 W*430 W*MethodWhen the add-on-fan blower is switched on, the T pr isdecreasing, causing an additional heat output since m sis kept constant. See Fig. 8.Temperature [C]6055504540353025T ssT sr,0T sr,FanOriginal cooling in radiatorAdditional cooling withadd-on-fan-operation200 0.2 0.4 0.6 0.8 1Relative heatloadFig. 8 Increased cooling of secondary system.25

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe relative heat output from the radiator with andwithout fan operation is calculated from equation 1.QQrel,0rel,Fan msCp( TssTsr0)QDOT mC( T T) Qspsssr,FanSince the temperature drop in the radiator is increasingwith the fan in operation, the radiator now could beconsidered oversized. Then, with the same type ofreasoning as in e.g. [2], the T ss program or m s needs tobe adjusted in order to avoid overheating of the building.In this paper, the m s has been considered constant,allowing us to compute the new relative space heatingload for a given T ss according to equation 2. Q rel,0 iscomputed using the original space heating temperatureprogram.Q( TT)DOT(1)ss sr,Fanrel,Fan Q(2)rel,0( TssTsr0)Knowing Q rel,Fan , a new temperature curve, which willresult in correct heat output from the radiator with thefan in operation, can be calculated. The curve appearsto the right in the diagram, see Fig. 9.T s[C]6055504540353025Panel radiator P fan= 2.7 WT ss,0T sr,0T ss,FanT sr,Fan200 0.2 0.4 0.6 0.8 1 1.2 1.4Q relFig. 10 Modified temperature program for panel radiator,P fan=2.7 W.60555045Panel radiator P fan= 1.9 WT ss,0T sr,0T ss,FanT sr,Fan6055T ss,0T sr,0T s[C]403550T ss,Fan3045T sr,Fan25T s4035200 0.2 0.4 0.6 0.8 1 1.2 1.4Q rel3025200 0.2 0.4 0.6 0.8 1 1.2 1.4Q relFig. 11 Modified temperature program for panel radiator,P fan=1.9 W.Column radiator P fan= 3.0 W60T ss,055T sr,0Fig. 9 Modified secondary temperature program.New space heating temperature program - resultsThe procedure described above has been applied to allcollected data. Results are shown in Fig. 10 and Fig. 11for the panel radiator, and Fig. 12 and Fig. 13 for thecolumn radiator.T s[C]5045403530T ss,FanT sr,FanAs seen in the figures, the temperature program is nowsignificantly reduced for both the panel and the columnradiator. The new temperature program shows a similarpattern for both types of radiators. For the panelradiator, the measured return temperature is moreconcentrated, especially at the higher fan speed. Thiscould be explained by a more favorable air flow patterndue to the physics of the radiator.25200 0.2 0.4 0.6 0.8 1 1.2 1.4Q relFig. 12 Modified temperature program for column radiator,P fan=3.0 W.26

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia60Column radiator P fan= 2.2 W110DH primary supply temperature5550T ss,0T sr,0T ss,Fan100T s[C]454035T sr,FanT ps[C]9080307025200 0.2 0.4 0.6 0.8 1601 0.8 0.6 0.4 0.2 0 0rel heatloadFig. 13 Modified temperature program for column radiator,P fan=2.2 W.Q relINFLUENCE ON DH NETWORKKnowing the reduced temperature level on thesecondary side of the HEX, the impact on the DHnetwork can be estimated. The impact is calculatedbased on two different strategies:1. Primary supply temperature (T ps ) is kept at thesame level as before2. The primary flow (m p ) through the HEX is keptconstantBy applying the first strategy, both T pr and the mass flowin the DH network is reduced. The second strategyresults in a reduced T ps and T pr without changing theflow rate in the DH network.Results so far will now be applied to a DH substationdimensioned as recommended by the Swedish districtheating association [1]. The calculations are made witha parallel connected DH substation serving a buildingwith 20 apartments. The substation is providing thebuilding with heat and domestic hot water (DHW) andDHW circulation. The assumed DHW usage is125 l/apartment&day, space heating load at DOT is3 kW/apartment. The heat loss from DHW circulation isassumed to be 0.1 kW/apartment. For each spaceheating load a flow-weighted mean value for T ps and T pris calculated for a time period of 24 h, including heatload from both DHW and DHW circulation. Thereference DH supply temperature, dependent on thespace heating load, is assumed as illustrated in Fig. 14.Fig. 14 DH primary supply temperature.ResultsThe first control strategy is in Fig. 15 – Fig. 20 noted as„T ps unchanged‟, and the second strategy is noted as„m p unchanged‟.In Fig. 15 and Fig. 16 the possible reduction of DHsupply temperature is shown.T pssaving [C]15105Panel radiator T pssavingP fan= 2.7 W m punchangedP fan= 1.9 W m punchanged01 0.8 0.6 0.4 0.2 0rel heatloadFig. 15 Resulting T ps reduction with panel radiator.T pssaving [C]15105Column radiator T pssavingP fan= 3.0 W m punchangedP fan= 2.2 W m punchanged01 0.8 0.6 0.4 0.2 0rel heatloadFig. 16 Resulting T ps reduction with column radiator.In Fig. 17 and Fig. 18 the reduction of T pr is shown. Asseen the reduction of T pr is of the same magnitudeindependently of which control strategy is used.27

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaHowever, by keeping the DH supply temperatureconstant (strategy 1) the flow rate in the DH network isaffected, see Fig. 19 and Fig. 20.2015Column radiator m psaving (%)15Panel radiator T savingprP fan= 2.7 W T psunchangedP = 1.9 W T unchangedfan psP = 2.7 W m unchangedfan pm psaving [%]10T prsaving [C]105P fan= 1.9 W m punchanged5P fan= 3.0 W T psunchangedP fan= 2.2 W T psunchanged01 0.8 0.6 0.4 0.2 0rel heatloadFig. 20 Resulting m p reduction with column radiator.01 0.8 0.6 0.4 0.2 0rel heatloadFig. 17 Resulting T pr reduction with panel radiator.T prsaving [C]15105Column radiator T prsavingP fan= 3.0 W T psunchangedP fan= 2.2 W T psunchangedP fan= 3.0 W m punchangedP fan= 2.2 W m punchangedAnnual gain in T ps and T pr during heating seasonIn order to evaluate the annual impact on the primarytemperature level, the outdoor temperature has to beconsidered. In this case measured values for theoutdoor temperature in Malmö have been used, seeFig. 21.T out(C)4035302520151050-501 0.8 0.6 0.4 0.2 0rel heatloadFig. 18 Resulting T pr reduction with column radiator.m psaving [%]2015105Panel radiator m psaving (%)P fan= 2.7 W T psunchangedP fan= 1.9 W T psunchanged01 0.8 0.6 0.4 0.2 0rel heatloadFig. 19 Resulting m p reduction with panel radiator.-10May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr MayFig. 21 Outdoor temperature in Mamö1 st May 2006 – 30 th April 2007.When calculating the annual gain for a DH-network acomparison of flow-compensated mean temperatureduring the heating season has been made. For thecalculations we assume a maximum heat output (Q DOT )at -15 °C and the balance temperature, when no spaceheating is needed, +17 °C.Table 2 Reduction in annual primary temperature levelduring heating seasonColumnradiatorT ps -m p -unchanged unchangedTΔTps T prpr2P fan= 2.2 W -2.2 °C -2.4 °CP fan= 3.0 W -5.7 °C -6.6 °CPanelradiatorP fan= 1.9 W -0.8 °C -0.9 °CP fan= 2.7 W -2.5 °C -2.7 °C28

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaNote that the annual flow-compensated meantemperature in Table 2 is based on results from the fieldstudy where measured values for low relative heat loadare missing, which makes the values in the tablesomewhat underestimated.CONCLUSION AND DISCUSSIONBy installing the add-on-fan blower application onexisting radiators the temperature level in the heatingsystem can be substantially reduced. This will also haveimpacts on the DH network and DH production units.The impact on the DH network can be applied based ontwo principles:1) DH supply temperature kept at the same level aswithout the add-on-fan blowers. This will result inreduced primary flow rate.2) Reduced DH supply temperature while primary flowrate is kept constant.The first strategy could be applied immediately, sincethe primary supply temperature is kept as the samelevel as before. This means that not all heating systemsconnected to the DH network need to be modified inorder to apply this method. The lowered secondarytemperature level results not only in reduced DH-returntemperature, but also in a reduction of the DH flow rate.The reduced flow rate could be used to increase thenumber of buildings connected to the DH network, or toavoid bottlenecks in the DH network. The magnitude ofthe reduction of the DH supply temperature is between9 and 12 °C at DOT and at the same time the flow rateis decreased with more than 10 %. On annual basis thepossible reduction of temperature level in the DHnetwork is in the magnitude of several degrees Celsius.In order to apply the second strategy the demand for ahigh temperature level in the DH network needs to bereduced for all the connected buildings. Otherwise theDH flow rate will increase. Calculations based on theresults from the field study in this paper shows that theDH supply temperature can be reduced with about10 °C at DOT without affecting the DH flow rate. At thesame time the DH return temperature will be reducedwith as much as 10 °C at DOT.The performance of the tested add-on-fan blowerscorresponds to the pattern of theoretical calculations.However, the results are not comparable since the airflow in the pilot project has not been measured.The results presented here are an important part in theevaluation of effects of improvements in consumerheating systems on primary energy efficiency in DHsystems including production plants, especially CHP.ACKNOWLEDGEMENTThis work is part of the Primary Energy Efficiencyproject of Nordic Energy Research.NOMECLATUREAbbreviationsCHPDHDHWDOTHEXVariablesCombined heat and power stationDistrict heatingDomestic hot waterDesign outdoor temperatureHeat exchanger (DH substation)α Heat transfercoefficient (W/m 2. K)β coefficient ofexpansion (K-1)δ Thickness (m)λ Conductivity (W/m . K)Gr Grashof number (-)h Height (m)k Heat transfercoefficient (W/m 2. K)L Length (m)ε Emissisivity (-) m mass flow (kg/s)ν Kinematic viscosity(m 2 /s)ζ Stephan-BoltzmanconstantΔθ Logarithmic meantemperaturedifference (K)A Area (m 2 )Nu Nusselt number (-)P Electric power (W orkW)Pr Prandtl number (-)Q Heat output (W or kW)c p Heat capacity (J/kgK) Ra Rayleigh number (-)C Constant Re Reynolds number (-)g Gravity force (N/s 2 ) T Temperature (ºC or K)Subscripts0 Design condition(without fan)Fan Add-on-fan blower inoperationprPrimary (side)Returni indoor rad radiationm Mean rel Relativeout outdoor s Secondary (side) orSupply29

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaREFERENCES[1] Svensk Fjärrvärme, ―District heating substationsdesign and installation, Technical regulationsF:101‖, The Swedish District Heating Association,2008[2] P. Ljunggren, P-O. Johansson, J. Wollerstrand,―Optimized space heating system operation with theaim of lowering the primary return temperature‖, inProc. of the 11th <strong>International</strong> <strong>Symposium</strong> onDistrict Heating and Cooling, 2008, Reykjavik[3] http://www.lenhovdaradiatorfabrik.se/display_sub.asp?apid=20, 2010-04-16, Downloaded spreadsheet for calculating heat output.[4] Swedish Energy Agency, 2008, ―Energy Indicators2008, Theme: Renewable energy‖, 2008[5] http://a-energi.jetshop.se/, 20010-04-20[6] EN 15316-4-5:2007, ―Heating systems in buildings.Method for calculation of system energyrequirements and system efficiencies‖, CEN,Brussels, 2007[7] J. A. Myhren, S. Holmberg, ‖Design considerationwith ventilation-radiators: Comparisons to traditionaltwo-panel radiators‖, Energy and buildings 41, p.92-100, 2009[8] A. Trüschell, ―Värmesystem med luftvärmare ochradiatorer, En analys av funktion och prestanda‖,Licentiate Thesis, Chalmers, Göteborg, 1999[9] S. W. Churchill, ―Correlating equations for laminarand turbulent free convection from a vertical plate‖,Int. J. Heat Mass Transfer, Vol. 18, p. 1323-1329,1975[10] J. P. Holman, ―Heat trans fer‖, 9th edition, 2002[11] Discussion with professor B. Sundén, April 2010[12] P. Selinder, H. Walletun, ‖Modell för ändradeförutsättningar i fjärrvärmenät‖, Rapport 2009:50,Svensk Fjärrvärme, 2009[13] S. Werner, FVB Sverige AB, ‖Nytta med svenskfjärrvärmeforskning‖, FoU – orientering 2004:9,Svensk Fjärrvärme, 200430

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaPRIMARY ENERGY EFFICIENCY AND SYSTEMS ENGINEERINGM.Berner 1 , R. Ulseth 1 , J.Stang 21 Norwegian University of Science and Technology (NTNU)2 SINTEF Energy ResearchABSTRACTThe revised Energy Performance of Building Directive(EPBD) [1] emphasizes that the energy performance ofa building shall be calculated by use of Primary EnergyFactors (PEF). Calculation of CO 2 emission will not bemandatory so far. Thus EPBD will reduce the use ofnon-renewable energy, incite the use of energy fromcombined heat and power generation (CHP) andreduce the energy consumption in the building sector.A simplified method that enables comparison of thePEF from different energy chains is required. However,calculation of all the parameters affecting the PEFvalues like energy used for extraction, transportation,power and heat generation etc. is time-consuming. Themethod described in EN 15603 [2] is rather general land provides PEF values for 13 energy carriers andchains. This is based on average European values. LifeCycle Assessment methods include several of therelevant steps, but a complete LCA often implycollection of more than 6000 parameters.The systems engineering method used here havedemonstrated the feasibility of developing a genericmethod that provides credible data for calculatingprimary energy efficiency. It applies the generic methodon energy chains in the Nordic region which is relevantto CHP plants utilising bio based fuel.INTRODUCTIONBackgroundThe terms Primary Energy, Primary Energy Efficiencyand Primary Energy Factors (PEF) are introduced [3] -[8] in order to compare different energy sources andchains based on losses and a calculated environmentalimpact.Primary energy is energy that has not been subject toany conversion or transformation process. The use ofprimary energy factors takes into account the energythat are used from the extraction of the energy carrierand all of the losses until energy is delivered to the enduse in the desired form such as heat, cooling orelectricity .The primary energy factor (PEF) expresses how muchprimary energy is needed to deliver 1 unit of power,heat or cooling to the end user. The term primaryenergy efficiency (PEE) therefore is used to describethe total use of energy from extraction to the end user.ExtractionProcessingFigure 1 A Typical energy chainMethodologyAn energy chain might consist of several elements orprocesses from extraction, through processes such asdrying, storage, transport, power/heat/cool generation,and distribution to the end user. In order to ensure thatthere is a correct PEF, all elements that influence theenergy flow have to be accounted for.The energy balance or calculation of the energyefficiency of a process focuses primary on the energyinput in the form of fuel and the output in kWh, andlacks information on the energy used to buildinfrastructures such as the power plant, distribution net,transportation and the extraction.Life Cycle Assessment (LCA) might contribute toprovide such information in a generic method.However, the number of input parameters, often morethan 6000 in an ordinary LCA analysis demonstratesthe need for an easily accessible method.Systems engineering is a method that has beendeveloped gradually with increasing complexity ofprojects and systems. Systems engineering is oftenconsidered to have started at Bell Laboratories in the1940s, later applied in organizations such as NASAand formalized as a separate engineering field with theformation of INCOSE [9] in 1990. The benefits ofsystems engineering is the possibility to treat complexsystems with several subsystems. Therefore, as a firststep in the development of a method a systemsengineering approach has been chosen. The mainobjective is to develop systems and methods thatenable a sufficiently reliable calculation to be made ofthe primary energy factor (PEF) in general and fordifferent energy chains with required level of details.At present systems engineering approaches have notbeen found to have been previously applied on thedevelopment of generic PEF methods for differentenergy chains.ObjectiveStorageTransportGenerationTransformationTransmissionDistributionThe objective of this paper is to show how systemsengineering can be used as a tool to reveal important31

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaparameters when a model for calculation of the PEE ofdifferent energy chains is developed. The paper willshow an overall approach and will not describe all thenecessary iterations in detail.SYSTEMS ENGINEERINGThe system engineering processA systematic approach such as systems engineering isessential to be able to develop a generic modeldescribing a complex system with several subsystems.The intention with the systems engineering process isto analyse and describe complex systems. Often themethod is used in the design process, to make surethat the subsystems are connected properly, that theprocess is optimized and that the different componentsare described, implemented and integrated precisely.A common feature of all systems engineeringprocesses is an indefinite number of iterations at alldifferent steps.Systems engineering principles are often applied whena new system or products are developed. Themethodology alters slightly between development andre-engineering.Re-engineering methods are applied when an existingsystem is described. The energy chains considered arealready designed and built, and a re-engineeringtechnique is selected in order to develop a method thatcalculates the PEF for different kind of energy chains.3. Measures of effectiveness (MOE)The definition of MOE are: ‖A small subset of therequirements that are so important that the system willfail if they are not met and will be a huge success ifthey are met‖ [11].4. Development of information modelsThe different information models describe the observedsystem in relation to legislation, physical architectureand a system interface model. Four separate modelsare developed5. Trade-offsRequirement traceability modelSystem architecture modelBehaviour modelSystem interface modelThe trade-off phase is essential in the development ofa method. Each of the steps is carried out in iterativeloops gradually increasing detailing level. Aftersatisfactory trade-offs have been performed andconsistent information models obtained, a theoreticalmethod is developed. Real data are collected and trade-off between the model and the gathered data areperformed.6. DocumentationThe developed method will be then documented byactual case studies before a final reporting.CHOSEN METHODOLOGYThe system re-engineering process consists of thefollowing six different tasks according [2]. Some ofthem might seem unnecessary, but they all contributeto the decomposing of a system and development of amethod.1. Establish problem statement;This comprises the definition of the problem approach,which includes development of a problem statementdescribing the problem/challenge, its importance and astate of the art. To be able to establish the problemstatement; four questions must be answered:What is the problem?Why is it importantWhat have others done?What must be done?2. Assess available information assessmentProvide available information including an overview ofpossible stakeholders.1EstablishproblemstatementIterate to find feasiblesolutionFigure 2 The system re-engineering process describedas a functional block diagram (FFBD), ref. [10]ESTABLISH PROBLEM STATEMENTWhat is the problem?and4Createrequirementtraceability modelCreate systemarchitect. modelCreate contextmodel2Asssessavailableinformation3DefineeffectivenessmeasuresTradeoffsUse of Primary Energy Factor (PEF) will provideinformation on the energy losses and consequently theenvironmental impact of different kind of energysources, power production processes and energy44Create behaviourmodel4Nofeasiblesolution5Feasiblesolution6Documentcurrentsystemdesign32

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniatransport systems. At preset there exists no easyaccessible calculation method.Different countries have different energy chains andenergy supply systems. Analysis of even the mostactual processes and process lines does not existneither for Norway or Europe [12] – [13]. In order tocompare and choose different energy chain there is aneed for standardized methods. The lack of objectiveand reliable data of the different elements in the energychain might prevent an efficient use of energy, andcontributes to wrong choices and unnecessary CO 2emissions. [14] – [15].The method is principally described in EN 15603 [2]and provides only single PEF values for 9 energycarriers and 4 energy chains, and is based on averageEuropean values. Without an easy accessible methodor methods is it not possible to compare PEF valuesand calculate the actual environmental impact ofdifferent energy chains. Some studies [16] -[21] havedescribed parts of this topic, but they lack a holisticview of the energy chains from cradle to grave, oftenthe chosen system boundaries are different, time scalevaries, detailing level different and the, approach/method varies. Results from different studies thereforeare not comparable.Why is it importantPEF is a key indicator to be able to evaluate energyuse (for different purposes) especially with regards tothe goals of the EPBD [1]. PEF is an over all energyefficiency indicator which makes it possible to compareand collocate different energy sources and energycarriers by a single number. The same method can beused to calculate the CO 2 emission.What have others done?Different CEN standards describe, and partly discuss,the theory. In the EC-mandated CEN standards relatedto EPBD mainly one single reference are referred [14]whilst the PEF values have been gradually changedover time. An extended literature survey has showeddiscontinuity between some of the studies performedand lack of details in the calculations.Methods developed to provide PEF values for heatingsystems in buildings might be useful, but they will nottotally comply with a whole energy chain approach. LifeCycle Assessment (LCA) might also contribute to ageneric method, but the vast number of inputparameters, often more than 6000 in a traditionally LCAdemonstrates the need for a more easy accessiblemethod.What must be done?In order to develop a method a systems engineeringapproach will be used. The most important task in thiscontext is the identification of relevant energy systems33and process lines (chains) primary in Norway and theNordic countries. Detailed data must be provided suchas efficiency and loss from the different systems andmix of systems, or at least provide the necessaryparameters. Since the systems engineering approachis chosen, the problem approach must be defined, atheoretical method developed and data collect. Thisincludes performing of a trade-off between thetheoretical model and available information. Themethod shall be tested by selected case studies andfinally adjusted.Main hypothesisAs a part of the systems engineering process, one orseveral (systems engineering) hypothesis is developed.The success of a system engineering process isrelated to the fulfilment of the hypothesis. In this projectthe system engineering method must prove two mainhypotheses;1. It is possible to develop a generic method thatprovides credible data for calculating primaryenergy use by use of PEF values.2. It is possible to apply the generic method onenergy systems in the Nordic region for CHPplants utilising bio based fuel.Stakeholder analysisA stakeholder is a party having a right, share or claimin the system [16]. The intention with the stakeholderanalysis is to reveal the different kinds of stakeholderssince they might have requirements influencing apossible method in a legal way. Stakeholders withmutual interest are aggregated in groups; some ofthem might not be in incompliance with each other.Energy producer, distributors, energy companies;Business profitability is the main issue byoptimizing production from different energy carriersaccording to cost-benefitInvestors (energy and building); The electricitymarkets are opening gradually throughout Europe,e.g. Nord Pool Financial Marked and investmentsin Power production and the introduction of socalled Green Electricity Certificates might be a newor extended business area. Investors in thebuilding marked might be interested in the actualPEF values and primary energy use whenchoosing between different investment objects.Building owners, end user; Correct calculation ofPEF values and primary energy use is supposed tohave significant importance for the choice ofenergy supply system, building services, insulationlevel, especially for new buildings and majorrehabilitation projects. Future operating cost might

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniadepend of PEF since taxes might be dependant ofthe primary energy use and/or the CO 2 emission.Developers& Building and construction industry;Technical equipment in the building and designstrategies depends on the actual use of thespecific energy carrier. The use of PEF values inthe primary energy calculations might change thevalue of traditionally installed equipment due tooverall energy costs and also create a demand formore energy flexible solutions.Politicians, government, Regulators, Communityplanning; Most European countries have affiliatedthe Kyoto Protocol, and a possible method toincrease the use of renewable energy policy toolsand subsidy schemes might be based on the useof PEF values for the different solutions, besidespossible tax on systems with high primary energyuse.National regulators mandatory monitor and reportemission levels and this influences nationallegislation, local and urban planningResearch groups, Universities; Different researchcommunities might be interested in development ofother PEF calculation methods or adjustments ofmethods and development of new solutions andsystemsDESCRIPTION OF MEASURES OF EFFECTIVENESSAs earlier stated the measures of effectiveness (MOE)should be independent of any solutions and notconcerned with internal details [22]. It might also befruitful to develop MOE for the different kind ofstakeholders since they often might have a differentopinion regarding MOE.In this context MOE are primarily described for theongoing Nordic PhD-project Primary Energy Efficiency(PEE). A further detailing level, by including thestakeholders, might provide valuable information, butthat is considered to lie outside the scope of this work.The methods (tools) developed during the projectshould be suitable for different kind of energy chains.The results should be utilised by the differentkinds of stakeholders e.g. the building owner, thearchitect/designers of the building, the energysupplier and producer and finally politicians andgovernments.The methods will enable the different stakeholdersto choose between different energy systems andfurthermore be able to reduce primary energy useand CO 2 emissions from stationary energypurposes.34INFORMATION MODELSThe requirement traceability information modelIn systems engineering shall the requirementtraceability information model “aim to show the breakdown of requirements from source documents to finalallocation functions to stakeholders ―[2]This model is an important tool to keep track of thedifferent requirements, source documents andeventually what the system accomplishes and who orwhat are in charge. Usually an Entity-Relationship-Attribute method is used [23], where the entities(objects) represent the legislation, requirements, etc.whilst the relationship shows the association betweenthe system/process.Planning andBuilding ActSourceBuildingregulation TEKSourceBuilding GuideRENSourceSpecifieNS3031 sfunctionAllocatedBuilding Permit tostakeholderDocumentsIncorporatesWorking Environm. ActSourceEN 15316-4-4:2007)SourceEPBD EnergyPerformance.SourceËN 15603:2007SourceLandfill DirectiveDocumentsSpecifieDisharge permit sRequirementsAllocated toSourceWaste regulationsSourceNOx emissionsBoilerFigure 3 Selected part of the requirement traceabilitymodel, case Norway96/62/ECAmbient AirSourceQualityDocuments1999/30/EC LimitvaluesRequirementsNOx…IncorporatesPollution ControlActRequirementsSpecifiesDischarge permitRequirementsSpecifiesInternal controlsystemfunctionAllocated toNOx emission1.2.2.1StakeholderMost EC directives are enforced and implemented inlaws, directions, regulations and guidelines, both in theEU and associated EEC countries, hence the order ofentities in Figure 3. Several directives influence thenational laws and regulations. Since the directivesusually are enforced through national laws, the lawincludes requirements from more than one directive likethe Norwegian Planning and Building Act [26], whichincludes requirements from EPBD [27], Directive on thepromotion of the use of energy from renewable sources[30], Directive on the landfill of waste [29], ThePollution Control Act [30] amongst others.The requirements traceability model provides importantinformation about constraints regarding an energychain. Some of the elements such as the the WorkingEnvironment Act [31] might seem irrelevant, but theregulations set limits for the pollutant inside the workingarea, introducing need for e.g. conveyor belts.Each function consists of several entities for instancethe discharge permit from regulators like TheNorwegian Climate and Pollution Agency will setrestrictions on the authorised discharge levels ofdifferent gasses not only NOx as illustrated in theFigure 3.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe requirements affect most of the stakeholders, forinstance the Planning and Building Act will affect bothend users, construction industry and energydistributors.Architecture information modelThe architecture information model shows the physicalcomponents of a system with subsystems.An energy chain consists of several sub systems asdescribed in Figure 4. A more detailed architectureinformation model is also developed.Each of those sub elements can be spilt up into subelements as shown in Figure 5. The final or basiselement can be described as Figure 6.In order to describe the possible physical systems ageneric model is developed [11], detailed description ofsome of the most relevant energy chains are carriedout in the actual PhD thesis, Figure 1 shows a principaldescriptions of a single energy chain. An end user willtypically be supplied with energy from a variousnumber of energy chains, and each element mightrepresent parallel processes.Consist ofEnergyFeedingsystemComponentEnergyTransformation systemConsist ofComponentEnergyProductionsystemComponentConsist ofConsist ofConsist ofCombustionEnergyprocessTransformation systemComponentComponentConsist ofConsist ofConsist ofWastehandlingsystemComponentA CHP utilizing biomass might consist of the followingelements:FuelinFertilizing, cultivation, logging, logging track, loopof twigs, trimming, transportChipping, packing, transport, local roadsIntermediate storage, transport regional roadsTransport central and regional roadsBuilding, operation demolition of power plant,technology, efficiency, part-load, size, Lifetime,waste treatment, gas cleaning supply of additives,internal transportTransformation to central net, building, operation,demolition of infrastructure, heat/power lossTransmission to local net, building, operation,demolition of infrastructure (pipes, high-tensionlines heat/power loss (insulation, temperaturelevels (supply, return, ground), twin/single pipes,length)Distribution to end user, building, operation,demolition of infrastructure (pipes, lines,substations) heat/power loss(1)FuelOther(chemetc)ConstructionElecprodConstructionHeatprodDismantlingElectricityDismantlingConstructionCoolingprodDismantlingHeatWastehandlingFigure 4 Architecture information model for a part of theenergy chain from generation including distribution, basedon [25]ColdConstructionDist.netColdHeatDismantlingConstructionSubstationDismantlingHeatCoolPurificationsystemComponentBuilt ofFilterComponentCombustioncamberComponentInternalControlSystemStakeholderElectricityproduction unitComponentHeatproduction unitComponentHeatTransportsystemComponentFigure 5 Segment/selection of part of the architectureinformation model.PEF inInfrastructure,buildings,machinery etc.Additional PEFOperation andDemolitionmaintenanceLossConsist ofPEF outFigure 6 Architecture information model, basis elementHeat StoragesystemComponentSince the Primary Energy Efficiency (PEE) of anenergy chain consist of all of the elements fromextraction to delivery the PEF for a chain can becalculated byEChain EFuelEExtraction EProcessing EStorage ETransport EGeneration ETransformation ETransmission EDistributionWhere E is the primary energy input to thesystemThe Power Bonus MethodIn [13] the power bonus method is applied to calculatethe PEF value for a district heating system with CHP. Edeli fP,del,i) (Eexp,iE f ) (2)P(, P,exp, iwhereE P – Primary energy input to the systemE del,I – Delivered energy, energy carrier i(1)35

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaf P,del,i – Primary energy factor, delivered energy carrier iE exp, – Exported energy, energy carrier if P,expl,i – Primary energy factor, exported energy carrier iPower exported from the CHP plant multiplied with thePEF value for the replaced power shall be subtractedfrom the delivered primary energy to the buildingswhen calculating the PEF-value for the for the districtheating system [7]. The power bonus method isenforced in order to promote CHP, and the subtractionof power produced and delivered outside the systemboundary significantly reduces the PEF value for theenergy chain. This implies that the PEF value for aCHP will be dependent on the power to heat ratio.Behavioural modelThe behavioural model is another information modelthe ―what it does“[10], but also described as ―the way inwhich an organism, organ, or substance acts,especially in response to a stimulus” [23].A behavioural model consists of functions, inputs andoutputs and control operators. This implies that it issupposed to provide information on what is happening,in which order and what kind of iterations areperformed.Energy sourceEstablish problemstatementAssess availableinformationDefine MOEAndEnergyproductionAndTrade-offDevelop genericmodelCollect datacase studyTest, evaluatemethodSimplify methodPublish modelEnergy transportsystemFigure 7 Simplified behaviour information model of themodel developing process.A more detailed partition of the energy chains havebeen applied in the development of the method. Theenergy chain is divided in subsystems as shown in36Figure 4 and Figure 5. This is an iterative process andthe detailing level is the first steps gradually increasing,until the analysis (trade-off) of the different factorsinfluencing the PEF value enables a removal of factorswith an impact of 1% or less.System interface information modelThe system interface model also denoted the contextinformation model shows the systems interface with itssurroundings and the environment. The model providesinformation on the core system and otherinterconnecting systems; this means a description onhow things relate to each other.The context is according to [10] ―the interrelatedconditions in which something exists or occurs‖.Energy Energyprod. RawmaterialEnergysourcecarrierRequirementsRequirementsRequirementsEnergyEnergyproductionEnergyEnergytransportsystemRequirementsFigure 8 A simplified context information model.EnergyEnd userThe system boundary is drawn with a dashed line, andthe system assessed lies within. Since this is asimplified model the relation towards investors, nationalregulators, constructors etc. are not shown. In thissystem energy source/carrier is closely connected toExtraction, Energy Source consists of storage andtransport, Energy production corresponds withGeneration and Energy transport system to Transformation,transmission and distribution in Figure 1.The main issue has been to show the connectionbetween the energy chains from production to end use.Politicians and national regulators might have specificrequirements on each level. The building industry,constructors may likewise have interest on several ofthe levels, but a final listing is not possible to providewithin this paper.Another important question is the definition of thesystem boundary. Precise definitions of the systemboundaries are essential in order to be able to comparedifferent studies. The system boundaries mustdistinguish between what is included and what liesoutside of the task of the LCA, since the method mustalso rely on data collected by other parties and the useof different constraints might influence the quality of themethod.In order to provide information of the whole energychain, all major elements have to be included i.e. theextraction phase is an integrated part of the chain.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTrade-offsIn the ongoing PhD-project of the first author a cut-offrule is set to 1%. That implies that factors with lessthan 1% impact on the final result can be removed.The trade-off considerations are still an ongoingprocess, and it will be presented and documented in alater paper. According to the main hypothesis this is notmeant be developed as an optimization tool, since theintention of the ongoing PhD-project is to develop amethod describing different energy chains.A complete trade-off could preferably [32] be performedby use of computerizes programme like CORE.[http://www.vitechcorp.com/solutions/]. The complexityof the different kind of systems shows the utility valueof more than manual tools, which has been applied.Document current system designThe results of the iterative process are described in thefigures mentioned above. Only selected parts of thechosen design are illustrated in this document due tolimitation in size.The final system design is carried out according toFigure 6 for each element.CONCLUSIONBy performing a system engineering processdescribing different energy chains an outline of themodel have been developed. The method has provento be efficient in structuring the thoughts and willhopefully reduce mistakes in the future development ofthe model.The decomposition process in different subsystems isvaluable, and the generic model will be able to treatdifferent kind of energy systems and chains.The systems engineering process have demonstratedthat;1. It is possible to develop a generic method thatprovides credible data for calculating PEF-valuesand the primary energy efficiency.2. It is e.g. possible to apply the generic method onenergy systems in the Nordic region with CHPplants utilising bio based fuelThe system engineering process provides a newapproach to the design and development of a genericmodel describing PEF-values for energy systems withdifferent kind of energy carriers. The method might beused for more than systems using CHP-technologysince the model development are generic and therebyutilizes different kind of energy carriers.The method can provide detailed data (e.g. efficiency,loss etc) from the different energy chains and mix ofchains. A major challenge is the data collection, someof the parameters lack standardization. The life time ofdifferent equipment varies, the economical lifetime isoften significant lower than the actual exchange ratee.g. pipelines might have a twice times higher - morethan 60 years. The use of yearly average efficiency andappurtenant power-to-heat ratio will often increase thePEF value for the whole system due to the impact ofthe power bonus method.The reliability of the method will be influenced bypossible lack of detailed data, but based of averagedata a reliable comparison of different energy chainsmight be performed.More standardized values for some the differentparameters needs to be developed, like lifetime, heatload curves and extraction of biomass. Someadjustment will be necessary for instance for extractionwhere the transport distances are an importantparameter. The resulting model can form a basis forfuture optimization tools, since only elements withmajor influence on the PEF-values are included.ACKNOWLEDGEMENTThis paper is developed as a part of the PhD-projectPrimary Energy Efficiency (PEE) and the work title forthee PhD-Theses is "System, methods and credibledata for calculating primary energy efficiency in generaland for energy systems in the Nordic region with specialfocus on energy systems applying CHP-technology withbio based fuel in particular".The project is financed by Nordic Energy Researchwith financial support from the industry and includes sixPhD-studies carried out in the respective countries;Estonia, Finland, Sweden, Iceland and Norway. Theprojects objective is to contribute to the effort ofenhancing the primary energy efficiency (PEE) andreducing CO2-emissions in the energy sector.FURTHER INFORMATIONPhD.student Monica Berner, Norwegian University ofScience and Technology (NTNU).Address: Monica.Berner@ntnu.noREFERENCES[1] Proposal for a Directive of the EuropeanParliament and of the Council on the EnergyPerformance of Buildings Recast SEC (2008)2820, SEC (2008) 2821)[2] EN 15603 – Energy performance of buildings –Overall energy use and definition of energy ratings[3] EN 15603:2007 Energy performance of buildings –overall energy use and definition of energy ratings[4] EN 15316-1: 2007 Heating systems in buildings –Method for calculation of system energyrequirements and system efficiencies – Part 1:General37

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia[5] EN 15316-2-1:2007 Heating systems in buildings –method for calculation of system energyrequirements and system efficiencies part 2-1space heating emission systems[6] EN 15316-2-3:2007 Heating systems in buildings –Method for calculation of system energyrequirements and system efficiencies – Part 2-3Space heating distribution systems:[7] EN 15316-4-4:2007 Heating systems in buildings –Method for calculation of system energyrequirements and system efficiencies – Part4-4Heat generation systems, building-integratedcogeneration systems[8] EN 15316-4-5:2007 Heating systems in buildings –Method for calculation of system energyrequirements and system efficiencies – Part 4-5Space heating generation systems, theperformance and quality of district heating andlarge volume systems[9] INCOSE, <strong>International</strong> Council on SystemsEngineering, A Consensus of the INCOSE Fellows,www.incose.org[10] Dahl H J, Information modelling and systems reengineering– an efficient approach to assessingcomplex current Norwegian natural gas transportoperations, Proceedings of the Tenth Annual<strong>International</strong> <strong>Symposium</strong> of the <strong>International</strong>Council on Systems Engineering (INCOCE), July2000[11] Olivier DW, Kelliher TP, Keegan JG, Engineeringcomplex systems with models and objects, ISBN048188-1, McGraw-Hill, 1997[12] Joelsson. A. Primary Energy efficiency and CO2mitigation in Residential buildings, Doctoral Thesis58/2008, Mid Sweden University (Dissertation3.October 2008)[13] Berner M., Ulseth R., The Primary EnergyConcept, The 11th <strong>International</strong> <strong>Symposium</strong> onDistrict Heating and Cooling, August 31 toSeptember 2, 2008, Reykjavik, ICELAND[14] Frischknecht, R, Jungbluth et al, 2007,Őkoinventare für energiesysteme –Grundlagen fürden ökologishen Vergleich von Energiensystemenund den Einbezug von Energiesystemen inŐkobilanzen für die Schweiz , ETH, Zürich 1996[15] CEN/ CLC BT JWG, Energy Management, 2005)[16] Nørstebø V., Application of systems engineeringand information models to optimize operation ofgas export systems, Systems Engineering archive,Volume 11 , Issue 4 (November 2008), p: 329-342, 2008, ISSN:1098-1241[17] Sæther S, Thermal Heat and Power Productionwith models for local and Regional energySystems, ITEV-Report 1999:06, Dr.ing Thesis1999:117, NTNU[18] Sarigiannis D.A., Triacchini G., Meso-scale lifecycleimpact assessment of novel technologypolicies: The case of renewable energy, Journal ofHazardous Materials 78, 2000 p. 145-171[19] Alanne K., Salo A., Saari A., Gustafsson S., Multicriteriaevaluation of residential energy supplysystems, Energy and buildings 39, 2007 p 1218-1226.[20] Eriksson O, Finnveden G, Ekvall T, Bjorklund A,Life cycle assessment of fuels for district heating: Acomparison of waste incineration, biomass- andnatural gas combustion, energy Policy 35, 2007p.1346-1362.[21] Mûnster M., Lund H., Use of waste for heat,electricity and transport – Challenges whenperforming energy system analysis. Energy 34,2009 p. 636-644[22] Lenzen M., Life cycle energy and greenhouse gasemissions of nuclear energy: A review, energyConversion &Management 49, 2008 p.2178-2199[23] Sproles N, Coming to Grips with Measures ofEffectiveness, John Wiley & Sons, Inc Syst Eng.3:50-58, 2000[24] Olivier, Merrian Webster 1981[25] Berner M, Primary Energy Concept and Life CycleAssessment (LCA), Report no: 2009/001, June2010, The Norwegian University of Science andTechnology[26] Act of 14 June 1985 No. 77 the Planning andBuilding Act, The Ministry of the Environment andthe Ministry of Local Government and RegionalDevelopment[27] Directive 2002/91/EC of the European Parliamentand of the Council of 16 December 2002 on theenergy performance of buildings.[28] Directive 2009/28/EC on the promotion of the useof energy from renewable sources and amendingand subsequently repealing Directives 2001/77/ECand 2003/30/EC[29] Council Directive 1999/31/EC of 26 April 1999 onthe landfill of waste[30] Act of 13 March 1981 No.6 Concerning ProtectionAgainst Pollution and Concerning Waste, [ThePollution Control Act][31] Act of 17 June 2005 No. 62 relating to workingenvironment, working hours and employmentprotection, etc. as subsequently amended, last byAct of 23 February 2007 No. 10, (The WorkingEnvironment Act)[32] Purves B, Information Models as a Prerequisite toSoftware Tool Interoperability, Incose Insight, 199838

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaENHANCED DISTRICT HEATING AND COOLING SYSTEMS– REALISATION OF THE LOW-EX CONCEPTStefan Bargel 1 , Clemens Pollerberg 1 , Armin Knels 2 , Li Huang 1 , Dirk Müller 2 and Christian Dötsch 11 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.de2 RWTH Aachen University, E.ON Energy Research Center - EBC,Mathieustr. 6, 52074 Aachen, Germany,Phone: +49 (0) 241-8049-780, Fax: +49 (0) 241-8049-769ABSTRACTSince heating and cooling represent low-exergy energystreams, high efficiencies can be obtained, if theenergy demand is covered by appropriate – meaningalso low-exergy (low-ex) – input energy flows.In order to be able to employ great potentials of lowexergyheat from many different sources, it is importantto develop technologies for the supply and the use ofenergy that allow network temperatures close toambient temperature in return as well as in supplypipes. Two possible technologies are phase changeslurries (PCS) and capillary tube mats (CTM).PCS are discussed as heat transfer fluid, which has anincreased heat capacity compared to water. The use ofPCS in energy supply networks instead of water leadsto an improved energy transport capacity, which resultsin a reduction of the necessary temperature differenceof the transfer fluid. To ensure the transfer of energyfrom the supply network into the building while thetemperature difference between network and building islow, large heat transfer areas are required, which canbe achieved by the use of CTM.This paper discusses opportunities for the realisation ofcold supply networks and low-ex systems and presentsexemplary technologies for their realisation.INTRODUCTIONTemperature levels in district heating and coolingnetworks have long been discussed. During the lastyears a tendency towards low temperature networkscan be observed. From a scientific point of viewanswers to the question for the optimal temperaturelevels can be given using exergy efficiencies as forexample discussed in [1]. The main advantage of thisevaluation parameter is the thermodynamically correctdistinction 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 byappropriate – meaning also low-exergy – input energyflows. For example a heating system based on a39domestic gas boiler used to provide space heatingwastes a huge amount of exergy, since the exergyefficiency of such a system reaches only approximately5%! This result is valid for arbitrary heating systems inthe supply target (room) itself. Therefore it ismandatory to use an integral evaluation approach todecide whether an energy system is efficient or not.With respect to district heating and cooling networks asenergy supply systems two findings are important.First, it can be shown that the network subsystem itselfas depicted in figure 2 reaches optimal exergeticefficiency at quite low temperatures since the heatlosses dominate the pumping electricity effort.Secondly the overall energy supply system efficiencycan be greatly enhanced by utilising low-exergy inputenergy flows such as industrial waste heat.In order to be able to employ great potentials of lowtemperaturewaste heat from many different sources, itis important to develop technologies for the supply andthe use of energy that allow network temperaturesclose to ambient temperature in return as well as insupply pipes.Today, district heating and cooling networks use wateras heat transfer fluid. The heat is transported assensible heat and the transport capacity of thenetworks is determined by the heat capacity of waterand the temperature difference between forward andbackward flow. In cold supply networks as well as inlow temperature heating networks, high volumetric flowrates are necessary to provide the required transportcapacity due to the comparably small temperaturedifference between forward and backward flow. Toovercome these restrictions, a new heat transfer fluidwith an increased heat capacity is under developmentas an alternative to water, phase change slurries. PCSare mixtures of dispersed phase change material and acontinuous liquid phase, which can be used as heattransfer fluid in district heating and cooling networks.PCS possess an increased heat capacity due toadditional latent heat of fusion occurring during thephase transition of the phase change material. The useof such a dispersion in energy supply networks leads toan improved energy transport capacity, which in turn

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaresults in a reduced temperature difference orvolumetric flow rate of the transfer fluid needed totransfer a given amount of heat. The application ofPCS for thermal energy transportation is investigatedand discussed for example in [2].An improved transport capacity is one important pointfor the realisation of the low-ex concept; anotherimportant point is the use of the energy on theconsumer side. To ensure the transfer of the energyfrom the supply network into the building while thetemperature difference between network and building islow, large heat transfer areas are necessary. Theseheat transfer areas can be realised by using capillarytube mats integrated into the walls, the floors and theceilings of buildings.The E.ON Energy Research Center of the RWTHAachen and Fraunhofer UMSICHT investigated thepossibilities to realise district heating and coolingnetworks as low-ex systems. These investigationsinclude system modelling and analysing as well as thedevelopment and testing of technologies.1. Exergy as evaluation parameter1.1. The low-ex conceptExergy can be understood as the theoretical maximumof mechanical work that can be utilised by equilibratingan energy flow whilst considering its ambientconditions.Consequently this property distinguishes betweentypes of energy that can theoretically be transformedinto each other without any losses - like mechanicalwork, electrical energy or combustible fuels - andthermal energy. The possibility to transform the latterinto any other type of energy is limited by the secondlaw of thermodynamics and therefore inevitablyconnected to losses.This distinction is of importance if one analyses asystem where both types of energy flows (thermal andnon-thermal) occur and have to be related to eachother – as is the case with heating and coolingapplications.The ultimate goal of heating and cooling is to keep atarget (room) at a constant temperature of e.g. 20 °C.As the outdoor temperature varies additional heat hasto be supplied or excess heat has to be disposed of tofulfil this task.Theoretically the supplied energy flow could betransferred to the room using infinitesimal smalltemperature differences between supply flow andtarget 2 . The real temperature differences occur due toheating and cooling techniques applied which aremainly limited by finite heat transfer areas. Keeping inmind that the annual average outdoor temperature forthe heating period e.g. in Germany is about 3.5 °C, itbecomes apparent that the exergy to energy ratio ofthe target energy flows - passing the building envelopeat 20 °C – is very small (approx. 7%). On the otherhand, exergy to energy ratios of conventional inputenergy flows are usually 100% as combustible fuels orelectricity is used.The low-ex concept acknowledges the fact thatdemand flows are ‗low-ex‘ - meaning that they possesssmall exergy to energy ratios. Hence the conceptdemands to supply energy on appropriate ‗exergylevels‘, instead of wasting exergy by transforming highexergy flows into low exergy ones. In doing so, thisapproach opens up a totally new dimension ofenhancement potential since it deals with the qualityaspect of the energy flows under consideration.Therefore, within the low-ex concept energy is nolonger one-dimensional. In addition to decreasing theamount of energy demanded by the consumers –leading to insulation efforts – a kind of exergeticsuitability has to be taken into account and the task athand becomes a two-dimensional problem (cf. fig. 1).Consequently, the concept aims at maximizing theexergy efficiency of an energy supply system, whichallows to utilize potentials in both dimensions, quantityAND quality.The exergy efficiency can be defined as:exergy ( demand )ex(1)exergy ( supply)In applying this efficiency the demand flows andparticularly the supply flows have to be definedcarefully (cf. chapter 1.2.).exergetic qualityexergetic suitabilityenhancementlow-exconceptinsulationenergy demand (quantity)Figure 1. Energy as two-dimensional concept. Orange (lightgrey): conventional system, green (dark grey): optimalsystem2 This statement is analogously true for cooling applications.40

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaconversionlossesnetw orkheatlossesbuildingheatlossesprimaryenergyheatgeneratorheatflowpumpingelectricitydistrictheatingnetw orkheatflowheatingsurfaceT = 20°Cheatflowpow er plantconversion lossesprimary energyFigure 2. Evaluation boundaries of an energy supply system. The blue (outer) dashed line marks the complete system; theblack (inner) dashed line marks the network subsystem1.2. Integrated system evaluationWhen evaluating a system it is important to specify theevaluation boundaries (cf. fig. 2). It has to be pointedout that an integrated system evaluation is mandatorysince otherwise results are ambiguous and misleading.This can be demonstrated by assuming e.g. evaluationof the building subsystem only. If two systems arecompared, one consisting of a target room equippedwith space heating and the other one with a targetroom equipped with conventional heating, one couldarrive at the conclusion, that the system utilizing spaceheating is more efficient. However, assuming bothsystems are also equipped with an identicalcondensing gas boiler providing the heat, an evaluationcomprising the total system (consisting of heatgeneration and heat transfer to the target) would arriveat a totally different conclusion. In this case, bothsystems possess the same exergy efficiency, which isapproximately 5% for the outlined case. This isbecause a potentially more efficient heating system isnot put to use as the same input and supply flowsoccur in both cases.1.3 Efficiency enhancement potentialsThe complete energy supply system can be dividedinto three subsystems – generation, distribution andbuilding (representing the consumption). Thesesubsystems possess different potentials to enhanceoverall system efficiency.Currently heating demands are met by burning highexergyfuels, great enhancement potentials areavailable within the generation subsystem. Firstly, fuelsshould not be used to directly satisfy thermal demandsat all since this embodies pure exergy destruction.41Instead thermal input flows as industrial waste heat3 orgeothermal energy should be applied. On the otherhand, if combustible fuels are used to meet thermaldemands, at least Combined Heat andPower generation (CHP) with a maximum electricaldegree of efficiency should be utilized. This allowstransforming part of the high-exergy fuel into highexergyelectric current. Heat is produced as ‗wasteproduct‘ of this conversion.Optimization potentials within the distributionsubsystem are basically indirect. At first glance, thedistribution system has no influence at all since thenetwork acts as connection between heat generationand heat consumption. Consequently, no thermal flowsexist that pass the overall system evaluationboundaries. However, two aspects remain and need tobe accounted for. One is heat losses occurringthroughout the network that have to be compensatedby additional heat generation. The other is pumping tomaintain the heat transfer medium circulation, which ismet by an unalterable high-exergy input (electricity).The main problem is that concepts, which lead todecreasing heat losses cause increasing pumpingefforts and vice versa. Nevertheless, heat losses arethe exergetically dominant influence, therefore thefocus should be to confine these losses. Heat lossesdepend on the driving temperature difference betweenmedium and surrounding ground and on surface area.Minimization of the losses can most easily be achievedby reducing the network temperatures since pipedimensions are affixed due to demands so that surfaceareas are not a modifiable parameter. This approach iseven more rewarding since it allows employing low3 Industrial waste heat in this sense is heat that can no more beput to any use within the industrial production process.

temperature thermal input flows and thereforerepresents the prerequisite for an efficient generationsubsystem.The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe last subsystem possessing enhancement potentialis the consumer. Since the target temperaturedetermines the exergetic quality of the thermaldemand, therein lays no significant optimizationpotential. However, as decreasing the amount ofenergy that has to be supplied is also part of the low-exconcept insulation can help to improve the system. Onthe other hand, benefits similar to those alreadydiscussed for the distribution subsystem can beidentified for the consumer system as well. By choosingappropriate heating and cooling technologies, as e.g.investigated in [3], the exergy destruction during heattransfer to the room air can be minimized. This isachieved by applying low-temperature heating andhigh-temperature cooling devices. Inlet and outlettemperatures of the heating/cooling devicesimultaneously define constraints for the distributionnetwork subsystem, which in turn set constraints for thegeneration. In the end supply temperatures close to thetarget temperature form the basis for a ‗low-ex ready‘consumer. Without this step an exergetically optimalenergy supply system would be greatly hindered.2. Applicable technologies for the realisation2.1. Phase Change SlurriesThe most used heat transfer fluid in district heating andcooling networks is water. In supply networks, the heatis transferred as sensible heat with a temperaturedifference between forward and backward flow. Theheat transfer capacity of a network is determined by thetemperature difference, the mass flow and the heatcapacity of the heat transfer fluid. The temperaturedifference and the temperature level of the network arelimited by technical restrictions and determine thenecessary mass flow of the heat transfer fluid. Toovercome these restrictions, fluids with higher heatcapacities than the heat capacity of water are underdevelopment. An alternative to water could be PCS.PCS are mixtures of dispersed phase change materialand a continuous liquid phase, which possess anincreased heat capacity due to the additional latentheat of fusion occurring during the phase transition ofthe phase change material. The PCS remainspumpable even when the phase change material isfrozen. Thus, the PCS can be used as heat transferfluid in supply networks. A promising PCS for heat orcold supply networks is paraffin/water dispersion.Figure 3 is a photograph of a paraffin/water dispersion.Paraffin is the phase change material, which can bechosen according to the desired temperature of thephase transition, and water is the continuous phase ofthe dispersion. In [4] paraffin/water dispersions areinvestigated and their properties presented42Figure3. Photograph of a paraffin/water dispersionThe increase of the heat transport capacity of a supplynetwork using a PCS instead of water can be describedby a thermal capacity enhancement factor (TCEF),which is calculated according to equation (2).TCEF PCSwh c T wcf, PCM p,PCM1p,w T(2) c Twp,wThe TCEF is a function of the densities of the PCS ρ PCSand water ρ w , the mass concentration of the PCM w,the specific heat capacity of PCM c p,PCM and water c p,wthe heat of fusion of the PCM Δh f,PCM and thetemperature change ΔT of the fluids. The TCEF iscalculated and plotted in the diagram figure 4 fortemperature differences ΔT between the forward andbackward flow of 10 and 15 K as function of the massconcentration w.TCEF [-]3.532.521.51delta T = 10 Kdelta T = 15 K0 0.2 0.4 0.6 0.8 1w [-]Figure 4. TCEF – PCS compared to water for temperaturedifferences 10 and 15 K, diagram calculated with theproperties of water and RT-42 of the companyRubitherm [5]Using PCS with a mass concentration w of 0.4 wouldincrease the heat transport capacity of the supplynetwork to 1.5 times of the value compared to water, if

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniathe temperature difference of the supply network is15 K, and even 2 times, if the temperature difference isonly 10 K. Furthermore, the diagram shows that withrising temperature difference the gradient of the TCEFis lower, which means that the advantage of the PCScompared to water disappear at higher temperaturedifferences. At the point where the gradient of theTCEF is 0, the water system and the PCS system havethe same transport capacity. At that point, the massconcentration of paraffin w has no influence on theTCEF.The use of PCS in energy systems leads to animproved energy transport capacity, which results in areduction of the necessary temperature difference orvolumetric flow rate of the transfer fluid needed totransfer a given amount of heat.Another technical issue of PCS systems is theincreased pressure drop in the pipes due to the higherviscosity of the PCS. A calculation methods andmeasurement data can be found in [6, 7 and 8]. Theviscosity of PCS is related to several influencequantities and can cause an incensement of thepressure drop up 100%. PCS are non-newtonian fluids.2.2. Capillary Tube MatsThe most often used heat exchanger type in heatingsystems is a convective radiator, which is installed inrooms close to the window. The size of a radiatorshould be small, so that also the heat exchangesurface is small and the heating system must beoperated on a high temperature level to ensure theheat transfer from the heating system into the room. Analternative to convective radiators are floor heatingsystems. Floor heating systems consist of a capillarytube mat, which is installed in the upper layer of thefloor. Because of the bigger heat exchange surfacecompared to the convective radiator, the temperaturelevel of the heating system is lower. A new approach torealise heating and cooling of buildings is via CTM,which are integrated in the floors of the building, as wellas in the walls and ceilings. This system offers a bigheat exchange area and allows the heating and thepassive cooling of the building. Due to the increasedheat exchanger area, a low temperature differencebetween the heating system and room is possible. Forthe further discussion, the following simple model isused to describe the heat release of the heating systemin the building. The heating release system isevaluated by the number of transfer units (NTU). Theheat capacity provided by the heating network Q iscalculated by equation (3) with the inlet and outlettemperature T in/out of the supply network, the mass flowm and heat capacity c p of the heat transfer fluid. mcp TinTout(3)QIn view of the heat release in the room, the heatcapacity Q can also be described by equation (4) andis related to the heat transfer coefficient U, the heatexchange area A and the temperature differencebetween the mean temperature of the heat release T mas well as the room temperature T r .Q U AT m T r (4)The mean temperature of the heat release T m iscalculated by equation (5).TT Tin outm (5)TinlnToutBased on the equations (3) to (5), it is possible tocalculate the NTU, which characterizes the heatrelease in the room, according to equation (6), which isonly a function of the inlet and outlet temperature T in/outof the heat supply, the mean temperature T m of theheat release and the room temperature T r .NTUUmAcmrin out (6)pTT T TThe NTU values have been calculated for a convectiveradiator system and a CTM system. The assumedtemperatures for the calculation and the results aregiven in table I.Table I. NTU for both heat release systems:conventional radiator and CTMparameter convective radiator CTM systemT in [°C] 80 37T out [°C] 60 31T r [°C] 20 20NTU [-] 0.4 0.43The NTU value of the CTM system is 0.43 and as highas the NTU value of the convective radiator. Thismeans that both systems have the same heat releasecapacity, although the inlet temperature T in of the CTMsystem is lower and the temperature differencebetween inlet T in and outlet T out of the CTM system issmaller.CONCLUSIONFrom the point of view of the low-ex concept the majortask en route to an exergetically efficient energy supplysystem is the replacement of the combustible fuelboiler by utilization of low temperature thermal input43

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaflows such as industrial waste heat or geothermalenergy. To achieve this goal it is necessary todecrease the medium temperatures within thedistribution networks first. A prerequisite is a low-exready consumer that allows meeting the thermaldemands applying low temperatures.A possible realisation employs CTM in the heatingsystem of the building that allows applying inlet andoutlet temperatures of approximately 37 °C and 31 °C,respectively. Within the district heating or coolingnetwork, the utilization of PCS instead of pure waterenables the application of small temperaturedifferences between forward and backward flow whileretaining the pipe dimensions. Since the backward flowtemperature mainly depends on the outlet temperatureof the consumer system, small temperature differenceswithin the network automatically lead to low forwardflow temperatures. Consequently, the exploitation oflow temperature heat sources as input flows for theenergy supply system is rendered possible.Moreover, the decreasing temperatures in both forwardand backward flows of the network reduce thetransportation heat losses. This leads in the end to areduction of energy input (quantitative aspect of thelow-ex concept) into the supply system.The only drawback suffered occurs in terms of anincreased pumping effort caused by a higher viscosityof the PCS in comparison with water. But, since heatlosses are the predominant factor over circulationpump energy, an overall benefit should beaccomplishable.Summarizing it should be pointed out that applyingtechnologies such as CTM in the building heating orcooling system and PCS as alternate heat transfermedium for the distribution networks the low-exconcept can be realised, thus greatly enhancing theefficiency of energy supply systems.ACKNOWLEDGEMENTThis study was supported by the Project ManagementJuelich (PTJ) and the Federal Ministry of Economicsand Technology (BMWi) under 0327471A.Comments of a highly constructive nature werereceived from Daniel Wolf, Jorrit Wronski and AstridPohlig.REFERENCES[1] C. Kemal et al., Evaluation of energy and exergylosses in district heating network, Applied ThermalEngineering, 24 (2004), pp. 1009-1017.[2] H. Inaba, New challenge in advanced thermalenergy transportation using functionally thermalfluids, <strong>International</strong> Journal of Thermal Sciences,39 (2000), pp. 991-1003.[3] M. Ala-Juusela et al., LowExergy Systems forHeating and Cooling of Buildings, final report of theIEA ECBCS Annex 37.[4] L. Huang et al., Evaluation of paraffin/wateremulsion as a phase change slurry for coolingapplications, Energy, 34 (2009), pp. 1145-1155.[5] Rubitherm RT-42, datasheet 08/20/2009,http://www.rubitherm.de, Rubitherm TechnologiesGmbH, Berlin (2010).[6] Yinping Zhang, et al., Experimental research onlaminar flow performance of phase changeemulsion, Applied Thermal Engineering, 26 (2006),pp. 1238-1245.[7] A., B. Metzner et al., Flow of Non-Newtonian Fluids– Correlation of the Laminar, Transition, andTurbulent-flow Regions, American Institute ofChemical Engineers Journal, Vol. 1, No. 4 (1955),pp. 434-440.[8] R. Rautenbach, Kennzeichnung nicht-NewtonscherFlüssigkeiten durch zwei Stoffkonstanten, Chemie-Ingenieur-Technik, 36 No. 3 (1964), pp. 277-282.44

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAPPLICATION OF EXERGOECONOMICS TO THE OPTIMIZATION OF BUILDINGHEATING SYSTEMS CONNECTED TO DISTRICT HEATING NETWORKSC. W. Snoek and S. C. KluitersRenewables and Integrated Energy Systems, CanmetENERGY, Natural Resources Canada,1 Haanel Dr, Ottawa, K1A 1M1, CanadaABSTRACTThe concept of energy efficiency, defined as usefulenergy output as fraction of required energy input, hasbeen used for years in technical systems assessments.In addition to energy efficiency, there are benefits tousing exergy efficiency to assess system performance.Whether systems will be installed or not is ultimatelydetermined by their economic performance. Thisperformance is usually determined by comparing initialinvestment cost and operational cost with revenuesthroughout a system‘s lifetime in terms of payback timeor net present value.This paper describes a novel methodology that usesthe concept of exergy and the thermoeconomic factor,a ratio that compares investment-related cost andexergy destruction cost, for the economic optimizationof a community energy system. It compares the cost ofexergy and the required capital and operational costsincluding carbon taxes to accommodate this low qualityenergy. In doing so it enables a quick way to properlyassess the value of a system‘s ability to use low exergyenergy inputs. The method is compared to a moretraditional economic analysis.INTRODUCTIONIn the last few years, we have become painfully awareof the effects of climate change. The burning of fossilfuels and the resulting emissions are thought to be amajor contributor to the apparent increase of adverseweather events. While people need energy for comfort,in some cases there may be a choice in the source andnature of that energy. In addition to climate change,there is also a concern about the rapid depletion of themore valuable of fossil fuels, natural gas and oil. Forthese reasons it makes much sense to re-evaluate thesources of the energy we use and the effect of usingthem has on the environment.To lower energy requirements, energy efficiency hasbeen practiced for many years. In terms of comfortheating in houses, most of the effort has gone intoimproving building insulation, better windows, buildingorientation with respect to the sun, shading from solarenergy etc. In terms of energy conversion equipment,improving the efficiency often meets ‗natural‘ limits,such as those expressed by Carnot‘s Law.45Often, omitted from consideration is the ―quality‖ of theenergy that is needed to provide comfort to theoccupants of a building. While the heatingrequirements of a building can be determined (in GJ orTJ), the nature or origin of this energy is not addressedin energy efficiency calculations. The total amount ofJoules can be provided by oil, natural gas, electricity orlow temperature ‗waste‘ heat. While the first threeenergy sources are considered high quality, and can beused to generate very high temperatures (over1000 °C), run equipment such as computers, radio andTV transmitters and receivers, ‗waste heat‘ is of lowquality and has no other use. Comfort heating does notrequire high temperatures and therefore using highquality fuel for low quality applications is consideredwasteful.Energy quality is often expressed as ‗exergy‘. Exergy isdefined as the maximum useful work possible during aprocess that brings the system into equilibrium with aheat reservoir. To illustrate the concept of exergy onecan compare two different forms of the same amount ofenergy: 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 ambienttemperature of 20 °C.Obviously, the energy contained in the battery isconsidered more useful and therefore has the higherquality or exergy.The ratio of Exergy (E) to Energy (Q) can be expressedas:EQTambient 1(1)Tsup plywhere T is given in K.Equation 1 shows that when the supply temperature ofan energy source is high, the exergy converges to theenergy value. Electricity and mechanical work are(nearly) perfectly convertible and the exergy content istherefore equal to the energy content. Conversely,when the supply temperature is closer to theenvironmental temperature, the value of the exergybecomes (much) smaller than that of the energy.Wall [1], in his paper on ―Exergy and Morals‖ quotesAlfven who claimed that energy accounting based onenergy only is like a bank teller counting by the amount

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaof coins and bills, but neglecting their value. In anethical society the value, worth and quality of differentenergy supplies should, as a minimum, be matched tothe requirements of the different energy applications.Methods to design low exergy buildings are availabletoday. For instance, Schmidt [2] developed a methodand pre-design tool for low exergy buildings in which hecompared different heating systems, such as boilers,condensing boilers, electric heating, GSHP and lowtemperature under-floor heating. However, this methoddoes not directly address the effect of system heattransfer surface area on the overall economics.Also, there is an additional benefit realizing that abuilding that can accommodate low exergy streams isready for future hook-up to other, perhaps renewableenergy sources: GSHP, solar, waste heat fromindustry, energy from thermal storage to name a few.This is a distinct advantage when the move to asustainable society gains momentum, and the conceptof low-temperature heating should be incorporated inbuilding codes.This paper considers the cost of using the low qualitypart of the energy source and the (increased) capitalcost and operating cost that are required to‗accommodate‘ low quality energy. A methodology hasbeen developed to determine the optimal cost ofoperation, based on the capital cost, operational costand the cost of the exergy.This type of analysis is considered part of the field ofthermoeconomics, more in particular exergoeconomics.Wikipedia defines thermoeconomics in a verytheoretical way as a school of economics that applieslaws of thermodynamics to economy. Valero et al. [3]operationalize this definition by describing two aims ofthermoeconomics, (1) optimization to minimize cost ofa system or component, and (2) cost allocation ofindividual outputs of a plant producing a number ofoutputs.Valero and coworkers [3] date this research field backas far as 1932, when Keenan apportioned cost of heatand work taking into account irreversibility andthermodynamic efficiency instead of enthalpy only.However, they go on to say that Gaggioli, and Tribusand Evans in the early 1960s started off realdevelopment in thermoeconomics. Ever since, thesefields have received tremendous attention. Valero andcoworkers identify that an important problem in thisbody of research is the variety of methodologies usedwith accompanying nomenclature. Between them andTsatsaronis [4] they already name a fair amount ofmethods. In doing so, Tsatsaronis introduces theexergoeconomic factor f, as a fraction that comparestwo sources contributing to cost increases, investmentrelatedcost and exergy destruction cost. Thisexergoeconomic factor is also found in other sources,such as Temir & Bilge [5].It is beyond the scope of this paper to provide acomprehensive literature overview of thermoeconomicpublications or even of the methods used in thesepublications. The aim of this paper is to apply one ofthese methods, using the above-mentionedexergoeconomic factor to optimize building heatingsystems connected to a district heating system. To thebest of the authors‘ knowledge, so far this method hasonly been applied to optimize individual components.This work ties in with research into advanced lowtemperaturedistrict energy systems currently carriedout at the CanmetENERGY laboratories of NaturalResources Canada in Ottawa, Canada.The system considered consists of buildings with theirheating system (radiators and cross-flow heatexchangers are considered), the energy centre withboilers and pumps and the pipeline to move the energyin the form of hot water to the community. Thedevelopment of the methodology was the main objectof the study, not the optimization itself.While the development of the optimization was relatedto economics, in other words, the least costly option, itshould be noted that the concept of ‗exergy‘ opens upthe notion of ―morals‖ and ―ethics‖. For newdevelopments, the costs of resource depletion andenvironmental destruction should be considered aswell. Just because a certain system is economic, it isnot necessarily the best moral or ethical choice. Justbecause a certain system does not cause localproblems, that does not mean that (environmental orother) problems caused by this system elsewhere canbe ignored.Traditional OptimizationsSystem optimization is often done by optimizingsystems separately, and not by considering the overallefficiency of integrated systems. Often, an integratedapproach leads to optimal solutions, as in electricitygeneration using a back pressure steam turbine.Accepting a lower efficiency of the turbine may lead tothe residual energy in the condenser being useful inother applications, whereas in the separately optimizedversion this thermal energy would be useless. In thelatter case, the turbine back pressure is kept as low aspossible, to extract the maximum electrical power. Thismakes the condensate of too low a temperature to beuseful in other applications. Optimizing integratedsystems as a whole avoids this problem.Exergoeconomic OptimizationIn an exergoeconomic optimization, the concept ofexergy is used to determine the best and mosteconomic solution to an energy conversion process or46

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniasystem. While the total quantity of transferred energyremains the same, the exergy that delivers this energymay vary. Analyzing the required exergy with respect tothe energy transfer equipment will result in an optimumeconomic solution allowing for integration of the systemwith other systems.In this exergoeconomic optimization, the system istreated as an integrated whole together with othersystems. While the (comfort) energy supplied remainsthe same in any given scenario, the exergy required forthis scenario varies and the cost implications of thisvariation are included in the analysis. Therefore, in thisanalysis the consumer of energy does not pay for theenergy, but for the exergy, the real value of the energysupplied.SYSTEM DESCRIPTIONThe system considered here to develop themethodology is a district heating system supplying hotwater for space heating to a 1000 home community inthe Ottawa area in Canada. It was modelled using theRETScreen clean energy project analysis software tool[6] and in-house spreadsheet based models. The hotwater is transported from an energy centre locatedcentrally in the community to the 1000 detachedhomes. Inside the buildings, radiators or cross-flowheat exchangers (water-to-air fan coils) are employedto provide space heating.Energy SupplyThe building temperature set point is kept constant at20°C. The hot water supply temperature is determinedby the outdoor temperature. If the outdoor temperatureis above 5 °C, the supply temperature is 70 °C. Whenthe outdoor temperature drops below -15 °C, thesupply temperature equals 90 °C. Between 5 °C and-15 °C, the supply increases linearly from 70 °C to90 °C. This is a common supply temperature profileused in many European district heating systems. Itprevents excessive flows in the pipes at high loads andpermits smaller heat transfer surfaces in the buildingsdue to the higher temperature difference betweenwater and building air. When the heat transfer surfacearea was varied to reach an optimum solution thesupply temperature was adjusted by a constant valueover the entire load range. The water returntemperature was set at 30 °C in all design calculations,but varied throughout the year according to the offdesigncharacteristics of the heating equipment used.The load of the buildings is related to the outdoortemperature. The annual heat consumption was set at100 GJ per house, a typical value for detached homesin this area. The instantaneous load throughout theyear is simply calculated as a linear relationshipbetween zero and the maximum capacity, when theoutdoor temperature varies between 20 °C and -28 °C,47the design temperature for Ottawa. While this is anover-simplification of reality, it neither hinders thedevelopment of the methodology nor introducesserious errors of consequence.Energy TransmissionThe pipe diameters were estimated using theRETScreen software tool. This means for diametersunder 400 mm the pressure drop is kept below 200 Paper meter of pipe and for larger diameters flow velocityis maximized at 3 m/s [6]. As RETScreen has a limit of13 sections for district heating systems, the 1000homes were assumed to be located along twelve 80-home streets. The final 40 homes were located in aseparate street.The energy transfer fluid is water. The pipes arepreinsulated steel or cross-linked polyethylene (PEX)pipes. The first iteration of the methodology accountedfor heat losses from the pipes. Since it was found thatthis heat loss was a negligible fraction of thetransmitted energy, it was omitted in subsequentversions. For a thorough analysis, it is recommended toinclude heat losses, especially if the piping system isextensive and the supply temperatures reach highlevels.A pressure drop analysis was used to determine therequired pump energy. The electric motor driving thepump was estimated to have 90% efficiency while thepump was assigned an efficiency of 85%.End-UseThe energy supplied to the pipeline was used to keepthe building temperatures at set point. Therefore,regardless of the (size of) building heating systemused, the same energy was used to keep the buildingswarm. However, the exergy used was dependent of thesystem in place and of its size. The larger the size, thelower the required water temperature and hence lessexergy was required to achieve the same end result.Two different technologies were used to model thetransfer of energy into the building space: cross-flowheat exchangers, and radiators. Simulations were donefor both technologies separately, and the technologieswere never mixed. This was done to simplify theanalysis. In reality, mixed systems will occur andshould be analysed as such. While this will increasethe level of modelling complexity, it is not difficult to do.DESCRIPTION OF MODELINGClimateThe local climate has a significant effect on the designof a building heating system. A maritime climate mayhave many degree days but not show the variability indemand that a building in a continental climate with anequal amount of degree days experiences. Even if both

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniabuildings use the same amount of energy per year, thedemand load in the building with the continental climatemay be far greater. Therefore, the climate plays animportant role in the design of a heating system andshould, therefore, be considered in this analysis.Supply TemperatureTo include the effect of the variability of the energydemand with time-of-day and the seasons, thestatistical average hourly temperatures for the city ofOttawa were used. These temperature values are realvalues, with realistic variability (high and lowtemperatures), using time-periods from different yearsto provide for a correct average. In total, 8760 hourlyvalues of temperature were used in the spreadsheet,as shown in Figure 1.Temperature (C)4030201000 1000 2000 3000 4000 5000 6000 7000 8000 9000-10-20-30-40Hour (-)Fig.1. Average hourly temperatures in OttawaSolar RadiationTo simplify the spreadsheet calculations, the effects ofsolar radiation, plug loads and occupancy gains wereneglected. When performing the optimization, theseeffects remain constant and so have little effect on thefinal outcome. When using this method for design,these contributors to the exergy balance should beconsidered.temperature. The pumping energy required wasincluded in the modeling. Since electrical energy isequivalent to exergy, the pumping energy calculatedfrom the pressure drop calculations (includingefficiencies), was numerically counted as exergy.Normally, during periods of no-load, the pumps keepoperating to keep a supply of design temperature waterclose to the load. This is done with a thermostaticallyoperated by-pass valve. Since this valve represents aconstant effect which does not affect the optimization, itwas not modelled for simplicity.Design of Cross-Flow Heat ExchangerThe design of the cross-flow (or fan-coil) heatexchanger was based on the assumption that theoverall 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 m2were required.Design of Radiator Heating SystemThe design of the radiators was done in a very simplemanner. It is acknowledged that better methods exist,but the development of the methodology did not sufferbecause of this simplification. For any optimization,actual modelling of the equipment should take place.To determine the heat transfer from the panels, thegeneral radiation equation4 4T panelT roomQ (2)was used with the ‗average‘ panel temperature. Thesurface emissivity ‗ε‘ was estimated at 0.9. Convectionfrom the surfaces was not separately considered.To meet the design load, 42,271 m2 of radiativesurface was required to meet the design load.Determination of Instantaneous LoadThe maximum thermal load of the community for spaceheating was determined using the average hourlytemperatures and the annual heat consumption perhouse of 100 GJ. It turned out to be 10,640 kW. Whenthe outdoor temperature reaches -28 °C, the Ottawadesign temperature, the community requires themaximum thermal load. At the ambient temperature of20 °C, the load is nil. The modelling is set up so thatbetween these ambient temperatures, the load varieslinearly. For instance, at -4 °C, the load equals 5.32MW.Pumping Power and ExergyTo meet the load, the water had to be pumped from thesupply source to the load. The amount of waterpumped varied with the load and the supplyDetermination of Exergy UseAs indicated in Section 1, the ratio of Energy to Exergycan be expressed as:EQTambient 1(1)Tsup plyWhere E is exergy, Q is energy and T is thetemperature given in K.For this study, knowing the energy supplied, Equation(1) was used to calculate the supplied exergy for eachhour interval. For each of the heating systems usedand each of their variations is size, the amount ofenergy supplied to the heated space remained thesame. However, due to the supply temperature48

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniarequirements, and the different flow requirements, theexergy used by each system was unique.Capital and Exergy Cost AssessmentsAll cost numbers reported in this paper are in 2009Canadian Dollars. An in-house costing tool was used toestimate cost for the district heating energy centre,containing the pumps and boilers, and the burieddistribution piping. For the two heating systemsconsidered, the costs were assigned as shown inTable I. For water-to-air fan coils (cross-flow heatexchangers) an installed cost of $250/m2 wasconsidered representative, and for radiators $200/m2was selected as a typical value.Table I. – Cost for heating technologiesCross-flow heatexchangerRadiative system$250/m 2 of heat transfer surface$200/m 2 of exposed panelFuture cash flows were discounted at a rate of 8% andsystem lifetime was set at 40 years. Annual operatingand maintenance (O&M) cost other than cost for heatand electricity were set at a fixed fraction of 1% of totalinvestment cost.To compare traditional optimization with exergoeconomicoptimization three types of analyses wereperformed. The ‗classical analysis‘ applies thetraditional optimization where heat is valued based onenergy content, at a rate of $5/GJ, which is consideredrepresentative for heat from natural gas combustion.Electricity cost has been set at $17/GJ (just over$60/MWh).In the exergoeconomic analysis heat and electricity arepriced based on the exergy content. The exergy chargewas determined at $30/GJ for thermal energy, basedon the above mentioned $5/GJ for heat, assuming a 1to 6 ratio of exergy to energy content (applies to atemperature around 80 °C). The electrical energy toexergy ratio was taken as one, resulting in an exergycharge of $17/GJ for electricity. At first glace it mayseem erroneous to charge more for exergy from thethermal source than that for the electricity for the pump,but it must be remembered that the (thermal) exergy isa fraction of the thermal energy.The third type of analysis is a classical analysiscorrected for the difference in value of low- and hightemperatureheat, by assuming energy under 60°C isavailable free of charge (as waste heat from a nearbyprocess). For energy over 60 °C the charge is still$5/GJ.To assess the influence of carbon taxes, two sets ofresults are presented. One assumes no carbon taxesare in place and the other assumes a carbon tax of $3049per ton CO2eq. Carbon intensity factors of 0.050 tonCO2eq/GJ were used for natural gas and 0.054 tonCO2eq/GJ for electricity (taken from RETScreen [6] asrepresentative for Canada). This results in a $6.5/GJenergy charge for heat, a $38.9/GJ exergy charge forheat and an $18.6/GJ energy (or exergy) charge forelectricity.Thermoeconomic FactorThe exergoeconomic or thermoeconomic factor ―f‖compares two sources contributing to cost, investmentrelatedcost and exergy destruction cost. It is definedhere as the ratio of Capital Cost Rate (CCR, whichincludes O&M cost, but excludes heat and electricitycost) and the sum of Exergy Destruction Cost Rate(EDCR) and CCR. The CCR equals the cost per unittime for the installation, depreciation, maintenance, etc,while EDCR is the cost of exergy.fCCREDCR CCR (4)Since CCR and EDCR have the dimensions of $/time,―f‖ is dimensionless.A high value for ―f‖ indicates that the capital andmaintenance costs are dominant. Also, a high f – valueindicates good use of the exergy in the fuel. On theother hand, a low value for ―f‖ indicates an inefficientuse of fuel resources. For each heating systemvariation, the average annual thermoeconomic factorwas calculated.MODELLING RESULTSBase case designTable II shows the main information for the base casedesigns for both the radiator and cross-flow heatexchanger systems. As expected, the distribution pipediameters, required pump capacity, annual space heatconsumption and annual heat cost are the same forboth systems.As the water return temperatures throughout the yearare generally lower for the radiator system, the requiredwater flows and consequently the annual electricityconsumption are lower for the radiator system. As bothsystems have a design supply temperature of 90 °C(and thus also the same off-design supplytemperatures throughout the year), the annual exergyconsumption is the same for both. The lower returntemperatures for the radiator system also show in thehigher fraction of energy provided under 60 °C. Interms of cost, the radiators are clearly more expensiveresulting in higher annual investment and O&M cost,which is not offset by the somewhat lower electricitycost. Overall the more capital intensive radiator systemhas a higher f-factor than the cross-flow heat

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaexchanger system. In a comparison between the twosystems, the cross-flow heat exchanger system worksout cheaper using all three types of analysis due to thelarge difference in investment cost.Table II. – Main information base case designs withoutCO 2 tax.RadiatorCross-flowheatexchangerSurface area (m 2 ) 42,271 17,152Distribution pipediameters (mm) DN80/DN65 DN80/DN65Required pump capacity(kW) 32.2 32.2Annual electricityconsumption (GJ) 115.2 147.6Annual exergyconsumption (GJ) 16,810 16,810Annual space heatconsumption (GJ) 100,000 100,000Fraction of energy < 60 °C 68.4% 64.4%Installed cost heaters $8,454,298 $4,288,066Investment cost districtheating system $11,556,386 $11,556,386Annual O&M cost $200,107 $158,445Annual charge investmentand O&M cost $1,878,206 $1,487,163Annual heat (energy) cost $500,000 $500,000Annual heat (exergy) cost $504,193 $504,193Annual electricity cost $1,958 $2,509Total annual cost classicalanalysis $2,380,164 $1,989,672Total annual costexergoeconomic analysis $2,384,357 $1,993,865Total annual cost heatunder 60 °C free analysis $2,038,164 $1,667,672f-factor 0.814 0.782Alternative designs – radiator systemFor both the radiator and the cross-flow heatexchanger system alternative designs with increasedand decreased surface areas were costed. The districtheat supply temperatures were modified accordingly,and as noted before, the required water flows and thusdistribution pipe diameters and pumping powerrequirements were modified too. The effects of thesevariations on cost were taken into account.The results of all the modelling runs are shown in thefigures below in the form of the relationship betweenthe annual cost (the sum of capital investment, O&Mcost and energy or exergy costs) and the f-factor. Anincreasing f-factor means increasing surface areas(and thus increasing capital and operating andmaintenance cost) and decreasing heat supplytemperatures (and thus decreasing exergy cost).Figure 2 shows results for the radiator based heatingsystem with no carbon taxes in place. The slight jumpin annual cost around an f-factor of 0.81-0.82 is causedby an increase in district heating piping diameter fromDN65 to DN80 for the 80-house streets and from DN50to DN65 for the 40-house. All lower f-factors shownhave piping diameters of DN65 and DN50 and allhigher f-factors shown have DN80 and DN65respectively.Annual cost ($)$2,900,000$2,700,000$2,500,000$2,300,000$2,100,000$1,900,000$1,700,000$1,500,0000.68 0.70 0.72 0.74 0.76 0.78 0.80 0.82 0.84 0.86 0.88f-factor (-)Classical analysis Exergo-economic analysis Heat under 60C free analysisFig. 2. Relation between f-factor and annual cost radiatorsystem, no carbon tax.The classical analysis shows a continuous increase inannual cost with increasing f-factor. 1 This makes sensebecause cost is not based on exergy but on energy.Therefore, an increasing surface area meansincreasing capital cost, but constant energy cost, so thelower exergy requirement does not offset the increasein capital cost. The classical analysis would tell us tooptimize the system with minimum capital expenses. Inreality there would be a limit as ever increasingtemperatures will mean that we are dealing with moreexpensive materials and at a certain stage steaminstead of hot water, requiring a more expensive districtheating system. Also, heat losses to the environment501 Although exergy is not explicitly costed in the classical analysis,we can still calculate an f-factor.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniawill increase with increasing supply temperatures,which is not modelled here.The exergoeconomic analysis does take into accountthe increasing exergy requirements for systems withlower surface areas. Consequently, there is a minimumannual cost (for an f-factor in the range 0.71–0.73).Going to lower f-factors, exergy cost significantlyincrease, which results in increasing annual overallcost. Going to higher f-factors, the annual cost increaseagain because the decreasing exergy cost are morethan offset by the increase in capital and O&M cost.The analysis discounting heat under 60 °C does takethe temperature level of energy supplied into accountwhile determining costs, though there is no explicitprice for exergy in the calculations. As a result, the linedoes not slope up as strongly with increasing f-factoras the line pertaining to the classical analysis. Going tohigher f-factors, eventually all heat will be deliveredunder 60 °C, and all heat provided will be free. Going tolower f-factors, eventually all heat will be supplied attemperatures over 60 °C and the green line willcoincide with the blue classical analysis line. The heatunder 60 °C free analysis does not show an optimumand would suggest minimizing the f-factor. Like theclassical analysis the economic analysis suggests thatcapital cost are dominant.Figure 3 shows the effect of introducing a carbon tax of$30/tonCO2eq on the exergoeconomic analysis. It isclear that the carbon tax leads to higher annual costand lower f-factors for the same systems, both causedby the increased exergy cost. Both lines show anoptimum for an f-factor in the range 0.71–0.73, but forthe case without carbon tax the corresponding surfacearea is lower than for the case with carbon tax. Thismakes sense as increasing heat and exergy cost meana shift to a system with higher surface areas and lowerheat and exergy requirements. As figure 3 shows, thecapital and O&M cost as a fraction of total cost (andconsequently also the exergy cost as a fraction of totalcost) remain in the same range.Annual cost ($)$2,750,000$2,700,000$2,650,000$2,600,000$2,550,000$2,500,000$2,450,000$2,400,000$2,350,000$2,300,000$2,250,000$2,200,0000.63 0.65 0.67 0.69 0.71 0.73 0.75 0.77 0.79 0.81 0.83 0.85 0.87f-factor (-)No carbon tax Carbon tax $30/tCO2Fig. 3. Relation between f-factor and annual cost radiatorsystem, with and without carbon tax.Alternative designs – cross-flow heat exchangersystemFigure 4 shows the results for the system with crossflowheat exchangers. Note again that the jump inannual cost at an f-factor around 0.78 is due to theincrease in district heating pipe diameter. As for theradiator system, the classical analysis shows a steepslope with increasing f-factors as capital cost aredominant and lower exergy requirements do nottranslate into cost savings. Again the classical analysiswould lead us to minimize the surface area (with thesame limitations as applied to the radiator).Annual cost ($)$2,200,000$2,100,000$2,000,000$1,900,000$1,800,000$1,700,000$1,600,000$1,500,000$1,400,0000.65 0.67 0.69 0.71 0.73 0.75 0.77 0.79 0.81 0.83f-factor (-)Classical analysis Exergo-economic analysis Heat under 60C free analysisFig. 4. Relation between f-factor and annual cost crossflowheat exchanger system, no carbon tax.The exergoeconomic analysis shows a downwardsloping line. This is caused by the reduced capital costand O&M cost compared to the radiator system andthus increased importance of exergy cost as fraction ofthe total cost. An increase in cost due to surface area ismore than offset by a decrease in exergy cost.Contrary to the radiator system, though, theexergoeconomic analysis does not show a clearoptimum, although it clearly levels off at higherf-factors. It is interesting to note here that the classicalanalysis and the exergoeconomic analysis lead tocontradictory recommendations as to optimization.The analysis with free heat under 60 °C shows a linegradually sloping up, though far less pronounced thanthe classical analysis line. Like the classical analysisline it would indicate that lower surface areas wouldoptimize this system.Figure 5 shows the effect of a carbon tax on theexergoeconomic analysis. As for the radiator systemthe carbon tax means higher annual cost and lowerf-factors for the same system due to increased exergycost. As there is not a clear optimum in either line, wecan not conclude that the optimum f-factor is the samefor both. However, it is clear that both level off in thehigher f-factors range.51

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAnnual cost ($)$2,500,000$2,400,000$2,300,000$2,200,000$2,100,000$2,000,000$1,900,000$1,800,000$1,700,000$1,600,0000.60 0.62 0.64 0.66 0.68 0.70 0.72 0.74 0.76 0.78 0.80 0.82No carbon taxf-factor (-)Carbon tax $30/tCO2Fig.5. Relation between f-factor and annual cost crossflowheat exchanger system, with and without carbon tax.From the foregoing, it is clear that useful comparisonscan be made using this methodology. The results fromexergoeconomic analyses can significantly deviatefrom those obtained with a classical analysis. Which ofthe two is the more relevant one will depend on thesituation. For non-integrated systems, the classicalanalysis may be the one to follow, but for integratedenergy systems, which are expected to become moreand more important, the temperature level of heatbecomes important, and the exergoeconomic analysisseems more appropriate. Using the f-factor will help infinding optimum solutions, especially for exergoeconomicanalyses.Variations in external factors, such as fuel costs orGovernment / utility incentives could change the shapeof the curves to make the minimum more pronounced.CONCLUSION AND SUGGESTIONS FOR FURTHERWORKFrom the results of testing the methodology ofexergoeconomic optimization using the f-factor, it isclear that it is a useful tool to determine the effects ofdifferent heating technologies and heat transfer surfacesizes of these technologies on the annual overalloperational costs. This is especially true if the heatingsystem is integrated with other energy systems. It isalso true if the temperature level of the heat isimportant for another reason. The methodology can beused to make informed choices regarding technologiesto be used for heating homes or buildings andregarding the size of these technologies.To continue this development work, it is recommendedthat more practical considerations will be incorporatedinto the models and analyses. Increasing temperaturesdo not just cost more in terms of exergy but also inmore expensive materials, and steam based districtheating systems are considerably more expensive thanhot water based systems. Heat losses from the pipelinewere small but may need to be considered in a followupstudy. Including passive heating of houses by solarradiation, plug loads and occupancy gains will alsoimprove model predictions. Also mixed systemscombining heating technologies and possibly includingother technologies such as under-floor heating providefurther opportunities to optimize system cost.In addition, the application of the methodologydeveloped in this study should be applied to a heatpump, where the variations in COP with supplytemperature would be included. This would result in theability to match the heating equipment to the heatpump, resulting in an optimum operation.ACKNOWLEDGEMENTDuring this work the authors have had very fruitfulconversations with many colleagues: Mikhail Sorin,Evgueniy Entchev, Libing Yang, Ibrahim Dincer, HajoRibberink and Kirby Wittich. These discussions helpedfocus the work and stimulated further thinking in thisinteresting area of science. This is to thank all thosewho spent their valuable time listening and providingvaluable comments.REFERENCES[1] G. Wall, ―Exergy and Morals‖, in Second lawanalysis of energy systems: towards the 21stcentury, E. Sciubba, M.J. Moran Eds, Circus,Roma (1995), ISBN 88-86662-0-9, pp. 21-29.[2] D. Schmidt, ―Design of Low Exergy Buildings –Method and a Pre-Design Tool‖, in <strong>International</strong>Journal of Low Exergy and Sustainable Buildings,Vol. 3 (2003), pp. 120-126.[3] A. Valero, L. Serra & J. Uche, ―Fundamentals ofExergy Cost Accounting and Thermoeconomics.Part I: Theory‖, in Journal of Energy ResourcesTechnology, Vol. 128 (2006), pp. 1-8.[4] G. Tsatsaronis, ―Application of Thermoeconomicsto the Design and Synthesis of Energy Plants‖, inExergy, Energy System Analysis, andOptimization, [ed. Christos A. Frongopoulos], inEncyclopedia of Life Support Systems (EOLSS),developed under auspices of the Unesco, EolssPublishers, Oxford, UK (2007).[5] G. Temir, D. Bilge, ―Thermoeconomic analysis of atrigeneration system‖ in Applied ThermalEngineering, Vol. 24 (2004), pp. 2689-2699.[6] Clean Energy Project Analysis – RETScreenEngineering & Cases Textbook, 3 rd edition,RETScreen <strong>International</strong>, Natural ResourcesCanada, Varennes (2005).52

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSLIMNET: AN INNOVATIVE INTEGRAL APPROACH FOR IMPROVINGEFFICIENSIES OF DISTRICT HEATING NETWORKSM. W. P. van LierStadsverwarming Purmerend B.V., the Netherlandsm.v.lier@svpbv.nlABSTRACTThis paper describes the innovative integral approachimproving district heating network efficiency, SlimNet.SlimNet consists of five phases which lead to annualenergy savings of about 227.000 GJ and almost 37.000ton CO 2 savings for the city of Purmerend in 2015.INTRODUCTIONCompany situationIn 2007 the new company StadsverwarmingPurmerend B.V. (SVP) took over the responsibilities ofthe district heating network from the municipality inPurmerend, the Netherlands. With 25.000 customersthe grid is the fourth largest grid of the Netherlands.District heating Purmerend started in 1980. Thenetwork expanded organically following the cityexpansions. While daily operations were outsourced toexternal and changing partners, the final responsibilitystayed with the municipality.A comprehensive business analysis performed by thenew management in 2008 showed severe problems. Inthe present state the company would remainstructurally loss giving, (future) heat delivery was notensured, and sustainability and customer satisfactionwere below benchmark standards. Fall 2009 a newbusiness plan was presented that sets course for afuture proof company, based on sustainable, costeffectiveand 80% renewable heat. On the technicalside this is achieved by two major project programs, a.improving network efficiency, SlimNet, and b.incorporation of sustainable energy sources, theEnergy transition. The company mission is to becomethe most sustainable district heating company of theNetherlands.installations. Specific to the secondary network are thepost-insulated steel distribution pipes and connectionsto customer installations hanging in narrow crawlspaces under blocks of buildings.In the distribution process no heat exchangers areused except from the production of hot tapping water inthe houses.Hydraulics are controlled by decentralized pressurizingvalves, differential pressure valves and pumpscompensating for hydraulic deficiencies.The supply temperature from production is directlyrelated to the ambient temperature (i.g. 95 C atT a =-10 C and 75 C when T a =15 C). The maximumsupply pressure to the primary network is 6,8 bars andto the secondary network 4,5 bars.NETWORK CONDITIONPart of the business analysis was an extensivetechnical research program covering all technicalaspects of the grid and finally entire district heatingchain. The main conclusions were:1. The network characteristic had becomeuncontrollable: Network builds out has occurredwithout a master plan. Effectively SVP had nocontrol on the characteristics of customerinstallations. Furthermore, hydraulic problems inthe grid had been masked with decentralizedpumps and control systems.2. Heat production capacity was critical, reaching acritical limit under the conditions of the winter of2008. There was certainly no spare capacity tofacilitate the planned expansion of the grid andthus the heat demand as shown in Fig 1.Network descriptionThe 520 km district heating network is fed by a CHP(CCGT) plant of 65 MWth and seven natural gas firedauxiliary boilers with a total power of 131 MWth. Duringthe last 6 years 64% of the total heat production camefrom the CHP plant. The heat sources are operated bya third party.The production units feed the heat to the network viabuffering tanks to the primary network. The heat is thendirectly transported through substations and asecondary network to the 25.000 customerFig. 1 Required heat production53

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia3. In 2008 the network showed a heat loss factor of33,6% (with a Dutch benchmark of 25%). requiring32.683 m3 of water replenishment in the sameyear.4. Parts of the network showed excessive heat lossand repairs, mainly due to high ground water table,exposing the pipes in crawl spaces directly towater for most of the year. Repairs with standardmaterial proofed insufficient and innovation onmaterial and building techniques was needed.Primary networkMost substations in the network are provided with aSCADA6 system. This data in combination with a newlydeveloped network model made it possible tocalculating annual heat loss at 100.706 GJ.According to [3] about 14% of this heat loss is causedby cross-linked polyethylene (PEX) piping materialused in the early 90‘s.SLIMNETSlimNet is part of a large restructuring program initiatedin 2008. SlimNet does contribute to stopping thenegative spiral glide of the above mentioned problemsSlimNet consists of the following phases:A. Knowing where the heat flowsB. Defining key performance indicators (KPI)C. Developing analyzing toolsD. Developing and defining measuresE. Quantifying KPI results from SlimNetIn the following those phases will be discussed.KNOWING WHERE THE HEAT FLOWSFor SVP the heat losses are defined as:Qloss Q Q(1)producedsoldThe heat losses in the network, Q loss , were 427.158 GJ(33,6%) in 2008. Causes for those losses 5 are:1. Losses in buffering tanks2. Losses in primary network3. Losses in secondary network4. Undefined lossesNone of the above can be determined exactly withinthe boundary conditions of the network but thefollowing describes the results of the researchperformed on this matter and the localization of―hotspots‖, parts of the grid with excessive losses.Buffering tanksIn [1] an estimated calculation was made for the heatlosses due to the buffering tanks, 5.562 GJ annually.There are four buffering tanks with a 4.000 m3 capacityin the network which are used for peak shaving. Acheck upon this calculation [2], based upon an IR-scanof one of the buffering tanks resulted in an estimate of14.032 GJ annually which is considered to be amaximum value.Fig. 2 IR scan of a PEX pipe constructed in 1990Considering that those PEX pipes are applied in only3,5% of the primary network, these may be referred toas ―hotspots‖.Secondary networkWith four public housing companies, SVP conductedresearch on failures in the district heating relatedsystems in Purmerend [3]. It became clear that duringthe period 2006-2008 74% of the unplanned repairswere caused by the high ground water level in thecrawl spaces where post-insulated steel pipes withArmaflex insulation are installed. In total researchidentified areas of 4000 houses, where heat loss wasextreme, i.e. ―hotspots‖.This research confirmed the conclusion of an earlierresearch [4] that the thermal conductivity k for the wetinsulation in the crawl spaces will be close to 0,1 W/mKand 0,2 W/mK instead of the 0,02 or 0,03 W/mK for thecurrent pre-insulated pipes. The total of heat losses inthe secondary network are estimated at 304.041 GJ.Conclusion addressing heat lossesTable 1 gives the overall results of the heat lossanalysis.Table 1: Overall results of heat loss analysisMain network part Loss(GJ) % of totalBuffering tanks 14.032 3,3 %Primary network 100.706 23,6 %Secondary network 304.041 71,2 %Undefined losses 8.170 1,9 %Total 427.158 100%5 Losses from heat plants are not taken into account.6 Supervisory Control And Data Acquisition54

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaIt was concluded that replacing the PEX pipes in theprimary network and post-insulated pipes in the crawlspaces of houses in the areas identified as ―hotspots‖was the most effective strategy for heat loss reduction.DEFINING KEY PERFORMANCE INDICATORSThe main goal of SlimNet is improving networkefficiency as part of the new business plan that setscourse for a future proof company which providessustainable, cost-effective and 80% renewable heat.The Key Performance Indicators (KPI‘s) can be dividedin four main criterions:1. Economics2. Sustainability3. Reliability4. Customer SatisfactionEconomicsEvery GJ of heat lost in the network cannot be sold andhas therefore a negative effect on the balance sheet.Consequently the heat loss in the DH-network is anobvious and important KPI.Another parameter that has a negative effect onprofitability is the amount of water that is replenished.SustainabilityThe avoided CO2-emissions are and should be animportant driver for DH grids. According to subsequentdirectives in the Netherlands for assessing energyperformance of buildings NEN 7120, the avoidedCO2-emissions has to be determined on the requiredprimary energy sources and by referring to commonstate-of-the-art technologies. The HR-107 type (107%LHV efficiency) is the required and accepted commonstate-of-the-art reference technology.ReliabilityThe condition of the network in terms of reliabilitypresents itself in the amount of times that mechanicshave to deal with unplanned repairs. It was apparentthat SVP was facing an increasing trend curve. Theactual deprecation of the replaced piping provided aanother criterion for assessing system degradation.Customer satisfactionReducing off time, during replacement was animportant element of the SlimNet approach.KPI summary1. Heat loss2. Water replenishment3. Avoided CO 2 emissions4. Unplanned repairs5. Network degradation6. Off-time during replacementDEVELOPING ANALYZING TOOLSResearch had located the ―hotspots‖ of unplannedrepairs and heat loss in an area of 4000 houses. These―hotspots‖ were responsible for 50% of the unplannedrepairs. In order to define and implement a suitable andcost-effective replacement strategy a set of tools wasdeveloped.Upgraded network diagramAnalyzing networks requires reliable andcomprehensive network diagrams. All requiredinformation such as dimensions, age, depth etc. shouldbe available in the diagram. Many network diagramsare drawn using CAD-software. Analyzing from thosedrawings is costly. It therefore was chosen to revise thediagram completely and apply the possibility to addelement attributes to the drawing connected to anintegral database system. The upgraded networkdiagram had a catalytic effect on two other models, thenetwork model and the grid valuation model.Network modelIn 2009 SVP replaced the outdated and inadequatenetwork with a validated dynamic model (TERMIS),developed by 7-Technologies with COWI as systemintegrator. With the upgraded network diagram SVPhad the first and validated model of the primarynetwork within five months.In combination with a new CRM system, operationalsince 2010, SVP will soon be able to tap into theinformation on customer behavior and consumption.This will allow SVP to dynamically calculate the currentstate of flow, pressure and temperature throughout thenetwork at a configurable cycle time. Additionally, everyreal-time model calculation cycle will include a forecastsimulation for a given period. This allows SVP to beabreast of demands, enabling optimization ofoperations and planning of the future.Valuation modelThe upgraded network diagram supplied databaseinformation on lengths, dimensions, age and type. Withthe following equations added to the database it waspossible to develop a valuation model, that could helpto prioritize and direct renovation efforts.networkX R L xx1network D AY x1 DxxRx Lx(2)(3)55

Z Y networkx1 D ADx1X = value network in new state (€)Y = current network value (€)Rx LZ = required annual maintenance costs (€)x = pipex(4)Rx = construction costs per meter pipe dimension x (€)Lx = length of pipe x (m)D = lifetime expectancy (year)Ax = age of pipe x (year)The network degradation is defined as factor :YX (5)From consultation with amongst others COWI, it wasconcluded that networks with a < 0.5 are in a criticalstage.For the entire grid the was above the threshold.Discriminating the for separate grid sections helpedto identify the hotspots and monitoring will help todetermine the effect of SlimNet.Sustainability assessment modelTo assess the current sustainability results of thenetwork, SVP developed a sustainability assessmentmodel in accordance with Dutch law and guidelines,resulting in Fig. 3 [5]. This model can also predict theeffects of optimization in the chain from production,distribution and delivery to customer installationsThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaStrategic meteringIt was concluded that actual data on heat loss onsmaller scale (houses and clusters of houses) wouldfacilitate decision making on future renovation projectsand grid management. To get hold of this informationSVP installed heat meters with radio transmissionmodules on strategic positions in the network. Togetherwith the metering data from heat meters in customerinstallations this firstly gives accurate data on the heatloss in the corresponding part of the grid. This setupwill also provide us with empirical data on the long termresults of network improvement measures.In order to make the data comparable, two areas wherechosen. One with the new SlimNet approach (Usingpolybutene pipes and new construction techniques)and one with conventional material and constructiontechniques. First comparative results will be availableby the end of 2010.Leak detectionMost producers of pre-insulated pipe systems offer thepossibility of leak detection wiring. Using a master planwith proper zero and recurrent measurements thiswould be a reliable method of leak detection.Unfortunately this is not applicable to the situation inPurmerend.With one of its partners SVP developed a method usingtracer gas to detect leakages. The detection devicesproofed to be very sensitive and with this methodalmost 2.500 houses have been inspected this yearand last year. Leakages were detected in 3% of thosecases, mostly in an early stage, that otherwise wouldonly have been detected through visual sighting ofdamp.DEVELOPING AND DEFINING MEASURESIt became clear soon that the only way to improvenetwork performance was to rigorously renovate thehotspots and to start implementing a structuralmaintenance program in accordance to Z, Eq. 4.In sum the challenge was: a.) cost effectively renewingthe steel pipes with wet insulation in narrow crawlspaces while b.) improving network efficiency.Fig. 3 CO 2 reduction DH-network Purmerend in pastIt appeared that the ratio of CHP operation to the totalof heat produced and the heat loss factor have thebiggest impact on the sustainability results.To meet (a), SVP started the first two pilots in 2008with pre-insulated steel flex piping material, using twodifferent construction methods. Both pilots met thetechnical requirements but were too time consuming,costly and, because access to the crawl spaces had tobe gained by digging in the gardens, meant hugeinconvenience for customers.Parallel to this SVP had challenged pipe manufacturesto come up with innovative material constructionmethods, suitable for the Dutch situation (groundwater56

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaand retrofit in narrow crawl spaces). The only viablesolution came from Flexalen of Thermaflex, usingflexible polybutene (PB) carrier pipes. The producer ofthe PB material offers a 50 years plus life guarantee [6]for the pressures and temperature profiles of the SVPnetwork.A pilot with Flexalen was conducted in September2009. The pilot used prefabricated joints of Flexalen,called Flexalinks, which were under research anddevelopment at that time. The pilot did meet all therequirements. Costs were reduced by 30% comparedto the steel flex pilots, 16 houses were overhauledwithin a week and access could be gained by the crawlspace hatches.On the basis of this pilot decision has been made toretrofit 4000 houses within four years. Works hascurrently started at the first 309 houses, at a speed of30 houses a week.The second part of the challenge (b): improvingnetwork efficiency, is furthered by SlimNet throughoptimizing pipe dimensions and lengths (smart gridredesign)SlimNet part I: Renovation and smart redesignApplying Flexalen means an improvement of k from 0,1of the wet post-insulated steel pipes to theoretically0,031 W/mK (manufacturer information, at 50 C).Key to the SlimNet approach was smart redesign.Calculations in TERMIS showed that many parts of theDH-grid in Purmerend are generally oversized, and thatthe common circular grid can easily be changed into astar shaped grid, whilst reducing pipe lengths. UsingTERMIS redesign focused on reducing radialdimensions and pipe lengths by deleting obsoletepipes.The results for the part of the grid that is replaced thisyear, Fig. 4 and Fig. 5, gave, Table 2 [8]:Table 2: Results from redesign 2010 areaHeat demandHeat lossCurrent situation 100,0 % 100,0 %New dimensions 93,0 % 76,3 %Finger system 91,0 % 69,5 %Heat losses can be reduced by optimizing:1. Thermal conductivity2. Pipe lengths3. Radial dimensions4. Fluid temperatureThese elements are captured in the following equationfor heat loss in a pipe [7]:Q( TTin outloss _ pipe 2 k L(6)routlnrin)Fig. 4 Existing network part to be renewedk = 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 canonly be changed by renewing pipes. The last can onlybe changed by chain modification, i.e. production andcustomer installations. SlimNet addresses both.Fig. 5 Redesigned and renewed network57

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe actual effect of SlimNet on heat losses will beclosely monitored in the grid, through the strategicmetering project.SlimNet part II: Smart chain managementThe last heat loss parameter, fluid temperature (Eq. 6),can only be changed by modification of the completechain.To start at the production side, the current supplytemperature 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 therequirements of customer installations, would reducethe average network temperature hence the heatlosses. It was calculated through the network modelthat the alternative temperature curve in Fig. 6 solelywould reduce the heat losses with 4%. Furtherresearch will focus on matching the most effectivetemperature curve with production characteristics.houses that have a 90–50 C characteristic duringdesign conditions (-10 C). In most areas before thattime SVP found return temperatures that arestructurally higher than the required 50 C. Hence theflows in those areas are also much higher thannecessary.The high return temperatures and corresponding highflows are caused by absence of pressurizing valves inthe customer installations and defective control valvesin the hot tapping water installations. By the end of2010 SVP starts a campaign to encourage houseowners to improve or renew their installations, also fortheir own benefit. This campaign will make use of localapproved installers of customer installations. Researchindicated that in certain areas the peak flow can bereduced with 60% [9].QUANTIFYING KPI RESULTS FROM SLIMNETSummarized, the measures that SVP takes before2014 to improve network efficiency:1. Renewing the distribution pipes and houseconnections in the crawl spaces of 4000houses, while optimized to dimensions andlengths.2. Replacing 4,0 km PEX-pipes in the primarynetwork, while optimized to dimensions andlengths.3. Doing this with a minimum of off-time forcustomers4. Implementing demand-driven heat productionFig. 6 Existing and alternative temperature curveThis research will also look upon the possibilities ofimplementing demand-driven heat production. This isachieved by using a real time network modelconnected to the substations and production SCADA.The model uses the weather forecast with customerinformation to adjust the temperatures and pressuresjust to meet the requirements of customer installations.It is expected that this will reduce the average fluidtemperature even more.Further research is done to implement cascadingheating services, i.e. using the latent heat in the returnpipes of the network with temperatures between 45 Cand 60 C to the customer installations. This ishowever only possible to implement in new houses withlow temperature heating installations. This research willfocus on further reducing the heat losses.5. Implementing cascaded heating installations6. Encourage house owners to improve or renewtheir installations in accordance with SVPguidelines.7. Eliminating arrears of maintenance animplementing a structural preventativemaintenance program.Heat losses will reduce from 33,6% in 2008 to 22,1% in2015. While heat consumption prognoses stays thesame, the corresponding required heat production falls,Fig. 8. This results in a energy saving of 227.000 GJthat year. In Fig. 9 the results of the sustainabilityassessment model are shown regarding CO2 savings.At the other end of the chain are the customerinstallations. Since 1996 the district heating company inPurmerend has only accepted installations in new58

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaREFERENCES[1] A. D. Heidweiller, B. C. Van Leeuwen and C. L.Paarmann, ―Systeemstudie StadsverwarmingPurmerend‖, Tebodin B.V., Den Haag, theNetherlands (2006)[2] A. E. Klop, B. P. Mensink and C. F. Dervis,―Transitiestudie Stadsverwarming Purmerend‖,DWA Installatie- en Energieadvies, Bodegraven,the Netherlands (2009)Fig. 7 Required heat production with SlimNet[3] A. L.J.A.M. Hendriksen and B. R.A. Brand,―Onderzoek naar storingen in hetstadsverwarmingnet van Purmerend (report 034-APD-2009-0021)‖, TNO Bouw en Ondergrond,Apeldoorn, the Netherlands (2009)[4] A. M. den Burger and B. D. Heidweiller,―Deelrapport 5: Warmteverliezen enmeetverschillen‖, Tebodin B.V., Den Haag, theNetherlands (2005)[5] F. Dervis, ―Nulmeting duurzaamheid SVP‖, DWAInstallatie- en Energieadvies, Bodegraven, theNetherlands (2009)Fig. 8 CO 2 savings with SlimNetReplacing the post-insulated steel and PEX pipestogether with a maintenance program including leakdetection will have a positive effect on the waterreplenishment. The leak detection actions have alreadyresulted in a 30.285 m³ replenishment in 2009, which isa 7% reduction compared to 2008.It is expected that al measures will result in a 50%reduction in 2015. Unplanned repairs will also reduce50% and consequently is expected to improvesignificantly.[6] J.J. Ribberink, ―Lifetime prediction of PB pipesused in a district heating network‖, KIWA N.V.Certification and inspection, Rijswijk, theNetherlands (2009)[7] A. D. A. Kaminski and B. M. K. Jensen,―Introduction to thermal and fluid engineering‖,John Wiley& Sons, Hoboken, USA (2005), pp 103[8] T.A. Østergaard, ―New dimensions for O16‖,COWI A/S, Aarhus, Denmark (2010)[9] A. B. Zitoony and B. E. Roukema, ―Rapportinregelstatus onderstations StadsverwarmingPurmerend 10.001.V2‖, Roukema B.V., Groningen,the Netherlands (2010)59

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaA DIRECT HEAT EXCHANGER UNIT USED FOR DOMESTIC HOT WATER SUPPLY INA SINGLE-FAMILY HOUSE SUPPLIED BY LOW ENERGY DISTRICT HEATINGMarek Brand 1 , Jan Eric Thorsen 2 , Svend Svendsen 3 and Christian Holm Christiansen 41Ph.D. student, Technical University of Denmark2Senior project manager, Danfoss District Energy, Nordborg, Denmark3Professor, Ph.D., Technical University of Denmark4Danish Technological Institute, DenmarkABSTRACTThe increasing number of new and renovated buildingswith reduced heating requirements will soon maketraditional District Heating (DH) systems uneconomic.To keep DH competitive in the future, the heat loss inDH networks needs to be reduced. One option is toreduce the supply temperature of DH as much aspossible. This requires a review of the behaviour of thewhole domestic hot water (DHW) supply system withfocus on the user comfort and overall costs. This paperdescribes some practical approaches to theimplementation of this Low Energy District Heating(LEDH) concept. It reports on the testing of the dynamicbehaviour of an Instantaneous Heat Exchanger Unit(IHEU) designed for DHW heating and space heating indetached family houses supplied by LEDH ensuring anentry-to-substation temperature of 51 °C. We measuredthe time it takes for the IHEU to produce DHW with atemperature of 42 °C and 47 °C when the tap isopened. Measurements were made for controlstrategies using internal and external by-pass and noby-pass. Our results show the importance of keepingthe branch pipe warm if comfort requirements are to befulfilled, but this involves higher user costs for heating.To increase user comfort without increasing costs, wepropose the whole-year operation of floor heating inbathrooms, partly supplied by by-pass flow.INTRODUCTIONDistrict Heating (DH) is a well known concept ofproviding buildings with heat for space heating (SH) andDomestic Hot Water (DHW) heating in economical andenvironmentally friendly way. Nowadays, buildingregulations have been introduced worldwide and arepushing to reduce energy consumption in buildings,because 40% of all energy consumption takes place inbuildings. The energy policy of European Union isrecently focused on energy savings, reducingproduction of CO 2 and increasing the ratio of renewableenergy [1]. DH is one of the most suitable solutions toachieve these goals for building sector and it gives highpriority for further development of DH. But recenlty usedtraditional high and medium temperature DH systemsare not optimal solution for the future. Sooner or later,energy consumption of all buildings will be inaccordance with low energy building regulations and it60will form areas with lower heat demand than nowadays.Currently used DH networks will not be able supplythese areas in economical way, because the ratiobetween network heat losses and heat consumption inbuildings would be unacceptable and thus cost of heatfor end users will increase and DH systems will looseconcurrency with other solutions, e.g. heat pumps.Recently, research in DH is focused to find the way howto use DH in areas with low energy buildings and how toincrease ratio of heat produced by renewable sources ofenergy as solar heat plants or heat pumps driven byelectricity from renewable sources.One of interesting application of renewable energy inDH is use of decentralised heat sources as e.g. solarcollectors installed on roofs of individual buildings,supplying heat to DH network, but it still needs moretime and work to develop new substations and newconcept of DH networks to be able to handle these newfeatures. The solution for future development of DH is toreduce heat losses of DH networks by means of pipeswith better insulation properties e.g. twin pipes, usebetter concepts of network design (circular networkconfiguration, possibility of using circulation line for mainpipes) and to reduce the supply temperature of districtheating water to lowest level as possible.The District Heating Systems designed due to thisphilosophy are called Low Energy District HeatingSystems (LEDH). The main focus in LEDH system is toreduce heat losses from network as much as possible,exploit more sources of renewable energy for heatsupply and still maintain or improve level of comfort forusers, because without high level of comfort thisconcept can‘t be successful. LEDH concept wasreported e.g. in project ―Development andDemonstration of Low Energy District Heating for LowEnergy Buildings [2], where theoretical case studydocumented, that LEDH concept is a good solution forfuture and even in sparse housing areas is fullycompetitive to heat pumps. This article is focused onapplication of LEDH for DHW heating. Considerationsrelated to use of LEDH for space heating will bereported in future in another article.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaLOW TEMPERATURE DISTRICT HEATINGCONCEPTReduced risk of Legionella by use of system withminimal volume of DHWSince LEDH is mainly developed for low energybuildings already designed with low temperature spaceheating, the lowest acceptable forward temperature ofLEDH system is defined by requirement for DHW supplytemperature. The hygienic requirement for heating ofDHW is due to recent standards 50 °C for single-familyhouses and 55 °C for multi-storey buildings [3] whereDHW circulation is used. In case of using circulation,temperature of recirculated water should never fallbelow 50 °C. These requirements are based on need toavoid Legionella growth in DHW pipes and storagetanks. It is widely believed, that Legionella grow intemperature range between 46 °C – 20 °C, in systemswith high volume of water. Mentioned temperaturelevels are made in order to assure comfort and hygienicrequirements in furthest tap away from a heat source. Itis important to say, that there is high level ofdiscrepancy among different results and nationalstandards focused on Legionella.Due to German Standard W551 [4], temperature ofDHW can be below 50 °C and not cause Legionellapromotion, if total volume of DHW system connected toone heat source is lower than 3 L. From literaturestudied, it can be concluded that requirements toproduce DHW with temperature higher than 50 °C aredefined for an old fashion DHW building installations,which can be characterized as systems with verticalriser, branched pipes with bigger diameter (increasingwater volume of the system), using DHW circulation.For new and renovated buildings, DHW installations aredesigned in much better manner, with individualconnection of DHW pipes between each tap and sourceof DHW and with maximally reduced pipe diameter,defined by requirements for noise propagation andpressure drop.Due literature, danger of Legionella growth in DHWsystem is influenced by temperature of DHW, nutrientsin DHW, laminar or turbulent flow in the DHW pipes andwater stagnation [5]. Several on site measurementswere performed in buildings using DH for DHW heating.From results of Martinelli [6] and Mathys [7] can beconcluded, that Instantaneous Heat Exchanger Unit(IHEU) tend to have much less problems with Legionellathan traditional units with DHW storage tank. Bothstudies concluded, that these findings are caused by thefact that in IHEU, DHW is produced with temperature60 °C, while in case of storage units only withtemperature 50 °C. But is necessary to mention, that incase of traditional DHW storage tanks, overall volume ofDHW in a system is much higher than in case of IHEUsystem. Due to our knowledge, there is not reported61investigation of Legionella in DHW system using IHEU,producing DHW with temperature below 50 °C andreduced volume of the system below 3L.For single family houses with appropriate close locationof tapping points, volume of DHW in IHEU and pipes willbe lower than 3 L and thus temperature of 50 °C onprimary side will not cause Legionella problems. Formulti-storey buildings, district heating substations foreach 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 hasincreased users comfort and no huge DHW systemswith circulation, where Legionella is forming andspreading [9]. The other advantage of using flat stationin multi-storey buildings is individual metering of eachflat and complete control over space heating and DHWpreparation, which is positively affecting energy savings.With properly designed DHW building installations,supply temperature of LEDH will be defined byrequirements for users comfort. These requirements arediscussed in following text.Users comfort in DHW supplied by LEDHAnother important question, when concerning DHWsystems is level of user comfort. From comfort point ofview, requirements for temperature and waiting time forDHW can be specified. Due to Danish Standard DS439―Code of Practice for domestic water supplyinstallations‖, [10] temperature of DHW should be 45 °Cin kitchen and 40 °C in other taps, provided withnominal flowrate and desired temperature reachedwithin ―reasonable‖ long time, without significanttemperature fluctuations. It is a question, if requirementof 45 °C degrees for kitchen tap is not too high, butargument of problems with fat dissolving from dishescan be objected and should be investigated. Based onmentioned standard, desired temperature of DHWflowing from fixture is 45 °C. But in order to definedesired forward temperature of LEDH system, weshould be aware of temperature drop in DH network, inuser‘s substation and in DHW installations in building.The temperature drop in DH network is not in focus ofthis paper, so our goal is to find needed temperaturelevel at the entrance of substation to produce 45 °Cfrom tap in building. Desired temperature will be foundby experimental measurement of LEDH substation laterin article.Beside temperature requirements, users comfort isinfluenced by time needed for DHW to reach a fixtureafter tapping was started. This waiting time is infollowing text called ―tap delay‖. Due to DS439,suggested value for tap delay is 10 sec and it is definedin order to avoid wasting of water and to protect usersagainst too long waiting times for DHW. In large multistoreybuildings with centralised preparation of DHW,short tap delay and measures avoiding Legionella

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniagrowth are assured by circulation line of DHW, but notproperly designed or maintained DHW circulation isquite often responsible for increased risk of Legionella[11]. Another disadvantage of DHW circulation is bigheat losses, sometimes even bigger than net energyneeded for DHW heating [8]. The 10 sec waiting time isnot rule and for some people it is a long time, for somepeople short, but this value is used to evaluate testedconcepts if they are fulfilling requirements for high levelof users comfort or not. An overall tap delay can bestudied from different angles. From dynamic point ofview, tap delay consists of transportation time neededfor ―new volume‖ of water travel to tap and dynamicthermal behaviour of passed components, i.e. pipes andsubstation. From point of view related to location, itconsists of three parts, tap delay in branch pipe (pipefrom DH pipe in street to users substation), in DHsubstation and in DHW system in building. A tap delayin branch pipe and substation are related to DH networkand substation‘s control system strategy, while tapdelay in DHW pipes in buildings without DHWcirculation are defined only by thermal capacity of pipes,volume of water in individual pipes, nominal flow and tosome extend also by their insulation.Tap delay in DHW system in buildingFor DHW systems with individual feeding pipes andoverall volume of pipes lower than 3 L, DHW circulationis not needed, because waiting time for DHW withdesired temperature is not critical. In Table 1, transportdelays for individual fixtures in typical house built in pilotLEDH project in Larch Garden - Lystrup, Denmark [11]are presented. It should be mentioned, that data areonly transport delay, without dynamic behaviour ofcooled pipe. From Table 1 can be seen, that reasonablydesigned close locations of fixtures, not so far awayfrom substation, lead to maximal transport delay around6 sec, for basin. The total volume of DHW systemconsists of 0.99 L in pipes and 1.1 L in HEX (typeXB37H-40). It means, that it is possible to install longerpipes or more fixtures and still fulfil requirement of DHWsystem with volume lower than 3 L. The velocity offlowing water is below 2 m/s and thus problems withnoise propagation during tapping are avoided.Table 1 – Transport delay for nominal flows for individualfixtures due to DS439, in DHW system in typical house inLystrup, for pipes with inner diameter 10 mmfixturenominalflow(L/min)lengthtofixture(m)volumeinpipes(L)velocity(m/s)transp.delay(s)shower 8.4 2.2 0.17 1.8 1.2basin 3.4 4.1 0.32 0.7 5.8kitchen 6 6.3 0.49 1.3 4.962Tap delay on primary sideA transport delay on primary side consists of delay inbranch pipe and delay in DH substation. While tap delayin DHW installations in building is for DHW systemwithout circulation uniquely determined, tap delay onprimary side varying as control strategies for substationcontrol 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 branchpipes is cooled down to temperature of ambient ground(if tapping wasn‘t performed for long time) and DH waterin substation to room temperature. In general, waitingtime for DHW is influenced by controller used insubstation. Basic principles of controllers areproportional flow controller and thermostatic controller.Each controller has own advantages anddisadvantages, thus best solution is to combine bothcontrollers [12]. In case of proportional flow controller,ratio between primary and secondary flow is fixed toprovide DHW with desired temperature and it means incase of using LEDH primary and secondary flow will bevery similar. If proportional flow controller is used forsetup without by-pass, user will face long waiting timefor DHW. Waiting time for this case can be seen fromTable 2. For branch pipe with inner diameter 15 mm (asis designed in Lystrup for IHEU), even transport delay toreach substation for nominal flow for basin, kitchen sinkand shower will be 31.6, 17.7 and 12.6 sec,respectively. This solution is from comfort point of viewand water savings completely unacceptable. If wedecrease inner diameter of branch pipe to 10 mm,transport delay is decreased roughly to one half of valuefor pipe with inner diameter 15 mm, but it is still longtime. In case of combined proportional flow controllerand thermostatic controller, from beginning of tappingthermostatic part assures opening of valve onapproximately full capacity until desired temperature ofDHW is reached.Table 2 – Transport delay for nominal flows for individualfixtures due to DS439, in branch pipe, 10 m long, for typicalhouse in Lystrup, data simulate using proportional flowcontroller without by-passfixturenom..flow(L/min)innerpipeØd(mm)volumein pipes(L)velocity(m/s)transp.delay (s)basin 3.4 15 1.77 0.3 31.6kitchen 6 15 1.77 0.6 17.7shower 8.4 10 0.79 1.8 5.6shower 8.4 15 1.77 0.8 12.6bath 12.6 15 1.77 1.2 8.4Full opening from beginning of tapping leads to muchhigher flow rate on primary side than on secondary andtime delay is decreased substantially. This solution canbe used for short branch pipes with reduced diameters.But it should be mentioned, that transport time in branchpipe will be always limited by maximal allowed flow on

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaprimary side defined by DH provider by means of flowrestrictor or by available differential pressure in DHnetwork. To reduce tap delay on primary side, controlconcepts with by-pass, avoiding cooling of DH water inbranch pipes and substations, and thus reducingsubstantially waiting time for DHW are available. Thereare two concepts of by-pass in relation to the heatexchanger: external and internal by-pass (see Fig. 1). Incase of external by-pass, DH water enters substation,but not enters heat exchanger and is sent back to DHreturn pipe and thus branch pipe is kept on desiredtemperature. Desired temperature is controlled bythermostatic valve situated in by-pass loop. Increasedlevel of comfort expressed by reduced tap delay can beadjusted independently on temperature of DHW onsecondary 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 isalways compromise between insufficient cooling of DHwater and additional heat consumed by customer andreduced waiting time for DHW. In case of operation ofspace heating system, the function of by-pass is tosome extend overtaken by space heating loop andthus heat for ―by-pass‖ operation is not wasted andtemperature of DH water returning to DH network iscooled sufficiently..In case of internal by-pass, by-passflow is passing through heat exchanger and keep itwarm (see Fig. 1). The benefit of this solution is evenmore reduced tap delay than in case of externalby-pass, but on the other hand, since heat exchangeris kept warm, internal by-pass solution has additionalheat losses. If substation is installed in room with needof space heating, heat losses are considered onlyoutside of heating season.Contrary to external by-pass solution, where it is not soimportant if space heating loop is installed in series orin parallel to DHW heat exchanger, in case of internalby-pass it is in importance. If space heating loop isconnected in parallel to DHW heat exchanger intraditional way, by-pass water just pass through DHWheat exchanger and is sent back to DH network withstill high return temperature, without any other use. Ifspace heating loop is connected in series to DHW heatexchanger or in parallel but with possibility to sent bypasswater flown through internal by-pass to spaceheating loop (see Fig. 2), this solution provides highlevel of comfort for users as well as proper use of heatneeded for by-pass operation.63Fig. 2 Combined by-pass concept, with possibility of useby-pass flow in space heating loopIn order to run by-pass without drawback of insufficientcooling of DH water and wasted heat also outside ofheating season, it is proposed to use by-pass flow forfloor heating, installed in bathroom and operate it allyear. From preliminary calculations it looks, that flowneeded to keep bathroom floor surface temperature on24°C will be enough as by-pass flow. Considering theuse of renewable sources of heat, the problem ofinsufficiently cooled DH water is related to reducedefficiency of these sources and whole year using offloor heating for comfort in bathroom is reasonable.Supply – supply recirculationAs an alternative solution for customers who don‘twant to use whole year bathroom floor heating,solution called supply-supply recirculation is apossibility how to use benefits of by-pass withoutwhole year heating of bathroom. In this case, district

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaheating water is supplied by pipe 1 to substation,circulated through HEX or external by-pass (seeFig. 3) and then sent back to district heating network(DHN) supply by pipe no.3. This concept is in earlystage of investigation but it looks promising. The mainquestion will be related to flow of DH water in branchpipe in order not cool it down too much before will besent back to DH supply pipe in the street.lower heat loss. On site measurements were started inLystrup to evaluate performance of both types of DHsubstations, but no detailed measurements requiringshort time steps are performed to evaluate level ofusers comfort. The measurements more focused touser‘s comfort are planed to be performed this year inDanish Technological Institute and TechnicalUniversity of Denmark (DTU) on DH systemssimulating the conditions in Lystrup. The DH systemswill consist of branch pipes, substation and DHWbuilding installations and different control approaches(external or internal by-pass, different set up by-passtemperatures, possibility of supply-supply recirculation,etc.) will be studied for DH substations supplied byLEDH. Measured data will be used for evaluation ofperformance of different control concepts, level ofusers comfort and lately also for validation ofnumerical model which is aimed to be developed foroptimization LEDH systems.Fig. 3 Supply – supply recirculation with external by-passThis solution is expected to be favourable mainly forcircular shapes of DH networks, but it should bementioned, that re-heating stations will be probablyneeded in point of DH network, where temperature ofDH water decrease bellow defined value.Full scale demonstration of LEDHFull scale demonstration of LEDH is recently running inLarch Garden in Lystrup, Denmark [11], where 40 lowenergy houses class 1 and 2 are connected to LEDHsystem, with designed forward temperature from heatplant 52 °C. For primary side of substation, forwardtemperature of 50 °C and return temperature of 25 °Care designed. The DH network is built from highlyinsulated single pipes (for main pipes) and main pipeswith smaller diameter, distribution and branch pipesare built from twin pipes. Two types of district heatingsubstations providing houses with DHW and spaceheating are tested by customers in real conditions. Thefirst concept is 29 Instantaneous Heat Exchanger Units(IHEU), second is 11 District Heating Water Units(DHWU). IHEU is classical concept of substation withinstantaneous heat exchanger, only with enlargednumber of plates. IHEU units have external by-pas,with set point temperature of 35 °C for customerssituated not at the end of street and 40 °C forcustomers situated at the end of the street. DHWU isnew concept of DH substation, reported e.g. byPaulsen [13]. DHWU consist of buffer tank for districtheating water and when DHW is needed, DHW isheated in instantaneous heat exchanger as in previouscase. Advantage of concept with buffer tank is peakshaveddemand of DH water during charging and useof branch pipes with lower diameter, connected with64TEST OF TEMPERATURE PERFORMANCEAs a first part of measurements planed to beperformed at DTU, the time needed for IHEU toproduce DHW with temperature of 42 °C and 47 °Cwas measured, after tapping of DHW was started. Thetap delay was investigated for two control strategies,one using internal and second using external by-pass.The measurements were performed for different initialconditions before tapping was started to simulate inrealistic way users behaviour. Finally, the periodbetween two by-pass flow operations was measured.Experimental setup and instrumentsTested DH substation was prototype of InstantaneousHeat Exchanger Unit (IHEU) developed specially forLEDH pilot project in Larch Garden – Lystrup,Denmark. The IHEU is a type of district heatingsubstation consists of a heat exchanger (HEX) withoutstorage tank. DHW is heated instantaneously in HEXonly when tapping is performed and then supplieddirectly to DHW taps by individual feeding pipes, whilespace heating is using direct connection without heatexchanger, i.e. concept typical for Denmark.Substation is same concept as regular IHEU fortraditional DH. The difference is in increased numberof plates in heat exchanger assuring better heattransfer. Water volume of primary and secondary sideis 1.1 L each and the heat exchanger is not insulated.The experiments were focused only on dynamicbehaviour of substation related to DHW heating andthus space heating loop wasn‘t connected and spaceheating valves in substation closed. Desiredtemperatures of DHW were chosen in accordance withrequirements in DS439 for temperature of DHW forkitchen sink and other fixtures. Required temperaturesmentioned in DS 439 are 45 °C and 40 °C. In order to

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniacover additional temperature drop in building DHWinstallations, 2 °C were added. This addition is basedon experience from previous measurements. Duringthe experiments, temperatures of four different flowspassing through the DH substation were measured.On primary side it was temperature of DH watersupplied to substation (T11) and temperature of DHwater returning back to DH network (T12) and onsecondary side it was temperature of cold potablewater entering substation (T21) and temperature ofheated DHW (T22). All temperatures were measuredby thermocouples type T, installed directly in pipes, inflowing water, so they do not have any practical timedelay for the measurements. The time constant toreach 90% of step change was less than 1 second.The distance of thermocouples from substation flangeswas 5 cm and thermocouples were previouslycalibrated. We also measured surface temperature ofHEX in upper (HEX-UP) and bottom part (HEX-DOWN) and temperature of air in the testing room.Temperatures were measured and collected bymultifunction acquisition unit every second. Forauthentic simulation of DH network, DH water withconstant temperature of 51 °C was necessary. It wassolved by connecting of IHEU to source of DHW inlaboratory of DTU, where DHW is supplied by DHsystem. DHW system of DTU is big enough, to assurestable temperature 51 °C without any fluctuations. Inorder to prevent cooling down of pipes supplying DHWto laboratory in periods when there was not flowthrough substation (stopped by by-pass controller),small guard flow, just before entrance to substationswas kept to maintain DHW always on 51 °C anddrained to sink.Experimental procedureAs a first step, both controllers were adjusted toprovide 47 °C on DHW side with supply temperature ofDH water 51 °C. Then we measured time delay in thesubstation, i.e. time needed for substation to produceDHW with temperature 42 °C and 47 °C on secondaryside outlet from the moment when DHW tap is opened.The measurements were performed for different initialconditions and secondary flowrate was always8.4 L/min, which is nominal flow for shower.1. For measurements of concept with external by-pass,substation was controlled by PTC2+P controller withby-pass set point temperature adjusted to 35 °C. Thissetup is exactly the same as is installed in Lystrup pilotproject. The testing procedure was made in followingsteps. Substation was left idle for long time in thetesting room, so all components and water in HEXwere on room temperature. Than we opened the valveon DH supply in substation and DH water withtemperature of 51 °C started to flow in the substationand flew through external by-pass, until closingtemperature was reached and by-pass flow stopped.Then we wait until by-pass was opened again. Timebetween two by-pass openings as well as volume andtemperature of DH water passed through by-pass waswritten down and after by-pass was closed again, wewaited a little bit shorter time than was needed to openby-pass flow again and we start tapping on secondaryside with flow rate 8.4 L/min. In this way, mostunfavourable condition for substation with by-pass, i.e.highest recovery time, was measured. After tapping ofDHW was finished, we wait 5 minutes and weperformed one more tapping to simulate short timestep between two subsequent tapping of DHW.2. For measurement of internal bypass concept, IHPTcontroller was used. In case of IHPT, by-pass set pointtemperature can‘t be adjusted independently and isdefined by desired temperature of DHW, i.e. 47 °C forour measurements. IHPT controller was developed fortraditional DH networks operating with forwardtemperatures around 70 °C. For traditional DH, bypassopens when temperature in HEX falls 5–7 °Cbelow set point of DHW, but in case of LEDH withforward temperature 51 °C, by-pass opens 1 °C belowDHW set point temperature, i.e. 46 °C in our case.The testing procedure was similar to measurementswith external by-pass. After supply valve on primaryside of substation was opened, DH water withtemperature of 51 °C started to flow in the substationand temper HEX, until by-pass closing temperaturewas reached. Then we wait until by-pass was openedagain and we performed tapping of DHW just beforenext by-pass opening was expected. In following stepswas procedure same as in case of external by-pass.Moreover, we also performed measurements of timedelay in IHEU for control concept without by-pass.RESULTSTime delay for IHEU with PTC2+P controller andexternal by-pass adjusted to 35 °C to start supplyDHW water with temperature 42 °C and 47 °C afterlong idling period just before opening of external bypasswas expected, can be seen from Fig. 4 and is 11and 22 seconds, respectively. This measurementrepresents condition with the longest time delay forPTC2+P controller. Temperature of room, where IHEUwas installed was 22.2 °C. For this case, temperaturesof produced DHW in first 10 sec after tapping wasstarted are listed in Table 3.65

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 4 Time delay for external bypass (PTC2+P), when tapping is performed just before expected start of by-pass flow, seton 35 °C.In case, when tapping of DHW was performed afterlong idling just after by-pass flow was stopped, timedelay decreased to 8,5 and 16,5 seconds. In thismeasurement, temperature of substation and thuswater standing in the HEX was little higher thanambient air temperature. It is expected that time delaywill be slightly longer, if substation will have realambient temperature but still shorter than in case 2.We also performed measurement of tap delay fiveminutes after previous DHW tapping was finished.In this case, tap delay in substation to produce DHWwith temperature 42 °C and 47 °C was shorter, 7 and14 seconds.For room temperature around 22 °C, external by-passwas opened roughly every 30 minutes. The by-passwas in average opened 2.5 minute and volume of DHwater needed to close the by-pass was in average 3 L,i.e. when substation is idle, by-pass uses 6 L of DHwater per hour.Table 3 – Temperatures measured for PTC2+P controller in first 10 sec after tapping was started for situation after longidling, just before by-pass was expected to run again(sec) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17T22 (°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 45.5Time delay in IHEU equipped with IHPT controller withinternal by-pass adjusted by requirement of DHW to47 °C was 6 and 14 seconds to reach 42 °C and 47 °Con outlet for situation when tapping was performed justbefore by-pass was expected to open. The internal bypassopens 3 minutes after previous tapping is finishedand when is once opened never closes, only whenanother tapping is performed, but again only on3 minutes.Table 4 – Overview of time delays for all measured casesThe average flow of internal by-pass was 24 L/hourand average return temperature to DH network was45 °C. When internal by-pass is once opened, the timedelay in substation decrease substantially to 1.5 and7 seconds to produce DHW with temperature 42 °Cand 47 °C. The condition with expected longest timedelay was solution without by pass. In this case timedelay to produce DHW with temperature 42 °C and47 °C was 12 and 25 sec. All measured results aresummarized in Table 4.NO BYPASSEXTERNALBY-PASSINTERNALBY-PASScase number and descriptionT 11(°C)τ 42(sec)τ 45(sec)τ 47(sec)T 12(°C)T 12AVG(°C)T HEX-UP(°C)T HEX-DOWN(°C)1 – after long idling, no by-pass (BYP)50.1 12 18 25 16.2 19.5 20.4 212 – after long idling, just before BYP wasexpected to open again49.6 11 16 22 30.1 19.3 21.5 21.43 – after long idling, just after BYP closed 50.6 8.5 12 16.5 42.6 19 29 264 – 5 minutes after previous tapping finished 50.8 7 10 14 25 19.1 22.3 37.45 – just before BYP was expected to open (3min after prev. tapp. finished))50.5 6 10 14 19.5 19.1 22.6 386 – anytime, when BYP was already inoperation49.3 1.5 3.5 7 47.3 18.4 44 45.566

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDISCUSSIONFocused on level of users comfort and proper coolingof DH water during idling, time delay of LEDHsubstation to supply DHW with temperature 42 °C and47 °C was measured. Three different control strategiesrelated to tap delay were investigated. Obtainedresults represent case of IHEU used in single-familyhouse in period when space heating is not inoperation. Explored concepts can be evaluated fromtwo different points of view, due to highest advantagesfor customer and for DHN.The solution without by-pass is from energy savingspoint of view very interesting because doesn‘t needany DH water for idling, but from users comfort point ofview is very poor because of reduced comfort andproblems with wasting of water during waiting for DHWwith desired temperature. Solution without by-pass canbe probably used for substations equipped withcombined thermostatic and proportional flow controller,for customers with short branch pipes or for customerswith low requirements for level of users comfort. Ifsolution without by-pass will be used for substationcontrolled only with proportional flow controller, eventransport delay in 10 m long branch pipe for nominalflow for basin will be 32 sec. For period when spaceheating is operated, branch pipe will be kept warmfrom flow needed for space heating and time delay forsolution without by-pass will be very similar to solutionwith external by-pass. Anyway, in non-circularlyshaped DH networks, by-pass should be installed atleast at the end of a street, so it is better to findsolution how to use by-pass flow in useful way thansent it directly back to DH return. Considering this, it issuggested to use by-pass flow for whole yearoperation of floor heating in bathrooms to increasecomfort for customers and at the same time solveproblem with by-pass flow which otherwise increasingreturn temperature to DH network.From user comfort point of view, better solution thansolution without by-pass, but consuming more energy,is substation equipped with external by-pass. Bycomparison of results of concepts without by-pass(case 1) and solution with external by-pass, for casewhen tapping is performed after long period of idlingjust before by-pass opens again (case 2), we can seethat time delays are almost the same (see Table 4).Difference is only that for external by-pass are pipes inDH substation kept on higher temperature and it madeslightly faster reaction. In the case 3, time delay iseven more reduced since pipes in substation werewarmer by just finished by-pass flow. For controlconcept with external by-pass and tapping repeated5 minutes after previous one, time delay is againreduced, since HEX is still hot from previous tapping.The time delay for case 4 and 5 are almost the same,only difference is that in case 5 (internal by-pass), tapdelay is again reduced because tapping wasperformed 3 minutes after previous (to preventinfluence of by-pass) and thus HEX was warmer.If the requirement is to fulfil 10 sec tap delay for lessfavourable fixture, i.e. in our case basin (see Table 1),DHW should leave DH substation with temperature42 °C in 4 sec after tapping was started, because it willtake 6 second to reach the tap. This requirement wasreached only by concept with internal by-pass and onlywhen by-pass was already opened. On the other handfrom Table 3can be seen, that even for concept withexternal by-pass and tapping after long idling and justbefore expected bypass opening, DHW at atemperature 26 °C leaving substation in 3 sec. DHWwith this temperature is not sufficient for taking acomfortable shower for which temperature 37±1 °C ispreferred, but for washing hands this temperatureshould be enough. The values in Table 3 are for flowrate used for shower, but it can be used to explain thatit is time to rethink the suggested value of tap delayfrom 10 sec to another value and consider alsonominal flows and use of tapped water. The differentstandards for the different use of DHW based on newsolutions in DHW supply systems and results from testpanels are needed, because it may have someinfluence on design of optimized DHW systems.Nevertheless, for customers requiring DHW in veryshort time e.g. continuously or discontinuously (onlyduring rush hours) operated trace heating elementscan assure almost no tap delay by keeping DHWstaying in pipes on desired temperature.CONCLUSIONBased on literature study it can be concluded thathygienic requirement of DHW with 50 °C on outlet ofDHW heater is not needed for systems with a totalvolume of the DHW lower than 3 L.From results of our measurements and evaluation ofIHEU supplied by LEDH, only substation with externalby-pass with set point 46 °C is able to produce 47 °CDHW in time bellow 10 sec. The easiest step how todecrease waiting time also for other concepts is toinsulate HEX. This measure will reduce time delay forDHW tapping and also will decrease heat losses fromDH substation. The lower waiting times for DHW canbe also achieved by further optimisation of HEX in wayof decreased number of plates reducing volume ofwater in HEX and thus transport delay, and byincreased thermal efficiency of HEX (followed on theother hand by higher pressure loss). Thesemodifications can lead for higher temperature of DHwater returning to DH network, but during all ourexperiments, average return temperature was below20 °C, what is 5 °C less than is designed for LEDH.67

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTraditional control concepts of DH substations arealways trade-off between users comfort and reducedcooling of DH water during idling and thus customershould have to some extent possibility to choose whichsolution prefers. In case of traditional concepts,decision is between longer waiting time for DHW andenergy savings or vice versa, if by-pass in substationis used. In non-circularly shaped networks, by-passshould be used anyway at least at the end of a streetline. The one of possible solutions how use by-passflow in better way can be proposed innovative conceptof whole year operated floor heating in bathrooms orsupply-supply recirculation. Both solutions will increaselevel of user comfort and at the same time also energyefficiency of DH system.LEDH is a promising solution for providing buildingswith DHW and space heating regarding fulfillingrequirements of modern society with reduced CO 2emissions and energy consumption. More detailedinvestigations by testing of different parameters andnumerical simulations are needed in order to optimizeLEDH concept.Future workIt will be very interesting to compare time delay ofsubstation for traditional DH with time delay for DHWproduced by LEDH substation. It is expected thattimed delay for LEDH will be higher because dynamicresponse is slowed down by lower temperaturedifference between DH water and desired temperatureof DHW, but on the other hand, lower temperaturedifference is in some extend compensated by biggerHEX. It is also suggested to rethink ―10 sec tap delaysuggestion‖ for different taping flows and purposes ofDHW use.REFERENCES[1] S. Froning, ―Low energy communities with districtheating and cooling‖, PLEA 2008 – 25thConference on Passive and Low EnergyArchitecture, Dublin,[2] Hovedrapport, ―Udvikling og Demonstration afLavenergifjernvarme til Lavenergibyggeri‖ 2009,(in Danish)[4] DVGW, ‖W551 - Trinkwassererwärmungs- undTrinkwasserleitungsanlagen‖ ,1993, Bonn, (inGerman)[5] Z. Liu, ―Effect of flow regimes on the presence ofLegionella within the biofilm of a model plumbingsystem‖, 2006, Journal of Applied Microbiology,Vol. 101, pp 437-442[6] F. Martinelli, ―A Comparison of Legionellapneumophila Occurrence in Hot Water Tanks andInstantaneous Devices in Domestic, Nosocomial,and Community Environments‖, 2000, CurrentMicrobiology, Vol. 41, pp. 374-376[7] W. Mathys, J. Stanke, et. al.,‖ Occurrence ofLegionella in hot water systems of single-familyresidences in suburbs of two German cities withspecial reference to solar and district heating‖,2008, Int. J. Hyg. Environ. Health, Vol. 211, pp.179-185[8] H. Kristjansson, ―Distribution Systems inApartment Buildings‖, Published at the 11th<strong>International</strong> <strong>Symposium</strong> on District Heating andCooling, August 31 to September 2, 2008,Reykjavik, ICELAND[9] T. Persson, ―District Heating for Residential Areaswith Single-Family Housing, paper IV‖, 2005,Doctoral Thesis, Lund Institute of Technology,Lund[10] Dansk Standard, ―DS 439 Code of Practice fordomestic water supply installations‖, 2009[11] P.K. Olsen, ―Low-Temperature District HeatingSystem for Low-Energy Buildings‖, 2009,http://www.fbbb.dk/Files/Filer/Peter_Kaarup_Olsen_-_COWI_29-10_2009.pdf[12] H., Boysen, J.E. Thorsen, ―Control Concepts forDH Compact Stations‖, Published in Euroheat andPower IIII 2004[13] O. Paulsen, ―Consumer Unit for Low EnergyDistrict Heating Net‖, Published at the 11th<strong>International</strong> <strong>Symposium</strong> on District Heating andCooling, August 31 to September 2, 2008,Reykjavik, ICELAND[3] EUROHEAT & POWER, ―Guidelines for DistrictHeating Substations‖, 2008, downloaded fromwww.euroheat.org in October 2009, pp 868

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaCHALLENGES ON LOW HEAT DENSITY DISTRICT HEATING NETWORK DESIGNM. Rämä 1 and K. Sipilä 11VTT Technical Research Centre of Finland PB 1000, FI-02044 VTT, FinlandABSTRACTWhile district heating is an energy efficient solution toprovide heating to areas with high heat consumption,mature systems extending out to more demandingoperational environment face challenges maintainingcompetitiveness over alternative heating systems. Asthe heat density falls below a certain level, districtheating is no longer economically feasible. Studying thepossibilities of extending this threshold by means ofdistrict heating system design and pointing out theoperational challenges while approaching it are themain topic of this paper.The problem is investigated in a representative case ofa low heat density area bordering a more extensivedistrict heating network. A node-and-branch typenetwork simulation model is used study the operation ofthe network and a simulation period of one year is usedto get a realistic view of the system in a normaloperational cycle.Not taking into account the characteristics of a low heatdensity area in network design can result in inefficientdistribution system. Operational problems, especiallymaintaining the temperature level in summertime, mustbe solved. Only concentrating on minimizing the heatlosses will not result in best possible design.The temperature level issue can be solved with a bypassvalve, auxiliary heating or accumulators, but inoverall more efficient system requires steps to be takenin the houses. Floor heating and a heat pump coupledwith an accumulator enables the use of low temperaturedesign where the heat losses can be cut significantly.heating. The expansion of mature and large scalesystems take place in areas with lower heatconsumption. This transition to more demandingoperational environment both technically and financiallyrepresents challenges to district heating network design.This is also true in small scale systems of limitedconsumption separated from a larger system.A careless network design in these circumstances canlead to deterioration of the advantages of districtheating; efficiency and reliability. An annual heat loss of5% in district heating distribution is considered a goodresult, but the case in question the heat losses caneasily reach 10% or even tens of percents if thecharacteristics of low heat density areas are not takeninto account in design.LOW HEAT DENSITY AREAA detached house area consisting of 56 identical 150m 2 houses with energy consumptions in compliance oftoday‘s building standards is studied. Dedicated heatexchangers between the network and the consumerexist for both heating and domestic hot water. Totalenergy consumption for the houses is 18.75 MWh/yearof which domestic hot water has a share of 20 percent.The district heating network studied is presented inFigure 1. The detached house connections are markedas green dots and the connection to the main districtheating network as a red rectangle. The connectionshave 1, 2 or 6 detached houses as consumers,indicated by the size of the dot.INTRODUCTIONDistrict heating remains to be one of the most efficientalternatives to provide heating mostly due to its hightotal efficiency especially when utilizing combined heatand power production or waste heat from industrialfacilities or other sources. A wide choice of productiontechnologies, based on fossil or renewable fuels orother sources of heat, provide flexibility to districtheating systems and enable the benefits from theeconomy of scale unlike most consumer specificheating systems. From the consumer point of view,district heating is considered as a reliable and carefreesource of heating energy and is also often aneconomically sound choice.Areas with high heat consumption i.e. economically themost attractive areas will be connected first to district50 mFigure 1. District heating network studied.The total trench length in the area is 2 390 m of whichthe service pipes (DN 15-25) account for 1 300 m. Thepipe size distribution is illustrated in Figure 2. The darkblue coloured bar (DN 65) represents the pipeconnecting the area to the main district heating network.69

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAs the pipe diameters are quite small, twin pipes withinsulation class IV are used in the area asrecommended by Energy Industry [1], [2] in Finland.The pressure drop design principle used here is roughly~1.5 bar/km.Pipe lenght (m)700600500400300200100015 20 25 32 40 50 65Pipe size (DN)Figure 2. Pipe size distributionThe linear heat density is 0.44 MWh/m which makes thearea a low heat density area by definition [3].The heat demand around the year is presented inFigure 3. The peak demand for the area is 507 kW. Asexpected, in the summertime the load consists almostsolely of domestic hot water consumption.Total heat demand (kW)60050040030020010000 50 100 150 200 250 300 350DaysFigure 3. Heat demand of the simulated area.SIMULATION MODELA node-and-branch type simulation model [4] was usedto study the case in hand. The model calculatestemperatures and pressures for the nodes and flowsand heat losses for the pipes, i.e. the branches. Fromthese results pumping power can also be calculated,although a constant efficiency of 0.5 is used for thepump. The pressures are calculated separately fromtemperatures. The temperature calculation is dynamicwhile the flow and pressure calculation is not. Aminimum 0.6 bar pressure difference over a consumeris assumed.When defining the network, each pipe is given a startand an end node, a pipe type (twin, single), aninsulation standard (class I to IV) and length.The consumptions for both heating and domestic hotwater use were given as hourly time series as well asthe radiator supply and return temperatures on thesecondary side.The heat exchangers were modelled with logarithmictemperature principle in a design point (described inTable 1) after which the conductance in W/K isassumed to be constant. When heat demand, bothsupply and return temperatures on secondary side andsupply temperature on primary side are given as input,the primary return temperature and district heating massflow can be calculated.Table 1. Design point for heat exchangers.DescriptionValuePrimary side temperatures 115/45 °CRadiator heating 70/40 °CDomestic hot water 55/10 °CDesign heating loadDesign DHW load8 830 W2 060 WThe design loads for domestic hot water are lowcompared to a real life design load of a heat exchangerin normal detached house in Finland, 50 kW is acommon choice. This is due to the simulation modeltaking hourly data originally calculated for a multifamilyhouse as input so the domestic hot water demand isalso flatter than it really is. However, from the networkdesign point of view hourly data is considered accurateenough.Other input data used were the undisturbed groundtemperature of 5 °C, assumed to be constant, and thesupply temperature from the main district heatingnetwork as a function of outdoor temperature. Theoutdoor temperature time series used described atypical year in Southern Finland. The supplytemperature reaches its maximum value of 115 °C in anoutdoor temperature of -26 °C and its lowest value of75 °C in 5 °C. Between these two points, the relation islinear.SIMULATION RESULTSThe most interesting results concern the heat lossesand the temperature variations within the network. Thepumping needed (less than 1 MWh) in a network of thissize is quite low and thus negligible.In the initial simulation runs it was noted that the systemwas struggling to maintain high enough temperaturelevel in the summertime when the load consist solely ofdomestic hot water demand. This problem was met bydefining a flow through valve at the consumer, opening70

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniawhen the supply temperature on primary side droppedtoo low (< 65 °C). The valve allowed a constant massflow (0.015 kg/s) to go past the heat exchanger on theprimary side. This solution helped the situationsignificantly although not without ill effects as can beseen from the heat losses presented below.The use of a by-pass valve to ensure the appropriatetemperature level for domestic hot water also meanhigher heat losses and pumping power and effectivelylower cooling; all of which are undesirable outcomes.One possibility to solve the problem is just to accept theflaw and to use additional electrical heating element toraise the temperature of domestic hot water to therequired level. As the temperature boost needed is formost of the time quite small and is only needed insummertime, the increase in electricity consumption isreasonable.Because of the high capital costs of district heating, thepipes should basically be sized as tight as possiblewhile keeping in mind the future demand for the pipelinein question. As the pipes are small, the volume of watercontained is also low. This leads to water cooling morerapidly than in larger pipes. The Figure 4 illustrates thiswith a simplified example by showing the temperatureon supply side service pipes if there is no flow for threedifferent pipe sizes. The temperature drop of 15 °C, forexample, takes 5 times longer with a pipe size DN 50than with a small DN 15 pipe. The calculations assumea constant return side temperature of 30 °C and aground temperature of 5 °C.Temperature (°C)70605040302010DN 15 DN 25 DN 5000 2 4 6 8Time (h)Figure 4. Temperature drop in three pipe sizes when noflow is introduced.The use of smaller pipes reduces the heat losses inW/m and this is accentuated if the temperature leveldrops as described above. As a result, looking solely onheat losses when designing a low heat density areanetwork on common design principles can lead toreliability issues as the system cannot supply the heatrequired by the consumers.The relative heat losses (that is, heat losses per neededproduction) for the simulated case are 13.8 % in a year.The monthly values can be seen in Figure 5. While the71relative heat losses in the heating season areacceptable, they reached 47 % in the summertime. Thehigh heat losses are partly because of the by-pass valveletting hot water past the heat exchangers. The by-passvalve is also responsible for small cooling, i.e. thedifference between supply and return temperatures,within the system in summertime (Figure 6).Relative heat losses (-)1.00.90.80.70.60.50.40.30.20.10.0I II III IV V VI VII VIII IX X XI XIIMonthFigure 5. Monthly relative heat losses.Cooling (°C)90807060504030201000 50 100 150 200 250 300 350DaysFigure 6. Difference between supply and returntemperatures at the border of the area.The most obvious way to cut heat losses in alreadyreasonable insulated network is to lower the supplytemperature. In the simulated system, this would causeproblems because aforementioned issues concerningdomestic hot water demand in summertime, and duringthe heating season because of the traditional radiatorheating design temperatures of 70/40 °C. However, ifmore significant changes would be possible, a floorheating system and a heat pump coupled with anaccumulator handling the higher temperature levelrequired domestic hot water would enhance theefficiency of the distribution system at a price of a verymodest increase in electricity consumption and higherinvestment costs for the consumer because of theaccumulator, heat pump and floor heating. If thedomestic 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

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia1.25 MWh with an average COP of 3. With this setup,supply temperature would need to be just 40 °C.CONCLUSIONSThe use of traditional district heating network designprinciples can lead to an inefficient area heating systemin areas with low heat density. Special attention must bepaid on operation of the system to ensure reliability, oneof the advantages of district heating.When aiming for an efficient system, one goal is tominimize the heat losses. However, concentrating solelyon this can make another problem, maintaining highenough temperature level for domestic hot water insummertime, even worse. The problem can be solvedusing a by-pass valve, but this causes unwantedeffects; worse cooling and an increase in heat lossesand pumping power. Other solutions are auxiliaryheating (electrical heating or a heat pump) or the use ofan accumulator and with it, aiming for a steady domestichot water load.REFERENCES[1] Lappeenranta University of Technology,Kaukolämpöjohtojen optimaalisen eristyspaksuudentarkastelu / Investigation of the optimalinsulation thickness on district heating pipes,Energy Industry, 2009, 36 p.[2] Preinsulated district heating pipes,Recommendation L1/2010, Energy Industry, 2010,44 p.[3] Zinko, H., Bøhm, B., Kristjansson, H., Ottoson, U.,Rämä, M., Sipilä, K., District heating distribution inareas with low heat demand density, IEA DHCAnnex 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 temperaturesignificantly and to use floor heating and heat pump withan accumulator for domestic hot water demand. This isnot suitable for existing areas with a heating systemalready designed, but for new areas it is a reasonableand, compared to the traditional district heating design,an efficient way to provide heating.72

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDESIGN OF LOW TEMPERATURE DISTRICT HEATING NETWORK WITH SUPPLYWATER RECIRCULATIONHongwei Li 1 , Alessandro Dalla Rosa 1 , Svend Svendsen 11 Civil Engineering Department, Technical University of DenmarkABSTRACTThe focus on continuing improving building energyefficiency and reducing building energy consumptionbrings the key impetus for the development of the newgeneration district heating (DH) system. In the newgeneration DH network, the supply and returntemperature are designed low in order to significantlyreduce the network heat loss. Meanwhile, the lownetwork operational temperature can make a betterutilization of renewable energy and further improve theCHP plant efficiency.Though the designed return temperature is low, it mayincrease considerably when the heating load becomeslow and the by-pass system starts to function. The aimof this paper is to investigate the influence of by-passwater on the network return temperature and introducethe concept of supply water recirculation into thenetwork design so that the traditional by-pass systemcan be avoided. Instead of mixing the by-pass waterwith return water, the by-pass water is directed to aseparated circulation line and returns back to the plantdirectly. Different pipe design concepts were tested andthe annual thermal performances for a selectedresidential area were evaluated with the commercialprogram TERMIS. The simulation program calculatesthe heat loss in the twin pipe as that in the single pipe.The influence of this simplification on the supply/returnwater temperature prediction was analyzed by solvingthe coupled differential energy equations.INTRODUCTIONIn European Union, one of the major energydevelopment targets is to reduce the building energyconsumption and increase the supply of renewableenergy. The introduction of European EnergyPerformance of Building Directive (EPBD) posesstringent requirement for the member countries toeffectively reduce their building energy consumption.According to the national energy policy, the buildingenergy consumption in Denmark will drop to 25% ofcurrent level by the year 2060, while the renewableenergy share will increase from 20% to 100% at themeantime [1].District heating (DH) benefits from economic of scalewith mass production of heat from central heatingplants. The significant reduction of building energyconsumption and wide exploitation of waste heat andrenewable energy, however, makes the current DH73technologies become barriers to further increase themarket share [2]. In order to sustain the economiccompetiveness and realize the long term sustainabledevelopment, the concept of design and operation ofDH system needs to be re-examined under the newenergy regulation and development trends. This is themain impetus for the development of the newgeneration DH system. Based on previous studies, in aproperly designed in-house substation system, thenetwork supply temperature at 55oC and returntemperature at 20oC can meet the consumer spaceheating and domestic hot water demand [3].The low return temperature has the advantages toreduce the network heat loss, increase CHP plantpower generation capability, and utilize direct flue gascondensation for waste heat recovery. However, thereturn temperature can become much higher than thedesigned value when the heating load becomes lowand the by-pass system at the critical user starts tofunction. In this paper, the influence of by-pass wateron network return temperature was examined for areference residential area. The concept of supply waterrecirculation was introduced to avoid the mixing of bypasswater and the return water. Three network designmethods were tested. The annual thermal performancewas evaluated with the commercial district heatingnetwork hydraulic and thermal simulation softwareTERMIS [4]. The simulation program calculates theheat loss in the twin pipe as that in the single pipe. Theinfluence of this simplification on the supply/returnwater temperature prediction was analyzed by solvingthe coupled differential energy equations.SUPPLY WATER RECIRCILUATIONThe solution to overcome the excessive temperaturedrop along the supply pipe due to reduced flow rate isto install by-pass system at the critical user in thenetwork. Figure 1 shows the principle of supply waterby-pass. Extra flow is called based on the temperaturemeasurement at the critical user until the minimumsupply temperature requirement is met. This extra flowis then ―by-passed‖ and sends back to the return pipe.As the by-pass flow rate may be considerable and itstemperature is high, the mixing with return water willsignificantly increase the return water temperaturewhich causes both increased heat loss in the returnpipeline and decreased power generation capability inthe CHP plant.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaA desirable design approach is to maintain the by-passsystem as the flow rate adjuster, while avoids themixing of the by-pass water and the return water. Thisdesign concept is schematically shown in Fig. 2, whichis realized through adding a third pipeline for supplywater re-circulation. When the by-pass water is called,the circulation line will transfer the extra supply waterback to the plant where it is re-heated up to the supplytemperature again. On the other hand, the addition ofthe 3rd pipeline provides the possibility to supply waterin two supply lines when the heat demand is high. Thenetwork, therefore, can be designed as two supplylines with reduced diameter together with one returnline.Fig. 3 Annual heating load (blue columns) and durationhours (red curve) at different ground temperatureNETWORK SIMULATIONFig. 1 Schematic for hot water by-pass systemFig. 2 Schematic for by-pass water recirculationHeating LoadThe simulation was performed for a reference area with81 low energy demand houses. The house wasdesigned based on the building standard Class 1,following the Danish Building Regulation. The domestichot water draw-off profile was designed similar to theDanish standard DS439 [5]. Detailed space heatingand domestic hot water heating load simulation can befound from [6, 7]. Figure 4 shows the averaged heatingload and the corresponding duration hours. The annualheating load is divided into 8 intervals, varying as afunction of undisturbed ground temperatures whichranges from 0 to 15 ºC. The summer season lasts 3281hours and the heating load comes only from thedomestic hot water demand. The space heating isrequired for the rest of the year.House InstallationsTwo house installations were considered in this study.Figure 4 shows the instantaneous heat exchanger (HE)in the DH system. Without a buffer tank, the branchpipe which connects directly to the HE installation musthave the capability to supply the instantaneous hotwater demand without causing significant pressuredrop, which otherwise can be compromised byinstalling a booster pump. The HE design load is 32kWper houses at the network supply temperature 55oCand return temperature 22 ºC. On the other hand,simultaneous factors which are the probabilities formultiple users‘ concurrent use of hot water areconsidered for the design of street pipes and mainpipes, as shown in Table 1 [3]. Fig. 5 shows thedomestic hot water storage tank (DHWS) in the DHsystem. The DHWS design load is 8 kW per house. Toavoid the legionella problem, the design temperaturefor DHWS is higher than HE, at 65 ºC /30 ºC for supplyand return respectively.74

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 4 In-house heat exchanger (HE) in DH systemTable 1 Simultaneous FactorsDistrict Heating NetworkThe DH network and the connection to the end usersare shown in Fig. 6. The house is designed to connectto the plant directly through different diameter pipeswhich were optimized with the simulation program. Thedirect connection allows the primary DH network tocirculate water directly into the end user installation. Itis suitable for a moderate pressure level network andthe differential pressure of DH network is sufficient tocirculate water to the house installation. The networksand house installations are assumed to withstandmaximum pressure 10 bar. The consumer differentialpressure is set as 0.5 bar. It is controlled at the enduser along the network critical route which is shown ingreen color.Three network design scenarios were investigated foreach house installation:Case 1: It is the reference case. The totalnetwork length is 3080 m and the network lineheat density is 177 kWh/year. Network wasdesigned in the traditional way for two pipeswith one supply and one return, respectively.The differential pressure is controlled at userA. Twin pipes were selected for theDH network. They are called ―reference pipe‖in this paper.Case 2: By-pass water recirculation. A thirdpipeline (Fig. 6 grey color line) was introducedto separate the by-pass water with returnwater and re-circulate the by-pass water backto the plant. The third pipeline was sizedbased on the summer by-pass water flow rate.The differential pressure is controlled at pointB.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 2functions all year round. It acts as supply pipeduring winter season and functions as supplywater recirculation pipe when there has bypasswater demand. In this case, the mainpipe in the reference case was resized as aportion of supply water is shared by therecirculation pipe. The connection ofrecirculation pipe to the reference pipe isshown with red color.Fig. 5 Domestic hot water storage (DHWS) in DH systemThe thermal by-pass temperature was set as 50 °C forHE and 60 °C for DHWS with dead band 2 °C. The bypassis placed on the end user at each street in case 1,while at the virtual point adjacent to the end user incase 2.75

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaNetwork Heat Loss CalculationThe reference network was designed with twin pipes byplacing the supply and return pipe in the same casing.Two types of twin pipes were considered in thesimulation: AluFlex multilayer flexible pipe and straightsteel pipe. The pipes were selected with continuousdimension ranging from Alx14 to 32 for AluFlex pipeand DN 32 to DN40 for steel pipe, based on the marketavailable products [8]. Single AluFlex pipe is selectedfor the 3rd recirculation line. This 3rd pipeline can beassumed being placed in the same trench along thetwin pipes. The thermal interaction between the twinand the single pipe is assumed negligible.The heat loss in the twin pipe was calculated accordingto the reference [7,9][1][2]coefficients corresponding to the temperaturedifference between the flow and the ground.The temperature variation along the pipeline wascalculated as internal flow with isothermal boundarycondition. The downstream temperature in the pipe isexpressed as [4]:T d , T u and T a represent the downstream fluidtemperature, upstream fluid temperature, and ambienttemperature respectively. M and K are parametersinclude the overall heat transfer coefficient. As theoverall heat transfer coefficients have to be calculatedbeforehand, the influence of flow temperature variationon U s and U r along the pipeline is neglected. It is areasonable assumption when the thermal by-passtemperature is set close to the plant temperature,however, may cause appreciable errors if thetemperature drop along the network is high.It is worth to be noted that though the design returntemperature (22 o C) is higher than ground temperature,the net heat transfer in the return pipe may absorb heatfrom surrounding which makes U r negative. However,negative U r has to be set to zero as the simulationprogram cannot handle negative heat transfercoefficient.[3]RESULTS AND DISCUSSIONFig. 6 District heating networkThe supply and return pipe are assumed identical andplaced horizontally in the same depth from the ground.The linear thermal transmittance U ij reduces toU 11 =U 22 =U 1 and U 12 =U 21 =U 2 . In addition, the thermalconductivity of insulation foam was assumed constant.U 1 and U 2 were then calculated with the analyticalsolution developed from the multi-pole method [10].The simulation program cannot handle two heattransfer coefficients in the same pipe, U s and U r werederived to represent the overall heat transferHeat ExchangerNetwork simulation starts from proper selection of pipedimension, based on the design condition and thedesign criteria introduced in the previous section.Table 2 shows the selected pipe types andcorresponding length for three different cases. Case 1is the reference case. Flexible twin pipe Alx 20 to 32and steel twin pipe DN32 and DN 40 were selected.The third recirculation pipe was designed in case 2based on the summer by-pass flow rate. Pressuregradient 1500 pa/m for street pipes and 500 pa/m formain pipes were set as the dimension criteria. Thoughsmaller pipe was suggested by the program, the Alx16single pipe was selected as the minimum diameter pipeavailable on the market. It was assumed that therecirculation pipe can be used as water supply in winterin case 3. Therefore, the main pipes in the referenceline were re-designed with considering that a portion ofsupply water goes through the recirculation line. It canbe seen that the supply pipe has smaller diameter thanreturn pipe in some sections in the twin pipe line.76

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 2 Selected pipe types and length in Case 1–3Figure 7 shows the pressure profile along the criticalroute. The network is designed for a 10 bar system.The minimum network static pressure is 2 bar and theminimum differential pressure at consumer is 50 kPa.The plant static supply pressure is 853 kPa in case 1 atdesign condition. In case 3, the designed plant supplypressure head rise to 917 kPa, which is due to theincreased flow rate indicated in Table 4. The pressuredrop along the reference line during summer is quitelow due to the reduced flow rate. However, extrapressure head has to be applied to overcome thepressure loss along the recirculation line in Case 2.The required static supply pressure is 800 kPa duringsummer as a result of small dimension recirculationline.Case 2 has higher return pipe heat loss comparing withcase 1 due to the introduction of recirculation line. Atconstant supply temperature 55 ºC, the heat transfercoefficient Us decreases with increase the return watertemperature. As shown in Table 3, the return watertemperature in case 2 (at 22 º C) is lower than that of incase 1. This leads to a higher heat loss in the supplypipe in case 2. As a consequence, more by-pass flowis required to compensate the extra supply pipe heatloss, therefore, the by-pass flow rate in case 2 is higherthan in case 1 in the summer season.Supply water in the recirculation pipeline in winterincreases the supply pipe heat loss in case 3. Theconcept of double pipe supply may not economicalfeasible, according to the simulation results. However,it may be used as an alternative solution to supplywater in the 3 rd pipeline under extreme whethercondition, which otherwise has to raise the plant supplytemperature to meet the increased heating demand.Furthermore, results in table 4 were limited to fixedrecirculation pipe diameters. The double pipe supplyconcept may be economical feasible by free selectionboth reference pipe and recirculation pipe diameterwith the objective to minimize the annual networkoperational cost or exergy consumption. This study isout of the scope of current paper due to the limitation ofthe simulation program.Table 3 shows the simulation results for case 1.By-pass is required when the heating load is smallerthan 1.53 kW. The return water temperature increasesalong with the increase of by-pass water flow rate. Insummer, the amount of by-pass water flow rateexceeds the actual flow rate passing through theconsumer, and the return temperature at the plantincreases up to 35.5 ºC. The heat loss in the returnpipe is accounted when the plant return temperature israised to higher than 30 ºC.Simulation results for case 2 and case 3 are shown inTable 4. They were put in the same table as case 2operates when there has by-pass requirement, whilecase 3 operates in the rest seasons. Italic is used forcase 3 to distinguish the two cases. Thanks to therecirculation line, the return temperature at the plant inthe reference line remains low at 22 ºC, while the returntemperature in the recirculation line can reach 44 ºC inthe summer, after deducting the single pipe heat loss.The low plant return temperature can help extract morepower in the CHP plant or be used in othercircumstance like direct flue gas condensation. On theother hand, high temperature return water in therecirculation pipe can be re-heated by an additionalheat exchanger or boiler with minimum energy input.Fig. 7 Pressure profile on the critical route in Case 1–377

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 3 Simulation results in Case 1Table 6 Simulation results in Case 1Table 4 Simulation results in Case 2 (First 5 rows) andCase 3 (Last 3 rows with italic)Table 7 Simulation results in Case 2 (First 5 rows) andCase 3 (Last 3 rows with italic)Domestic Hot Water Storage TankTable 5 shows the pipe types and corresponding lengthin the DHWS installation. Alx 14 was selected asbranch pipe due to the smaller design heating load.Similar to the HE, the by-pass flow rate exceed theactual flow rate through the consumer in summerseason. The plant mixed return water temperature incase 1 is 46 o C. The introduction of the recirculationline can keep the plant return temperature in referenceline as low as 30 o C, while increases the returntemperature in the recirculation pipe to 54 o C at theplant. Extra heat loss has to be tolerated due to therecirculation pipe in both case 2 and case 3.Table 5 Selected pipe types and length in Case 1–3Further Discussion on Heat TransferAs shown in Eq. 1–3, the simulation program simplifiesthe calculation of the heat loss in the twin pipe as thatin the single pipe. The influence of the adjacent pipewas accounted through converting the linear thermaltransmittance U ij to the overall heat transfer coefficientsU s and U r , with pre-assumed constant networksupply/return temperatures. To assess the influence ofthis simplification on the temperature predication, thethermal interaction between the supply and returnpipes was calculated by solving the coupled pipe heattransfer differential equations. The governing equationsfor supply and return pipes can be expressed as:[4][5]The boundary conditions can be expressed as:The dimensionless temperature is introduced with:[6][7]78

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe governing equations then change to:[8]Table 8 Pipe temperature predication comparison (supplyoutlet temperature is controlled at 50 o C)WhereThe boundary conditions change to :[9]The system linear ordinary differential equations can besolved with Eigen value method or with Laplacetransformation. The Laplace transformation wasapplied in this study. Eq. 8 is transformed to:The final solutions are given as:Where :[10][11][12], [13], [14]Tws- DN32, which is the longest main pipe in HE ofcase 1, is selected for the assessment with U 1 =0.141and U 2 =0.0523. The pipe length is assumed 500 m.Ground temperature ranges from 0 to 15 o C. The inletof supply and return temperatures are known as 55 o Cand 22 o C respectively. The outlet temperature ofsupply pipe is controlled as 50 o C and 45 o C,respectively.Table 8 shows the temperature prediction based onsingle pipe simplification and the coupled pipeequations. T_Difference represents the coupledsolution minus the single pipe solution. When thetemperature drop along the supply pipe is controlled at5 o C, the prediction between the single pipe and thecoupled pipe is very close. The prediction errorsincrease with increase the ground temperature. Thesingle pipe approach predicts lower supply watertemperature and higher return temperature than thoseof coupled pipe solutions. It was also observed thatwhen the ground temperature is higher than 4 o C, thenet heat transfer effect in the return pipe is to absorbheat to the surrounding.79The by-pass water temperature in this study was set ina conservative way. In many practices, the by-passwater can be set 10 °C lower than the supply watertemperature. Even lower by-pass temperature isproposed for the low temperature district heatingnetwork [3]. Table 9 shows the simulation results basedon a10 °C temperature drop along the supply pipe. Itshows the prediction errors increase in both supply andreturn pipes. The heat transfer was predicted in areverse trend in the return pipe at 4 °C. Considerableprediction error was found in the return pipe at highground temperature.It is worth to be noted that the increase of supplytemperature drop has more influence on the return pipetemperature prediction than that of supply pipe. Thereason can be explained from the expression of U s andU r in Eq. 1–2. As the magnitude of T s -T g is higher thanT r -T g , the same amount of return water temperaturevariation will have more influence on U r than U s ,therefore causes a larger prediction error in the returnpipe than in the supply pipe.Table 9 Pipe temperature predication comparison (supplyoutlet temperature is controlled at 45 o C)CONCLUSIONIn this paper, a preliminary study was conducted on theinfluence of by-pass flow on the network return watertemperature in a designed low temperature DHnetwork. The concept of supply water recirculation was

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaintroduced to avoid the mixing of by-pass water to thereturn water. Double pipe water supply concept wastested to use the recirculation pipe supply water duringwinter season. Two different house installation modeswere considered in the analysis.The by-pass water significantly increases the returnwater temperature in the traditional design. The mixedreturn temperature can reach 35.5 o C for HE and45.6 o C for DHWS. With applying the by-pass waterrecirculation, this return temperature can be maintainedat 22 o C, while the re-circulated by-pass water can bekept as high as 44 o C and 53.5 o C for HE and DHWS atthe plant, respectively. It was found that the doublepipe supply leads to the highest network heat loss.However, the conclusion that whether the concept ofdouble pipe supply is inferior to other network designmethods can only be drawn after further networkthermal-economic optimization.The simulation program simplifies the twin pipe heattransfer prediction as a single pipe, and neglects thereturn pipe heat loss when the return pipe absorbs heatfrom the surroundings. The temperature predictionerrors due to the single pipe assumption were analyzedthrough solving the coupled supply/return pipedifferential energy equations. The prediction errorsincrease with increase the allowable temperature dropin the network. Considerable error was found for thereturn pipe at high ground temperature.NOMENCLATUREc p = specific heat capacity [ J/kg.K]q = Heat transfer rate [kW / m]s = Laplace transform variableT = Temperature [ K]U = Overall heat transfer coefficient [ kW /m.K]U ij = Linear thermal transmittance [kW/m.K]= mass flow rate [ kg/s]Greek Letter = Dimensionless temperatureSubscriptsg = Undistributed groundr = Returns = Supplyu = Upstreamd = DownstreamAbbreviationDH = District heatingHE = Heat exchangerDHWS = Domestic hot water storage tankREFERENCE[1] H. Lund, B. Moller, B. V. Mathiesen, A. Dyrelund, ―The role of district heating in future renewableenergy systems‖, Energy, 35, pp. 1381-1390,2010.[2] Charlotte Reidhav, Sven Werner, ―Profitability ofsparse district heating‖, Appliced Energy, 85, pp.867-877.[3] ―Udvikling og Demonstration af Lavenergifjernvarmetil Lavenergibyggeri‖, EFP 2007.[4] TERMIS Help Manual, Version 2.093,7-Technologies A/S.[5] Dansk Standard DS 439, 2000. Norm forvandinstallationer, Code of Practice for domesticwater supply installations, 3. udgave, www.ds.dk.[6] Otto Paulsen, Jianhua Fan, Simon Furbo, Jan EricThorsen, ―Consumer Unit for Low Energy DistrictHeating Net‖, The 11th <strong>International</strong> <strong>Symposium</strong>on District Heating and Cooling, 2008, Iceland.[7] P. K. Olsen, et.al, ―A new low-temperature districtheating system for low energy buildings‖, the 11th<strong>International</strong> <strong>Symposium</strong> on District Heating andCooling, Iceland, 2008.[8] Logstor. http://www.logstor.com/[9] Benny Bohm, Halldor Kristjansson, ―Single, twinand triple buried heating pipes: on potentialsavings in heat losses and costs‖, <strong>International</strong>Journal of Energy Research, 29, pp. 1301-1312,2005.[10] P.Walleten, ―Steady-state heat loss from insulatedpipes‖, Thesis, Lund Institute of Technology,Sweden, 1991.80

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSTEADY STATE HEAT LOSSES IN PRE-INSULATED PIPESFOR LOW-ENERGY DISTRICT HEATINGA. Dalla Rosa 1 , H. Li 1 , S. Svendsen 11 Technical University of DenmarkABSTRACTThe synergy between highly energy efficient buildingsand low-energy district heating (DH) systems is apromising concept for the optimal integration of energysaving policies and energy supply systems based onrenewable energy (RE). Distribution heat lossesrepresent a key factor in the design of low-energy DHsystems. Various design concepts are considered inthis paper: flexible pre-insulated twin pipes withsymmetrical or asymmetrical insulation, double pipes,triple pipes. These technologies are potentially energyefficientand cost-effective solutions for DH networks inlow-heat density areas. We start with a review oftheories and methods for steady-state heat losscalculation. Next, the article shows how detailedcalculations with 2D-modeling of pipes can be carriedout by means of computer software based on the finiteelement method (FEM). The model was validated bycomparison with analytical results and data from theliterature. We took into account the influence of thetemperature-dependent conductivity coefficient ofpolyurethane (PUR) insulation foam, which enabled toachieve a high degree of detail. We also illustrated theinfluence of the soil temperature throughout the year.Finally, the article describes proposals for the optimaldesign of pipes for low-energy applications andpresents methods for decreasing heat losses.INTRODUCTIONThe energy policy on energy conservation posesstringent requirements in the building energy sector, sothat the entire DH industry must re-think the way districtenergy is produced and distributed to end-users [1, 2].This is a requirement to be cost-effective in low heatdensity areas. Low-energy DH networks applied to lowenergybuildings represent a key technology to matchthe benefit of an environmentally friendly energy supplysector and the advantages of energy savings policy atthe end-users‘ side. Future buildings with a highperformance envelope will lead to reduced spaceheating load and therefore to a lower requireddistribution temperature for heating. The introduction oflow-energy DH networks is an appropriate and naturalsolution to enhance energy and exergy efficiencies.Distribution heat losses represent a key-point fordesigning low-energy DH systems, due to the criticalrole they have in the economy of the system. Theindustry could meet the requirements of higher81insulation series to reduce heat losses and thus savingoperational costs; however, this option would increaseinvestment and installation costs. The design principlesfor DH networks could instead be changed towards theuse of media pipes with small nominal diameters, witha higher permissible specific pressure drop. All-yeararound lower supply temperature and returntemperature constitute an effective option to reduceheat losses [3]. These principles have a big potentialfor heat supply to low-energy buildings, as explained in[4] and they are investigated in this paper.The total length of branch pipes can be significant inproportion to the total length of the network, above allin areas with a low-energy demand density. Moreoverthe temperatures in the critical service lines affect thetemperature level in the whole network, so that the heatlosses and the temperature decay in buildingconnection pipes are decisive for the overallperformance of the system. In this paper particularfocus was given to branch pipes.State-of the art of district heating pipesAt present time DH distribution and service lines arebased either on the single pipe system, where thesupply/return water flows in media pipes with their owninsulation, or on the twin pipe system, where both pipesare placed in the same insulated casing, or in a mixtureof them. All plastic pipe systems are characterized byhaving the water medium pipe made of plastic (crosslinkedpolyethylene (PEX) or polybutylene (PB)). Theyare covered by insulation, usually polyurethane foam,but in some cases of PEX foam or mineral wool; theouter cover is formed by a plastic jacket. Durability ofplastic pipes is not a real issue, since it has beenproved that the expected life of PB pipes and PEXpipes is, respectively, more than 40 years and approx.100 years [5]. As consequence of even lower averageoperational temperature, longer lifetime can bepredicted according to Annex A in [6]. Studies haveindicated that cross-linked polyethylene (PEX) pipeshave a cost advantage over steel pipes at pipedimensions less than DN60, due to their greaterflexibility since the joints do not require welding [7].Alternative design concepts must be considered inbranch pipes from street lines to consumers‘substations: a pair of single pipes, twin pipes or triplepipes. Traditionally most DH branch connections havebeen built with two single steel pipes: one supply pipeand one return pipe. Twin pipes can be made of steel,

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniacopper or PEX, with the supply and return pipe in thesame casing. The heat losses from twin pipes arelower than from single pipes, considering samedimensions and temperatures.Furthermore commercially available twin pipes, withdimensions up to DN200 for traditional steel media pipeor up to DN50 for PEX media-pipes are usually lessexpensive to install than single pipes [7]. Thistechnology 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, dueto flexibility in the way the system can operate andlower heat losses in case of optimal configuration. Thechoice of house connections depends mainly on thelength of the branch pipe, on supply and returntemperatures, building heating load and type ofsubstation. The latter is decisive with regard to energyperformance and thermal comfort. The types ofsubstations are typically divided into three concepts:unit with domestic hot water (DHW) storage tank,where the tank is the secondary-loop and consumerunit with DH water tank, where the tank is placed in theprimary loop. In this paper branch pipe solutions areconsidered for the concept of a consumer unit with heatexchanger and no storage tank. Two possibleconfigurations of user connection to the distribution lineare shown in Figure 1.demand, although a non perfect cooling of DH wateroccurs when tapping of DHW starts. The conceptbased on twin pipes and a substation withinstantaneous production of DHW in a heat exchangeris an optimal solution, if certain conditions arerespected. The first requirement is that the controlmethod gives priority to DHW preparation over spaceheating; the second condition is that the space heatingload during summer, to keep a high level of comfort inbathrooms for example, has to guarantee a sufficientcooling of the return water. As a result media pipes withinner diameters as small as 10 mm can be applied inthe primary loop and the water return temperature canbe kept sufficiently low, even in summer conditions.The triple pipe system is applicable in three differentoperational modes. The first one (mode I) occurs incase of DHW demand, when pipe 1 and pipe 3 both actas water supply pipes; the second operational mode(mode II) is activated when an idle water flow issupplied by pipe 1 and pipe 3 acts as re-circulation lineto the supply distribution line, while the return line (pipe2) is not active: this is often the case when there is nodemand for space heating, but a small amount of watercirculates in the DHW heat exchanger, keeping theloop warm to satisfy the instantaneous preparation ofDHW in the required time. This system avoids anundesirable heating of the water in the returndistribution line. The third operational mode (mode III)occurs during the heating season when there is onlydemand for space heating and no tapping of DHW:pipe 1 and pipe 2 operate as a traditional supply-returnsystem, while there is no water flow in pipe 3. Thedifferent modes are summarized as follows: Operational mode I: DHW tapping, pipe 1, 2, 3active.Figure 1: Sketch of a user connection with heatexchangers: twin pipe connection with/ without boosterpump (1–2) and triple pipe connection (1-2-3).1: supply2: return3: supply/re-circulationA simple and cost-effective configuration is composedof the control system and two heat exchangers for,respectively, space heating (SH) and domestic hotwater (DHW). The main disadvantage of such type ofsubstation unit is that only rather short lengths ofservice pipes can usually be applied; otherwise it wouldnot be possible to assure the required DHWtemperature at tapping points in the required time, dueto the unsatisfactory transportation time. A modifiedunit is therefore proposed and it is equipped with abooster pump which assures quicker response to DHWOperational mode II: supply-to-supplyre-circulation, pipe 1, 3 active; pipe 2 not active.Operational mode III: space heating demand, pipe1, 2 active; pipe 3 not active.METHODSTheory of steady state heat loss in buried pipesIn order to calculate steady-state heat losses in DHburied pipes there are analytical methods [8] andexplicit solutions for the most common cases [9]. Acomplete review of the available literature aboutsteady-state heat losses in district heating pipes hasbeen carried out in [10]. Here the methods arepresented with reference to the present status of thetechnology in the district heating sector. Furthermorekey-points and critical aspects are discussed; finally,improvements in the methodology of how to calculatesteady-state heat losses are proposed, with particularfocus on low-temperature and medium-temperature82

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaapplications. Low-temperature district heating systemsare defined as networks where fluids at a temperaturebelow 50 °C are used, while a medium-temperaturedistrict heating system is defined as using fluids attemperatures not higher than 70 °C [11, 12].Steady-state heat losses from pre-insulated buriedpipes are generally treated by use of the followingequation [10], which is valid for each pipe-i:where Uij is the heat transfer coefficient between pipe-iand pipe-j, Tj is the temperature of the water in pipe-jand T0 is the temperature of the ground. In case of twoburied pipes, which is the most common application inthe DH sector, the heat losses can be calculated asfollows, respectively for the supply pipe and the returnpipe, where T1 is the supply temperature and T2 is thereturn temperature.(1)Supply pipe: (2)Return pipe: (3)Equations (2) and (3) show how the heat transfer fromeach pipe can be seen as linear superimposition of twoheat fluxes, the first one describing the heat transferbetween the pipe and the ground, the second onerepresenting the heat transfer between the supply pipeand the return pipe. The equations can also bere-arranged in the following way:thermal coefficient, which is function of the temperaturein this case. U-values are dependent both ontemperature and time. If the time-dependency due tothe ageing of the foam can be restrained by introducingeffective diffusion barriers, that is not true for theintrinsic dependency on temperature. It is practice toevaluate the steady state heat loss applying a thermalconductivity value that corresponds to a hypothesizedmean temperature of the insulation. Nevertheless weneed models based, for example, on the finite elementmethod (FEM) when complex geometries or a highdegree of detail are requested.Temperature dependant thermal conductivity ofPUR insulation foamIn this paragraph the authors want to explain anddemonstrate the importance of taking into account thetemperature-dependency of the thermal conductivity ofthe insulation (lambda-value). The temperaturegradient in the insulation foam in the radial direction isoften higher then 10 °C/cm, meaning that the thermalconductivity of the material locally varies remarkably. Inthe example shown Figure 2, it varies more than 10%of the prescribed mean value. This affects themagnitude of the heat transfer. Considering a life cycleassessment of a DH system, the main impact to theenvironment is represented by heat losses [13]. Thethermal conductivity of the insulation material inpre-insulated DH pipes is usually stated at atemperature of 50 °C. The lambda-coefficients werechosen according to the available data at the end of2009; 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 2010new results are available [14]. It is preferable to have amodel that takes into account the temperaturedependencyof the thermal conductivity of theinsulation foam. The calculations in this paper use thefollowing expression, if not differently stated. It derivesfrom experimental data [15]:λ(T) = 0.0196734 + 8.0747308 . 10-5.T [W/(mK)] (1)Supply pipe: (4)Return pipe: (5)Equations (4) and (5) show how the heat transfer fromeach pipe can be calculated by use of only one linearFigure 2: Thermal conductivity in the insulation, horizontalcross-section of the pipe. Pipe: Aluflex 16-16/110,temperatures supply/return/ground 55/25/8 °C.83

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTemperature field in the soil around the pipeIn this paper we address the question of how to createa simple yet detailed FEM model for steady state heatloss calculations. The overall heat transfer resistancebetween the DH water and the environment is mainlycomposed of the thermal resistance of the insulationand the thermal resistance of the soil; compared tothese two factors, the thermal resistances of the pipewall and the convective resistance at the surface waterpipeare in practice negligible. The insulation foamalways offers the greatest share in the overallinsulation effect. The contribution of the soil is smalleron small-sized pipes than on large-sized pipes. Theshare is smaller in Insulation Series 2 and 3 [3]. Theheat conductivity coefficient of the soil is the mainparameter affecting the thermal resistance of the soilitself, and its value is often unknown in practice.In thecalculations we chose a value of 1.6 W/(m.K). The soiltemperature influences heat losses from DH pipes. Thesoil layer around the heating pipes slightly warms uparound the pipes. The evaluation of the temperaturefield in the soil is a prerequisite to create a realisticmodel for calculations of heat losses. Finite ElementMethod (FEM) simulations were carried out andtemperature conditions in the soil around a typical DHservice pipe, suitable for low-temperature applicationswere evaluated over a 10-year period.Table 1: Thermal properties of materials.λ [W/(m∙K)] ρ [kg/m³] C p [J/(kg∙K)]λ soil 1.6 ρ soil 1600 C p_soil 2000λ PE 0.43 ρ PE 940 C p_PE 1800Combined heat and moisture transfer is disregarded.The material properties are homogeneous and phasechanges, i.e. freezing and thawing were notconsidered. Table 1 lists the material properties, usedas input values also for the following models; a sketchof the slab-model, where the boundary conditions aredescribed, can be seen in Figure 3.Figure 3: Sketch of the model. Dimensions are in [mm].FEM modelA rectangle representing a semi-infinite soil domain(width: 10–20 m, height: 20–40 m) is the most usedgeometry to model the ground in heat loss calculations[18, 19]. In this paper a finite, circular soil domain wasapplied, instead. Its diameter is 0.5 m and it is equal tothe distance between the surface and the centre of thecasing pipe. Calculations show that the introducedsimplification hardly affects the accuracy of the results.The mesh model and an example of the temperaturefield in a small size twin pipe are shown in Figure 4.λ PUR 0.023λ PEX 0.38ρ PURρ PEX60 C p_PUR 1500938 C p_PEX 550λ Steel 76ρ Steel8930C p_Steel480λ Cu 400 ρ Cu 8930 C p_Cu 385The simulation calculated the soil temperature atvarious 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 wasassumed to be 14.6 W/(m2K), including convection andradiation [16]; we set the outdoor air temperatureduring the year according to the harmonic function validfor the Danish climate [17]: M T air 8.0 8.5 sin2 (6) 12 Figure 4: Mesh model of a pre-insulated twin pipeembedded in the ground (top and left). Temperature fieldin Aluflex twin pipe 16-16/110 (bottom-right); temperaturesupply/return/ground: 55/25/8 °C.In [3], where FEM simulations were performed, it isstated that for media pipes size from DN 50 to DN 400,the deviation of the lineal thermal coefficient betweenthe piggy-back laying (arranging the supply pipe belowthe return pipe) and the traditional system (horizontallaying) is less than 1%. The same conclusion can bestated for twin pipes; this is confirmed by calculations84

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniawith the multipole method in [20] for two examples oftwin pipe (DN 20 and DN 80) and by [10]. For twinpipes of even smaller size, such as in branchconnections, the heat losses occurring in case ofvertical layout are only slightly more favorable than thelosses occurring on horizontally arranged pipes; thisresult is shown with an example in the results section.RESULTS AND DISCUSSIONIn this section we discuss the influence of the soiltemperature on heat losses; next, we present thevalidation of the FEM models; finally we apply themethod to show the potential for energy saving in thecase of asymmetrical insulation of twin pipes, in thecase of double pipes and triple pipes.Temperature field in the soilTemperature conditions in the soil around a typical twinpipe, type Aluflex 16–16/110, were evaluated over a10-year period. Figure 5 shows the all-yeartemperature profiles of the outdoor air and of theground at depth equal to 0.5 m, at three horizontaldistances from the centre of the casing, during the firstyear of operation. No notable differences in the yearlyprofile were noticed in longer periods of time.We found that in state-of-the-art well insulated twinpipes (series 2 or 3) a certain amount of soil is slightlyheated up by the warm twin pipe; nevertheless thelevel of such heating can be neglected because itseffect is not noticeable in comparison to the fact thatthe uncertainties about the thermal properties of thesoil usually have a bigger impact. Considering yearlyaverage temperatures, the magnitude of the soilheating is about 1 °C for distances of around 0.2-0.3 mfrom the centre of the casing, and less than 0.5 °C by0.5 m. The temperature raise is considered incomparison to the undisturbed temperature of theground at a distance of 10 m.Figure 5: All-year temperature profiles of the outdoor airand of the ground at depth equal to 0.5 m and 3 horizontaldistances from the centre of the casing.85FEM model: geometry of the ground and of thepipesWe considered the geometric model of the preinsulatedAluflex twin pipe type 16-16/110; thetemperatures of supply/return/ground are 55/25/8 °C.We calculated the heat losses for vertical or horizontalplacement of the media pipes inside the casing, whichwas embedded in a rectangular or a circular model ofthe ground. The same calculations were repeated forother twin pipe size, up to DN 32 and other mediumpipe materials, i.e. steel and copper. The resultsconfirm that the vertical placement of the media pipesinside the insulation barely affect the heat transfer,being the difference between the two configuration lessthan 2% for the considered cases.Table 2: Heat loss for various placements of the mediapipes and various model of the ground.GroundmodelMediapipeslayoutHeat losssupply[W/m]Heat lossreturn[W/m]Heat losstotal[W/m]A Vert. 3.79 -0.17 3.62A Horiz. 3.80 -0.18 3.62B Vert. 3.84 -0.18 3.66A: Semi-infinite, rectangular (width x depth: 40 m x 20 m)B: Finite, circular (diameter: 0.5 m)Steady-state heat loss in commercial pipesThe model was validated by comparing the results fromFEM simulation to the analytical calculation for preinsulatedpipes embedded in the ground [14].Calculations were carried out for four different sizes ofAluflex 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. Theselected pipes are suitable to be used as branch pipesin low-energy demand areas. There is a goodaccordance between the two methods, the deviationbeing lower than 1%. Figure 6 gathers the values oftotal heat loss for the Aluflex twin pipe category; fourdifferent approaches are reported. The term ―standard‖is used when the effect of the temperature on thethermal properties of the insulation is neglected and thethermal conductivity of the PUR foam is thus constant.This is in accordance with [21]. The term ―advanced‖ isused when the calculation method takes into accountthat the thermal conductivity of the insulation dependson the temperature. Based on the temperaturescalculated for a number of points in the insulation theprogram calculates an average temperature for thematerial; the lambda-value of the insulation is thencalculated as a function of such temperature. Anaverage temperature of the ground is similarlycalculated. The calculation is repeated until the meantemperature difference for the insulation material, pipe

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniashell and surrounding soil is less than 0.005 °C for twoconsecutive calculations. The‖standard‖ and―advanced‖ model are available online [14]. In the―FEM advanced‖ model we directly implementedequation (1) in the insulation domain, instead. Theresults indicate that in case of low-temperatureoperation, lower total heat losses are calculated if thetemperature-dependency of the insulation lambdavalueis taken into account. Moreover the heat transferbetween the pipes in twin or triple pipes can beproperly evaluated.Total Heat Loss [W/m]7.06.05.04.03.02.01.00.0DN 14 DN 16 DN 20 DN 26 DN 32 DN 40Standard 3.61 4.24 4.62 5.71 6.45FEM Standard 3.34 3.68 4.33 4.80 6.02 6.76Advanced 2.86 3.36 3.69 4.55 5.10FEM Advanced 3.19 3.51 4.14 4.59 5.75 6.47(Aluflex: ≤ DN 26, steel: ≤ DN 50) the best design is toput the supply pipe in the centre of the casing, assuringthe best possible insulation for the supply pipe. Thisstrategy guarantees also the lowest temperature dropin the supply side, which is a critical figure inlow-temperature applications.For bigger sizes (Aluflex: ≥ DN 26, steel: ≥ DN 50) thebest design is achieved by ―moving up‖ the media pipelayout and at the same time by keeping the samedistance between the media pipes as in thesymmetrical case.Double pipesA double pipe consists of a pair of media pipes ofdissimilar size, co-insulated in the same casing. It is afurther development of the twin pipe concept. A sketchof a possible application of the double pipe concept isshown in Figure 7. Though these measures, networkheat loss reduction is possible, in case of operationduring low heating load periods.Figure 6: Comparison of 4 different approaches for steadystateheat loss calculation. Aluflex twin pipe series,supply/return/ ground temperatures: 55/25/8 °C.Asymmetrical insulation in twin pipesThe results show that improvements are possible,thanks to asymmetrical insulation (see Table 3). Weproved that a better design leads to lower heat lossesfrom the supply pipe (leading to a lower temperaturedrop); next, the heat loss from the return pipe can beclose to zero, maintaining isothermal conditions in thereturn line. If commercial available casing sizes arekept, we suggest two design strategies, depending onthe size of the pipes. For small pipe sizesTable 3: Comparison between asymmetrical and symmetrical insulation in twin pipes.The centre of the casing is the origin of the Cartesian system.Size(DN)14Coordinates(x; y) [mm]Heat loss[W/m]86Figure 7: Sketch of the possible application of the doublepipe concept in a simple district heating network.The space heating demand in summer is diminished,except for the energy requirement in bath roomheating. According to the energy balance, the reducedasymm.-symm. [%]Mat. Sup. Ret. Sup. Ret. Tot. Sup. Tot.(0; 0) (0; 27) 3.24 0.01 3.25 -7.6 2.016 (0; 0) (0; 28) 3.56 -0.01 3.55 -5.1 1.120Alx.(0; 0) (0; 30) 4.16 -0.04 4.12 -4.2 -0.326 (0; 0) (0; 36) 4.67 0.00 4.67 -5.1 1.932 (0; -16) (0; 28) 5.54 0.00 5.54 -5.8 -2.550(0; -25) (0; 55) 5.69 -0.03 5.66 -7.7 -2.4Steel65 (0; -36) (0; 60) 6.70 -0.02 6.68 -7.8 -3.2heating load requires lessnetwork flow rate as far as thedesigned building temperaturedrop is sustained. However, thereduction of network flow ratewill increase the supply watertemperature drop along thepipeline due to heat loss. As aconsequence, the supplytemperature at the end usermay lower down below theminimum requirement. Thisproblem is relevant to towenergyDH systems with analready low supply temperature.This design is based on the factthat the supply line acts also asre-circulation line during low

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaheating load periods; hence by-pass at the criticalconsumers are not necessary and the exergy loss dueto the mixing of warm water into the return line isavoided. Furthermore the water flow in the return linehas the same direction as in the supply line (clockwisein the example), so that the smallest size for the returnpipes are expected in correspondence to the biggestsize for the supply size, and vice versa. This results inlower local pressure differences between supply andreturn lines and savings in operational costs, thanks tolower heat losses. This is shown in Table 4 and Table5, by means of two examples: the first one refers to asmall to medium-size distribution network, the secondone to a bigger one, being capable to supply four timesmore energy than the previous one.Triple branch pipesThe development of an optimized triple pipe solution forlow-energy applications is reported to show thepotentiality of utilizing detailed models for steady-stateheat loss calculation. In this survey focus was given onthe choice of media pipes diameters as small aspossible. The triple pipe geometry is based onmodifications of the 14-14/110 (outer diameters in [mm]of respectively supply pipe, return pipe, casing) twinpipe design which has been reported in [18]. Fourgeometrical variations have been considered (seeFigure 8) and the Cartesian coordinates describing theplacement of media pipes inside the casing are listed inTable 6.Table 4: Comparison between a distribution networkbased on twin pipe (DN40-40 and DN80-80) with adistribution network based on double pipe (DN40-80 andDN80-40). Supply/return/ground temperature: 55/25/8 °C.Size(DN)Heat loss [W/m]Sup. Ret. Tot.Total(system) [%]40-40 -6.24 0.04 -6.20 Twin:80-80 -7.66 0.07 -7.59 -13.7940-80 -5.55 0.05 -5.5880-40 -7.41 0.05 -7.36Double:-12.94Table 5: Comparison between a distribution networkbased on twin pipe (DN100-100 and DN200-200) with adistribution network based on double pipe (DN100-200and DN200-100). Supply/return/ground: 55/25/8 °C.Size(DN)Heat loss [W/m]Sup. Ret. Tot.100-100 -7.83 -0.55 -8.39200-200 -8.92 0.24 -8.68Total(system)Twin:-17.06100-200 -6.4 0.08 -6.36 Double:200-100 -8.07 -0.03 -8.69 -15.056.1[%]11.8We considered an optimal placement of the mediapipes in case of double pipes, thus asymmetricalinsulation is applied. The total amount of insulation isused both in the twin pipe-based distribution networkand in the double pipe-based one, so that theinvestment costs are equal in both cases. Results showthat the heat loss can be reduced by 6% by means ofdouble pipes instead of twin pipes for the low tomedium-size distribution network. Even higher energysavings (around 12%) are possible in the case of thelarge-size distribution network.Figure 8: four different geometries for a triple service pipetype Aluflex 14-14/110.Table 6: placement of media pipes inside the casing forfour triple pipe geometries, type Aluflex 14-14-20/110.VariationPipe 1(Sup.)Coordinates (x, y) [mm]Pipe 2(Ret.)Pipe 3(Sup. orre-circ.)A (14;-14) (0;20.5) (-14;-14)B (10;-14) (0;20.5) (-21;-7)C (3;-14) (0;20.5) (-21;-7)D (0; 0) (0;25) (0;-28)The results of FEM simulations are listed in Table 7 forthe four geometries (A, B, C, D) and the threeoperational modes (I, II, II), previously described. Sincemode II occurs in case of no demand of space heatingand then outside of the heating season, simulationswere additionally performed with a more realistictemperature of the ground during that period(14 °C),considering Danish weather. This gives also aninsight in the effect of ground temperature throughoutthe year.87

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 7: Steady state heat losses of triple pipes typeAluflex 14/14/110 for 4 geometries and 3 operationalmodes. Temperature supply/recirculation/return/ground:55/55/25/8 °C.ModeI(DHWtapping)II(supply-tosupplyrecirculation)III(spaceheating)Table 8: Steady state heat losses of triple pipes typeAluflex 14/14/110 for 4 geometries and operational modeII. Temperature supply/recirculation/ return/ ground:55/55/25/14 °C.II(supply-tosupplyrecirculation)Geom.Geom.Pipe1Pipe1Heat loss [W/m]Pipe2Heat loss [W/m]Pipe2Pipe3Pipe3Tot.A 2.67 -0.08 2.67 5.30B 2.91 -0.29 2.75 5.38C 2.52 -0.22 2.74 5.06D 2.46 0.05 2.74 5.24A 2.67 / 2.67 5.34B 2.69 / 2.85 5.55C 2.48 / 2.70 5.18D 2.49 / 2.75 5.25A 3.46 0.48 / 3.95B 3.39 0.43 / 3.83C 3.41 0.35 / 3.76D 3.53 -0.01 / 3.53Tot.A 2.35 / 2.35 4.70B 2.37 / 2.51 4.88C 2.39 / 2.63 5.02D 2.20 / 2.42 4.62We conclude that an absolute best design for theservice triple pipe does not exist, but it depends on theoperational mode that is chosen as critical. In fact theresults reported in Table 7 and Table 8 show thatgeometry C gives the lowest total heat loss foroperational modes I and II, while geometry D has thebest thermal performance for operational mode III andfor operational mode II, if a temperature of the soil of14 °C is considered. It has to be underlined that,considering the operational mode III, geometry Dshows no heating of return water; this is a situationalways desirable, although it has a slightly higher heatloss from the supply pipe than the other geometries. Itis proved that usually operational mode I occurs forless than 1 h/day [20]. Moreover the temperature dropin the supply pipe to the DHW heat exchanger is criticalin low-temperature applications, so that it is stronglyrecommended to minimize the heat loss from thismedia pipe. Considering all this and the fact that modeIII is the most likely during the heating season and88mode II is the most likely outside heating season, theconclusion is that geometry D is preferable.CONCLUSIONSThe soil temperature at 0.5 m below the surface variesbetween 2 °C in January-February and 14 °C inJuly–August, for Danish conditions. This knowledgecan be used to better predict the winter peak load andthe temperature drop in the distribution line duringsummer.The slab-model for steady state heat loss calculationscan be replaced, in case of small sizedistribution/service pipes, by a model where the effectof the soil is represented by a circular soil layer aroundthe district heating pipe.The results confirm that the vertical placement of twinmedia pipes inside the insulation barely affects the heattransfer, in comparison to the horizontal placement; thedifference between the two configurations is less than2% for the considered cases.We proposed a FEM model that takes into account thetemperature-dependency of the thermal conductivity ofthe insulation foam; in this way we enhanced theaccuracy of the heat transfer calculation among pipesembedded in the same insulation.We applied the model to propose optimized design oftwin pipes with asymmetrical insulation, double pipesand triple pipes. We proved that the asymmetricalinsulation of twin pipes leads to lower heat loss fromthe supply pipe (from -4% to -8%), leading to a lowertemperature drop; next the heat loss from the returnpipe can be close to zero.It is possible to cut the heat losses by 6–12% if anoptimal design of double pipes is used instead oftraditional twin pipes, without increasing the investmentcosts.The development of an optimized triple pipe solutionwas also reported. It is suitable for low-energyapplications with substations equipped with heatexchanger for instantaneous production of DHW.REFERENCES[1] S. Froning, ―Low energy communities with districtheating and cooling‖, 25 th Conference on Passiveand Low Energy Architecture, Dublin (2008).[2] S. F. Nilsson et al., ―Sparse district heating inSweden‖, Applied Energy 85 (2008), pp. 555–564.[3] F. Schmitt, H.W. Hoffman, T. Gohler, Strategies tomanage heat losses – technique and economy,IEA-DHC ANNEX VII, (2005).[4] P.K. Olsen, B. Bøhm, S. Svendsen et al., ―Anew-low-temperature district heating system for

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonialow-energy buildings‖, 11 th international symposiumon district heating and cooling, Reykjavik (2008).[5] M. Klompsch, H. Zinko, Plastic pipe systems forDH, handbook for safe and economic application,IEA-DHC ANNEX V (1999).[6] DS-EN 253:2009, District heating pipes – Preinsulatedbonded pipe systems for directly buriedhot water networks - Pipe assembly of steel servicepipe, polyurethane thermal insulation and outercasing of polyethylene.[7] H. Zinko, GRUDIS-tekniken för värmeglesfjärrvärme (The GRUDIS technology for low heatdensity district heating), Swedish District HeatingAssociation, Stockholm (2004).[8] J. Claesson, J. Bennet, Multipole method tocompute the conductive heat flows to and betweenpipes in a cylinder. Department of BuildingTechnology and Mathematical Physics, Lund(1987).[9] P. Wallenten, Steady-state heat loss from insulatedpipes, Lund (1991).[10] B. Bøhm, ―On transient heat losses from burieddistrict heating pipes‖, <strong>International</strong> Journal ofEnergy Research, 2000, Vol. 24, pp. 1311-1334.[11] Terminology of HVAC, ASHRAE, Atlanta (1991).[12] I.B. Kilkis, ―Technical issues in low to mediumtemperaturedistrict heating‖, <strong>International</strong> Journalof Global Energy Issues, 2002, Vol. 17,pp. 113-129.[13] J. Korsman, S. de Boer and I. Smits, ―Cost benefitsand long term behavior of a new all plastic pipingsystem‖, DHC ANNEX VIII (2008).[14] www.logstor.com (March 2010).[15] Udvikling og demonstration af lavenergifjernvarmetil lavenergibyggeri (development anddemonstration of low energy district heating for lowenergy buildings), 2007.[16] B. Kvisgaard, S. Hadvig, Varmetab frafjernvarmeledninger (Heat loss from pipelines indistrict heating systems), Copenhagen (1980).[17] DS418:2002, Calculation of heat loss frombuildings.[18] H. Kristjansson, F. Bruus, B. Bøhm et al.,Fjernvarmeforsyning af lavenergiområder (Districtheating supply of low heat density areas), 2004.[19] T. Persson, J. Wollerstrand, ―Calculation of heatflow from buried pipes using a time dependentfinite element model‖, 45th <strong>International</strong>Conference of Scandinavian Simulation Society,Copenhagen (2004).[20] B. Bøhm, H. Kristjansson, ―Single, twin and tripleburied heating pipes. On potential savings in heatlosses and costs‖, <strong>International</strong> Journal of EnergyResearch (2005), Vol. 29, pp.1301-1312.[21] EN 13941:2003, Design and installation of preinsulatedbonded pipe systems for district heating.89

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTRANSIENT THERMAL CONDUCTIVITY OF FLEXIBLE DISTRICT HEATINGTWIN PIPESC. Reidhav and J. ClaessonDepartment of Civil and Environmental Engineering, Division of Building Technology,Chalmers University of Technology, Göteborg, Sweden.ABSTRACTThe standardized methods to measure the thermalconductivity of straight district heating pipes are notapplicable on flexible district heating pipes. This paperpresents a transient method determining thetemperature dependent thermal conductivity of flexibletwin pipes.A transient method to determine the temperaturedependentthermal conductivity of flexible single districtheating pipes is presented in this paper. A flexible pipecoil is immersed into cold water. Hot water isdistributed in the coil. The temperature decline of thecoil water is measured and calculated. Minimizing thedifference between the calculated and measuredtemperatures gives λ(T) of the flexible polyurethanefoam. The method gives small errors.INTRODUCTIONDistrict heating is supplied to the customers in one pipeand returned to the heat generation plant in anotherpipe. 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 heatingsupply (~80-110ºC) and return temperature (~40-50ºC)gives an internal heat flow from the supply pipe to thereturn pipe in a twin pipe. The total distribution heatloss from a twin pipe is lower than that of comparablesingle pipes due to this internal heat flow. Whendistributing district heat to areas with single-familyhouses to heat sparse areas, the issue of distributionheat losses is of special importance. The relativedistribution heat losses are considerably higher insparse areas than in more heat dense areas due to lowheat densities. Flexible district heating twin pipes arewidely used when single-family houses are connectedto district heating systems due to their light weights,flexibility and long lengths. In the efforts of minimizingdistribution heat losses, the possibility of determiningthe insulation capacity of flexible twin pipes is animportant issue.The standardised method used for determining thesteady-state thermal conductivity of district heatingpipes, the guarded hot pipe method, is only applicableon straight single pipes. The method is based on [1],described in [2] and [3]. A heater pipe is placed insidethe service pipe and the heat transferred through theinsulation is measured. The measurements must beconducted with a constant distance between the heaterpipe and the service pipe along the test specimenwhich can not be achieved with flexible pipes. Analternative method was presented in [4] and applied in[5] where the thermal conductivity of flexible districtheating single and twin pipes can be determined. Thetemperature decline of hot water pumped in a flexiblepipe coil is measured. A long pipe coil is needed tohave a sufficient temperature decline along the pipe.The Danish method is based on steady-statemeasurements at different temperatures to get thetemperature dependence of the decline. The FiniteElement Method is used to determine the thermalconductivity λ(T) of twin pipes.A transient method to determine the temperaturedependentthermal conductivity of flexible single districtheating pipes was presented in [6]. A pipe coil isimmersed into cold water and the temperature declineof hot water inside the coil is measured. The measuredtemperatures are compared to numerically calculatedvalues to characterize λ(T). A Kirchoff transform isused to simplify the calculations. Finally, the meansquare difference of the measured and calculatedtemperatures are minimized which gives λ(T). In thispaper, a similar experimental set-up is used for aflexible twin pipe. The numerical and mathematicalmodel developed in [6] cannot be used for twin pipesdue to the complicated geometry of twin pipes. In [7]and [8] a method was presented where heat lossesfrom district heating twin pipes were calculated withconformal coordinates describing the twin pipegeometry. In this paper, the conformal coordinatemodel is used to calculate the temperature decline in aflexible twin pipe. The calculated temperatures arecompared to experimentally measured temperatures.This gives the temperature-dependent thermalconductivity of semi-flexible polyurethane foam of thestudied flexible twin pipe.Fig. 1 Cross-section of single (left) and twin (right) districtheating pipes90

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaEXPERIMENTAL MEASUREMENTSThe experimental set-up is similar to that used whendetermining the thermal conductivity of single districtheating pipes in [6]. A flexible twin pipe of about 18meters coiled with a diameter of 1.8 meters isimmersed into a pool with circulating water. In thisexperiment, the pool water was about 17 ºC. Previoustests show that air is unsuitable as surrounding mediadue to difficulties in keeping stable temperatures. Thesupply and return service pipes are connected in a loopcirculating water at a temperature of about 80ºC. Whensteady-state is established in the insulation, at timet = 0, the circulation is stopped. Then, the temperaturedecline of the stagnant loop water is measured at oneposition in the coil. The thermocouples are placed atthree positions of each service pipe, see Fig. 2. One isplaced on top of the service pipe, one on the side andunderneath the service pipe. The insulation is peeledoff at the positions of the thermocouples and then putback and sealed to be water proof. The reattachment ofthe insulation was probably insufficient and it appearsas if pool water permeated after about 5.5 hours anddisturbed the measurements.The twin pipe studied in this paper has two copperservice pipes, semi-flexible polyurethane foam and aslightly corrugated LDPE casing. The pipe is ofdimension DN 20 with the pipe dimensions described inTable 1. The pipe producer declares a thermalconductivity of the semi-flexible polyurethane foam ofλ50 = 0.0255 W·m-1·K-1of a newly produced pipe ofthis kind. The pipe has no diffusion barrier. The densityof the polyurethane foam was ρ = 60 kg/m 3 .Table 1. Dimensions of the twin pipe studiedCasing outer diameter (mm) 91Casing thickness (mm) 2.2Service pipe outer diameter (mm) 22Service pipe thickness (mm) 1.0The initial coil temperature was T 0 = 81.3ºC. The watertemperature at the service pipe (T w,meas (t)) decreasesduring the 16 hours of measurements. The pooltemperature was initially T 1 = 17.4 ºC and increasedslightly to T 1 =17.9 ºC during the 16 hours. Themeasured coil and pool temperatures are showed inFig. 3.A sawtooth disturbance of about T = 0.07 ºC and asmall noice of about T = 0.015ºC can be seen in thepool water measurements in Fig. 4. A detailed study ofthe coil temperatures T w,meas (t) shows that thetemperature of some thermocouples decreases abruptoccurred at about t = 5.5 hours. The marked choseninterval in Fig. 3 and Fig.4 is chosen to minimize theerrors.Fig. 2 Experimental set-up and positions of thermocouplesat the service pipes91

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia9080Chosen interval70Temperature (C)60504030Tw, meas2010T1, meas00 2 4 6 8 10 12 14 16Time (h)Fig. 3 Measured coil T w, meas and pool temperature T 1, meas.18,017,9Chosen interval17,8Temperature (C)17,717,617,517,40 2 4 6 8 10 12 14 16Time (h)Fig. 4 Measured pool temperature T1, meas92

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaMODELLING USING CONFORMAL COORDINATESIt is rather complicated to calculate the temperaturedecline in twin pipes due to the pipe geometry. A socalled conformal mapping presented in [8] was used tomap the twin pipe geometry onto a rectangulargeometry. In the experimental measurements, thesupply and return service pipes were assumed to haveequal temperatures in the test-procedure. Then,symmetry is assumed between the four quarters of apipe cross-section. A quarter of a twin pipe is studied,see Fig. 5. In the x,y-plane, the temperaturedevelopment is described by the heat equation:T T T c ( ( T) ) ( ( T) )t x x y y(1)The (x,y)-coordinates ( z x i y)are transformed tosuitable conformal coordinates ( w u i v)with theaid of line sources and so called multipoles.water and the right-hand boundary against the pollwater. The heat flux in the vertical v-direction is zero onthe horizontal boundaries due to symmetry.Fig. 6 Initial temperature distribution in the crosssectionof a pipe quarter in the u, v-plane.In the numerical solution, the region is divided into arectangular mesh. The area factor is now the area ofeach of the cells shown in Fig.5. They are shown inFig. 7. The largest cell is the one in the lower left cornerin Fig.5 near the stagnation point (usp). The areas areused to calculate the heat capacity of each cell in the u-v-plane.Fig. 7 Areas of the computational cells in the x, y –planetransferred to a u, v-plane. The stagnation point is denotedusp.Fig. 5 A quarter of a twin pipe in x-y-plane geometryThe heat equation in the conformal coordinates is:T T T c A( u, v) ( ( T) ) ( ( T) )t u u v vHere, A(u,v) is the area factor in the conformaltransformation.The considered region shown in Fig. 5 is transformedto a rectangular region in the u, v-plane, see Fig. 6. Inthe figure, the left-hand boundary lies against the coil(2)The initial steady-state condition for a twin pipe withcoil water temperature T w = 81.3ºC immersed into poolwater at T0=19.7ºC is showed in u-v coordinates inFigure 6. Then, the temperature decline of stagnantwater in the twin pipes are calculatedThe density ρ and the heat capacity c of thepolyurethane foam are assumed constant in thetemperature interval studied. The boundarytemperatures at the casing are given by the pooltemperature.The thermal conductivity λ(T) of the polyurethane foamis determined by the thermal conductivity at 50ºC λ 50(W/m·K) and a coefficient λ‟ to account for a lineartemperature dependence.50 50( T) 1 ' ( T T )(3)93

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaEVALUATION OF MEASUREMENTSThe temperature-dependent thermal conductivity of thepolyurethane foam is obtained by calculating thetemperature decline of the coil water. Certain values ofthe thermal conductivity of the polyurethane arechosen, λ 50 and λ‟. The actual λ(T) are obtained byminimizing the difference between the measured andcalculated coil temperatures, (4).The difference between the calculated and measuredtemperatures for the optimal parameter values of λ 50 , λ‟and c are showed in Figure 9. The saw toothdisturbance from the measurements is seen. Thedifference giving the best fit lies in the interval -0.20 to0.25 (ºC). The error is small.The heat capacity c (J·kg -1·K -1 ) of the polyurethanefoam is input to the calculations. Literature referencesfor the heat capacity of polyurethane foam varies, 1300J·kg -1·K -1 at 50ºC in [9], 1400 J·kg -1·K -1 in [10], 1400-1500 J·kg -1·K -1 for rigid polyurethane foam in [11].The densities and heat capacities of water, service pipeand insulation were assumed to be constant in thetemperature interval studied.The optimal parameter values of λ 50 λ‟ and c wereobtained by minimizing the difference D (ºC) betweenthe calculated T w (ºC) and measured coil temperaturesT w,meas (ºC).D max T ( t) T ( t)for t t t (4)w, calc w, meas1 2A certain time interval, 0.5

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe sawtooth variation seen in the measurements in [6]is seen in this measurement as well and should befurther investigated.A large difference between the pool and coiltemperature is desirable to minimize the relative errors.It is also important to assure that steady-stateconditions are established before starting thetemperature decline.The final result, the obtained thermal conductivity:5( T) 0.0235 10 10 T ( ºC), 500.0285 (5)is in reasonable agreement with the declared λ 50 =0.0255 W·m -1·K -1 for a newly manufactured pipe. Thispipe piece had been in store for some time and had nodiffusion barrier. The temperature-dependent part ofthe thermal conductivity is in well agreement with [12].REFERENCES[1] U. Jarfelt, Test apparatus of pipe insulation.Doctoral thesis. Chalmers University ofTechnology, Göteborg (1994)[2] European standard EN 253:2009, District heatingpipes - Preinsulated bonded systems for directlyburied hot water networks – Pipe assembly of steelservice pipe, polyurethane thermal insulation andouter casing of polyethylene, Brussels, Belgium.(2009)[3] European committee for standardization. Europeanstandard EN ISO 8497:1996, Thermal insulation-Determination of steady-state thermal transmissionproperties of thermal insulation for circular pipes,Brussels, Belgium. (1996)[4] Danish District Heating Association. Developmentof an experimental set-up for measuring the heatconduction properties of flexible pipes, Project nr.2006-05, Århus, Danmark. (2006), In Danish,available at Dansk Fjernvarmes F&U-Konto,www.danskfjernvarme.dk[5] District Heating Association (2008), Heat planDenmark, Ramboll Danmark A/S and AalborgUniversity, (2008), In Danish, available at DanskFjernvarmes F&U-Konto, www.danskfjernvarme.dk[6] C. Reidhav and J. Claesson, A transient method todetermine temperature-dependent thermalconductivity of polyurethane foam in district heatingpipes, Building Physics 2008 - 8 th Nordic<strong>Symposium</strong>, Copenhagen, Denmark, (2008)[7] C. Persson and J. Claesson, Prediction of heatlosses from district heating twin pipes, The 11 th<strong>International</strong> <strong>Symposium</strong> on District Heating andCooling, August 31 to September 2, Reykjavik,Iceland, (2008)[8] C. Persson and J. Claesson, Numerical solution ofdiffusion problems using conformal coordinates.Application to district heating pipes, ReportDepartment of Civil and EnvironmentalEngineering, Chalmers University of Technology,Göteborg, Sweden (2008)[9] S. Peng, P. Jackson, V. Sendijarevic, K.C. Frisch,G.A Prentice, A. Fuchs, Process Monitoring andKinetics of Rigid Poly(urethane-isocyanurate)Foams, Journal of Applied Polymer Science,(2000) Vol 77, 374-380[10] R. Zevenhoven, Treatment and disposal ofpolyurethane wastes: options for recovery andrecycling, Helsinki University of Technology,Report TKK-ENY-19, Espoo, Finland, June (2004).[11] BING, Federation of European Rigid PolyurethaneFoam Associations, Thermal insulation materialsmade of rigid polyurethane foam (PUR/PIR),ReportNo1 October (2006)[12] U. Jarfelt and O. Ramnäs, Thermal conductivity ofpolyurethane foam – best performance,10 th <strong>International</strong> <strong>Symposium</strong> on District Heatingand Cooling, Sept 3-5, Hanover, Germany, (2006).95

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDISTRICT HEATING PIPES 200 MM BELOW SURFACEIN A STREET WITH HEAVY TRAFFICAnders Fransson 1 and Sven-Erik Sällberg 21 Göteborg Energi AB, Sweden2 Building Technology and Mechanics, SP Technical Research Institute of SwedenABSTRACTThis article reports the results from a field experimentinitiated by Göteborg Energi AB with an extremeshallow burial of district heating pipes 162/76,1 (DN 65)casaflex under a street with heavy traffic designed foran average of 2000–4000 passes of vehicles a day andline. The pipes were laid only 200 mm below thesurface. The backfill was of 0–40 mm particle size.Several consecutive measurements were done to studythe effects from instant and long term loads from thetraffic. The tests were done on a test pipe preparedwith displacement gauges and on operating pipes.The aim is that the results will inspire and give input formaking district heating and cooling more cost effective.The tests showed that both the instant and long turndeformation of the pipes are small at the actual layingdepth and also that the acceleration in the ground asheavy vehicles passes does not seem to be alarming.sales (i.e. less district heating, d.h. to be sold). Toconnect new district heating customers in the future,with the competition of other heating suppliers, it is notenough to use just smaller pipes because of thesmaller demands. Building the grid and maintaining thegrid needs to become more cost efficient.The purpose of this article is to inspire and if possiblehelp whoever is interested in making district heatingand cooling in the world more cost effective using theideas or test results from this article.1.2 Cost-cutting due to shallow burial in roadsWhen reducing costs, it is important to maintain thequalities that are required. The road owner needs theroad to be functional and has its standards. The districtheating supplier is responsible for its pipes anddeliveries of heat and has its standards. Finally thereare workers (contractors and maintaining staff) whoneed acceptable working conditions.The conclusion is that shallow burial is technicallypossible if the road and backfill is done properly.1. INTRODUCTION1.1 New conditions for district heatingThe branch of district heating is in need of a newgeneration of district heating pipes.The conditions for selling district heating are slowlychanging due to new legislation, harder competition,new technique and climate changes. Since 2003Göteborg Energi AB is connecting more and morecustomers but is selling less and less energy. Newlegislation from 2006 allowed new buildings inGothenburg to use a maximum of 110 kWh/m2externally supplied energy for heating (,cooling) andproducing domestic hot water. Today the municipalityof Gothenburg wants new buildings to use 60 kWh/m2at most.These changes are not unique. New houses are usingless and less energy per square meter. There arealready households that are not using but producingenergy. The former energy suppliers in Europe arefinding themselves not as suppliers but distributors,buying and selling energy. Climate changes are globaland have already measurable effects on district heating96Fig. 1 Standard shaft sectionIn a standard shaft section the drainage may be takenaway in roads. A properly built road has a hard top andis drained as it is. You do not need to drain it anymore.It is also possible to make the shaft more narrow andmaintain acceptable working conditions if either longpipes with no joints are used or if the joints are weldedon top of the shaft.Less coverage is also an alternative. Earlier studies[1]–[3] shows that the pipes are solid enough to beplaced with very little coverage (180 mm) and in roughmaterials. It is also shown that there is less settling in

the street the more shallow the shaft is [3]. On theother hand if too little coverage over the pipe is chosen,it may get hit by a rock curb or a supporting leg from atruck with no cover.The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 4 Marking of the location of the gauges.Fig. 2 Left; Supporting leg with no cover, Right; Rockcurb.If the existing fraction is used as backfill transports canbe reduced which lower the costs and theenvironmental influence.Normally the district heating pipes have no problemscoping with the traffic load. The extra pressure and themovement in the soil are making extra loads that arequite negligible compared to the thermal load, the innerpressure load and the load from the outer pressurefrom the soil.To get an idea of if the graphite gaskets used in thepipe joints (for casaflex) can stand the traffic load200 mm below the surface in a street with heavy trafficthe traffic load was empirically measured in the testarea.The test shaftThree displacement gauges were placed in a test pipecontaining air (see section 2.1) beside two operatingpipes (see Fig. 3–5). Two accelerometers were placedin separate boxes near the test pipe. The gauges weremonitored through wires at the bicycle path beside thestreet.The test was done with two single district heating pipes162/76,1 (DN 65) casaflex buried in a fraction withgrain size 0–40 mm. (Normal standard is a fraction ofsand 0.2–16 mm.) The distance from the top of thepipes to the top of the fraction was 60 mm. (Normalstandard is 460 mm.) The distance from the top of thepipes to the top of the asphalt was 200 mm. (Normalstandard is 600 mm.)As extra protection, the operating pipes were wrappedin a grid of polyethylene, PE.EXPERIMENTALThe tests were done during 2009–2010.The test siteAn industrial street classified as a street with2000–4000 passes of vehicles for every lane and dayas an average through the year was chosen as the testsite. The extension of the test area was 8 meters as thepipes crossed a street.The gauges were placed in one of the lanes close tothe centre of the street. The location of the gaugeswere 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 compressedwith a 500 kg plate compactor.The pipes are designed for 1.6 MPa but the localhydraulic pressure is approximately 1.4 MPa. Thedesigned temperature for the district heating water is110 °C. The real temperature varies between 70 and100 °C in the supply pipe and between 40 and 60 °C inthe return pipe.Fig. 3 Drawing over the test area.97About casaflexThe type of pipe, casaflex, was chosen to overcomethe thermal loads and the working conditions (i.e. the

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonialength 140 m of a single pipe means that there is noneed to work with the pipe in the shaft).Casaflex is a type of pipe that differs from ordinarydistrict heating pipes in several ways. It is comparedwith ordinary d.h. pipes in earlier studies [4]. This typeof pipe is not particularly common in the Nordiccountries. The pipe is supposed to be used as ordinarypipes with sand as backfill.The media pipe is made of corrugated stainless steeland surrounded by CFC-free polyisocyanurate foam.The foam is wrapped in a multi layer barrier foil at theoutside covered with a corrugated low densitypolyethylene casing. Inside the insulation along thepipe there are three surveillance wires. The casaflexpipe can be delivered in very long lengths. The pipesused in this test were 140 m.To connect different casaflex pipes a system withflanges, bolts and gaskets are used. The gaskets aremade of graphite.Fig. 6 Left; Casaflex pipe, Right; Casaflex pipe with ajoint.2.1 Deformation of the pipe over timeThe test pipe was 1.66 m long and prepared with threedisplacement gauges inside to measure the radialdeformation in three directions. The displacementgauges were installed at the half length of the pipe witha distance of 100 mm in between. One displacementindicator measured on the upper side of the pipecasing, the second on the underside of the pipe casingand the third at the side of the pipe casing. Thedisplacement gauges were fixed to the media pipe tomeasure the changes in the pipe casing.Before the test pipe was installed referencemeasurements were done at the laboratory to createzero values for the displacement gauges.100100Indicator 3 Indicator 2Indicator 1Fig. 8 Locations of displacement gauges inside test pipe.During the test period that lasted for one yearindications from the displacement gauges weremeasured twelve times. During the test period thetemperature varied between summer temperatures towinter temperatures.2.2 Instant deformation of the pipe andaccelerations from traffic loadTwo accelerometers were placed 200 mm respectively600 mm below the asphalt surface (see Fig. 5), close tothe test pipe, to measure the vibrations in the roadstructure when heavy vehicles pass over the test pipe.The measure equipment used were a signal analysator01dB Harmonie, ser. nr 4227 and accelerometers ofthe type ST Microelectronincs type LIS2L02AL with asampling rate of 3200 Hz and resonance frequency ofat least 2 kHz. The accelerometers were installed insmall boxes and calibrated within the frequency interval4–6 Hz. The calibration is traceable to the Swedishnational centre for acceleration metering.It was arranged so that a heavy lorry passed over thetest area several times at different speed (20 and 40km/h) while the vibrations in the road structure wereregistered with the two accelerometers. The weight ofthe lorry was 26 400 kg.To investigate the instant deformations in the test pipewhen heavy vehicles pass over the pipe the indicationsfrom the displacement gauges (see section 2.1) weremeasured at the same time as the vibrations in theroad structure were registered.2.3 Radial and axial stiffness of pipeIn laboratory the physics of the test pipe were testedconcerning radial stiffness and axial stiffness. The testswere done on a 165 mm long test specimen from thesame pipe as the test pipe. The arrow in Fig. 9 showsthe direction of the applied load during the test.Fig. 7 1.66 m long district heating test pipe of typeCasaflex 162/76.1 (DN 65).Fig. 9 Arrangement for test of radial stiffness.98

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe axial stiffness in the pipe was tested in three ways.Fig. 10 describes the three arrangements for applyingthe load in the tests, (a) the applied load acts on thewhole cross section, (b) the applied load acts on onlythe steel pipe and (c) the applied load acts on the outersteel net including the pipe casing. The arrow in thefigures indicates the direction of the load.2.5 Leak test of the pipe casingTo discover moisture or even water in the insulation,there are different indicators on the market.The typical indicator system used in Gothenburg is theso called Nordic System. The Nordic System is asystem which is using two naked cupper wires insidethe insulation along the pipe at 10 am and 2 pm.(a)(b)(c)The casaflex pipe uses the Hagenuk System. Thatsystem uses three wiresa) Ni Cr,b) Cu, insulated andc) Cu, not insulated.165In this test different pipes and different systems wereconnected. The Ni Cr wire in the Hagenuk System wasleft disconnected.90 140Fig. 10 Three types of arrangement for applying the loadfor test of axial stiffness.2.4 Pipe prolonging while pressurizingA casaflex pipe does not expand because of thethermal load. It is self compensating. But because ofthe geometry of the media pipe it expands when it getspressurized. On the other hand the multi layer barrierfoil in the pipe holds the expansion back. Because thepipe is flexible, it will still be able to expand, but onlyuntil the multi layer barrier foil stops the expansion.To see how much the pipe expands because of thepressurization, a distance indicator, Hilti PD4, was fixedon the pipe before it was installed and pressurizedwhile it was still on the ground. The distance wasmeasured three times against an iron angle which alsowas fixed on the pipe. After the pressurization thedistance was measured again three times.The resistance was measured with an ordinary ohmmeter, BM 400.The pipes were also three times tested with a, Statemeter, Time Domain Reflectometer (TDR) fromStateview.2.6 Test of degree of compaction of the streetTo get the permission from the road owner to do thistest in the street there were certain standards to follow[5] and [6].Before the asphalt could be put on the shaft there wereto be some tests of the degree of compaction of thestreet with certain limits. It is a German test that is alsoused in Sweden [7]. Basically the soil gets compressedwith a known load over a known area and onemeasures the Young‘s modulus Ev for the soil twotimes. The demands were that;a) E v2 / E v1 < 2,8b) E v2 > 50 MPac) At least 4 out of 5 tests should be correct.Fig. 11 Left; Fixed distance indicator, Right; Fixed ironangle.To see with which force the casaflex pipe wasexpanding, the following equation was used:F p = PA (1)Where Fp is the prolonging force [N], P is the internalover pressure [Pa] and A is the maximum inner area ofthe pipe [m2].Fig. 12 Test of degree of compaction of the street2.7 Visual control of the surface of the streetAs an extra precautionary measure, the street wasoptically inspected every month through a year. Duringthe first month, the street was inspected every week.And there was an extra inspection in spring in order to99

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniafind potential frost action damages. The inspectionswere documented with photos.3. RESULTS3.1 Test results from deformation of the pipe overtimeThe measured pipe deformations during the test periodturned out to be very small. The diagram in Fig. 13describes the measured changes in the casing sinceinstallation and average air temperatures during thetest period. All three displacement gauges were set tozero before the installation. The diagram shows thatthe casing of the test pipe during the installation wassqueezed out up to 0.5 mm at the three measurementpoints. The deformations in the casing are most likelycaused by the packing of the backfill surrounding thepipe.After the installation during the test period the resultsindicate that the upper side (violet curve in thediagram) of the test pipe casing have been pressed in0.2 mm. The side of the test pipe casing havesqueezed out approximately 0.1 mm. The under side(red curve) was squeezed out approximately 0.1 mmduring the period between the first and secondmeasurement results. During the rest of the test periodthe casing have been pressed back in 0.1 mm.It is to be observed that these measured changes arevery small relative to the test pipe casing diameter.Compared to the zero values in the laboratory themeasured changes are not more than 0.3 % relative tothe casing diameter.In Fig. 16 and 17 the diagrams show the vibrationvelocity (m/s) in the ground when a heavy lorry passover the test area at a speed of 40 km/h. The vibrationvelocity is calculated from the acceleration signal byintegration.The diagrams in Fig. 18 and 19 show the maximumamplitude of the acceleration in the ground as afunction of the speed of the lorry when it passes overthe test area in 20 km/h and 40 km/h, respectively themaximum vibration velocity as a function of the speedof the lorry.200 mm below surface, vehicle speed 40 km/hTime (s)Fig. 14 Vertical acceleration 200 mm below the roadsurface when a lorry passes at 40 km/h.600 mm below surface, vehicle speed 40 km/hmm0,90,80,70,60,50,40,30,20,10-1000 100 200 300 400Days since installationAverage temperatures (°C)Under SideUpper SideFig. 13 Average air temperatures and changes in casingat installation and during test period.3.2 Test results from instant deformation of thepipe and accelerations from traffic loadThe diagrams in Fig. 14 and 15 describe the vibrationsprocess 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 thetest area at a speed of 40 km/h.Side200-20-40-60-80Time (s)Fig. 15 Vertical acceleration 600 mm below the roadsurface when a lorry passes at 40 km/h.Velocity (mm/s)200 mm below surface, vehicle speed 40 km/hTime (s)Fig. 16 Vibration velocity 200 mm below the road surfacewhen lorry passes at 40 km/h.100

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaVelocity (mm/s)600 mm below surface, vehicle speed 40 km/hover the test area. At next instant it is squeezedtogether approximately 0.07 mm at the same time asthe blue curve indicates that the pipe goes eccentricapproximately 0.04 mm.0,120,100,08Side (mm)Up (mm)0,06Down(mm)Time (s)Fig. 17 Vibration velocity 600 mm below the road surfacewhen lorry passes at 40 km/h.mm0,040,020,00-0,02-0,04-0,060,00 1,00 2,00 3,00 4,00time (s)Fig. 20 Instant deformation 200 mm below the roadsurface when lorry passes at 40 km/h.3.3 Test results from radial and axial stiffness ofthe pipeThe test pipe was compressed 1.8 mm two times with afeed speed of 1 mm/min. In Fig. 21 it can be seen thatthe maximum force at 1.8 mm turned out to be 1.4 kN.Fig. 18 Maximum amplitude of acceleration as a functionof speed.Using this result to look at what the correspondingforces should be in the test when a heavy lorry passesover the test area (see Fig. 20) it can be establishedthat the instant forces from passing vehicles is small.1600140012001000Load, N800600400Test 1200Test 200 0,5 1 1,5 2Fig. 19 Maximum amplitude of vibration velocity as afunction of speedDeformation, mmFig. 21 Diagram radial stiffness of a casaflex pipe.Feed speed 3 mm/minThe acceleration of the ground increases with thespeed of the traffic. And the effect is more sensitive thecloser you are to the surface (see Fig. 18).201816Load case (a)The vibration velocity also increase with the speed ofthe traffic. The effect is not as sensitive as for theacceleration when it comes to the coverage (seeFig. 19).Axial load, kN141210864Load case (c)Load case (b)In Fig. 20 the diagram describes the instantdeformations in the test pipe when the heavy lorrypasses over the test area at a speed of 40 km/h. It canbe seen from the red and the violet curves that the pipecasing is squeezed together approximately 0.17 mmfrom top to bottom at the instant when the lorry passes200 0,5 1 1,5 2Axiell compression %Fig. 22 Axial stiffness of a casaflex pipe101

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia3.4 Test result of pipe prolonging whilepressurizingThe test results are as follows:Table 1. – Test result of pipe prolonging while pressurizingDistancebeforepressurizingDistance afterpressurizingTests1 2 3 Average[mm] [mm] [mm] [mm]4 360 4 360 4 360 4 3604 365 4 365 4 365 4 365The diagram (Fig. 23) below contains two different TDRmeasurements. It is one graph per wire and test. Ifthere are no changes in the impedance there are nochanges in the profile in the graph. And there are nochanges in the profiles.The pressure that was used was approximately1400 kPa. The test results show that a single pipecasaflex prolongs itself 100*5/4360 = 0,11%.his could be compared to the more common steel pipefor district heating. If that pipe would be loaded with athermal load of 100 ºC it would prolong itself 0,12%.The force with which the casaflex pipe is expandingbecause of the inner pressure would for 11400 kPa be7,3 kN according to the supplier. That would mean thatthe diameter would be 81,5mm. In real life the diameterwas measured to be 83,9 mm. The corresponding forcefor the diameter 83,9 mm would be 7,7 kN.If the pressure would have been 1 600 kPa and thediameter would have been 81,5 mm then thecorresponding force would have been 8,3 kN.This could again be compared to the steel pipe with thethermal load of 100 ºC. This pipe would prolong itselfwith the force of 164,9 kN.So the casaflex pipe expands with a force that isapproximately 100*8,3/164,9 = 5,0% of the force from asteel pipe when heated 100 ºC.Looking at Fig. 22, case a), one sees that the innerforce (axial load) that the expanding force has toovercome is negligible.3.5 Test results from the leak test of the pipecasingDifferent TDR graphs have been made in May 2009, inJune 2009 and in April 2010.Through metering the resistance and making TDRgraphs it is proven that:a) It can be done to connect the two differentsystems (The Nordic System and the HagenukSystem).b) There are no leaks in the test area, neither inthe supply pipe nor in the return pipe.Fig.23 TDR graph for the supply pipe.3.6 The results from the test of degree ofcompaction of the streetThe different tests were plotted in diagrams and gavethe different Young‘s modulus E v1 and E v2 for differentplaces. The places were documented in a photo. Theresults can be read in the table below.Table 2. - Results from test of degree of compactionSpotE v1(MN/m 2 )E v2(MN/m 2 )E v2/E v11 38,75 91,44 2,362 18,23 25,25 1,393 61,29 126,4 2,064 51,53 100,65 1,955 29,30 70,08 2,39For every test the division E v2 /E v1 are approved. Spot 2E v2 is to low but the other four spots are approved sooverall the test is ok.3.7 Results from the visual control of the surface ofthe streetThe strength of the roadThere was no change in the surface of the street due tothe shallow laying of the district heating pipe what soever the first eleven months. In spring after anunordinary cold winter one could see a small crack(approximately 12 cm) in the street along the pipesextension. As this article is getting written it is notinvestigated why the crack has appeared nor of theimportance of it. The street has much worse injuriesfrom frost action damages outside the test area.102

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaHeat lossThe snow did not melt over the pipes in the test area. Ifone would study other older district heating pipes theywould reveal themselves by melting the snow overthem. This effect never happened in the test area.The acceleration and vibration velocity are alsonegligible under the traffic load from e.g. a heavy lorry.Probably there will be no problems using graphitegaskets also with only 200 mm of coverage.The casaflex pipe is prolonging itself if it may but theforce with which it is prolonging itself is but a fraction ofwhat comparable ordinary steel pipes uses. This effectmakes it more suitable for shallow shafts.It is possible to combine different leak indicatorsystems and still get the TDR-graphs. The graphs donein this test indicates that there are no leaks in theoperating pipes after one year.Fig. 24 Left; Test area in January, Right; Test area inFebruary.The demanded levels for the degree of compaction ofthe street are possible to reach also with a d.h. pipe60 mm below the surface as it gets compressedwithout hurting the pipe.There are still other issues that can be considered thatare not included in the tests presented in this paper,aspects as e.g. heat losses.5. ACKNOWLEDGEMENTFig. 25 Left Test area in March, Right; D. H. chamberrevealing itself in February.However the heat loss is of course bigger comparedwith normal standard because of that the pipes areplaced closer to the air.Frost action damagesIn theory one could imagine that the street on bothsides of the district heating pipe would erect during thewinter if there were soil that could frost heave. Thiscould of course damage the asphalt. But streets are notsupposed to be built with soil that could frost heave. Sothere should not be any problem.There was no notable difference in the height of thestreet over the district heating pipes compared to thestreet beside the test area during the winter.4. CONCLUSIONThe article probably describes the first operating d.h.pipes placed in backfill of 0-40 only 200 mm below thesurface in a street with heavy traffic. As expected, thepipes are working nicely. The loads that have beenmeasured are acceptable or even low for the d.h. pipe.As it seems also the street is satisfactory working eventhough there are d. h. pipes close to the surface.The pipe deformations are negligible with respect to thepipes function both over time and under an instanttraffic load.The authors would like to express their appreciation toa couple of key persons. There had been no test ofthis kind without their support and permitting. Thepersons are:– Mr. Bo Andersson Planing Manager atTrafikkontoret Göteborgs Stad,– Mr. Lars Ljunggren Manager at GöteborgEnergi AB and– Mr Göran Johnsson Technical Manager atPowerpipe Systems AB.6. REFERENCES[1] Molin J., Bergström G. and Nilsson S. (1997).Kulvertförläggning med befintliga massor, SwedishDistrict Heating Association FOU 1997:17, (inSwedish)[2] Bergström G., Nilsson S. and Sällberg S-E. (2001),Täthet hos skarvar vid återfyllning med befintligamassor, Swedish District Heating Association FOU2001:58, (in Swedish)[3] Nilsson S, Sällberg S-E, Bergström G, (2006)Grund förläggning av fjärrvärmeledningar, SwedishDistrict Heating Association FOU, FOU Värmegles,2006:25, (in Swedish)[4] Gudmundson T. ÅF-Processdesign AB, (2002),Casaflex-rör i Malmö 2001,. Swedish DistrictHeating Association, FVF 021241, (in Swedish)103

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia[5] Trafikverket, (2005) Allmän teknisk beskrivning förvägkonstruktion ATB Väg,http://www.vv.se/Startsida-foretag/vagar/Tekniskadokument/ATB-Allmanna-tekniskabeskrivningar/Vagteknik/Aldre-versioner/ATB-Vag-2005/, visited 2010-04-27, (in swedish)[6] Göteborg Stad Trafikkontoret, Bestämmelser förarbeten inom gatu- och spårområden i Göteborg,http://www.goteborg.se/wps/portal/!ut/p/c0/04_SB8K8xLLM9MSSzPy8xBz9CP0os3gjU-9AJyMvYwMDSycXA6MQFxNDPwtTo2Anc_2CbEdFABCTfUM!/?WCM_GLOBAL_CONTEXT=/wps/wcm/connect/goteborg.se/goteborg_se/Foretagare/Upphandling_staden%20som%20kund/Specifik%20upphandlingsinformation/art_N400_FOR_Up_SU_Trafikkontoret, visited 2010-04-27, (in swedish)[7] Trafikverket, (1993) Publikation 1993:19Bestämning av bärighetsegenskaper med statiskplattbelastning Metodbeskrivning 606:1993, (inswedish)104

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSTUDY ON THE HEAT LOSS REDUCTION METHOD FROM THE SECONDARYPIPELINES IN THE APARTMENT COMPLEXByung-Sik Park 1 , Yong-Eun Kim 2Sung-Hwan Park 1 , Yong-Hoon Im 1 , Hyouck-Ju Kim 1 , Dae-Hun Chung 1, Mo Chung 31 Building Energy Research Center, Korea Institute of Energy Research,102 Gajeong-ro, Yuseong-gu, Daejeon 305-343 KOREA, bspark@kier.re.kr2 Energy System Engineering, University of Science and Technology,113 Gwahangno, Yuseong-gu, Daejeon 305-333 KOREA, rainyday@ust.ac.kr3 Dept. of Mechanical Engineering, Yeungnam University,214-1 Dae-dong Gyeongsan-si Gyeongsangbuk-do 712-749 KOREAABSTRACTThis study aims to suggest better methods for reducingheat losses from the pipelines installed as secondaryheating pipes in the apartment complex in which hotwater is being supplied for space heating and hot waterby a district energy supply company. Right now thedistrict heat supplier is responsible only for the primarydistrict heating pipelines just before the substations inthe apartment complex. That is why the heat lossreduction becomes more important in the secondarypipelines after the substation in the Korean apartmentcomplex.Several methods to reduce the heat loss from thesecondary pipelines were set up and compared by asimulation technique. One of the methods is tocombine the hot water heating pipes and space heatingpipes. Another method is to install a small heatexchanger in each house to supply hot water from thesingle space heating pipeline. In this case we caneasily change the means of heat supply and the rightchoice of end users can be ensured for the means ofheat supply.In this study the preferable method to reduce the heatloss in the secondary pipelines has been suggested.The simulation result has shown about 30% heat lossreduction compared to the existing scheme for thesimple change of methods and much more reductionfor the optimization of pipe diameter and insulationthickness or surface enhancement by low emissivity.INTRODUCTIONKorea is characterized as having four distinct seasons.Apartment complexes became a typical type ofresidence in urban areas after the recent rapidindustrialization of last 30 years. At the moment overhalf of the population chooses to live in apartmentsrather than in individual houses and the trend willincreasingly continue in the future. There are threetypical heating methods for apartment complexes -individual heating, central heating and district heating.At the moment there is a lot of potential for districtheating and cooling. Korea has seen about a 10%supply of DHC among total residential houses which isvery low compared to that of European countries whichsupplies over 50% DHC.If we are to increase green growth with low carbon, it iscrucial to supply DHC, which has higher energyefficiency than any other method, in dense regions ofpopulation. The recent Korean government has showneffort in making a point of energy efficiency throughoutmain energy consuming sectors including buildingarea. However, the supply policy of DHC is now beingcrippled due to various reasons. Makers or consumersof individual heating devices do not have positiveattitudes toward DHC. Therefore it is important to drawattention to the multitude of benefits and merits ofDHC.There is certainly some heat loss from the pipelinesinstalled under the ground to supply the district energyfrom the power plant to the consumers. To reduce theheat loss from these primary pipelines manyinnovations and advancements have been made for along period since the district energy was supplied in thenorthern European countries. The heat loss generallydiffers according to the network type of pipelines. Themore compact that the network is, the less heat lossoccurs. But the type of network cannot be madearbitrarily by the designer. The designer can simplyoptimize the network in view of geological andenvironmental conditions, such as population density,not the type of network. Although heat loss exists withthe primary pipelines, it can be controlled andmaintained effectively by the district heat supplier. Onthe other hand, the heat loss from the secondarypipelines cannot be controlled properly by the buildingowners who are responsible for.The apartment complex is a unique housing system inKorea. It contains many high rise buildings of over 10,often over 20, stories high. In many cases it has over105

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaone thousand homes. However, the basic structure isalmost the same as that of western apartment buildingsexcept for the pipeline network between buildings andsubstation. In the past there was one substation in oneapartment complex. The substation has a minimum oftwo heat exchangers which are in general shell andtube type or plate type. Nowadays the number ofsubstations grows bigger and bigger. That means thatthe designer plans to install the heat exchangerseparately and respectively according to the buildingswhich stand nearby each other. The secondarypipelines have been said to have much heat loss inKorea. There have been a few studies related to heatloss from the secondary pipelines. It is very hard todistinguish between positive heat gain and heat lossfrom the pipelines installed within the buildings. If thepipelines are installed in the center of the building, theheat loss from the heating pipes or hot water supplypipes can be regarded as positive heat to theconsumer. But if the pipelines are installed near thebuilding surface, the heat from the pipes can beregarded as loss.Measurement and analysis of the heat loss fromseveral apartment complexes in Korea has been tried.The heat loss data from the several sites has beenstored and accumulated throughout the year. Asimulation method has been set up and the accuracy ofthe simulation has been investigated. Somealternatives to reduce the heat loss have beenprepared from the existing scheme. The simulationmethod and results have been presented in this paper.TYPE OF APARTMENT COMPLEXThe apartment complex was built and opened inNovember 2007. It has 8 buildings which arecomprised of 518 homes. Each home has 112 m 2 ofheating area. Fig. 1 shows the location and overallshape of the apartment complex which was chosento be measured and evaluated on the heat loss fromthe secondary pipelines. Many thermocouples andflow-meters were installed in the region of the pipelinesto collect information on heat demand pattern,temperatures and heat loss from certain regions to beevaluated.City water201202SupplementaryWaterHWRHWRHWRHWRHWS207205203HWRHWRHWRHWRHWSHWRHWRHWRHWRHWSCalorimeterM208206204Fig. 2 Secondary pipeline network from the substationANAYSIS OF HEAT CONSUMPTION PATTERN1) Space heating water flowrateSpace heating amount is being measured daily. Toomtemperature does not differ much between the homesin the apartment complex. Thus the temperaturedifference (ΔT) between inlet and outlet of the pipelineof individual homes remains fairly constant except inthe summer season. Therefore the heating water flowrate can be estimated from the following equationQ=CmΔT. In other word, the flow rate could beevaluated from the measured calorific amount. A goodexample of this is shown in Fig. 3.DHWRDHWSCHINASouthKOERANorthKOREAJAPANFig. 1 The location and the shape of the apartmentcomplexFig. 3 Temperature difference between supply and returnIn this study the flow rate was measured for the twomonths of November and December 2009 using theflow meter installed in the space heating water pipeline.And the hourly heating water flow rate of individualhomes for the year 2009 was extracted from thecomparison of the total measured amount and theindividual house measurement. Fig. 4 represents theannual heating water flow rate. Some differences existduring the cold winter season.106

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia<strong>12th</strong> floor11th floor10th floor9th floorT highT out8th floor7th floor6th floor5th floor4th floorT lowPrimary pipe lineSecondary pipe line3rd floor2nd floor1st floorFig. 4 Heating water flow rate2) Hot water flow rateHot water consumption is measured by ton from thegeneral water flow meter. Hot water supply line isdesigned to have a supplementary recirculation line inorder to supply instant hot water. By adding the waterwhich was used by individual homes to the heatexchanger, the flow rate can be constantly maintained.DHWSDHWRT baseFig. 6 Category of temperature charicterization.Tout:Temperature outside the buildingT sbTbase: Temperature of the underground spacefrequently open to the outside surrounding.Tsb:Temperature of the underground space closedto the outside surrounding.Tlow:Temperature of low-rise region in the buildingThigh: Temperature of high-rise region in the buildingFig. 5 Hot water flow rate3) Various temperatures outside pipesThe temperatures outside the pipeline werecategorized into four cases taking into considerationthe atmosphere outside the pipeline. The first one isthe underground space which is fairly open to theoutside of the building. The second one is theunderground space which is not so open to the outsideof the building. The other two are the spaces of low andhigh regions of the building which is not open to theoutside of the building. These temperatures, measuredaccording to the categories, were applied in thesimulation in view of the pipelines outsidecharacteristics.107Fig. 7 Varous temperature of the surroundings ofpipelines.SIMULATION METHODFor the heat loss simulation commercial tool,―Flowmaster‖ of 1 D system analysis has been used.Flowmaster is a program which can analyze thethermo-hydrodynamic characteristics of pipe systems ifthe following items are given such as the physicalproperties of pipes, flow rate and outside temperaturethrough the following equations. Annual heat loss canbe simulated by using the information such asmeasured temperatures, flow rates and variousphysical properties of pipes and insulation materialsaccording to the drawing of all the pipelines which are

E-2P-1P-2P-4E-1P-6P-3P-9E-3The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniainstalled underground between buildings andsubstation as well as in the buildings themselves in theapartment complex. The simulation was performeddaily in view of calculation time.VARIOUS SCHEMES FOR SIMULATION1) Present Scheme<strong>12th</strong> floor11th floor(1)(2)HWSHWRHWRHWRHeating water supplyHeating water returnHot water supplyHot water return10th floor9th floor8th floor7th floor6th floor5th floorHWSHWRHWRHWR(3)Primary pipe lineSecondary pipe line4th floor3rd floor2nd floorDHWSDHWRHWSHWRHWRHWR1st floorfluid temperature, °Cambient temperature, °Cconvection heat transfer coefficient,radiation heat transfer coefficient,internal heat transfer coefficient,insulation thermal conductivity,pipe thermal conductivity,external pipe diameter,external insulation diameter,internal pipe diameter,For accurate simulation, individual flow rate was usedrespectively and differently based on total measuredflow rate and read amount of individual flow meter of518 homes throughout the year. By doing this, the flowrate in the individual pipelines can be determinedaccording to the usage amount of heating water andhot water. This similarly leads to actual flow rate in thepipelines. From this complicated process, the numberof individual flow rate of hot water and heating watercomes to 378,140. Macro which was combined withExcel and Flowmaster was used for a 365 dayanalysis. This process requires 32 hours for 8 buildingsfor only one case.Fig. 8 Pipeline configuration of present schemeThe present scheme is composed of 4 pipelines, two ofwhich are for heating water supply and return and theother two of which are for hot water supply and return.Heating water is supplied to each home and returnedfrom each home and resultantly the flowrate of supplyand return are equal. On the other hand hot water hasa certain amount of recirculation in order to keepsupply water hot. Hence the same amount ofconsumed hot water in each individual home should besupplied to the hot water heat exchanger to guaranteeinstant hot water supply.2) Alternative AAlternative A is a scheme which removes the hot waterpipelines and combines with the heating waterpipelines. Thus there are only two pipelines of supplyand return from substation to each building. Thesesupply and return pipelines have two functions ofheating and hot water supply and return. By reducingfrom 4 to 2 pipelines the heat transfer surface area canbe decreased. However, this scheme hasdisadvantages in summer when the heating watersupply has been closed. If the pipelines should be usedfor the supply of hot water, the resultant water speed inthe pipelines would be very small.HWSHWRHWRHWRHeating water supplyHeating water returnHot water supplyHot water return<strong>12th</strong> floor11th floor10th floor9th floor8th floor7th floor6th floorHWRHWSHWR5th floor4th floor3rd floorPrimary pipe lineDHWSSecondary pipe lineHWS2nd floor1st floorDHWRHWRFig. 9 Pipeline configuration of Alternative A108

P-1P-4P-9P-7E-5P-14P-8The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFor hot water, this scheme is effective in heatexchange because the heat exchanger acts for theconsumed hot water only and can avoid extrarecirculation pipelines. But it has the drawback ofsupplying cold water or non heated water whenintermittently using hot water.3) Alternative BHWSHWRPrimary pipe lineDHWSDHWRE-1Heating water supplyHeating water returnSecondary pipe lineHWSHWR<strong>12th</strong> floor11th floor10th floor9th floor8th floor7th floor6th floor5th floor4th floor3rd floor2nd floor1st floorFig. 10 Pipeline configuration of Alternative BAlternative B is the same as Alternative A in the pointof unifying the heating water and hot water pipelines.But it is different in the point of the individual hot waterheat exchanger being installed in each home amongthe heating water pipelines. This scheme is said toresemble the pipelines installed in the apartments ofEuropean countries.HWSE-10E-11E-15E-14E-12E-13E-7E-8E-9E-6E-5E-4HWR5) Use of PEX PIPESPEX is being used in western countries as districtheating pipes for low temperature service fromrenewable resource application. For this reason PEXcan be used as secondary pipelines which normally areunder service of low temperature.VALIDATION OF SIMULATION RESULTTo validate the simulation result, the measurementvalue was compared with the prediction result by thepresent simulation method. The measured value ofDecember 2009 was used for heat loss referencevalue. As seen in Table.1 the simulation result fairlyagreed with the measured data. And the simulationmethod can be used without much modification. Formore accurate prediction it needs slightly moresupplementation in the numerical modelling of heattransfer phenomena of outer pipe surface andenvironment.Table1 Comparison of measured and predictedvalues[unit: MWh]Measurement heat Hot waterHeat loss rate9.44%loss 20.2 11.2supply 264.7 67.8Simulation heat Hot waterHeat loss rate9.90%loss 17.0 14.29supply 252.6 63.534) Alternative CHWS Heating water supplyHWR Heating water returnHWRHot water supplyHWR Hot water return<strong>12th</strong> floor11th floor10th floor9th floor8th floor7th floor6th floorHWRHWRHWSHWRSIMULATION RESULTFig. 12 shows typical heat supply and heat loss for the24 hours of 11.11.2009. The simulation result of eachscheme is shown from Fig. 13 to Fig. 17. Each graphshows similar patterns and the heat loss comparison ofeach scheme is summarized in Table 3.5th floor4th floor3rd floor2nd floorPrimary pipe lineDHWSSecondary pipe lineHWS1st floorDHWRHWRFig. 11 Pipeline configuration of Alternative CAlternative C is a variation of Alternative A to make upfor the defect of cold water supply when intermittentlysupplying hot water. In this alternative the recirculationpipelines are equipped in the buildings.Fig. 12 Hourly heat supply and heat loss of 11.11.2009109

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 13 Annual heat supply and heat loss (Presentscheme)Fig. 17 Annual heat supply and heat loss (PEX Pipe)The heat loss comparison of each scheme can besummarized in Table 2.Table 2 Heat loss comparison of each scheme[unit:MWh]Supply Loss Heat loss rateFig. 14 Annual heat supply and heat loss (Alternative A)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%Fig. 15 Annual heat supply and heat loss (Alternative B)Alternative B should supply heating water in thesummer season when it is not required for the supply ofheating water in order to supply hot water to theindividual home. From the comparison of Fig. 14 and15, Alternative B is more efficient than Alternative A inthe cold region.HEAT LOSS COMPARISON DUE TO THE CHANGEOF INSULATION THICKNESS,NOMINAL DIAMETERAND INSULATION MATERIALTable 3 shows a comparison among parameters whichaffect heat loss. There is frequent excessive design forthe nominal diameter of pipelines which are installed inboth the underground and the buildings. In thiscomparison the increases of insulation thickness by10 mm and 20 mm were considered. Also the decreaseof nominal diameter by 1 level was considered. It wastaken into consideration of insulation material changeand each combination of affecting parameters. In thiscomparison, the present popular design method ofpipelines and insulation was regarded as a referencefor 100% of heat loss and other alternatives wereevaluated from the reference heat loss relatively.Table 3 shows the result of relative comparison ofannual heat loss from the pipelines of the apartmentcomplex.Fig. 16 Annual heat supply and heat loss (Alternative C)110

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 3 Heat loss comparison due to variousparametersHeat loss comparisonPresent 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 + insulationthickness 50 mmPipe diameter downsizing + insulationthickness 60 mmCONCLUSION76.7%67.6%Present scheme for the secondary pipelines isevaluated to have 14% annual heat loss based on thetotal heat supply to the apartment complex. This is avery large amount when we consider that the primarydistrict heating pipeline has only about 4 to 5% annualheat loss in dense population urban areas.Heat loss by Alternative A can be reduced about 30%compared to that of the present scheme which hasbeen widely adopted in Korea until now. However,Alternative B has more heat loss compared to that ofAlternative A, which was not the common expectation.The main reason was the increase of the pipeinsulation surface area which acts as a heat transferarea. Alternative C and PEX system can bereplaceable when they have merits in the point of initialcost.More enhancements in heat loss can be extracted fromheat loss reduction by the selection of optimum pipediameter, good insulation material, increasinginsulation thickness and changing surface emissivity ofinsulation material.ACKNOWLEDGEMENTThe financial support from KDHC made this workpossible. This paper is based on the results of anongoing research project which will be completed at theend of 2010.REFERENCES[1] W. F. STOEKER, DESIGN OF THERMALSYSTEMS 3rd edition, Mc Graw Hill, pp. 53-160[2] Incropera, HEAT TRANSFER 5th,WILEY[3] Byung-sik Park et al, Study on the Reductionmethod of Heat Loss from the Secondary Pipelinesinstalled in the Apartment Complex, 2008[4] Byung-sik Park et al, Study on the Reductionmethod of Heat Loss from the Secondary Pipelinesinstalled 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, IEAR&D Programme on District Heating & Cooling111

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaHEAT LOSS OF FLEXIBLE PLASTIC PIPE SYSTEMS,ANALYSIS AND OPTIMIZATIONEJ.H.M. van der Ven 1 , R.J. van Arendonk 21Thermaflex <strong>International</strong> Holding B.V2 Liandon B.V.ABSTRACTA newly developed, in-house, test rig for measuringheat loss of pipe systems allows the user to analysevarious systems in a short timeframe. This allows quickinsights into heat loss variables and mathematicalanalyses. The effect of alternative compositions ofinsulation and other layers can be evaluated within ashort time span.This already led to improvements of the productionprocess and of the insulating foam.INTRODUCTIONLiandon developed a heat loss testing rig forThermaflex to test their produced flexible pipe systems.Within a short time the pipe system, undergoing a heatloss test, tends towards the controlled temperature inthe sections of the sample, the added power reachesequilibrium and the test results can be collected.Due to the short time required for testing, the results ofalternative production methods are easily available.Due to the short response time the test is a great helpin the search for product and production improvements.The objective of this paper is to present the results ofthe research to the overall heat loss performance of aflexible plastic pipe product, Flexalen 600.The objective of the research is:1 Find correlations between heat loss and otherparameters of the pipe system such as outerdiameter, inner diameter, foam surface and foamstructure. These correlations are determined by themathematical analysis of practical heat lossmeasurements.2 Find possibilities for the improvement of the pipeparameters by analysing the heat loss correlations.NOVELTY AND MAIN CONTRIBUTIONThe actual heat loss of pre-insulated pipe products isdetermined under controlled, similar conditions for anentire diameter range. This range comprises variousouter diameters, various inner diameters and variouscompositions in materials and pipe systems. The timerequired for one single test run varies from half an hourfor a small-sized pipe to about eight hours for thelargest sized diameter.The novelty of the test rig is described in the paper‗Verification of heat loss measurements conducted on(semi) flexible pipe systems‘ [3].The novelty for product improvement is that due to thereduced time required for a test run the effect ofalternative systems can be mathematically analysedand evaluated in a short time. In this way the analysesof alternative production methods has a shortfeedback. Optimization of the product can be effectedin a short time.In the near future the test rig will be used for qualitycontrol of the production process. This test will partlyreplace other currently applied standard tests, such asdensity and cell size measurements.METHOD DESCRIPTIONIn the Flexalen 600 pre-insulated pipe a PB mediumpipe is encapsulated in insulating foam, which isprotected against wear and tear in a corrugated hardcover pipe. The pipe product has a solid bondingbetween the insulation and cover and no bondingbetween the insulation and the medium pipe.According to EN 15632, the European Standard forpre-insulated flexible pipe systems, this pipe system isclassified part 3: Non bonded system with plasticservice pipes. The Flexalen 600 plastic pipe systemdiffers in some areas significantly from most othersystems in this class:1 Physical bonding between foam and outer casing,2 One layer of foam, filling the complete spacebetween service pipe and cover,3 Next to other connection methods the service pipescan also be connected by welding.Annex D of part 1: Classification, general requirementsand test methods give rules for calculation of the heatflow to ambient (heat loss) from measured values,making the heat flow of various parameterscomparable.The heat loss calculations of annex D are based on thethesis of Wallentén as published in Steady-state heat112

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonialoss from insulated pipes; Lund Institute of Technology,Sweden [2].The described heat loss calculation is valid for the preinsulatedpipes. Branches and connections areexcluded.systems. Test spools can be extracted directly fromproduction. In this way tests can be executed with freshproduct.Also testing of cured piping and piping that is aged andhas been degassed during storage or in hightemperature aging is possible.Knowing the relationship between heat loss andvarious parameters, the most prominent parameters forheat loss can be evaluated.The most prominent parameters may lead to theimprovement of the flexible pipe system to ensureoptimal performance with minimal heat loss. Theinfluence of several prominent parameters isdetermined and recommendations are given in order tooptimize the insulation performance.Figure 1, Thermaflex heat loss equipmentIn the new developed test rig (figure 1) a test spool(figure 2) is put in a slim fitting sleeve.The test spool is heated internally in three sections.The middle section of the spool is the test section.Heating in this section is controlled to obtain therequired test parameters. The two ends are heated tocompensate for the heat loss from the ends of themiddle test section. In this way an endless pipe isimitated.The outer side of the sleeve is water-cooled to obtainheat transport from the test spool.During the start of the test, heat is lost into the heatingof the pipe system and into the surrounding coolingwater. When heat losses have reached equilibrium, thesteady state heat transfer can be measured.Figure 2, Longitudinal section guarded end heating probeThe time span required for testing in the test rig israther short. The time to reach equilibrium lies in theorder of hours, depending on the diameter andinsulation thickness. Comparable tests often requiretime spans in the order of days.Containing various diameters, which are based on thestandard production outer diameters, the test rigenables heat loss tests for various diameters of pipingReliability and reproducibility of the test rig is discussedin [3] Verification of heat loss measurements conductedon (semi) flexible pipe systems (van der Ven et al).MANUFACTURING PROCESSThe Flexalen 600 product has been developed byThermaflex, located in The Netherlands. Thedevelopment started in 2002 and resulted in a firstsmall-scale commercial production in 2005. During theproduction of Flexalen 600 four different productiontechniques are combined, partly simultaneous andpartly sequential:1 Production of PB service pipes optionally coveredwith an EVOH oxygen barrier layer.2 Production of LDPE insulation foam to fill the areabetween medium pipe and outer casing.3 Production of outer casing of HDPE.4 Assembly of the different elements (1, 2 and 3) witha 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 equipmentcombined with technology developed in-house. Thecomplete production is a (semi)-in-line production. Allpipe systems are produced at Waalwijk in theNetherlands. Unique for the process is the ability toproduce continuous lengths. For practical reasons thelengths produced depend on the outer casing of theproduct and the size of the reel. The maximum lengthproduced can reach up to 2000 meters.The complete Flexalen 600 pipe system includes preinsulatedpipes, couplings, sleeves, pre-insulated T-connections, etc. The production range is described byEngel and Baars. [5]113

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaProduction of PB service pipesAll service pipes are made of Poly-butene. This is aplastic with a special combination of properties. Polybutenehas excellent heat and creep resistance,flexibility and strength at a long lifespan and is fullyrecyclable. In accordance with the temperatureduration profile mentioned in the BRL5609/EN15632,PB is suitable up to a maximum temperature of 95 °C.All PB-pipes are weldable by socket fusion, electrofusion and butt welding, which allows for an all plasticdistribution system without metal parts that are prone tocorrosion. The service pipes are produced via state-oftheart extruders.The production line consists roughly of an extruder,calibration tools for adjusting pipe size, cooling baths,and marking and cutting equipment.Pipe dimensions are checked inline every secondduring and after production with ultrasonicmeasurements. PB-pipes can be produced up to anouter diameter of 225 mm.If desired an outer oxygen barrier layer may be appliedvia co-extrusion up to an outer PB pipe diameter of 90mm.After production the PB-pipes are stored for a minimumperiod of 5 days and cured to create the correctcrystalline polymer structure. After curing the pipe isused for the Flexalen 600 production.Every batch produced is verified by the in-house QCdepartment according to Dutch directives- BRL-K5609, for PB pipes with oxygen barrier or- BRL-K17401 for PB pipes without oxygen barrier.Product and manufacturing processes are checked 6-8times a year and certified by independent agenciessuch as Bureau Veritas, KIWA and CSTB and leCentre Scientifique et Technique du Bâtiment. Thequality of the QC department is validated by thesechecks and by internal and external audits.Production of LDPE insulation foamThermaflex has now approximately 35 years ofexperience in the production of LDPE foam viaextrusion techniques.Most of the raw materials that are used are tailor-mademixtures according to Thermaflex specifications.Through these specifications and proper productionquality control the company‘s philosophy related tocore business is also realized for raw materials.During the heating, melting and mixing of the rawmaterials the foaming agent is injected into theextruder. This foaming agent is a hydrocarbon thatcauses the expansion of the LDPE.The quality of the foam, moreover the insulationproperties of the foam, depends on parameters such asdensity, cell size and chemical composition. Allparameters are measured and adjusted within limitedtolerances to meet the specifications. The bestefficiency is further improved when the space betweenthe service pipe and the foam is filled better.As there is no bonding between the service pipe andfoam, there is no risk of damaging the foam andproperties by expansion of the pipes due to thermalfluctuations in the applications.Production of corrugated outer casing (HDPE) andassembly of complete productThe outer casing is applied by an extrusion processand thermally welded to the foam after the pipe hasbeen inserted. The outer casing is also corrugated tooptimize the flexibility of the finished product. Now theproduct is ready for coiling.After production the final product must cure for 5 days.During this curing period the degassing of the foamingagent starts, while the insulation foam is still stabilizingProduction testing and controllingDuring the production, parameters are checked andcontrolled, such as:1 Chemical composition of the foam.2 Settings of all extruders involved (foam extruder andextruder for corrugated outer casing).3 Density and cell size of the foam.4 Dimensions of the foam (outer diameter and innerdiameter).5 Line-speed of all involved products (Foam / PB-pipe/ End-product).6 Thickness of the corrugated outer casing and theconnection of the corrugated outer casing to thefoam.7 Since a few months: in-line production control ofHeat Loss of the pre-insulated pipe system.The most prominent factors to influence heat loss areparameters one to four. The foam lambda is directlyaffected by these factors as represented in equation 1. total convection conduction radiation blowingagent(1)114

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaRESULTS AND CORRELATIONSIt is again the company´s philosophy that paves theway for innovations. One of these innovations is thedevelopment of an in-house device for testing pipesystems, directly from production, cured and degassed.This proved to be a suitable device for productioncontrol, but also for gaining more insights into productparameters.The objective of testing is to improve the knowledge ofthe produced pipe systems in order to optimize:Production methods:Machine data can be adjusted based on test results.Test results may lead to new production methods withnew equipment.Chemical and physical composition of layer material:Knowledge of various composition materials may leadto an improvement of insulating values.Cell structure and gap between insulating foam andmedium pipe:Cell size influences values, test result basedimprovements are possible. The gap is a bad insulator.The quest for a minimal gap started with testing.Improvement and minimization of this gap was anachieved challenge in the testing period.Thermaflex is a lean and mean organisation thatresponds quickly to new insights. Therefore newinsights were applied even before the complete rangeof production testing was performed.For the company, improvements of product andproduction have the highest priority. Although theproduction range is wide, the insight into specific andgeneral parameters increased considerably.The research provides the prominent variables toimprove insulation performance. Practical heat lossdetermination, in combination with analytical studies,results in a clear understanding of heat loss behaviourin single and twin flexible pipe systems during theirentire lifetime.As a result of the tests the manufacturing process isimproved in two steps.The emphasis of the first step was to diminish the cellsize of the foam. This succeeded in a decrease of cellsize by some 20%.DISCUSSION OF PARAMETERSTable 1 summarises the results of measurements andcalculations of tests on Flexalen 600 pipes directly fromproduction. Various diameters are tested andcalculated according to EN 15632 for a surfacetemperature of 10 °C and a common mediumtemperature of 70 °C (Instead of the maximum mediumtemperature of 95 °C).The table indicates the relationships between product,cross sectional area of the foam, foam density, cell sizeof the foam, remaining foaming agent, calculatedthermal conductivity and the calculated heat loss of aburied piping system.The products 50A25, 63A32, 75A40 and 90A50 arenewly developed. These products are not necessarilyDistrict Heating products. However, they are producedusing the same process and have their application inthe connection between the district heating networkand the building or house. It is also applicable in caseof low temperature differences, cooling or in-househeating or cooling.Table 1: Test results of fresh, uncured piping systemsProductFoamsection density cell size agent 50,calc Heat Loss*mm² kg/m³ (mm) % mW/m.K W/m50 A 25 1.473 50,0 0,47 52 39 15,363 A 32 2.313 34,0 0,50 52 38 15,275 A 40 3.044 38,0 0,40 46 44 17,890 A 40 5.105 42,0 0,80 64 51 17,190 A 50 4.398 39,0 0,80 62 55 23,0125 A 63 9.155 39,0 0,88 70 56 22,0160 A 75 15.688 40,3 1,20 81 54 21,0160 A 90 13.745 35,0 1,30 85 61 25,2200 A 110 21.913 45,0 1,60 81 68 27,4*) calculated heat loss of buried system at temperature difference of 60 KIn Graph 1 foam density and cell size are related to thecross sectional surface.In Table 1 the foam density varies from about 35 kg/m³to about 50 kg/m³. Graph 1 shows hardly anyrelationship with the surface of the cross section.Table 1 shows that cell size varies from 0.47 to 1.60mm. Graph 1 shows that cell size is directly related tothe cross sectional surface, however less than 1 to 1.This relationship is influenced by physical productionparameters.The latest step is altering production such that thecontent of anti-radiation agent increases. The initialresults are promising but are not yet conclusive as theanti-radiation agent is also a good heat conductor.115

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAverage cell size [mm]60504030201001,81,51,20,90,60,30,0- 5.000 10.000 15.000 20.000 25.000Cross section of foam [mm²]Graph 1: Density and cell size in relation to foam crosssectionHeat loss of buried piping [W/m]40,035,030,025,020,015,03050 100 150 200Outer diameter pipesystem [mm]Graph 2: 50,calculated and calculated heat loss of buriedpipe in relation to outer pipe size8070605040Conductivity [mW/m.K]Density [kg/m³]Graph 3 shows the relationship between outer pipesystem diameter size and the percentage of foamingagent directly from production. With increasingdiameter the foaming agent increases, possibly to anasymptotic value.An additional interesting factor is the ongoing processin the foam during and after production. As describedbefore the final step in the production is a 5-day curingstage.During the curing stage the foam expands and part ofthe foaming agent releases from the foam. As the foamis locked by a hard outer shell, expansion is directedinwards. By this the gap between the foam and the PBmedium pipe, typical for our production method, isdecreased.Table 2 shows the effect of curing and degassing onboth the contents of foaming agent and the calculatedheat loss.Even when forced, degassing takes time. The numberof degassed samples manufactured in the same wayas the fresh samples is therefore limited. Table 2 isshort due to a lack of adequate and comparablesamples.Table 2: The effect of time on curing and degassingFresh6 days curing DegassedProduct AgentHeatHeatHeatAgent AgentLossLossLoss% W/m % W/m % % W/m %63 A 32 52 15 23 17,5 15 0 17,1 1275 A 40 46 18 40 16,9 -5 0 20,1 1390 A 40 64 17 53 18 5 0 18,5 8Graph 2 shows the influence of outer pipe size tocalculated conductivity 50 and heat loss of buried pipesystems.It also shows that part of the increase of the heat losswith the diameter is caused by increase of conductivity.Percentage foaming agent90807060504050 100 150 200Outer diameter pipesystem [mm]Graph 3: Foaming agent content in relation to outerdiameter sizeHeat loss of buried piping [W/m]2019181716150 10 20 30 40 50 60Percentage of foaming agentGraph 4: Relationship between heat loss and foamingagentGraph 4 shows that there is a tendency of decreasingheat loss with increasing foaming agent. This tendencyhas seems weak. The spread is large over the entiregraph.116

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaRESULT BASED FUTURE DEVELOPMENTSThe tests have led to improvements to the flexibleplastic pipe products produced by Thermaflex. In thenear future we expect improvements in:Production methods:Starting June 2010 a change in the machineconfiguration will be implemented as an extra step. Thenew configuration improves temperature control in theextruder, which leads to better cell structure.Chemical and physical composition of layer material:Up to a certain degree the anti-radiation agentimproves isolating values. With trial and error the antiradiationagent content is increased. Up till now themaximum content has been limited by productionmethods. Research is required to investigate maximumdesired value for insulating effects.Cell structure and gap between foam and medium pipe:In the tests we see variations in cell structure and gapwidth. Future research will aim at acquiring moredetailed knowledge of these phenomena.Up till now the heat loss performance on single pipeshas been measured and analysed. This has resulted inan understanding of the heat loss principles in districtheating systems. Twin pipe systems will soon betested, analysed and evaluated.In this paper only the heat loss of the Flexalen 600 preinsulatedpipe product has been handled. Informationabout the system can be read in [4] Heat lossoptimization of flexible plastic piping systems, life timeheat loss performance (Korsman et al) and [5] Neweconomical connection solutions (Engel).CONCLUSIONThe results of testing are reliable. Knowledge of theproduct and production has led to promisingimprovements of both.Further research will certainly lead to furtherdevelopments.Heat loss of buried piping [W/m]31.029.027.025.023.021.019.017.015.0Heat lossHeat loss Degassed50 70 90 110 130 150 170 190Outer diameter pipe system [mm]Graph 5: Degassed heat loss values of a buried system ata temperature difference of 60KExamples of product improvementFurther research has led to product improvements.Based on these improvement proposals Thermaflexhas been able to produce new pipe system samples.As represented in graph 5 the new samples have aheat loss decrease up to 16 percent compared to theprevious results.Heat loss of buried piping [W/m]40.035.030.025.020.015.0Heat lossHeat loss New50 70 90 110 130 150 170 190Outer diameter pipe system [mm]Graph 6: New heat loss values of a buried system at atemperature difference of 60KADDENDUMDegassingThe Thermaflex pipe system is liable to the process ofdegassing. Degassing causes the heat loss values torise over the products life time. Extra research on thissubject shows an average heat loss increase of9 percent (range 5–13 percent) (graph 5). Heat lossesare calculated according to EN 15632.117

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaACKNOWLEDGEMENTWe would like to acknowledge P. van Rijswijk for hisdedication to all the heat loss measurementsperformed during this research.FURTHER INFORMATIONQuestions concerning the paper may be addressed to:Thermaflex <strong>International</strong> Holding B.V,Veerweg 15145NS WaalwijkThe Netherlandswww.thermaflex.comLiandon B.V.Dijkgraaf 46920AB DuivenThe Netherlandswww.liandon.comREFERENCES[1] EN 15632 District heating pipes, Pre-insulatedflexible pipe systems, Requirements and testmethods[2] P. Wallentén; Lund Institute of Technology,Sweden; 1991, Steady-state heat loss frominsulated pipes[3] E. van der Ven, F. Duursma, H. Korsman, I. Smits;Paper on DHC, Tallinn; 2010, Verification of heatloss measurements conducted on (semi) flexiblepipe systems[4] H. Korsman; Paper on DHC, Tallinn; 2010, Heatloss optimization of flexible plastic piping systems,life time heat loss performance[5] C. Engel and G. Baars, ―New economicalconnection solution for flexible piping systems‖,12 th ISDHC 2010.118

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaCOMPARISON OF COMPETITIVE (SEMI) FLEXIBLE PIPING SYSTEMS BY MEANSOF HEAT LOSS MEASUREMENTI.M. Smits 1 , J. Korsman 1 , J.T. van Wijnkoop 1 and E.J.H.M. van der Ven 21Liandon B.V.2 Thermaflex <strong>International</strong> Holding B.V.ABSTRACTDifferent types of pre-insulated pipes are tested ontheir heat loss values. Three flexible pipes and a rigidpipe are tested. The different heat loss values arecompared not only on absolute heat loss, but also ontheir performance relative to the insulation surface. Theheat loss values are measured according to EN 15632and published as ―declared values‖. The declaredvalues are calculated according to EN 15632 AnnexD1-D3.The flexible pre-insulated systems, with PE and PE-Xfoams, show a variance of up to 5 W/m in the heat lossvalues. These absolute differences in the system arecaused by the outer casing dimensions of the preinsulatedpipes. Recalculation to the same outer casingdiameters shows a slight advantage for the PE systemin service pipes of 32 and 63 millimetresThe flexible piping system with the PUR insulationfoam on the other hand performs better compared toequally dimensioned flexible PE and PE-X insulationfoams.Flexible pre-insulated pipes have a higher heat losscompared to rigid pre-insulated pipes. Recalculation tothe same transport capacity [kg/s] and the same outercasing diameter also shows that rigid pre-insulatedpipes perform better. However the fact that smallerdiameters show a smaller heat loss difference betweenrigid and flexible pre-insulated pipes is interesting.INTRODUCTION & OBJECTIVEThe objective of this research is to compare differenttypes of competitive (flexible and rigid) pre-insulatedpipes on their differences in heat loss values. Thecomparison is based on an overall heat lossmeasurement under similar conditions. Overall heatloss is determined for different samples of pre-insulatedpipes, by using newly developed heat loss testingequipment as described in ‗Verification of heat lossmeasurements‘ by J.T. van Wijnkoop et al. [1]. Theheat loss data of these flexible pipes will be comparedwith practical measurement on a rigid pre-insulatedpipe.119Notice that this study does not compare entire districtheating systems. For system comparisons see „Heatloss analysis and optimization of a flexible pipingsystem‟ by J. Korsman et al. [2].NOVELTY AND MAIN CONTRIBUTIONWhere most studies only focus on one product thisstudy compares different types of flexible pre-insulatedpipes on their practical heat loss values and gives anexplanation of the practical heat loss values. It alsocompares flexible pipes with rigid pre-insulated pipeson an equal basis.BRIEF METHOD DESCRIPTIONFirst a brief description of different types of flexible preinsulatedpipes and a rigid pre-insulated pipe is given.This chapter highlights the differences and similarities.The different types of foam for plastic pre-insulatedpipes are described in a separate paragraph.Secondly, the method of testing is briefly addressed.Thirdly, the different types of flexible pre-insulatedpipes are tested on their absolute heat loss just afterproduction.Since heat loss of pre-insulated pipes can increaseover time due to degassing of the insulation foam, agas analysis is performed on all test samples.In the second paragraph the absolute heat loss valuesof the different types of flexible pre-insulated pipes arecompared on the basis of service pipe dimensions. Thethird paragraph defines a comparison on the basis ofinsulation surface and service pipe dimension.In the fourth paragraph the comparison of flexiblepre-insulated pipes versus a rigid pre-insulated pipe isdescribed. The comparison in the third and fourthparagraph is based on ―declared values‖. The definedconditions are: (1) thermal conductivity of soil: 1.0W/m.K, (2) thermal transmittance factor of earth toambient air: 0.0685 m 2 .k/W and (3) soil covering:0.8 m.In the first paragraph absolute heat loss values arecompared on the basis of corresponding service pipedimensions. The second paragraph gives a comparisonbased on equal transport capacity for flexible and rigid

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniapre-insulated pipes. The third paragraph founds thesecond comparison by adding a heat loss value basedon insulation surface.Finally, both flexible and rigid pre-insulated pipes arecompared, resulting in conclusions concerning flexibilityversus heat loss behaviour.barrier (EVOH) is placed in Polyurethane (PUR)insulation foam with a corrugated outer casing of highdensity poly-ethylene (HDPE). This product shows atight connection between the service pipe, the foamand the outer casing. This product therefore cannot bere-used once it is formed.PRE-INSULATED PIPESThis paper compares different types of pre-insulatedpipes and highlights their mutual similarities anddifferences.The flexible pre-insulated pipe systems are; Two different types of Cross linked Polyethylene(PEX) service pipe with Cross linkedPolyethylene (PEX) insulation; One type of Cross linked Polyethylene (PEX)service pipe with Polyurethane (PUR) insulation;One type of Polybutene (PB) service pipe withPolyethylene (PE) insulation.The rigid pre-insulated pipe system is;One type of Steel (ST) service pipe withPolyurethane (PUR) insulation.Firstly all types of pre-insulated pipes are functionallyexplained. Secondly the different kinds of foamproduction methods are described. All types of preinsulatedpipes described are commonly availableproducts used for district heating purposes in Europe.PEX service pipe with PEX insulationA short description of the PEX/PEX systems is given.Figure 1 shows the cross section view of thePEX/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-densitycross linked poly-ethylene (PE-X) insulation foam witha corrugated outer casing of high density Poly-ethylene(HDPE). Because of the cross linking in both servicepipe and foam this product cannot be re-used.Fig. 2. Section view of PEX/PUR pipePB service pipe with PE insulationThe third type of pre-insulated pipe is a flexible PB/PEpipe. Figure 3 shows the cross section view of thePB/PE/PE pre-insulated pipe. A Polybutene (PB)service pipe with anti-oxygen barrier (EVOH) is placedin a low-density poly-ethylene (LDPE) insulation foamwith a corrugated outer casing of high density Polyethylene(HDPE).The PB service pipe makes it possible to use electrofusion welding with a PB coupling. This makes a strongbond. Corrosion is not an issue, because PB is inertwith water.There is no connection between the PB and foam plusouter casing. Therefore it is possible to re-use bothelements in own production. The complete product canbe re-used.Information regarding the use of this product is given in‗‘New economical connection solutions for flexiblepiping systems‘ (Engel) [7].Fig. 3. Section view of the PB/PE pipeSteel service pipe with PUR insulationFig.1. Section view of the PEX/PEX pipePEX service pipe with PUR insulationThis paragraph describes the flexible PEX/PUR preinsulatedpipe. Figure 2 shows the cross section viewof the PEX/PUR/PE pre-insulated pipe. A cross linkedpoly-ethylene (PE-Xa) service pipe with anti-oxygenThe last system described in this paper is the rigidsteel/PUR system. Figure 4 shows the cross sectionview of the ST/PUR/PE pre-insulated pipe. A steel (St)service pipe is placed in Polyurethane (PUR) insulationfoam 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.120

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThis system is called a rigid system with a tightconnection between the service pipe, the foam andouter casing. Once it is formed, this product cannot bere-used in own process.Fig. 4. Section view of Steel/PUR pipeextruder. Once the material flows out of the extruder,the pressure drop causes the expansion of thehydrocarbon. The aggregation state of the moleculechanges from liquid into gas.Examples of hydrocarbon gases that can be used are:LPG, Butane or Isobutane. And just like the PUR foamthere is a degassing effect: the exchange of theblowing agent with air will increase the heat loss of theproduct. This effect is shown in ‗Heat loss of flexibleplastic pipe systems, analysis and optimization‘ byE.J.H.M. van der Ven et al. [4].Foam production processesThe insulation foams described in this paper are madeof Poly-urethane (PUR) foam, Poly-ethylene (PE) foamor cross linked Poly-ethylene foam (PE-Xa). Thesefoams have different properties. Some of theseproperties influence the heat loss properties of thecomplete product.Polyurethane (PUR) foamPUR foam is a thermo-set foam. It is made out of twochemicals, a Poly-alcohol and an Iso-cyanate. Thesematerials react and the Polyurethane is formed. Thisreaction is irreversible, so the material can never returninto its original chemicals. The blowing agent for thiskind of foam can be Carbon Dioxide, Nitrogen orHydrocarbon molecules, for instance Cyclopentane orButane.If Hydrocarbon gases are used, these gases stronglyinfluence the heat loss performance of the preinsulatedsystem. These gases have different thermalconductivities compared with air. After production of thefoam an exchange with air starts. A product that isfreshly made contains a high percentage ofHydrocarbon gases. At this point in time the productwill have the lowest heat loss possible. If the sameproduct is for instance three years old it contains moreair and less Hydrocarbon gases due to gas diffusion.And so the product will have a higher heat losscompared to the fresh product. The process ofdegassing is described in research papers ‗Long termheat loss of plastic Polybutylene piping systems‘ by S.de Boer et al. [3].Poly-ethylene (PE) foamPE foam is a thermoplastic foam. Once it is formed, itcan go back to its original state by heating it above itsmelting point. Because of this property, it is possible tore-use these kinds of foams.The foaming process to make PE foam is called the―physical foaming process‖. A Hydrocarbon molecule ismixed into the PE matrix under high pressure in an121Cross linked Poly-ethylene foam (PE-Xa)Although cross linked PE foam (x-PE) is also made ofPE, there is a big difference compared to PE foam: thetype of Blowing agent.The foaming process to make x-PE foam is called the―chemical foaming process‖. In this case a chemical ismixed into the PE matrix. The blowing agent can forinstance be Azodicarbonamide. While heating thematrix, the chemical starts decomposing and gases arereleased. These gases are Carbon dioxide andNitrogen. The thermal conductivity of these gases ismore or less equal to the thermal conductivity of air. Sothe aging effect of this product in relation to the heatloss is less.To make this foaming process possible, it is necessaryto connect the Poly-ethylene chains with each other.This is called the cross link. The complete process tomake x-PE foam is called the ―chemical foamingprocess with cross link‖. To make the comparison withPE foam complete: this process is called the ―physicalfoaming process without cross link‖.The blowing agent is not the only additive thatinfluences the thermal conductivity of the foam andtherefore the heat loss properties of the pre-insulatedsystem. Also other additives can influence the thermalconductivity of PE foam.Nucleating agents will influence the cell structure of thefoam. As a basic rule: the finer the foam the lower thethermal conductivity. With this additive the convectionpart of the insulation material will be influenced.Another additive that influences the thermalconductivity is an anti-radiation additive. By using thisspecial kind of additive it is possible to create areflection of radiation energy.HEAT LOSS TEST METHODThis chapter briefly describes the test rig and testmethod used to determine the absolute heat losses ofthe different types of pre-insulated pipes.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTest rigThe test rig is used to determine the absolute heatlosses. The test rig has been designed in compliancewith EN 15632 and the tests are carried out accordingto ISO 8497 and EN 15632.The physical part of the Thermaflex heat lossequipment consists of three sections. The first is thewater cooled compartment in which all the tests areperformed. This compartment is kept at a constant23 °C during each measurement.Heating probes are used as a heat source. Theseheating probes are custom made by preparing a twometer Thermaflex piping segment of all availablediameters. The third part of the heat loss equipment isthe control unit. This unit powers the probes andregulates the temperature and reads out thetemperature and power values.Method of testingDifferent heating probes are used for the testing. Theprobe with the appropriate diameter is inserted in a testsample and inserted in the cooled test section.The heat loss measurement is done by measuring theenergy required to keep the probe at a constanttemperature, by measuring the current at constantvoltage in the heating coils and calculating the powerconsumption. Since the middle/testing coil is exactlyone meter in length the required energy represents theexact heat loss through one meter of piping andinsulation in W/m. For this paper the heat loss isdetermined for multiple probe temperatures.InformationFor more information concerning the test rig andmethod of testing see the paper ‗Verification of heatloss measurements‘ (J.T. van Wijnkoop et Al. [1])Blowing agent analysisAll products that are involved in this paper have beenanalyzed on quantity of blowing agent and type ofblowing agent. The following results were found:The samples of PEX/PEX I and PEX/PEX IIdid not show any amounts of hydrocarbonblowing agents;The PB/PE samples contained a quantity ofhydrocarbon blowing agent over 50 percent;The samples of ST/PUR and PEX/PURproducts contained a mixture of hydrocarbongases. These gases were analyzed. Bothproduct types contained approximately 95% ofblowing agent.Measured samplesThe table below contains all products and dimensionsof outer casing and service pipe that are treated in thispaper.Table 1: Measured productsST/PURDc/DsPEX/PURDc/DsPB/PEDc/DsPEX/PEXIIDc/DsPEX/PEXIDc/Ds-- -- 62A32 -- --90DN25 -- 90A32 90A32 90A32125DN50 -- 125A63 160A63 175A63162DN80 162A110 200A110 200A110 200A110Results of testingAll results have been extracted from measurementscarried out by the Thermaflex testing rig. The newEuropean standard EN 15632 has been used.This standard describes in Annex D a method topresent the results of testing in end-use condition. Thismeans: the product is buried in soil. According AnnexD.3 the following general values are used for thecalculation:Soil coveringo0.8 mThermal transmittance factor of earth-airo0.685 m 2 .K/WThermal conductivity of the soilo1.0 W/(m.K)The heat loss is calculated using the followingformulas:(1) Q (2) Z T flow T surroundingd 42R soil R flow H(3) Z c Z R 0 soil1(4) R soil ln2 soil4Z cd 4WmmmmKAll results are presented in W/m, measured andcalculated at a temperature difference of 60 Kelvin.This temperature difference is derived from innerservice pipe temperature minus surrounding ambientW122

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniatemperature (70 degrees Celsius minus 10 degreesCelsius).Table 2: Results according to EN 15632 at a temperaturedifference of 60 Kelvin.Product Type Heat LossBuriedsystemW/mΛ System(W/m.K)ST/PUR 90DN25 11.6 0.042ST/PUR 160DN80 16.0 0.033PEX/PUR 162A110 22.3 0.049PB/PE 63A32 15.2 0.038PB/PE 90A32 12.8 0.044PB/PE 125A63 22.0 0.056PB/PE 200A110 27.4 0.068PEX/PE II 90A32 16.6 0.057PEX/PE II 160A63 17.6 0.055PEX/PE II 200A110 31.1 0.073PEX/PE I 140A32 12.5 0.057PEX/PE I 175A63 17.6 0.059PEX/PE I 200A110 28.8 0.051COMPARISON OF FLEXIBLE PLASTICPRE-INSULATED PIPESThis chapter compares the flexible pre-insulated pipes.The comparison is based on three diametersrepresenting the entire diameter range for plastic preinsulatedpipes.The comparison is expanded by evaluating the heatloss in correlation to the outer casing diameter (resp.the foam area).In Table 1 the flexible plastic pre-insulated pipes aredefined. These are the products PEX/PEX, PEX/PURand PB/PE.For more information concerning the PB/PE preinsulatedpipes see „Heat loss of flexible plastic pipesystems, analysis and optimization‟ by van der Ven etal. [4].The 32, 63 and 110 millimetre service pipesFirst the absolute heat loss is displayed, followed bythe insulation area analysis.Absolute heat lossIn this paragraph all absolute heat loss values arecompared for the 32, 63 and 110 millimetre servicepipes.In Graph 1 the results are displayed for temperaturedifference of 60 Kelvin.Heat Loss [W/m]35.030.025.020.015.010.05.00.0PB/PE PEX/PEX I PEX/PEX II PEX/PUR90/32 160/63 200/110Diameter service pipe [mm]Graph 1 Absolute Heat Loss 32, 63 and 110 mm servicepipe (dT = 60 K)The products based on PE or PE-x foam show higherheat losses for the 110 mm service pipe than thesystem based on PUR foam. The difference isapproximately 20 percent.The different test samples show a wide variance in thediameter of the outer casing.Therefore, only the results for the 32 millimetre servicepipe are comparable for PB/PE 90A32 and PEX/PEX II90A32. For the 110 mm service pipe, a comparison canbe made between the PEX/PUR 200A110, PB/PE200A110, PEX/PEX II 200A110 and PEX/PEX I200A110.Another difference in this comparison is the use of aPB pipe or a PE-x pipe. PB and PE-x have differentthermal conductivities (0.19 W/m.K versus 0.40W/m.K). However, this effect is already corrected byusing the Wallentén equation [5], as shown in (1).Insulation areaTo compare the different kinds of flexible pre-insulatedpipes on their performance, all outer diameters arealtered towards 90, 160 and 200 millimetres. The123

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniacorresponding heat loss is calculated using the thesisof Wallentén [5], as in (5). i Where:1ln sd 2d 1 2 T probe T casing1 ln id 3d 21 ln c T probe /T casing = Probe / Casing temperature d 1 to d 4 = inner/outer diameters of servicepipe and casingd 4d 3λ s , λi, λ c = heat coefficient of service pipe,insulation and casingGraph 2 represents the comparison on the basis ofthe same outer casing.35.030.0PB/PE PEX/PEX I PEX/PEX II PEX/PUR(5)is expanded by evaluating the heat loss incorrelation to the foam area.Rigid Pre-insulated pipesThe different systems and their correspondingdimensions are represented in Table 1.The rigid pipe product that has been testedaccording to EN 15632 was the ST/PUR product.First the absolute heat loss is displayed, followed by arecalculation towards transport capacity and finally theinsulation area analysis.Absolute heat lossIn this paragraph all absolute heat loss values arecompared. The rigid DN25 pipe service pipe iscompared with a flexible PB/PE-x service pipe with anouter diameter (OD) of 32 mm. DN50 is compared withOD 63 mm and DN80 is compared with OD 110 mm.In Graph 3 the results are displayed for temperaturedifferences of 60 Kelvin.Heat Loss [W/m]25.020.015.010.05.00.090/32 160/63 200/110Diameter casing/service pipe [mm]Graph 2 Relative Heat Loss 32, 63 and 110 mm servicepipe, all with an equal outer casing (dT=60K).Result analysisThe flexible pre-insulated systems, with PE and PE-xfoams, show a variance in heat loss values. Theabsolute differences in the system are caused by thedimensions of the pre-insulated pipes and the quantityand type of blowing agent that has been used. Also therecalculation to the same outer casing diametersshows an advantage for the PE foamed system in PBservice pipes of 32, 63 and 110 millimetres.COMPARISON OF FLEXIBLE PLASTIC PRE-INSULATED PIPES VERSUS A RIGID PIPINGSYSTEMIn this chapter the flexible pre-insulated pipes arecompared with a rigid piping system. Thecomparison is based on diameter. The comparison124Heat Loss [W/m]PB/PE PEX/PEX I PEX/PEX II PEX/PUR ST/PUR35.030.025.020.015.010.05.00.0DN25-PB32 DN50-PB63 DN80-PB110Diameter service pipe [mm]Graph 3 Absolute Heat Loss DN25/PB32, DN50/PB63 andDN80/PB110 mm service pipe (dT = 60 K)).The different test samples show a wide variance in thediameter of the outer casing.The heat loss for ST/PUR 160DN80 is much lowercompared to the heat loss of the 200A100 flexiblepiping products.Even the difference with the PUR based PEX/PURsystem is high (28 percent). For the PE and PE-x foambased products the difference is even higher(42 percent)The heat loss for ST/PUR 90DN25 is more or lesscomparable with the heat loss for PB/PE type 90A32(9 percent). So it seems that for smaller sizes the

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniadifference in absolute heat loss is lower, compared tothe absolute heat loss difference for larger sizes.Transport capacityWhen comparing rigid steel service pipes with flexibleplastic service pipes there is a difference in transportcapacity for comparable diameters.This paragraph calculates the amount of heat losswhen transporting water of 70 degrees Celsius throughone meter of steel DN25, DN50 and DN80(k-factor = 0.07 mm, velocity = 1.0 m/s). Subsequentlythe same amount of heat loss is used as a referencefor calculating the amount of water that can betransported through a plastic pipe 32 and 63(k-factor = 0.007 mm, velocity = 1 m/s). For thesecalculations the thesis of Colebrook and White [6] isused. The results of this calculation are displayed inTable 3.Table 3:Velocity[m/s]HeadLoss[Pa/m]Flow[kg/s]DN25Calculation results transport capacityPB32DN50PB63DN80PB1101.0 1.072 1.0 0.81 1.0 0.76425 188 1120.64 0.57 2.33 1.68 5.35 4.86Next the absolute heat loss is recalculated to an equalflow per diameter. The basis is 0.57 [kg/s] for theDN25/PB32, 1.68 [kg/s] for the DN50/PB63 and4.86 [kg/s] for the DN80/PB110.Insulation areaTo compare the flexible pre-insulated pipes and therigid piping system on their performance, all outerdiameters are altered towards 90 and 160 millimetresand compared on the same transport capacity. Thecorresponding heat loss is calculated using the thesisof Wallentén, as in (5). The steel DN80 and PB 110has the same outer casing and is not recalculated.Overall comparisonThe comparison on basis of the same outer casingand transport capacity is shown in Graph 4.Heat Loss [W/m]PB/PE PEX/PEX I PEX/PEX II PEX/PUR ST/PUR40.035.030.025.020.015.010.05.00.090/DN25-PB32 160/DN50-PB63Diameter service pipe [mm]200/DN80-PB110Graph 4 Relative Heat Loss, all with an equal outercasing and transport capacity.Result analysisFlexible pre-insulated pipes have a higher absoluteheat loss compared to rigid pre-insulated pipes.Recalculation to the same transport capacity [kg/s] andthe same outer casing diameter shows that rigid preinsulatedpipes perform better.The reason for this difference is the relative small innerdiameter of the plastic service pipes. The low k-factorcan not compensate for the smaller diameter. Table 4shows the steel versus plastic service pipe diameterdimensions.Table 4: Service pipe diameter dimensionsID[mm]DN25 PB32 DN50 PB63 DN80 PB11028.5 26.0 54.5 51.4 82.5 90.0However the fact that smaller diameters show a smallerheat loss difference between rigid and flexible preinsulatedpipes is interesting.CONCLUSIONSThis chapter briefly addresses each chapter andoutlines its conclusions.Test samplesThis paper compares different types of pre-insulatedpipes that have been randomly taken from the market.125The flexible pre-insulated pipes compared in this paper; PB/PE, PEX/PEX I, PEX/PEX II, PEX/PUR.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe rigid system in this paper; ST/PUR.Method of testingAll heat loss tests are performed on a test rig that hasbeen designed in compliance with EN 15632. Thetests are carried out according to ISO 8497 andEN 15632.Blowing agent analysisAll measured products are checked on type of gas andgas content. The ST/PUR and PEX/PUR productscontain approximately 95 percent of blowing agent.The PB/PE product range has a quantity over 50% ofblowing agent.In the products of PEX/PEX II and PEX/PEX I noHydrocarbon gases were detected.Comparison of flexible pre-insulated pipesA fair comparison is difficult because of differences inouter casing and other dimensions. These conclusionsare therefore only valid for the products that have beentested for this paper.In a buried condition the PB/PE pre-insulated pipeshows for equally dimensioned pipes 90A32 and200A110 the lowest absolute heat loss values for allpre-insulated pipes based on PE or PE-x foam.As mentioned before, the absolute differences in thesystem are caused by the dimensions of the preinsulatedpipes. Recalculation of the same outer casingdiameter shows also an advantage for thePB/PE system in service pipes of 32, 63 and 110 mm.See Graph 5.Heat Loss [W/m]PB/PE PEX/PEX I PEX/PEX II PEX/PUR ST/PUR40.035.030.025.020.015.010.05.00.090/DN25-PB32 160/DN50-PB63Diameter service pipe [mm]200/DN80-PB110Graph 5 Relative Heat Loss 32, 63 and 110 mm servicepipe, all with an equal outer casing and transport capacity(dT = 60 K)The flexible piping system with the PUR insulationfoam on the other hand performs better compared toflexible PE and PE-X insulation foams with equaldimensions.Comparison of flexible piping system versus therigid pre-insulated pipesFlexible pre-insulated pipes have a higher absoluteheat loss compared to rigid pre-insulated pipes.Recalculation to the same transport capacity [kg/s] andthe same outer casing diameter shows that rigid preinsulatedpipes perform better.However the fact that smaller diameters show a smallerheat loss difference between rigid and flexible preinsulatedpipes is interesting.To be comparable in heat loss some dimensions of theflexible piping systems range need to be optimized.However, other advantages of flexible pipe systems, forinstance the potential decrease of service metersbecause of a curved layout-design, can partlycompensate the higher heat loss compared to the rigidsystem (see ‗Heat loss analysis and optimization of aflexible piping system‘ by J. Korsman et al. [2]).ADDENDUMSignificant product improvement of the PB/PE/PE pipesystem has led to a decrease in heat loss [4]. Graph 5is updated with these improvements resulting in thecomparison displayed in Graph 6. The new samplesare displayed under the name of PB/PE II.Heat Loss [W/m]40.035.030.025.020.015.010.0PB/PE PB/PE II PEX/PEX I PEX/PEX IIPEX/PUR5.00.0ST/PUR90/DN25-PB32 160/DN50-PB63Diameter service pipe [mm]200/DN80-PB110Graph 6 Relative Heat Loss 32, 63 and 110 mm servicepipe, all with an equal outer casing and transport capacity(dT = 60 K)126

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFURTHER INFORMATIONQuestions concerning the paper can be addressed to:Thermaflex <strong>International</strong> Holding B.V.Veerweg 15145NS WaalwijkThe Netherlandswww.thermaflex.comLiandon B.V.Dijkgraaf 46920AB DuivenThe Netherlandswww.liandon.comACKNOWLEDGEMENTWe would like to thank all involved employees ofThermaflex Isolatie B.V. who made this researchpossible (especially H. Leunessen and M. van Doorn).Special thanks go to P. Blom and P. van Rijswijk for thededication they showed in carrying out all the heat lossmeasurements during this research.REFERENCES[1] J. T. van Wijnkoop and E.J.H.M. van der Ven,―Verification of heat loss measurement‖, in Proc. ofthe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on DistrictHeating and Cooling, Tallinn, Estonia (2010).[2] J. Korsman, I.M. Smits and E.J.H.M. van der Ven,―Heat loss analysis and optimization of a flexiblepiping system‖, in Proc. of the <strong>12th</strong> <strong>International</strong><strong>Symposium</strong> on District Heating and Cooling,Tallinn, Estonia (2010).[3] S. de Boer, J. Korsman and I.M. Smits, ―Long termheat loss of plastic Polybutylene piping systems‖,in Proc. of the 11th <strong>International</strong> <strong>Symposium</strong> onDistrict Heating and Cooling, Tallinn, Reykjavik(2008).[4] E. J .H. M. van der Ven and R.J. van Arendonk,―Heat loss of flexible plastic pipe systems, analysisand optimization‖, in Proc. of the <strong>12th</strong> <strong>International</strong><strong>Symposium</strong> on District Heating and Cooling,Tallinn, Estonia (2010).[5] P. Wallentén, ―steady-state heat loss frominsulated pipes‖, Lund Institute of Technology,Sweden, 1991[6] C. F. Colebrook, "Turbulent flow in pipes, withparticular reference to the transition regionbetween smooth and rough pipe laws", February1939[7] C. Engel and G. Baars, ―New economicalconnection solution for flexible piping systems‖, inProc. of the <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> onDistrict Heating and Cooling, Tallinn, Estonia(2010).127

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaEFFECTIVE WIDTH – THE RELATIVE DEMANDFOR DISTRICT HEATING PIPE LENGTHS IN CITY AREASUrban Persson 1 , Sven Werner 11 School of Business and EngineeringHalmstad University, PO Box 823, SE-30118 Halmstad, SwedenABSTRACTOne key concept when assessing network investmentcost levels for district heating systems is the linear heatdensity. In contrast to a traditional way of expressingthis quantity entirely on the basis of empirical data, arecently developed analytical approach has made itpossible to estimate linear heat densities on the basisof demographic data categories. A vital complementingquantity in this analytical approach is the concept ofeffective width.Effective width describes the relationship between agiven land area and the length of the district heatingpipe network within this area. When modellingdistribution capital cost levels by use of land areavalues for plot ratio calculations, there is a potentialbias of overestimating distribution capital cost levels inlow dense park city areas (e < 0.3).Since these areas often include land area sectionswithout any housing, avoiding overestimations ofnetwork investment costs demand some kind ofcorrective mechanism. By use of calculated effectivewidth values, a compensating effect at low plot ratiolevels is achieved, and, hence, renders loweranticipated distribution capital cost levels in low densepark city areas.INTRODUCTIONOne key concept when estimating investment costlevels for district heating systems is the linear heatdensity, i.e. the quota of annually sold heat in a districtheating scheme and the trench length of the pipingsystem in this scheme (Q s /L) [1]. In contrast to atraditional way of expressing this quantity entirely onthe basis of empirical data, a recently developedanalytical approach has made it possible to estimatelinear heat density on the basis of demographic datacategories [2]. A vital complementing quantity in thisanalytical approach is the concept of effective width.BACKGROUNDEffective width is a stand alone concept within districtheating theory, describing the relationship between agiven land area, A L , and the length of the districtheating pipe network, L, within this area. Hence, theeffective width becomes the width of an analogousrectangle with the trench length as the length andwhere the rectangle area is equal to the given landarea.The concept was introduced by Werner [3] and hasbeen further elaborated recently in model estimationsof distribution capital cost reactions to decreased heatdemands in four north European countries [2].Essential for calculations of anticipated investment costlevels for future district heating systems, the effectivewidth constitutes an important model parameterindicating levels of network extensions in given landareas.Since the concept of effective width itself is rather new,with no previous analytical or statistical use, data oneffective widths are in principal non attainable withinnational statistical sources. Effective width might beregarded as an innovative model quantity with noprevious representation in the field of district heatingresearch.AIMThe aim of this paper is to describe the concept ofeffective width and outline the basic properties of thisquantity. On the basis of, although sparse, empiricalobservations, preliminary statements concerning theproperties of effective width are made. The aim isfurther to enlighten the theoretical environment in whicheffective width contributes when applying demographicquantities for estimations of district heating networkinvestment costs.LIMITATIONSDue to a limited amount of empirical data, in principalless than 100 observations, the specific result valuesand relationships accounted for in this paper must beconsidered as preliminary. Although thorough in theory,the concept of effective width needs to be supportedfurther by extended empirical data gathering. In orderto be able to produce solid and reliable estimations ofeffective width values in different kinds of city areas,such information is considered vital for future use of theconcept.128

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaEFFECTIVE WIDTHEffective width is a measure indicating the districtheating network extension level within a given landarea. The quantity effective width, which is symbolisedby use of the letter w, with the unit metres, expressesthe ratio between land area and the total trench lengthof the distribution network within a district heatingsystem [3]w = A L / L [m] (1)Being in this way the result of explicit area and gridproperties, effective width can be used to describetypical district heating properties in different populationdensity areas and hence, give information onprerequisite conditions for future district heatestablishments.THE CONCEPTIn order to introduce the concept of effective width, it isnecessary to first understand some basic principalsregarding the linear heat density. The concept of linearheat density, being the division of total annually soldheat in a district heating system and the total length ofthe district heating piping network, indicates the level ofdistrict heat distribution system utilisation. Furthermore,linear heat density is a denominator parameter whencalculating district heating network capital costs.p = P/A L [number/m 2 ] (7)α = A B /P [m 2 /capita] (8)P = Total populationA L = Total land area [m 2 ]A B = Total building space area [m 2 ][number]The concept of effective width hereby plays a key rolein the reformulation of the traditional expression forlinear heat density, and hence, constitutes a centralquantity in model estimations of the feasibility andviability of future district heating network. If linear heatdensity can be said to indicate the level of district heatdistribution system utilisation, the effective widthindicates the distribution system coverage of the landarea at hand.THE PROBLEMFrom a district heating distribution point of view it isrelevant to distinguish between two kinds of land arealow plot ratio situations. The land areas can, principally,consist of either a wide dispersion of householdsspread out over the whole area (A), or households canbe closely limited to only a fraction of the land area (B),see figure 1.LinearHeatDensityQs [GJ/m] (2)LAs has been put out in [2], this traditional presentationof the concept of linear heat density offers ―no entrancefor estimations of future district heating systems, sincenone of the two quantities can be known for yet notbuilt systems‖, which is the fundamental reason forreformulation of the expression by use of demographicquantities. If combining the two concepts of populationdensity (p) and specific building space (α) into the cityplanning quantity plot ratio (e), which is suggested in[2], the concept of linear heat density can bealternatively expressed as;Q sLq e w [GJ/m] (3)The three new parameters, specific heat demand (q),plot ratio (e) and effective width (w), are defined as:q = Q/A B [GJ/m 2 a] (4)e = p α [1] (5)w = A L /L [m] (6)where129Figure 1. Low plot ratio land areas, scenario A with widedispersion of buildings and scenario B with highconcentration of buildings.In the first case (A), a district heating distribution gridwould have to cover all of the land area at hand inorder to deliver heat (at very low linear heat density),while in the latter case (B), the grid could be narroweddown to the limited area fraction. If, when conductingdistrict heating feasibility model analysis, plot ratios areextracted by means of (5), it would be relevant andrecommended to somehow adjust the land areamagnitude in order not to include non-targeted areafractions. An adjustment to reach this purpose can beachieved in several different ways, of which EffectiveWidth compensation suggested in this paper is oneoption.

DATA AND VALUESIn the spring of 2009, the authors, both being lecturersat Halmstad University in Sweden, initiated a pre-studyto be carried out by two Bsc-students at theirdepartment [4]. The study was two-fold in regard ofgathered data. Partly it delivered previously assembledand crucial data on plot ratios, land areas and trenchlengths in 39 detached house districts heating schemesin Sweden [5], allowing estimations of effective widthsin these districts, see Figure 2, and partly own collecteddata.Effectivewidth (w)[m]250200150100500y = 27,802x -0,37310,0 0,1 0,2 0,3 0,4Plot ratio (e)DetachedhousesPower(Detachedhouses)Figure 2. Effective width as a function of plot ratio in 39district heating schemes in detached house districts inSweden. Source: [5]The own collected data of the study refers to data from34 district heating schemes in multi-family housingdistricts in the Swedish cities of Halmstad andGothenburg, see Figure 3.Effectivewidth (w) [m]400350300250200150100500y = 56,622x -0,410 0,5 1 1,5Plot ratio (e)MultifamilyhousesPower(Multifamilyhouses)Figure 3. Effective width as a function of plot ratio in 34district heating schemes in multi family housing districts inSweden. Source [4]On the basis of these results, and when combined inone common graph, see Figure 4, a power functionwere established and presented in [2]. Note that (e)refers to plot ratio values, not to the natural logarithmbase (e);0.15w 61.8 e[m] (9)As can be seen in Figure 4, the graph suggests aconvergence at effective width values at 60 meters forThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia130plot ratio values above 1. This would indicate that therelationship between high dense inner city land areasand the length of the required piping grid in such areasis constant.Still, if plotted explicitly, the function does not convergeat any effective width value, no matter how far the plotratio value is extended, but the rate of divergencedecreases with higher plot ratio values. Since plotratios 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 befound in the interval of 50 < w < 60 meters.Effectivewidth (w) [m]400350300250200150100500y = 61,838x -0,14950 0,5 1 1,5Plot ratio (e)Detachedand MFhousesPower(Detachedand MFhouses)Figure 4. Effective width as a function of plot ratio,combination of 39 district heating schemes in detachedhouse districts and 34 in multi family housing districts inSweden. Datapoints merged from figure 2 and 3.EffectiveWidth (w) [m]1401301201101009080706050400 0,5 1 1,5 2 2,5 3Plot Ratio (e)Figure 5. Effective width as a function of plot ratio by useof eq. (9).For plot ratio values below 0.5, on the other hand(outer city area and park areas), the relationship is byno means constant, but diverges rapidly with increasedeffective width values as a consequence. At a plot ratiovalue of 0.04 the effective width reaches a value of 100meters, and the curve reveals that the increase ofeffective width values at even lower plot ratio valuesbelow 0.04 renders values above 100 meters andbeyond.The graph characteristics of Figure 5 has significancefor estimations of district heat distribution capital cost

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonialevels in park areas, since these areas often alsoinclude land area fractions without any housing, i.e. notto be targeted by district heating networks. When usingcrude statistical land area values for plot ratiocalculations, there is a potential bias of overestimatingdistribution capital cost levels in these suburban areas,since actual habitations plausibly only occupy parts ofthe land area at hand. In these occasions, effectivewidth values arrived at by use of eq. (9). have acompensating effect by rapidly increasing it‘s value atlow plot ratio levels, and, hence, rendering loweranticipated distribution capital cost levels.CONCLUSIONThe main conclusion from this analysis is that theconcept of effective width offers a new simple shortcutfor quick estimations of capital investments for heatdistribution in virgin urban areas.This conclusion is especially valid if the effective widthhas almost a constant value over a plot ratio of 0.5 aspreliminary stated from Figure 4. Further data collectionwill show how true this new finding will be.REFERENCES[1] Frederiksen S. and Werner S, Fjärrvärme – teori,teknik och funktion (District Heating – theory,technology and function). Studentlitteratur, Lund1993.[2] Persson U. and Werner S, The FutureCompetitiveness of District Heating, to bepublished.[3] Werner S, Fjärrvärme till småhus – värmeförlusteroch distributionskostnader (Sparse district heating– heat losses and distributions costs). Report1997:11, The Swedish District Heating Association.Stockholm 1997.[4] Netterberg H and Isaksson I, District Heating inSlough. BSc thesis from Halmstad University,Halmstad 2009.[5] [Andersson S et al, Nuläge Värmegles Fjärrvärme(The current situation for sparse district heating),The Swedish District Heating Association, researchreport FoU 2002:74. Stockholm 2002.131

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaINTEGRATING RENEWABLE ENERGY INTO LARGE-SCALE DISTRICT HEATINGSYSTEMSPeter Begerow, Dr. Stefan HollerMVV Energie AG, Mannheim, GermanyABSTRACTRenewable energy for heating is mostly used in smallsystems for single-family houses. The existing districtheating networks are generally run by large heatingplants or combined heat and power plants fired withfossil fuels.To combine these two systems, a feasibility study wascompleted with a focus on the district heating grid inMannheim, Germany, and with a focus on solar thermalheat. Other renewable energy heat sources,geothermal heat and heat from biomass, are includedfor a comparison.The study focuses on the heat price as a key figure toanalyse the economic feasibility. The technicalfeasibility has been evaluated by using a simulationmodel of a secondary district heating grid, which isoperated on a low flow temperature level of 70 °C andwhich is connected to a central solar thermal energyplant. The paper describes which technical andeconomic framework conditions are necessary forimplementing renewable energy into large-scale districtheating systems. The calculations show that incomparison with other renewable heat sources solarheat has the highest heat costs ranging from7,7 ct/kWh to 14,5 ct/kWh depending on the plant size,the solar fraction and the use of a storage system. Themajor technical problems for integrating solar heat intoa heat grid are the pressure difference between theflow pipe and return pipe and the low temperature theflat plate solar collectors are working with.INTRODUCTIONBased on the protocol of Kyoto and Europeanregulations a high reduction of CO 2 emissions inGermany is necessary. To achieve those goals, anexpansion of renewable energy in the heat market isrequired. In Germany the major aim to reach is a shareof 50 % renewable energy in the heat market by 2050.Furthermore, 50 % of the renewable heat is supposedto be contributed by a heating grid. [13]To achieve these goals, different governmental as wellas local support mechanism and financial subsidies areavailable. If the heat production is combined with anelectricity production, the major financial support isbased on the EEG (German law for renewable energy).If the used technology just produces heat, a financialaid from the BAFA (Federal Office of Economics andExport Control, Germany) or KfW (bank under controlof the Federal Republic, Germany) is possible. Thereare different regulations which have to be fulfilled bythe project in order to be eligible for those subsidies [8].Main criteria are the size and type of the investor andplanner, the type of technology and the size of the heatplant and of the storage tank.Previous studies [2], [12], [16], [18] showed that solarthermal energy is mostly used in single-family housesand smaller heating grids combined with seasonal heatstorage systems. Those systems are still indevelopment and need financial support to be realized.Most of these heating grids run with a lower flowtemperature and use either fossil fuels or heat frombiomass for an auxiliary heat generation. The largestsolar thermal district heat system in Germany is locatedin Crailsheim. It covers an area of approx. 7300 m² ofsolar collectors with two buffer tanks with a combinedvolume of about 500 m³ as thermal storage. In additiona seasonal geothermal storage has been built whichwill cover 50 % of the heat demand for about 2000residents. This research project has heat productioncosts without any financial support of about 19 ct/kWh.This sum will be reduced depending on the possiblesubsidies. [12] Reported technical difficulties weremostly in the thermal storage technology. There werelittle problems with the collectors as common flat platecollectors were used which are commercially availableand used in large numbers in smaller systems.The project in Crailsheim has shown the technicalfeasibility of a system with a seasonal thermal storage,but it also shows that considerable costs are involved.Furthermore, if a single-family house will install aseasonal storage to get a solar fraction above 50 %, itneeds more than 10 m³ of hot water storage(depending on the building type and planned solarfraction). However, in common buildings there isn‘tenough room for that size of storage [17]. Those twoaspects show that the use of a heating grid couldsignificantly reduce the costs of the solar thermalsystems and could save space otherwise necessary fora storage tank.The paper will give an overview of how the expansionof renewable energy in the heat market will be possible132

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaby integrating renewable heat into a district heatingnetwork. A detailed simulation for a solar thermalintegration was done by using RETscreen [14] assimulation software.MATERIAL AND METHODSEvaluation of heat demandFor planning a new heat production facility, the heatdemand of the connected consumers is necessary. Ifthose are existing households, the heat demand fromthe past can be used for calculations. For newly builthouses the heat demand should be exactly calculatedwith the standards named in DIN V 4108-6.If this is not possible, the yearly heat demand can beassumed by the given figures:Table 1: heat demand [15]Building size[housingunits]Heat demand(room heating)[kWh/m²a]heating demand(hot tap water)[kWh/m²a]1-2 72,3 20More than 3 55,3 20These figures can be realized in buildings constructedbetween 2011 and 2020. [15]Another factor for the planning of a heating grid is theoutlook into the future, because the payback period ofa renewable heat production facility is very long.The following graph shows the expected change inheat demand for Germany focusing on different factorsof influence:is not part of the simulated area and on the reductionthrough population, which has not a direct effect on onespecial housing area. The reduction through influenceof temperature has a share below 5 % within 15 yearsand is therefore not included within the simulation.Existing SystemsBetween newly built and existing heat networks thereexist some main differences which have to beconsidered. If the network is designed especially for therenewable energy source, it can be technicallyspecialized (e.g. forced low return temperature forbuilding owners; special isolation of the used pipes).Older heating grids on the other hand are normallyconstructed for the heat production with fossil fuels andare normally designed for higher temperatures.Furthermore, in some heating grids a high temperatureis necessary either for thermal cooling systems (e.g.absorption chillers) or for the heat transfer stationswithin the houses which are built for high temperatures(low flow temperatures need optimized heat transferstations [12]. In the following, the main aspects for theintegration of different sustainable heat generationtechnologies are described.Heat grid for renewable energyFor the integration of renewable energy into heat grids,different possibilities for the connection exist.Especially for the solar thermal energy production it isassumed, that more than one heat plant will beconnected.The three options are:1. Taking water from the return pipe, heat it andreturn it into the return pipe2. Taking water from the flow pipe, heat it furtherand return it into the flow pipeFig 1: development of heat demand [11]Within the following simulation this development is notfurther regarded. The major reduction within wholeGermany is based on renovation of old buildings, which1333. Taking water from the grid out of the returnpipe and rise the temperature to the necessaryflow pipe value [3]All of those options have some obstacles. The firstoption is normally not welcome by the grid operatorbecause of higher losses in the system. The secondoption is almost impossible for the use of flat plate solarcollectors; because the high flow temperature cannotbe further heated.The third option shows the best possibility forintegration but has the obstacle with high pressuredifferences between the flow pipe and the return pipe.To evaluate the necessary pump work a first estimationcan be done with equation (1). It gives the pump workW depending on the necessary heat flow ∆Q, the

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniapressure difference ∆p and the temperature difference∆T between flow and return. Included in this equationis, the pump efficiency η as well as the density ρ andthe thermal capacity of water c p .(1)For the integration of renewable energy into heat gridssome aspects have to be regarded.For solar thermal energy the flow and returntemperature of the network is a major problem. Thisshould be lower than in the existing district heating gridand run with a temperature of about 60 °C / 40 °C. [6]For using deep geothermal heat it depends on the usedtechnology. If it is combined with the electricityproduction, the waste heat after the power plant isnormally below 80 °C.Another major obstacle is the variation of the heatproduction and the demand if using solar thermal heat.During the summer months, the solar radiation is at itspeak, but the heat demand has its peak during thewinter months. To cover a heat grid with a high solarfraction, a long term thermal storage system isnecessary.Description of selected siteFor modeling the integration of solar thermal energyinto a district heating network, a yet to be built housingestate was selected. This housing area is planned witha district heating grid running at a flow temperature ofabout 70 °C. This area is connected with a heatexchanger to the central heating grid of the city, whichis run with flow temperatures between 90 °C and130 °C.Table 2: heat demand selected siteBuilding size[housing units]Number ofbuildingsTotal heatingdemand [MWh/a](room heating +hot tap water)1-2 135 1561More than 3 111 585Sum 246 2146Table 2 gives an overview of the planned houses andtheir heat demand. The whole heating grid will have anlength of about 1,3 km and the total heat demand willbe 2146 MWh/a.Used SoftwareRETscreen is a program to make first feasibility studiesof all kind of green energy projects. In terms for solarthermal heating, it uses an included weather databaseto calculate the expected heat production. Furthermorea product database is included with the necessarytechnical parameters for many different solar thermalcollectors. The needed amount of heat can either becalculated other ways or assumed by the softwaredepending on the amount and size of buildings.Combining those input factors with others, thesimulation tool gives a recommendation of the usednumber of solar collectors and the size of a thermalstorage system. If all input factors are included theprogram calculates the yearly heat production and thesolar fraction. Beyond that, the program can be used toinclude a second heating system for the remainingneeds to get the final payback period and the totalemissions. For the simulation of this paper the version4 (November, 2009) of the named software was used.Financial CalculationThe calculation of the heat costs is based on the netpresent value method. For the internal rate of return thegiven value was used, all other costs included and theheat costs varied to get a net present value of zero.This method gives the current heat price and a furtherincrease during the next years is included. This makesit possible to compare the actual heat price to the givenvalues of other systems. For the economical calculationin the conclusion of this paper, a competitive heat pricefrom now on was realized.The named financial support which is included in thecalculations are subsidies on the capital cost. Theydepend, like mentioned in the introduction, on differentaspects. A research project like the one in Crailsheim,can get a higher support than commercial ones run bylarge companies. [1]Solar thermal heat productionFor the heat supply of the given housing area, differentscenarios based on solar thermal energy weredeveloped. Using the RETscreen software tool thetechnical parameters of the flat plate solar thermalcollectors, weather data from a climate database andthe given heat demand of the area was combined foreach scenario.The scenarios differ in the necessary amount ofcollectors needed to achieve a solar fraction of the totalheating demand of 50 % [scenario 1], a 100 % solarheat production of the used hot tap water (which staysconstant throughout the whole year) [scenario 2] and a134

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia50 % solar fraction of the total heating demand withouta thermal storage system [scenario 3].For the thermal storage a hot water system isassumed, because those are state of the art and canbe used in most applications. Other systems havemore specific requirements to the geological situationof the area. A geothermal heat storage for exampledoes not work in an area with a flow of the groundwater. For the simulation of scenario 2 a smallerthermal storage compared to scenario 1 was assumed,because there is no necessity for a seasonal heatstorage system.For the simulation model a commercial solar collectorwas taken (s. Table 3). It is a flat plate collector with aanti-reflection glass and a gross area of about 2,6 m².Its efficiency is 84,4 % (calculated according to EN12975). The simulation uses specific given parameters.Those are shown in Table 3.Table 3: used input parameters for solar simulationAnnual heating energy(calculated with givenmethod)Flow temperature 67 °CReturn temperature 45 °CSlope of collector 55°Azimuth of buildingType of collectorStorage capacityHeat exchangerefficiencyMiscellaneous lossesPump efficiency(for grid integration)Time periodScenario 1: 1076 MWhScenario 2: 493 MWhScenario 3: 1071 MWhScenario 4: 11,6 MWhScenario 5: 11,6 MWh-45° (southeast)WagnerSolar L20 ARScenario 1: 1000 l/m²Scenario 2: 100 l/m²Scenario 3: 1 l/m²Scenario 4: 100 l/m²Scenario 5: 10 l/m²80 %5 % if storage is used(smaller grid)8 % if integrated intolarge grid40 %20 yearInternal rate of return Scenario 1: 8,5 %Scenario 2: 8,5 %Scenario 3: 8,5 %Scenario 4: 5,0 %Scenario 5: 5,0 %Increase of heat price per 2 %yearFinancial support 30 %For a comparison of the different scenarios the heatcost per kWh were calculated.The calculation of the emissions is based on theoperation of the system and not on its total life cycle.For solar thermal heat the CO 2 emissions only arisefrom the used electricity for the necessary pumps.Included in the calculation is only the pump energy forthe solar thermal collectors and, if necessary, toincrease the pressure for the integration into theheating grid flow pipe. The CO 2 emissions for theGerman electricity grid are given with 506 g/kWh.For the calculation without a thermal storage system[scenario 3] it was assumed that the produced solarheat can directly be distributed throughout a districtheating network. This would make it possible to savethe investments of a seasonal heat storage system andalso reduce the losses within the thermal storagesystem.For those calculations the same heat amount was usedthan in scenario 2. But in this case it is not possible tocover 50 % of the heat demand of the total grid. Just asmall amount, for example the losses of the grid andthe base load, can be produced with solar thermaltechnologies without a thermal storage.Another option would be to integrate small systems intothe district heating grid. In this case the operator of thegrid would not run the facility by itself. The heatproducer could use a solar thermal collector for its ownheat demand but without a thermal storage system.Instead of using an in-house thermal storage (what isgetting very large if a seasonal heat storage system isused) the heating grid could be used. For the singlehouse technology an internal rate of return of 5 % wasused for the economic calculation (average percentageof building credit [4]). Furthermore the financial supportis a little different because of different regulations forlarge and small systems. In the following those twocalculations are named ―scenario 4‖ for the heatproduction of a single-family house with a thermalstorage and ―scenario 5‖ for the calculation without athermal storage.Summary of different scenarios:Scenario 1 Solar fraction of 50%Seasonal thermal storage includedScenario 2100% heat production of hot tap waterBuffer heat storage included, but no seasonalthermal storage135

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaScenario 3 Same amount of heat produced than in scenario 1No storage; connected to large district heating gridScenario 4Solar fraction of 100% for a single family houseSeasonal thermal storage includedScenario 5 Same amount of heat produced than in scenario 4No storage; connected to large district heating gridGeothermal heat productionThe geothermal heat can be used in various ways forroom heating. Using the shallow geothermal heat isonly possible in combination with a heat pump.Therefore, for a large integration into heating grids thedeep geothermal energy is the favoured one.Furthermore in the upper valley of the river Rhein(Oberrheingraben) the geothermal heat can be used fora combined heat and power production because of itshigh temperature. In Germany this gives the possibilityto get a payment for the electricity based on the EEGwhich grows for 3 ct/kWh if the heat is used as well.For a comparison to the solar thermal heat ageothermal power plant in Landau, Germany is used asa reference.This project began in 2004 and at the end of 2007 thepower plant started its first electricity production. Thefirst heat output was planned for 2009.The power plant uses the ORC process (OrganicRankine Cycle) to generate electricity. A drill hole witha depth of 3000 m connects to thermal water with atemperature with up to 160 °C which is cooled downduring electricity production to 70 °C. The whole yearlyenergy output of the power plant is planned to be22.000 MWh electricity and 9.200 MWh heat. One ofthe major benefits of the geothermal heat production isthe base load which is always available. On the otherhand this gives the problem that the heat is alsoavailable in the summer time and needs to be cooleddown in other ways.The calculated emissions of the power plant are0 g CO 2 /kWh because the electricity production has noemissions and for the pumps the own electricity can beused. [6]Currently the power plant runs with a limited output dueto small earthquakes in the area of the drilling hole anddoes not deliver heat until now. Additional geologicalstudies are done right now and a heat output shouldstart after they are finished.Heat production with biomassThe heat production from biomass is technically verysimilar to the fossil fuel powered heating plants.Therefore the integration into existing district heatinggrids is the easiest way compared to the otherrenewable energy sources.The exact technology depends on the used fuels andtherefore the economic calculation is mainly based onthe price development of the biomass.The emissions of such a system are by way ofcalculation zero, because the emitted CO 2 was firstlybound by the biomass during its growing period. If thebiomass is planted in an area which was deforested forthat, the emissions are not zero any more. The formerforest was a CO 2 sink which does not exist anymoreand should be included in the calculation. Furthermorethe transport and processing of the biomass should beincluded. [19]Fossil fuels for comparisonIn our days the district heating grid in Mannheim is fedwith heat from a fossil fuel fired CHP plant. The heatprices from that system are much lower than therenewable heat. Looking into the future it mainlydepends on the price development of CO 2 emissionsand the coal price. [8]The emissions of such a system are very high, even ifthe used heat is more or less waste heat. To reducethose, a CCS technology can be implemented in thefuture.RESULTSThe results of the simulation are shown in Tab. 4 andFig. 2 and 3.In conclusion the heat price is lower if the collector areaincreases (economy-of-scale). Furthermore the use ofa district heating grid instead of a thermal storagelowers the heat cost extremely.For scenario 1 it is necessary to install a gross area of3080 m² solar thermal collectors. 1076 MWh heat canbe produced in combination with a 2820 m³ hot waterstorage. The heating costs calculated with the givenframework conditions are 11,2 ct/kWh. To operate thecollector area, pumps are needed which consumeelectricity. The emissions of that electricity are, basedon the produced heat, 7,9 g CO 2 /kWh.In scenario 2, 1916 m² solar thermal collectors need tobe installed. Combined with a hot water buffer storagewith a volume of 175 m³, 494 MWh of heat can beproduced. The financial calculation over 20 years lead136

to heating costs of 14,5 ct/kWh. The emissions of sucha system are 8,9 g CO 2 /kWh.The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaIf the heating amount of the 50 % scenario is used butwithout a storage [scenario 3], and therefore withoutthose losses, a much smaller collector area iscalculated. However, the produced heat has to be useddirectly within a large heating grid. For such a system2542 m² of solar thermal collectors are needed whichproduce 1071 MWh/a. The smaller collector area andthe elimination of a storage system give heating costsof 7,7 ct/kWh. On the other hand the emissions of sucha system are higher because of the necessary pumpenergy for the pressure compensation. The totalspecific emissions of that system are 12,2 g/kWh.Fig 2: CO 2 emissions of different system [1], [5]Under consideration of scenario 4, the heat costs are13,8 ct/kWh within a single family house. If a heatinggrid would be used for storage and therefore no largethermal storage is necessary, the heat costs can godown to about 11 ct/kWh. If the losses of the storagesystem are included in the calculation, a smaller grossarea of collectors can be used. Combining all thosesavings, the heat cost for a single-family house can godown to 7,2 ct/kWh (scenario 5). This shows, that thereis a wide margin and a high potential of cost reductionif a heat grid is used. But it has to be said, that thoseheat costs are still much higher than from other heatgenerating systems.Table 4 shows the technical results and parameters foreach calculated scenario. Based on those figures thefinancial and ecological calculation where made. Thoseresults are shown in Figure 1 and 3. For the renewabletechnologies the increase of 2% of the heat costs caneasily be included in the calculation. For the fossil (andthe biomass) use, the heat price is highly dependent onthe fuel price development. Therefore a price range isgiven on those systems. The reason for the range forfossil CHP heat CO 2 emissions is that differentreferences are used.Table 4: output parameters of simulationscenarioCollector grossarea [m²]Producedheat[MWh/a]Storagesize[m³]1 3080 1076 28202 1916 493,6 175Fig 3: heat costs of different system [1], [20]CONCLUSIONTo make the solar thermal heat production economicalcompared to the other systems, different aspects haveto be changed. In order to show the potential of costreduction for solar thermal heat generation a sensitivityanalyses has been carried out. The followingparameters have been varied in order to reach a heatprices of around 3,5 ct/kWh in the beginning year. Thisis the actual heat price for private customers inMannheim.Future heat price developmentChange of investmentAmount of financial supportInternal Rate of Return3 2542 1071 -4 44 11,6 45 29 11,6 -137

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 5: parameters for economical operationScenario 1 Scenario 3Heat costs 3,5 ct/kWh 3,4 ct/kWhIRR 5 % 5 %Capital cost 70 % 90 %Financial support 40 % 30 %Heat pricedevelopment8 % p.a. 8 % p.a.Fig 4: influence on the heat costs of different factorsFigure 4 shows the influence of the different factors, ifthe others stay the same. But it also shows, that bychanging just one aspect, a reduction of the heat costto 3,5 ct/kWh is only possible with lowering the capitalcosts 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 financialsupport und were calculated separately.In order to reduce the heat costs down to about3,5 ct/kWh in scenario 1, a reduction of the capital costby 40% combined with financial support of 50% isnecessary, if the heat price will rise with 8% per year.This is higher than shown in figure 4 and is based on ahigh price assumption as reported in [10].If a lower IRR is assumed (5 %), the capital costs haveto go down to 70 % and a financial support of 40% ofthe investment is necessary.For scenario 3, a rise of the heat price and the lowerIRR (5 %) just need a reduction of capital costs of 10 %to achieve heat costs of 3,4 ct/kWh. In this case theassumed financial support of 30% stays the same. Thisshows that an economical use of solar thermal energywithin district heating could be achieved. The financialsupport of 30% can be possible based on a KfWprogram, the capital costs of 90% of the baseinvestment is possible within a feasibility study and the5% IRR is an average figure for building loans.The high requirements to make the solar heat profitableshow, that this technology is not advisable if otherrenewable energy sources are available. Furthermorefor the technical integration a low temperature heatinggrid is necessary.For the near future it might get more interesting to lookon the biomass and geothermal heat, particular if theheat is needed in a region where high temperatures inthe depth could be exploited or cheap biomass sourcesare available. Further research in the solar collectortechnology is needed to lower the capital costs andequally within the thermal storage technology, as itmight get interesting in the future to include those evenin district heating grids with fossil fuels as heat sourceto cover peaks in the demand and transfer a surplusheat production from the summer into the winterseason.NOMENCLATUREW [J]∆Q [J]∆p [Pa]∆T [K]cp [J/(kg*K)]ρ [kg/m³]ηREFERENCESwork of pumpheat flowpressure difference (flow / return)temperature difference (flow / return)heat capacitydensitypump efficiency[1] Begerow, P.; Integration von erneuerbarenEnergien in Fernwärmenetze – Eine technischeund wirtschaftliche Analyse aus Sicht einesFernwärmeversorgers, Diplomarbeit an derUniversität Flensburg, MVV Energie AG,Mannheim; 2010.[2] Bodmann, M.; Mangold, D.; Nußbicker, J.; Raab,S.; Schenke, A.; Schmidt, T.: Solar unterstützeNahwärmeversorgung und Langzeit-Wärmespeicher; Forschungsbericht zum BMWAVorhaben; Universität Stuttgart; 2005138

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia[3] Bucar, G.; Schweyer, K.; Fink, C.; Riva, R.;Neuhäuser, M.; Meissner, E.; Streicher, W.;Halmdienst, C.; Dezentrale erneuerbare Energiefür bestehende Fernwärmenetze;Bundesministeriums für Verkehr, Innovation undTechnologie; Wien, 2005. Page 15-16[4] Deutsche Bank; Zinslandschaft;https://www.deutsche-bankbauspar.de/de/media/Zinslandschaft.pdf;2010.[5] Fielenbach, H.; Ohl, G.; Schwarzburger, H.:Effiziente Wohnwärme und hoher Komfort; GBG –Mannheimer Wohnungsbaugesellschaft mbH;2009.[6] Frey, M.; Milles, U.: GeothermischeStromerzeugung in Landau; BINE Projektinfo14/07; Karlsruhe; 2007.[7] Heidemann, W.: Solare Nahwärme und saisonaleSpeicherung; FVS LZE Themen; Berlin; 2005.Page 36[8] Kaltschmitt, M.; Streicher, W.; Wiese, A.:erneuerbare Energien; Springer Verlag; Berlin;2006. Page 29[9] KfW: Programm erneuerbare Energien;http://www.kfw-mittelstandsbank.de/DE_Home/Service/Kreditantrag_und_Formulare/Merkblaetter/KfW-Programm_Erneuerbare_Energien_ 270_271_272_281_282.jsp; 2010.[10] Klöpsch, M.; Besier, R.; Wagner, A.: Reicht fürKunststoffmantelrohre die Standarddämmung?;Euroheat&Power 38. (2009); issue 12[12] Mangold, D.; Riegger, M.; Schmidt, T.: SolarNahwärmeversorgung und Langzeit-Wärmespeicher; Forschungsbericht zum BMUVorhaben; Solites; Stuttgart; 2007. Page 14, 20[13] Nitsch, J.; Wenzel, B.: Langfristszenarien undStrategien für den Ausbau erneuerbarer Energienin Deutschland; Leitszenario 2009; BMU; Berlin;2009. Page 53-57[14] RETscreen Version 4; Natural Resources Canada;http://www.retscreen.net; 2009[15] Smolka, M.: Ökologisch-technische Auswirkungendezentraler Energieversorgungsszenarien mitBlockheizkraftwerken in elektrischen Verteilungsnetzen;Verlagshaus Mainz GmbH; Aachen; 2009.Page 18[16] Solarge: Marstal district heating Plant; ProjectSummary; http://solarge.org/index.php?id=1235&no_cache=1; 14.03.2010[17] Sonnenhaus-Institut e.V.; http://sonnenhausinstitut.de/wohnhaeuser.html;2010.[18] Ulbjerb, F.: Large-Scale Solar Heating; Hot|Cool;3/2008; DBDH; Frederiksberg; 2008[19] Watter, H.; Nachhaltige Energiesysteme;Vieweg+Teubner; Wiesbaden, 2009. Page 168[20] Voß, A.: Das Wachstumspotential der Nah- undFernwärme - wirtschaftliche und gesetzlicheVoraussetzungen für den Ausbau; aus: Forschungund Entwicklung Heft 10; AGFW; Frankfurt, 2005.[11] Lutsch, W.: Neue Wege zur Marktumsetzungsolarer Nah- und Fernwärme; Fernwärme-, KälteundKWK-Versorgung: Entwicklungsstrategie;AGFW; Frankfurt; 2009.139

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSOLAR DISTRICT HEATING (SDH): TECHNOLOGIES USED IN LARGE SCALE SDHPLANTS IN GRAZ – OPERATIONAL EXPERIENCES AND FURTHERDEVELOPMENTSM. Schubert 1 , C. Holter 1 and R. Soell 11S.O.L.I.D. Solarinstallationen und Design GmbH, Puchstr. 85, A-8020 Graz,m.schubert@solid.atABSTRACTS.O.L.I.D. installed three large scale solar plants forfeeding into the city‘s district heating in Graz in recentyears. These three solar plants have an annual heatproduction of 15,8 PJ, the city‘s grid delivers 2800 PJper year. Therefore the integration of solar thermal in atechnical and economical feasible way has to meet therequirements of Graz‘ existing district heating grid,which is one of the largest in Austria.The first plant, at stadium Graz-Liebenau with1.420 m², has been now for seven years in reliableoperations, with very good power output data.AEVG Graz, the largest plant in Graz at 4.960 m²,feeds into the gas power station (maximum power of250 MW) and from there the heat is distributed throughthe district heating grid.The latest plant, at Wasserwerk Andritz with currently3.860 m², has a buffer storage of 60 m³ and theplanning for installation of a heat pump is completed.The plant feeds into the district heating grid andsupports the room heating of a large office building.This paper presents operational experiences aboutthree different ways for feeding solar thermal energyinto a large city‘s district heating grid. Recentdevelopments like buffer management for combineddistrict heating and room heating and integration of aheat pump are outlined.First solar thermal plants for district heating were builtin the 1970‘s in Sweden. Since then, various plantshave been built mainly in Austria, Denmark, Germanyand Sweden.Most of these solar plants feed into rather small heatinggrids or sub-grids with an annual heat delivery below50 GWh th (180 TJ). In Denmark, this market wasgrowing rapidly in recent years and is now bigger thanthe market for small-scale solar systems for singlefamilyhouses.In Graz, Austria, solar thermal plants feed into a largescale heating grid with an annual heat delivery of830 GWh th (2,99 PJ) and a maximum power of382 MW th . Technical parameters and operationstrategies in large scale heating grids are different tothose in small scale grids and solar thermal technologyhas to adopt to these circumstances.Three solar thermal plants in Graz are presented andthe way they are integrated into the city‘s heating grid.SDH PLANT DESIGNS IN GRAZ1. Feeding directly into the district heating grid– plant at stadium Graz-LiebenauThis plant is located on the roof of an ice-skating hallnext to the city‘s football stadium (Fig. 1).INTRODUCTIONFor reasons of energy security and environmentalprotection, the European Union has set a target of 1%solar fraction in district heating in 2020 and of 5% in2050 [1].Solar thermal technology is widespread in the singlefamily house sector in most European countries. Mainlyfor domestic hot water preparation (DHW), but also forroom heating (RH).In multi-family houses and for heating grids, there arenot yet as many solar thermal plants and the marketbegins to develop.Fig. 1: Aerial view of solar plant Stadion Liebenau140

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe return medium of the heating grid is heated up andtransferred to the flow (Fig. 2) [2]. The adaption of solarthermal technology for the temperature and pressurelevels of the district heating grid were challenging. Thisproject was realized with standard large scalecollectors (1420 m² collector area) of the Austrianmanufacturer Ökotech and temperature levels in thedistrict heating flow of above 70 °C have to be reacheddependant on the ambient temperature.During first operation years, detailed monitoring wasdone on the plant‘s performance. Dependant onclimate condition, the annual yield of the plant wasbetween 521 MWh/a and 569 MWh/a. Thiscorresponds to a specific yield of 370–404 kWh/a persquare meter collector area. Also the returntemperature of the heating grid is of great importancefor the performance of the solar plant.Fig. 2: Hydraulic scheme of solar feed-in at Stadion Liebenau2. SDH connected to a large scale fossil fuel firedstation – plant AEVG GrazThis is the largest solar thermal plant in Austria and it isinstalled on four different buildings of the localcollection and recycling station (Fig. 3).Situated next to the central heating plant, pressureparameters are favourable for feed-in. Pressure ishigher in return and thus only valves are necessary andno additional pumps for integration into the districtheating grid.Fig. 3: Solar plant AEVG Graz141

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia3. SDH for combined room heating and districtheating with buffer and heat pump – plantWasserwerk AndritzAs solar thermal systems can‘t always generate thehigh temperatures as required for the district heatinggrid, other applications were found for temperaturelevels below 75 °C (Fig. 4).Solar heat at low temperature level is stored into a60 m³ buffer tank and later used for room heating of anoffice building (low temperature floor heating). Thebuffer is also fed by district heating and thus decreasesthe required connected load of the office building.Even lower temperature levels in shoulder seasons andin winter can be raised by a heat pump. The installationis planned for the end of 2010. COPs above 4 areexpected, i.e. when heat from the collectors of 26 °C isheated up to 55 °C for room heating.ACKNOWLEDGEMENTThis 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 fossileoder teilfossile Fernwärmenetze – Einbindung vondezentraler Energie aus ErneuerbarenEnergieträgern – Chancen und Hemmnisse,Endbericht zu „Energie der Zukunft―Forschungsprojekt No 807718 im Auftrag desBMVIT, publisher: Grazer EnergieagenturGes.m.b.h.Fig. 4: solar thermal plant Wasserwerk Andritz142

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaBIOENERGY COMBINES IN DISTRICT HEATING SYSTEMS:PROSPECTS FOR A FUTURE GROWTH INDUSTRY?E. Axelsson 1 , A. Sandoff 2 , C. Overland 21 Profu, Gothenburg, Sweden.2 Department of Business Administration, University of Gothenburg, Sweden.ABSTRACTDistrict heating offers opportunities for integration ofbioenergy production (e.g. of biofuel). The aim of thispaper is to assess the environmental benefit and theeconomic value of such integration, in order to evaluatethe prospect for bioenergy combines in district heatingsystems. Since the detailed characteristics of thedistrict heating system are crucial for the feasibility forintegration of bioenergy production, the assessment isbased on four real district heating systems. Theenvironmental evaluation shows that the decrease ingreen house gas emissions from a combine are inproportion to the increase in output of CO 2 neutralenergy products. However, the CO 2 reduction per usedquantity of biomass is higher in conventional combinedheat and power production as long as marginalelectricity is related to high CO 2 emissions. Also theeconomic evaluation show ambiguous results: twocases had negative net present value even for lowdiscount rates, while the two other cases showed to bemore economically robust. In addition to this, a moredetailed analysis of the industrial conditions for theintegration shows a need for achieving a fit regardingseveral operational, strategic and economiccircumstances for this type of business ventures. Twoimportant conclusions that can be drawn from this isthat: 1) not all district heating systems are suitable forbioenergy combines 2) there are many barriers for awide spread adoption of bioenergy combines.INTRODUCTIONDistrict heating is a technology that receives increasinginterest as it has great potentials in several ways. Oneunique characteristic of the district heating technologyis the use of low temperature energy flows for largescale energy distribution. In contrast to other energytransformation technologies (e.g. condensing power ordistributed gas heating), district heating can interactwith energy flows that otherwise do not have anyalternative use (e.g. industrial residual heat). Althoughthis is one of the competitive advantages of thetechnology and a fundamental platform for its businessmodel, this can further enhance the scoop of thebusiness: by backward integration it is possible toincrease profitability in other industrial processes withwaste heat as a by-product.143One industrial branch that shows promising prospectsin this respect is bioenergy production, i.e. productionof various kinds of biofuel, biogas and solid biofuel.Integration of bioenergy production to district heatingproduction eventuates in a bioenergy combine were theresidual heat from the bioenergy production can beutilised for district heating. Moreover, the integrationcan, in many cases, offer additional positive synergies,e.g. regarding the use of steam and combustibleby-products.The fact that worldwide bioenergy production as well asthe number of bioenergy products offered is increasingis a result of changing demand, which in turn offersnew business opportunities. However, one of the greatissues with large-scale production of bioenergyproducts is the growing concern over the negativeexternalities (social and environmental aspects as wellas resource efficiency). Since energy production andconsumption shows strong path dependence [1], thereis an urgent need to develop and establish productiontechnologies that help minimize the negativeexternalities. Utilizing the taiga and deciduous forestresources in the Northern hemisphere for this purposesis, arguably, a promising alternative. The majority ofthese natural resources exist in harvested forests,typically found in regions with, or suitable for, districtheating.This paper investigates the prospects of using districtheating production as a base for bioenergy productionand its potential to become a wide spread technology.For this purpose, we use data from four existing districtheating companies to which a bioenergy productionunit is fitted. By acknowledging the complexity of thisintegrative business venture, it is possible to getcredible assessments of the magnitude in energyefficiency, environmental gains and economic profits.Equally important is the possibility to detect potentiallimitations for bioenergy combines to become acomplement to district heating. Finally, conclusions aremade to acquire clues to important restrictions to awide spread adoption.RESEACH DESIGNWe argue that prospects for becoming a future growthindustry are dependent on the environmental benefits,economic attractiveness and fit with existing businesscontext. Hence, these three aspects of joint production

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaare analysed. The environmental benefits are analyzedwith a system perspective on greenhouse gases (GHG)emissions, taking into account both on and off siteconsequences of introduction of an energy combine;see Environmental evaluation below. Moreover, theresource efficiency in the form of CO 2 reduction perused quantity of biomass is evaluated for eachcombine.The economic benefits of the ―joint production‖ set upare analyzed through both a short and long-termcommercial lens. By using discounted cash flowtechniques as a base for this analysis, it is possible toaccount for both the yearly consequences as well aslong term economic value; see Economic evaluationbelow.Fit with existing business context is analysed withrespect to input/output markets, production and systemconfiguration and general business conditionsdominant in the host industry. The analysis focus onrestrictions for short term fit; see Business contextevaluation.Since the detailed characteristic of the district heatingsystem is paramount to the feasibility for integration ofbioenergy production, we base our investigation on fourreal district heating systems in Sweden with differentcompositions. The chosen systems are all of equal size(500-600 TWh of yearly heat deliveries) established intowns with 40 000 to 80 000 inhabitants. Thesesystems are in turn equipped with a bioenergyproduction unit that best suits ruling company strategyas well as operational characteristics and maximizesenergy efficiency. In order to capture the additionalvalues of these investments, evaluation of eachcombine configuration is made in relation to areference case consisting of the existing system(complemented with investments to maintain acomparable level of production quality). The referenceand combine cases are further described in theDescription of the cases below.Much effort was put into indentifying efficient technicalsolutions that best take advantage of the site-specificconditions in each system. This work includedeverything from choice of equipment, appropriate sizeof the integrated production unit and productionstrategies over the year regarding output of heat,electricity and other energy products. To identifyefficient technical solutions an integrative computerizedprocess was applied, including both the district heatingsimulation software MARTES [2], and detailed spreadsheet calculations. In order to guarantee high qualityinput data, representatives from these four companiesgave access to technical, environmental as well aseconomic data.Below follows a description of the environmental andeconomic evaluation procedure. It is important to stressthat the input data for these assessments only includethe change resulting from the integration of thebioenergy production. One implication of this approachis that the environmental benefit of the heat produced(for district heating) is not included, since one basecondition is that the heat deliveries are the same withand without bioenergy production. Another implicationis that production units in the district heating systemthat are not affected (e.g. base load and peak loadproduction units) are not included. This systemboundary is also pervading for the Description of thecases to follow.Description of the casesThe four district heating systems with reference andcombine cases, respectively, are presented in briefbelow. The four objects for the evaluation are alsosummarized in Table I. A more comprehensivedescription can be found in ref. [3].Table I. Overview of the reference and combine cases inthe four district heating systems. Economic and energydata are given for both the reference and combine case,separated with a slash (ref./combine).CONFIGURATIONHeat deliv.(TWh/y)1 2 3 4500 530 560 620Ref. inv. Bio CHP None Bio CHP Bio CHPCombinetechnologyPyrolysisEnzymatichydrolysisAcidhydrolysisGasificationProducts Bio oil Ethanol 1 Ethanol FTdiesel 2ECONOMIC DATA, reference/combine1 2 3 4Inv. (M€) 74/60 0/144 116/310 146/473O&M (M€/y) 2.3/2.8 0/8.8 3.6/15.8 6.1/11.1ENERGY CONSUMTION, (GWh/year), ref./combine1 2 3 4Biomass 397/244 730/1537 470/1271 362/2970Others 74/135 3 - - -ENERGY PRODUCTION (GWh/year), reference/combine1 2 3 4Electricity 125/0 218/209 145/55 99/78Biofuel 0/90 0/444 0/294 0/1336Others - - 0/384 4 -1 Besides ethanol also biogas and pellets is produced.2 Also kerosene and nafta is produced.3 Fuel oil (21/15) and industrial waste heat (53/120).4 Biogas (0/114) and Pellets (0/270)144

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSystem 1In the current configuration of this system 15-20% ofthe energy demand is covered with fuel oil, whichneeds to be reduced. One interesting option could beto convert biomass into bio oil by pyrolysis and thenuse the bio oil in the existing oil boilers. Bio oil that isnot used within the system can be sold (e.g. summertime). If no pyrolysis reactor is built, a conventionalbiofuel fired combined heat and power plant (bio CHP)will be invested in, building up the reference case.System 2In this system, there is no need for new productionunits, rather there is a high production capacity,allowing for integration of a bioenergy production unit.System 2 has good access to biomass, but might havedifficulties to find a market for large quantities of byproducts.Based on these prerequisites, a suitablecombine technology could be cellulose ethanolproduction with enzymatic hydrolysis aiming at highyield and in-house use of energy by-products.Regarding the O&M cost for the enzymatic process inTable I, future enzyme price are assumed [4], Withtoday‘s prices, the enzymatic process will not beprofitable.System 3In System 3 there is a need for new productioncapacity, which is represented by a bio CHP in thereference case. This system has good access to alarge energy market, which enables output of otherenergy products. Hence, a cellulose ethanol plantbased on acid hydrolysis can complement thereference case investment to build up the combinecase.System 4This system is in many aspects similar to System 3, butethanol production is not in line with company strategy.Moreover, System 3 has good access to peat, whichcould supplement biomass for a large scale productionunit. Hence, gasification of biomass for production ofsynthetic biofuel is evaluated for this system.Environmental evaluationThe assessment of the environmental implication ofintroducing a bioenergy production in an existingdistrict heating system focuses on changes inemissions of green house gases (GHG). A systemapproach for analysing the changes of GHG‘s isapplied. This means that besides changes of the directemissions on site, also the changes of emissions inaffected parts of the energy systems are included; seeFigure 1. For instance, production of biofuel in thecombines ads to the environmental benefit since fossilfuels can be replaced, while reduced electricityproduction has a negative impact to the environmentalbenefit in accordance with marginal electricityproduction.Production,distribution anduse of biomassDirect GHG emissionsGHGDH system withor without bioenergyproductionGHGPowersystemProduction,distribution and useof transportationfuelGHGFig.1. Illustration of the applied system approach forassessing the changes of GHG‘s.In the assessment, all GHG‘s of significance areincluded [3]: carbon dioxide (CO 2 ), dinitrogen oxide(N 2 O) and methane (CH 4 ). For all energy carriers, lifecycle emissions are considered, i.e. both combustionemissions and well-to-gate emissions such asemissions from fuel extraction, processing andtransportation. Also leakages are considered whenapplicable. How the GHG‘s for the relevant energycarriers are assessed are described in brief below, amore thorough description can be found in [3].Theadopted life cycle GHG emissions associated withchanges in consumption/production of the energycarriers are summarized in Table II.Table II. Emission factors for included energy carriers.ENERGY CARRIERBiomass 14-17 1High emission elec. (E1) 800Low emission electricity (E2) 260Pyrolysis oil 292Ethanol 307FT diesel 277Fuel oil 312Biogas 207Pellets 286LIFE CYCLE EMISSION(kg CO 2 eq./MWh)1 The lifecycle emission of biomass is dependent on howthe biomass is used in the energy combines (e.g.hydrolysis for fermentation or gasification)BiomassThe energy input in all four combines is in the form ofbiomass. Production, distribution and use of biomass isrelated to GHG emissions. The GHG emission from theuse of biomass differs depending on how the biomassis used. Combustion raises emissions of both methane145

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaand N 2 O (the CO 2 emission are assumed to be neutralfrom a climate perspective), while hydrolysis andfermentation is not assumed to raise these emissions.Hence, the net lifecycle emission of biomass differsbetween 14-17 kg CO 2 eq./MWh fuel.ElectricityIn all district heating systems, the electricity productiondecreases as a consequence of introducing thecombine (see Description of the cases). Any change inelectricity production is assumed to be compensated bychanges in marginal electricity production. Forinstance, if the electricity production decreases by 85GWh/year, it is assumed that other producers willincrease their production by 85 GWh/year. To assessthe environmental impact of this, the decrease has tobe multiplied with a emission factor for marginalelectricity.There are many opinions regarding the emissions ofmarginal electricity. Here we have used a high and alow level, based on dynamic response for electricityproduction with two different developments over a longtime period [5]. By using a high and low figure, theimpact and importance of changes in electricity can beillustrated in a clear way. For the high figure, thereference case in [5] is used where lifecycle emissionsof marginal electricity are about 800 kg/MWh el . Thismarginal electricity is denoted E1 hereon. With morestringent environmental targets the electricityproduction can be carbon lean [5] implying that the longterm lifecycle emissions would be about 260 kg/MWh el ,denoted E2 hereon.BiofuelAs seen in Table I, the evaluated bioenergy combineshave various biofuel products as output. In System 1pyrolysis oil is produced. The pyrolysis oil is assumedto replace fossil fuel oil (but is categorized as an biofuelherein). If lifecycle emissions are regarded according tothe approach in ref. [6] for both pyrolysis oil and fossilfuel oil, the net GHG reduction for replacing fuel oil withpyrolysis oil is 292 kg per MWh of pyrolysis oil exportedfrom the combine. Also the amount of fuel oil useddiffers in the combine case from the reference case inSystem 1 (see Table I). The net life cycle GHG of thisfuel oil is set to 312 kg/MWh.In systems 2 and 3 ethanol is produced, which isassumed to replace gasoline with net GHG reduction of307 kg per MWh of ethanol reaching the market.In System 4, three biofuels are produced: FischerTropsch (FT) diesel, nafta and kerosene. All threeproducts are assumed to replace fossil transportationfuel with the net GHG reduction of 277 kg/MWh. Thepossible leakage of methane from the gasificationprocess is assumed to be negligible.Biogas and pelletsIn the energy combine of System 3, also biogas andpellets are produced. The biogas is assumed to beused as a transportation fuel to replace both petrol anddiesel. The net GHG reduction for replacing fossiltransportation fuel with biogas is set to 207 kg/MWhincluding life cycle emission and gas leakage in theproduction. The pellets are also assumed to replacefossil fuel, in this case oil with a net GHG reduction of286 kg/MWh pellets.Resource efficiencyWith the emission factors in Table II and the energyflows of the reference and combine case in Table I, theenvironmental benefit of the energy combine can beassessed. However, if biomass is assumed to be alimited resource from a sustainability point of view, itmakes sense to evaluate the use of biomass from anefficiency perspective. Hence, the resource efficiency isassessed as the net GHG reduction potential (in kgCO 2 eq.) per used quantity of biomass (in MWh). Bycomparing this key figure for the reference case withthe combine case for each system, the resourceefficiency of the combines can be evaluated.Economic evaluationIn order to analyze whether an investment addsfinancial value we rely on a standard discounted cashflow (DCF) model estimating the net present value(NPV) for each project so that:NPVntCF t1 rt0/ (1)where CF t denotes the net cash flow in year t, r is thefuture weighted cost of capital and n is the number ofyears included in the cost-/benefit analysis. The cashflow at year 0 indicates the initial outlay. Concerning r,the weighted cost of capital (WACC), we do notpredetermine a specific hurdle rate; instead we analyzevalue added for three different levels of discount rates.We do so because any statements on the actualriskiness of the project or an estimation of the WACCfor the companies are outside the reach of this study.As stated before, when estimating cash flows the pointof departure is a reference object. That is, our NPVcalculations only address the differences in cash flowsbetween the reference and the bioenergy combine; thisfor two reasons. First, only the incremental cash flowsare relevant in a DCF analysis. For instance, in thecase of System 3 they already decided that they wouldat least build a combined heat and power (CHP)facility, and the question is if they gain from makingadditional investments in a bioenergy production unit.Second, by focusing on the differences we do not needto consider the cost structure in the reference case, it istreated as a given. Besides simplifying the analysis,146

academic access is facilitated as there is no need toreveal sensitive information.Table III. Assumptions made for non-site idiosyncratic inputand output prices (€/MWh).Ethanol 78 Biomass 19FT-diesel 78 Fuel oil 57Kersone 78 Pellets 25Nafta 52 Electricity 47Biooil 47 Electricity excise 0.5Biogas 68 Electricity certificate 1 211 Premium paid to producers of renewable electricity.Cash flowsThe initial outlay is assumed to take place in full at year0. Yearly operational cash flows are projected by firstestimating an operational cash flow for the first year. Ascash flows are the products of price and quantity, thisestimation is based on the technical analysis in order toobtain energy flow estimates (see Table I), and thenmultiply them with price estimates, to which we addout-payments for operation and maintenance. Weextrapolate this operational cash flow over the 20 yearlong investment horizon with a three percent yearlygrowth rate (adjusted for the fact that green certificatesare obtained for fifteen years only). All cash flows areconservatively assumed to occur at the end of eachyear. Next, we add tax payments (assuming aneffective tax rate of 26,3%), tax discounts fromdepreciation (according to Swedish tax code), changesin working capital (approximated by dividing thedifference between in-payments and out-payments ofyear t by 12 and subtracting the corresponding valuefrom year t-1, save for the last year where thedifference is set to zero) and a terminal value (5% ofthe initial outlay). Initial outlays are determined byconsulting [7]– [19]. Our price assumptions for non-siteidiosyncratic inputs and outputs are presented inTable III. For translation between different currenciesthe following exchange rates were used: 9.6 SEK/€ and6.5SEK/USD.Sensitivity analysisWe then control the robustness of the NPV estimatesthrough sensitivity analysis; that is, we examine howthe cost-/benefit analysis is affected when changing avariable at the time, holding all else equal. We do thisin two steps for each system. First, we illustrate thechanges in estimated NPV by changing yearly inpayments,yearly out-payments, initial outlay andterminal value respectively. Second, we show howyearly in-payments and out-payments respond to pricechanges.By this sensitivity analysis, we can to some degreecompensate for the uncertainty that surrounds ourestimates of initial outlays and terminal value, and weThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia147can see for what potential price changes extra concernis warranted. Certainly, a drawback with the sensitivityanalysis is that it is just a ceteris paribus analysis anddoes not take into consideration the potentialcovariance of variables, for instance between ingoingbiomass and outgoing biofuel.Business context evaluationThe environmental and economic analyses of a jointproduction operation act as a starting point for thebusiness context analysis. A wide-spread adoptiondemands not only indications of environmental benefitsand economic profits, but must also offer a fit with theexisting business context. Even though the degree of fitis defined on company level we will not analyze it assuch. Rather we use the business context of thestudied systems in order to put together a compilationof restrictions and barriers to a wide-spread adoption.The magnitude and importance of these will giveimportant indications of the short term possibilities ofrealizing environmental benefits and economic profitsin making bioenergy combines a future growth industry.The restrictions and barriers are identified through thefit with existing input/output market situation, productionand system configuration and general businessconditions, (i.e. strategic focus and capacity to absorbadditional risk) dominant in the host company.ENVIRONMENTAL BENEFITSAs already stated in the Research design, theenvironmental benefit from integrating bioenergyproduction into an existing district heating system isassessed as the reduction of GHG‘s from a systemperspective. As also explained, the net differencedepends on the reference case as well as thecomposition of the energy combine. In Figure 2, theGHG reduction for the included parts of the referencecase and energy combine case of System 3 isdisplayed. In the reference case (left bar in Figure 2)– a combined heat and power (CHP) plant – biomass isconverted into heat (for district heating) and electricity.The amount of heat is the same in both the referenceand combine cases and, hence, not considered in theevaluation of GHG reduction. However, the productionof electricity will change and the system consequencesof that is, as stated, considered by including twodifferent assumptions for marginal electricity. Assumingthat marginal electricity is related to about 260 kg CO 2eq./MWh el (E2), the electricity produced in thereference case results in a yearly reduction of 38Mtonne (dark blue bar to the left in Figure 2). If theemissions of marginal electricity instead is assumed tobe 800 kg/MWh el (E1), the emission reduction wouldincrease by 78 Mtonne/year (light blue bar) to be intotal 116 Mtonne (dark + light blue bar = E1). Thehandling of the biomass is related to GHG emissions

(see Environmental evaluation) and, hence, there is anegative bar of 8 Mtonne for biomass. To sum up, thenet GHG reduction in the reference case is 30 or 108Mtonne CO 2 equivalents depending on assumptions forthe marginal electricity.The combine case of System 3 has lower electricityproduction than in the reference case (see Descriptionof the cases). Consequently, the GHG reduction fromthe electricity production is also lower, which is seen aslower dark and light blue bars for the combine case;middle stacked bar in Fig. 2. Moreover, the negativebar for biomass is larger for the combine since morebiomass is used in this case. In the energy combine,however, bioenergy products such as biofuel (ethanolin this system), biogas and pellets are produced. Asalready explained, these energy products are assumedto replace fossil fuels and the resulting GHG reductionfrom the combine is significant: 188 or 217 Mtonne CO 2eq. with carbon lean (E2) and carbon intense (E1)electricity production, respectively.GHG reduction (Mtonne CO 2 eq./yr)3002001000-100-200Net reduction (E2/E1):30/108 188/217 158/109Reference Combine DifferenceElectricity, E1-E2*Electricity, E2PelletsBiogasEthanolBiomass* additonal emissionreduction/change ifelectricity is relatedto high CO 2 emissionsFig. 2. GHG reduction in System 3 for the reference case,combine case and the net difference for converting to thecombine.The dark blue bars are related to marginal electricityassociated to low GHG emission (E2). The additionalemission reduction/change if electricity is related tohigh GHG emissions (E1–E2) is indicated by the lightblue bars. The total emission/change for E2 is given bythe sum of light blue and dark blue bar.The implication in terms of GHG‘s of integratingbioenergy production in System 3 can be visualised bymoving from the left bar in Figure 2 to the middle bar.Consequently, the difference of the two bars shows theGHG implication of converting to an energy combine inSystem 3, which is presented in the right hand bar inthe figure. The change from the reference to thecombine case gives rise to GHG reduction from the fuelproducts (green bars) However, the electricityproduction decreases, implying decreased reduction(emission increase) and, hence, negative bars forelectricity. As can be seen in the figure, the net GHGThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia148reduction from introducing an energy combine inSystem 3 is 158 or 109 Mtonne/year depending on theassumption for marginal electricity (E2 and E1,respectively).The equivalents to the right hand bar in Figure 2 for allfour 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 lowemissions (E2, dark blue bar only). In System 1, theenvironmental benefit is negative, even if the marginalelectricity is CO 2 lean.Significant environmental benefits, as displayed forsystems 2-4, are expected since the combines in thesesystems use more biomass, which eventually replacesfossil fuel in the system approach applied (in system 1less biomass is used which explains the negativeresults for this system). However, if biomass isassumed to be a limited resource from sustainabilitypoint of view, the use of biomass should also beevaluated from an efficiency point of view. As explainedin the Environmental evaluation, one measure ofresource efficiency is the GHG reduction potential perused quantity of biomass. This key figure is presentedin Figure 4 for both the reference case and thecombine case for the four district heating systemsevaluated.GHG reduction (Mtonne)400350300250200150100500-50-100-150Net reduction (E2/E1):-2/-69 124/119 158/109 321/309System 1 System 2 System 3 System 4Others*BiofuelElec., E1-E2Elec., E2Biomass* biogasand pelletsFig. 3. Environmental benefit from introduction of energycombines.As seen in Figure 4, the energy combines are lessresource efficient than the reference cases (generally abiomass fired CHP plant) if the marginal electricity isassociated with high CO 2 emissions (E1, dark + lightblue bar). However, if the marginal electricity isassociated with low CO 2 emissions (E2, dark blue baronly), the combines are more resource efficient thanthe reference cases. As also can be seen, the resourceefficiencies do not differ dramatically betweensystems 2–4. System 1, however, shows lowerresource efficiency, which can be explained by the factthat a major part of the produced pyrolysis oil isconsumed internally in the system instead of replacingfossil fuel off site.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaResource efficiency(kg CO 2 eq./MWh biomass)250200150100500Ref.Comb.Ref.Comb.Ref.Comb.Ref.E1-E2E2Comb.System 1 System 2 System 3 System 4Fig. 4. Resource efficiency of biomass quantified asGHG reduction per used quantity of biomass.ECONOMIC VALUEWhether the cost/benefit analyses return positive NPVsdepend largely on the hurdle rates assigned to them. InTable IV a summary of the economic results arepresented including the initial outlay, the expected freecash flow for the first year and estimated NPVs for 4, 7and 10% discount rates, respectively. With theexception of System 1, where the bioenergy combine isactually cheaper than the reference plant, marginalinitial outlays vary between M€ 140 and 330, andexpected cash flows for the first year of operationsbetween M€ -3 and 57. The largest addition to existingcash flow (both in absolute and relative terms) comesfrom the bioenergy combine investment in System 4.Table IV. Summary of cost/benefit analyses for adding abioenergy combine to the reference investment in thestudied systems.1 2 3 4Initial outlay (M€) - 13.9 144 194 327Cash flow (M€y) -3.4 18.8 15.7 57NPV (M€) for different discount rates4% -40 76 -62 3627% -27 29 -89 20710% -19 -4 -108 101As also can be seen in Table IV, only two projects arevalue adding at a 4% discount rate, and System 4 isthe only one that can bear a 10% discount rate. Theresults for System 1 are a bit upside down, sincecompared to the reference case the investment costand net cash flows are negative for the combine.System 3, perhaps being the weakest of casesanalyzed, will not show positive figures for any positivediscount rate.For robustness control purposes, sensitivity analysesare performed, here presented for System 3. Figure 5illustrates the estimated NPV consequences fromchanges in marginal cash flows, disaggregated into inpayments,out-payments, initial outlays and terminalvalue.Change in NPV (M€, 10% disc. rate)100500-50-100-150-200-250-300In-paymentsInitial outlay-30% -20% -10% 0% 10% 20% 30%Change in cash flowsOut-paymentsTerminal valueFig. 5. Estimated changes in NPV (M€) for System 3 asa result of percentage changes in cash flows assuming a10% discount rate.A percent change in either of these, results (ceterisparibus) in a NPV change, as indicated in the figure. Itis clear that the project is most vulnerable for changesin in-payments followed by out-payments. Assuming ahurdle rat of ten percent, a 20% average increase inyearly in-payments would result in an increase in NPVof about € 100 million. Correspondingly, a 20%increase in yearly out-payments result in a NPVreduction of € 84 millions. Fig. 5 also show that thecost/benefit analysis is not very sensitive to changes ininitial outlay and leave no visible mark for changes interminal value. The order of importance of NPV impactof cash flow changes are similar in the other threesystems, where in-payments being the most importantones.Change in marginal in-payments25%20%15%10%5%0%-5%-10%-15%-20%-25%Ethanol Biogas Pellet-30% -20% -10% 0% 10% 20% 30%Price changeFig. 6. Estimated percentage changes in in-payments forSystem 3 as a result of percentage changes in inputprices.149

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaHaving established the sensitivity to changes in cashflows it follows naturally to examine also to whatdegree different cash flows changes with respect tochanges in underlying prices. In Figure 6, the relationbetween marginal in-payments and prices of ethanol,biogas and pellets are shown for System 3. It is clearthat ethanol is by far the most important bioenergyproduct, where a 20% increase in prices renders a 12%increase in in-payment.Change in marginal out-payments20%15%10%5%0%-5%-10%-15%-20%Biomass Electricity O&M-30% -20% -10% 0% 10% 20% 30%Price/unit cost changeFig. 7. Estimated percentage changes in out-payments forSystem 3 as a result of percentage changes in inputprices/unit costs.Similarly, Figure 7 shows how out-payments vary withinput prices. Inputs included in the figure are biofuel,operations and maintenance (O&M) and electricity 7 .Not surprisingly, biofuel is the key input, where a 20%price change results in a 10% change in out-payments,which in Figure 5 translates to a € 42 million change inNPV.The sensitivity analyses of System 3 show that minorchanges in underlying factors can result in significantchanges in the NPV estimates. However, a notinsignificant part of the indicated variability in cashflows should be hampered by the offsetting effectsdriven by the probable covariance between prices forbiomass and bioenergy products. To be noticed is thatthe order of importance of the inputs in the other threesystems show a similar ranking, where biofuel andbiomass price being the two most important ones.FIT WITH EXISTING BUSINESS CONTEXTThe environmental and economic evaluations indicatethat the integration of bioenergy production intomedium sized district heating systems can beassociated with both environmental and economicbenefits, but the picture is mixed and ambiguous. Froman environmental point of view, the results are coherentacross all systems: the absolute environmental benefitof bioenergy production is in proportion to the use ofbiomass, since increased use of biomass impliesincreased output of CO 2 neutral energy products.However, from a resource efficiency point of view,biomass should not be used to replace transportationfuel as long as the marginal electricity is related to highCO 2 emissions. One important explanation to thecoherent environmental profiles of the differentbioenergy combine solutions is similar resourceefficiency for the four technologies evaluated. Hence,our results suggest that it is possible to find differentenergy combine with similar resource efficiency.However, these similarities in resource efficiency donot indicate similarities in economic attractiveness. Infact, the economic evaluation seems to suggest thatsome bioenergy production technologies are notcurrently economic viable for integration with districtheating system. Furthermore, the results indicate thatnot all district heating systems are suitable forintegration with a biofuel production unit. Despite beingof the same size, use the same raw material and beingevaluated only on marginal effects on the economicsituation, differences in district heating systemcharacteristics have a profound impact on theeconomic possibilities of energy combine integration. Inthis study we have matched every system with acombine solution in order to maximize the site-specificopportunities in each system. This opens of course thepossibility that there exist other matches with lessresource efficiency but higher economic profitability.Even if this can be the case, we would like to point outthat one of the starting points of this study was to basein-data on the conditions of real systems. This includestaking various kinds of restrictions into consideration.Even though these restrictions vary, the onesprominent in this study can be grouped into fourdifferent categories:Proximity to input resourcesProximity to customers or infrastructure fortransporting the finished productsExisting production and system configurationDominant business conditionsProximity to input resourcesSome combine solutions (such as the one for System4) demand huge amounts of biomass. This requireslarge areas of regional biomass recourses and little orno competition over it. Import by sea is an alternativebut it requires production sites close to a harbour.Proximity to market for the finished productThe production of biogas is one example of both theimportance of proximity to customers and to1507 The electricity in out-payments corresponds to the electricityused in the bioenergy production unit. In Table 1, only the netelectricity export is displayed.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniainfrastructure. Only relying on local demand for biogasis considered too challenging at present time.Yearly Cash Flows (M€)100Existing production and system configurationInvestments in bioenergy combines are seldom greenfield but, as we have shown earlier, have to be adaptedto suit existing heat volumes, demand curves, systemconfigurations and also production site layout. In one ofthe systems, the production site was too small to housethe large amounts of biomass necessary for achievingan economic profitable size of an ethanol operation.80604020In-paymentsFree cash flowOut-paymentsBiofuelBiogasPelletsBiomassIndustrial waste heatElectricityElectricity certificateDominant business conditionsThe results of the study show that two business areashave an evident influence on the type of bioenergycombine investments the companies carry out: 1) thestrategic framing of the district heating company and 2)the risk that these investments innate. Concerning thefirst, many of the municipally owners use the utilities toenhance and to some extent even realize theenvironmental visions that are formed and expressedon the political level. Examples of these found amongthe companies represented in this study include;phasing out fossil fuels, use of local waste resourcesand visions of a fossil free cities based around locallyproduced bioenergy fuels. When present, strategicframing has a visible effect on limiting the number ofavailable alternatives for integrates production.As stated, the second area that has an significantinfluence on the type of bioenergy combine that thesecompanies consider is the risk that these investmentsinnate. Due to the municipal ownership, thesecompanies are inherently dependent on stablebusiness conditions. The ability to absorb negativeresults is strongly limited. The added business risk ofbioenergy production must, if needed, be able to beabsorbed by cash flows from existing operations or astrong capital base. In principle, this can be done in twoways, either by keeping the investment relatively small,or by only accepting business propositions with cashflows that can be made relatively stable.In Fig. 8, the operational risk of the investment can tosome extent be visualized by the size of the marginalcash flows of the different investments. The investmentin system 4 stands out not only because it is the largestone but also because its in-payment comes from onesource only. If the price correlation with biomass ishigh, this might not be a large problem. However, it isinteresting to note the relatively small positive cash flowavailable from existing operations in Systems 4, andalso for System 3. If the company carries through withthe evaluated investment, it will dramatically change itsoperational risk profile and over-all business focus.0-20System 1 System 2 System 3 System 4O&MFig. 8. Marginal cash flows (in-payments/out-payments)for each system in comparison to free cash flow fromexisting operations in 2007 (shaded bar).The considerable positive free cash flow of system 2from its existing operations is explained by thecompany‘s sell of hydropower. Although irrelevant forthe value of this investment, it could function as ageneral safeguard against negative results, due tounfavourable relation between biofuel and biomassprices.The investment in system 1 was not profitableaccording to the valuation earlier. Despite this, it isworth pointing out that the risk of this investmentshould be low since it uses its own products as input. Ittoo has, relatively speaking, a strong free cash flowfrom its current operation that will decrease the risk ofending up in the red.CONCLUSIONSThe results of the bioenergy combine analyses showthat there are indications for both environmental gainsand added economic value of such investments.However, these benefits seem to be limited by severaloperational, environmental and economiccircumstances present in these systems. First, theseinvestments are dependent on the need for makingmajor changes in current production layout, typicallythe need for new or altered production plants. Thislimits the available window of opportunity. There arealso several limitations related to operationalcharacteristics, availability of input resources andsuitable product markets. A closer investigation ofexisting governance situation also shows that theseinvestments often are made to fit owner strategiesregarding environmental goals of the local energysystem. Finally, the municipally ownership typicallylimits the risk appetite which also limits availableinvestments. The doubtful short term environmentalbenefit is a more general objection based on thevaluation of the current marginal power production.151

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaNever the less, it will hamper the potential for widespreadadoption of bioenergy combines.These circumstances lead us to conclude that not allbiofuel production technologies are suitable for alldistrict heating system. Our economic analyses alsoindicate that not all district heating systems are suitablefor bioenergy combine production. In fact the barriersare so many that it is reasonable to assume they willeffectively reduce the number of systems adopting thisoperational design in the near future.ACKNOWLEDGEMENTSThe main funding for this project is provided byFjärrsyn, which is a research program organized by theSwedish district heating branch agency. Additionalfunding is also received from the project ―Pathways toSustainable Energy Systems‖.We kindly thank the representatives from each districtheating system for a good cooperation and forproviding us with technical and economic data of theirsystems. Without these inputs, the work would nothave been as solid as it is.We also thank Karolina Nilsson and John Jonsson(both at Profu) for their valuable contribution to thework.REFERENCES[1] M. Odenberger, F. Johnsson, ―Pathways for theEuropean electricity supply system to 2050‖, Int. J.of Greenhouse Gas Control, 2010, Vol. 4:2, pp327-340[2] J.Sjödin and D. Henning, ―Calculating the marginalcosts of a district-heating utility‖, Applied Energy,2004, Vol. 78:1, pp 1-18.[3] E. Axelsson, C. Overland, K. Nilsson, and A.Sandoff, ‖Bioenergikombinat i fjärrvärmesystem‖,Fjärrsynsrapport 2009:11.[4] T. Brandberg, Senior researcher at SEKAB E-technology, Personal communication, 2009.[5] H. Sköldberg and T. Unger, ‖Effekter av förändradelanvändning/elproduktion‖. Elforsk report (2008).[6] IVL, ―Miljöfaktabok för bränslen‖, IVL Rapport B1334B-2 (2001).[7] Svebio, ―Kraftvärmeutbyggnad 2007-2015‖, Svebiorepport 2008-03-31.[8] H. Hansson, S-E. Larsson, O. Nyström, F. Olssonand B. Ridell, ―El från nya anläggningar - 2007‖,Elforsk repport no 07:50 (2007).[9] M. Zakrisson, ―Internationell jämförelse avproduktionskostnader vid pelletstillverkning‖,Master‘s thesis no 29 2002, SLU.[10] A. Hang and S. Ilic, ‖En förstudie för bioetanolproduktion i Borås‖, Master‘s thesis at InstitutionenIngenjörshögskolan, Högskolan i Borås (2008).[11] M. Lantz, ―Drivmedelsproducentersbetalningsförmåga för energigrödor‖, Miljö- ochenergisystem, LTH (2006).[12] J. Benjaminsson and A. Dahl, ―Uppgradering avbiogas‖, Presentation at ―Temadag uppgraderingav biogas‖, Göteborg (2008).[13] I. Granberg, Project leader at Jönköping Energi,Personal commication (2008).[14] M. Tijmensen, A. Faaij, C. Hamelinck, and M. vanHardeveld, ―Exploration of the possibilities forproduction of Fischer Tropsch liquids and powervia biomass gasification‖, Biomass and Bioenergy2002, Vol. 23.[15] I. Johansson, S. Larsson and O. Wennberg,―Torkning av biobränslen med spillvärme‖,Värmeforskrapport 881 (2007).[16] E. Sandvig, G. Walling, R. Brown, R. Pletka, D.Radlein, and W. Johnsson, ―Integrated PyrolysisCombined Cycle Biomass Power Systems‖,Repport of Alliant Energy, Iowa, USA (2003).[17] H. Thunman, F. Lind, and F. Johnsson Delstudieenergikombinat, Elforskrapport, 2008.[18] NREL, Research Advances Cellulosic Ethanol,NREL (2007).[19] P. Sassner, M. Galbe, and G. Zacchi, ―Technoeconomicevaluation of bioethanol production fromthree different lignocellolosic materials‖, Biomassand bioenergy 2008, Vol 32.152

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSEA WATER DISTRTICT COOLING FEASIBILITY ANALYSIS FOR TALLINNA. Hani 1 , I. Britikovski 1 , H. Voll 1 and T.-A. Kõiv 11Tallinn University of Technology, Department of Environmental Engineering, EstoniaABSTRACTIn this paper sea water district cooling feasibilityanalysis for Tallinn is presented. It has become moreand more interesting to study alternative solutions forpublic buildings A/C cooling due to relatively highelectrical energy prices. Besides economical aspectstechnical and environmental sides must be considered.INTRODUCTIONalso important to locate the district cooling station nearto energy source.Typical SW district cooling system principle is indicatedin Figure 1. The system consists of three mainsections:Cold sea water pumping;Cooling plant with heat exchangers;Standard cooling distribution network.The first large district cooling systems were developedduring the 1960‘s in Hartford (1962) and California(1965) in United States [10]. The first systems inEurope were La Defense (1967) in France and inHamburg (1968) Germany [1]. In the beginning of the70`s the first system in Japan (Shinjuku) was built [3].However, because of the energy crises in the end of70`s, the District Cooling development was slow and nonew large systems were built. Until the end of 80`swhen many new large systems were opened forexample Kioi-cho, Nishi-Shinjuku in Japan and TrigenTrenton in United States. Also the first district coolingsystem of the Nordic countries was installed in Norway.Operation started in 1989 in Baerum, near Oslo. Thefirst system in Sweden was built in 1992 in Västerås [2]and since then the district cooling in Sweden hasdeveloped rapidly. Since the 1990‘s the establishmentof commercial district cooling systems has increasedrapidly worldwide. Nowadays, more than 20 countrieshave a commercial district cooling system and this isexpected to increase rapidly [4].The sea water (SW) district cooling is based on largenatural cold water source. Enough cold water isaccumulated in lakes, seas, oceans, rivers, etc [8].Lowering the coolant temperature with sea water is analternative to conventional electrical energy consumingchillers [5]. The system working principle is quite similarto geothermal energy production which is used inheating systems [6]. Until now the sea water districtcooling is quite conservatively expanded around theWorld.SW DISTRICT COOLING PRINCIPLEThe temperature in conventional cooling water networkis between +4…+7 o C so applicable the sea watertemperature should be below +5 oC . Despite thatcompressor based cooling can be used in case coolingwater temperature is higher than mentioned [9]. It is153Fig. 1 SW district cooling principle schematicHeat exchangers allow usage of the soft water indistribution network while problematic salty sea waterhandling will be done in open central circuit.Environmental impact study is required before any ofthe projects will be executed. Large sea waterquantities have to be available to minimize pumpingimpacts. In addition to evaluation of the deep zone coldwater pumping, the analysis of recycling the sea waterback to lower sea water zone with higher temperaturesshould be made.Following factors shall be considered before systemdesign [7]: Minimum altitudes between heat exchangers andwater resource level should be designed; Centralized district cooling plant (heat exchangers,pumping station and chillers) is less expensivethan decentralised system; Centralized system has less maintenanceproblems.DESIGN PARAMETERSTemperature of the sea water varies during differentseasons and distance from the coast.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaIn following Table 1 and Figure 2 the relations of theSW parameter can be found.cooling network with total capacity of 19 MW. Project isinteresting to public buildings which have lower balancetemperature and due to that higher cooling demand.Study was carried out to construct: Cooling plant with 4 water chillers; Sea water pumping station (free cooling, precooling)with 5 heat exchangers; District cooling network to customers.In Tallinn costal area is 21 potential customers whosecooling demand is app. 19,2 MW. Simultaneous factor0,85 is assumed. Cooling demand will be covered withwater chillers and SW free cooling. Calculations of 21public buildings information are presented in Table 2.Cooling load is calculated 120 W/m 2 (building no 17cooling load 60 W/m 2 ). In calculations was notconsidered residential area cooling load due to differentusage profile compared to public areas.Fig. 2 Temperature and SW depth relationTab. 1. SW parametersDist.fromcoast, mDepth(sea),mAnnualaver.temp, o CMintemp,o CFrom previous studies has been found that coolingdemand exceeds significantly when the outdoortemperature exceeds 16 ºC (see Figure 3).Fig. 3 Ambient temperature and cooling power relationCASE STUDY – TALLINN COSTAL AREAMaxtemp,o C500 20 7,5-8,5 2,5 17,51500 25 5,5-6,5 1,5 173200 30 4-5 1 14-164000 35 3,5-4

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTab. 3 Main technical parameters for design the systemMax cooling demand19,2 MWAmbient temp. calc.27 o CSimultaneous factor 0,85Cooling station capacity18 MWAnnual average cooling consumption 21600 MWhSupply water temp6 o CReturn water temp (max consumption) 16 o CReturn water temp (min consumption) 13 o Ccase sea water temperature is below 5 o C. Maximalpressure drop in both circuits is selected 0,85 bar. Heatexchanger parameters are indicated in Table 5.Tab. 5 Free-cooling heat exchanger parametersHeat exchangers capacitySea water (SW) supply temp.SW return temp.SW flowDistrict cooling supply temp.District cooling return temp.5x3600 kW4,5 o C10 o C130 l/s6 o C16 o CWater chiller coolingCentralized cooling plant contains up to 4 water chillersto gain flexibility of the system. Also it is possible toconstruct the cooling plant step by step according toconsumers‘ interest and cooling energy demand.System contains four 4500 kW water chillers withcentrifugal compressors. It is possible to adjust thecooling power of the unit between 300–4500 kW whichmakes the system more energy (el) efficient during thepartial load period. The condenser has to be producedfrom titan or similar resistant material due to fact that itis being cooled with sea water. In the following Table 4are indicated technical parameters for water chillers.District cooling flowMax pressure dropTab. 6 Coolant parametersSea water (SW)SW tempMax pressureDistrict cooling liquidTemperatureMax pressure72 l/s0,85 bar1,5-18 o C6 bar10-18 o C10 barTab. 4 Water chiller parametersCooling powerRefrigerantCondenser temp.Seawater (SW) supply temp.SW return temp.SW flow (each unit)4x4500 kWR-134a28 o C18 o C24 o C215 l/sSea water coolingThe sea water is supplied through insulated 800mmpipes to pumping station using sea water gravity. Threepumps (max 1080 m3/h) with frequency converters areinstalled using parallel scheme to suction pipe. Seawater pressure is ca 1,5 m and pumps will add 2 barsto overcome self-cleaning filters, heat exchangers andcondensers pressure drop. Frequency converters areused to lower energy consumption during partial load.Evaporator temp.COP full load 73 o CCOP partial load 12District cooling supply temp.District cooling return temp.District cooling flow (each unit)Free-cooling6 o C16 o C115 l/sWhen sea water temperature is lower than returntemperature from the network free-cooling through heatexchangers can be used. Optimum logarithmictemperature difference shall be app. 1,5 o C. Five heatexchangers with capacity of 3600 kW are selected,which assure whole cooling plant capacity (18 MW) inCooling plant operation modesCooling plants are designed to have three differentoperation modes: SW temperature < 5 o C. Completely free-cooling; SW temperature 5–12 o C. Pre-cooling with SW +compressor cooling;SW temperature > 12 o C Only compressor cooling(free cooling heat exchangers are equipped withbypasses).District cooling networkSupply (forward) water temperature is designed 6 o C.Return water temperature between 13–16 o C (seeFigure 4).155

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDue to fact that summer period soil temperature in1,5 m depth is 10 o C it is not necessary to insulate thereturn pipe of the district cooling network. Supply pipeis insulated with 10 cm nowadays heat insulationmaterial.The cooling plant shall have three operational modes:Free-cooling;Pre-cooling + compressor cooling;Compressor cooling.Optimization of the proposed system should be carriedout in further studies.REFERENCES[1] Vadrot, A. and Delbes, J, (1999). District CoolingHandbook a Survey of Techniques, equipment andChoice of System. European market Group.Number of pages 208.[2] Feldhusen. H, Francesc. M. R, (2001). "DistrictCooling-Present Market Assessment," Master,Kungl Tekniska Högskola, Stockholm division ofApplied Thermodynamics and Refrigeration. pp 52.Stockholm.Fig. 4 District cooling network tempCONCLUSIONThe sea water (SW) district cooling has until year 2000quite modestly developed among different countriesaround the World. Due to the fact that energy priceshave raised rapidly more and more researches for freeenergy resources are carried out. Wind power, heatpumps, solar energy and sea water have obtainedhuge attention.SW district cooling is centralized and will haveadvantages like less pollution, less maintenanceproblems and in perspective also economic benefits.Current feasibility analysis was done in Tallinn costalarea to define possible cooling plant load, potentialconsumers and technical possibilities.Due low costal area it is possible to locate the coolingplant near to sea water. Further studies should addsome more economic aspects to the technical solution.Problematic is to develop the district cooling network inTallinn area (existing tunnels and subways will easethe process).Most of the new built or renovated public buildingshave high cooling demand due to glass walls and highinternal heat loads. In present research 21 buildingswith only public area were included (total coolingdemand 19,2 MW). The cooling demand risesconsiderably when ambient air temperature exceeds16 o C. Sea water temperature 5 o C can be found indepth of 35–40 m.[3] Euroheat and Power, (2003). District Heat inEurope Country by Country/2003 Survey. BrusselBelgium.[4] Mildenstein, B. S. P, (1999). District Heating andCooling Connection Handbook.[5] Gosney. W.B, (1982). Principles of Refrigeration.Cambringe University Press. Published by thepress syndicate of the University of Cambridge.[6] Westin, P. E. H., (1999). Production Technologiesin District Cooling Systems and the Importance ofLocal Factors. New Energy Systems andConversion-NESC 99.). pp 6.Osaka.[7] Westin, P. E. H., Karlson, B., and Lundqvist, P,(1999). Straategies and Methods For Increasingthe Capacity of District Cooling Systems.20th<strong>International</strong> Conferenss of Refrigeration, IIR/IIF.).pp 1-8. Sydney.[8] Nordell, B., and Skogsberg, K, (2002). Snow andice storage for cooling applications.Winter Cities2002.Japan Aomori. Luleå University ofTechnology[9] Eliadis, C, (2003). Deep Lake Water Cooling ARenewable Technology. Number of pages 3.[10] Morris, A.P, (1995). The Road to Lockport:Historical Background of District Heating andCooling. Ashrae Transactions: Symposia.[11] Arvidson, J, Asplund, A-L, Birgerrson, E, (1997),Cold production uning low temperature wasteheat,. Kungl tekniska högskolan Kemiskapparatteknik. Pp 54, Stockholm156

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaABSTRACTANALYSIS FOR THE OPERATION BEHAVIOR AND OPTIMIZATION OF CHPSYSTEM IN DISTRICT HEATING AND COOLING NETWORKYong Hoon Im 1 , Hwa-Choon Park 1 , Byung-Sik Park 1 and Mo Chung 21 Cogen. & Boiler Research Group, Building Energy Research Center,Korea Institute of Energy Research, Korea3 Mechanical Eng. Dept., Yeungnam Univ., KoreaA simulation program for analyzing the effects of thenetworking operation of existing DHC system inconnection with CHP system on-site is to be discussedin this study. The practical simulation for arbitrary areaswith various building compositions is carried out for theanalysis of operational features in both systems, andthe various aspects of thermal network operation arehighlighted through the detailed assessment ofpredicted results. The intrinsic operational features ofCHP prime movers, gas engine, gas turbine etc., areeffectively implemented by realizing the performancedata, i.e. actual operation efficiency in the full and partloads range.For the sake of simplicity, a simple mathematicalcorrelation model is proposed for simulating variousaspects of change effectively on the existing DHCsystem side due to the networking operation, instead ofperforming cycle simulations separately. The empiricalcorrelations are developed using the hourly basedannual operation data for a branch of the KoreanDistrict Heating Corporation (KDHC) and are implicit inrelation between main operation parameters such asfuel consumption by use, heat and power production. Inthe simulation, a variety of system configurations areable to be considered according to any combination ofthe probable CHP prime-movers, absorption or turbotype cooling chillers of every kind and capacity. Fromthe analysis of the thermal network operationsimulations, it is found that the newly proposedmethodology of mathematical correlation for modellingof the existing DHC system functions effectively inreflecting the operational variations due to thermalnetwork operation. The effects of intrinsic features ofCHP prime-movers, e.g. the different ratio of heat andpower production, various combinations of differenttypes of chillers (i.e. absorption and turbo types) on theoverall system operation are discussed in detail withthe consideration of operation schemes andcorresponding simulation algorithms. The variousaspects of system configuration in terms of CHPsystem optimization are also discussed.INTRODUCTIONIn Korea, the district heating and cooling (DHC) systemgains share of the market steadily and it amounts to15712.3% on the basis of the total number of householdsat the end of 2008 [1]. The annual heat sales, via DHCnetwork, in 2008 have reached 16,676 thousand Gcaland it increased by about 5% on average after 2001.Considering the trend of new-town development inmetropolitan areas and newly developing residentialareas on a large scale, it is generally expected to showa clear increasing trend of DHC systems on the marketfor the time being. Furthermore, the relevant changesof circumstances such as the long-term expectation forhigh prices of fossil fuels and the imminent realizationof UNFCCC around the world will help the CHP andDHC system tighten its grips on the forthcomingheating and cooling market [2]-[3]. Among the severalmerits of DHC systems against separate heat & power(SHP) or central heating system, the distinctive featureof being able to construct the networking system withthe neighbouring DHC systems certainly deserves toreceive attention from the view point of efficient use ofenergy resources and operation costs reduction [4].However, the effectiveness of networking operation ofCHP and neighboring DHC systems is stronglyinfluenced by the conditions of energy consumptionbehaviours and corresponding operation scenarios onboth sides. The different pattern of energy consumptionin new demand areas is highly desirable for creatingsynergy effects by networking operation. In addition,the different operation strategy of CHP system with thatof DHC network can also improve the effectiveness ofnetworking operation. The optimal system configurationof the CHP system with networking operation certainlydiffers from that of stand-alone CHP system not tomention the operation characteristics. Since the heatflows in the network are bi-directional, the appropriatemodelling for the mutual effects on each system ishighly required for the accurate estimation of thenetworking operation.The main purpose of this study is to examine thefeasibility of the network operation of the CHP systemon-site with the existing DHC system in terms ofefficient use of primary energy and reduction of theoperation cost. In this study, a simulation program isdeveloped for analysing the thermal networkingprocess between the existing DHC system and theCHP system for the newly developing area. The effectsof thermal networking on the existing DHC systemoperation are implemented using mathematical

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniamodelling with empirical correlations for main operativeparameters. The intrinsic features for the CHP primemovers is modeled using the actual performance dataof operation efficiency in full or part load conditions.The specific features of the newly developed programin simulation of thermal networking process in districtheating is described in terms of the energy loadprediction and operation simulation of various systemconfigurations with CHP prime movers and types ofcooling chillers. The unit energy load model for variousbuildings by use, e.g. apartment, hotel, hospital,buildings for business and commercial use etc, isintroduced for the accurate prediction of energy loadsfor newly developing area. The effects of intrinsicfeatures of CHP prime movers, e.g. the different ratio ofheat and power production, various combination ofdifferent types of chillers (i.e. absorption and turbotypes), on the overall system operation are alsodiscussed in detail in the following.Fig. 2 and Fig. 3 show examples of the daily unitenergy load model of heating for the apartment andhourly unit energy load model of electricity for the officebuilding respectively.The annual hourly unit energy model can be obtainedby synthesizing the daily and hourly unit energy loadmodels [5]. The final annual hourly energy consumptionfor given building compositions and correspondingscale is to be predicted with the input of the total areasfor respective buildings since the unit energy loadmodels have been developed by normalizing thestatistical energy consumption measurement data withthe corresponding building areas. The example ofannual hourly energy consumption for the apartment isshown in Fig. 4.MODELLING FOR NET-WORKING OPERATION1. Modelling of CHP system operationIn the previous studies [5]–[9], a simulation tool for theoptimal design of the CHP system had beendeveloped, which is composed with three differentmodules of energy load prediction, operationsimulation, and economic analysis modules as shownin Fig. 1. The main goal of the simulation is to draw anoptimized system configuration for a given target areaby the systematic analysis of the physical andmechanical behaviour of the CHP system andcorresponding operational cost structure. In principle,the analysis is performed on hourly basis for a year.The unit energy load model for a variety of buildingtypes (e.g. apartment, commercial building, officebuilding, department store, hospital etc.) has beendeveloped for different types of energy loads, i.e.heating, cooling, electricity and hot water [10]-[13]. Inenergy load prediction module, the hourly, annualenergy demand for a target area is predicted using theunit energy load models.Fig.1. Relationship between load, operation and economicanalysis modules [9]Fig. 2. Daily unit energy load model for the apartmentFig. 3. Hourly unit energy load model for the office buildingIn the operation simulation module, a variety of CHPsystem configurations can be considered in terms oftypes of prime-movers for the CHP system (e.g. gasengine, gas turbine, combined CHP, flexible electricitygas turbine), its capacity, and facility types for cooling(if cooling load is available) [6]. In the operatingsimulation of the CHP system, it is noted that thephysical or mechanical operation results such as fuelconsumption, heat supply, electricity produced by CHPetc. are calculated by using the operation performancedata for the real products of CHP system, or coolingfacility instead of performing thermodynamic cyclesimulations for respective facilities separately. In order158

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniato implement the schemes, the performance data forthe commercial products, operation efficiency in fulland part load condition, has been extensivelyinvestigated and the database has been realized on thesimulation program.One can consider a variety of CHP systemconfigurations with various CHP prime movers andtypes of cooling chillers. If the type of CHP primemovers is being selected, the capacity of it is to bedetermined in the form of any percentage on the basisof the maximum value of annual hourly electricitydemand. Then, the feasible options, which can matchthe condition entered by the user, are compiledaccording to the relevant algorithm as shown in Fig. 5.(a) Heating load(b) Electricity loadFig. 4. Prediction of annual hourly energy consumption forthe apartmentthe number of units, and the load factor in terms of unitcapacity. When an option is selected by the user asdescribed above, its corresponding technical data forCHP product will be linked automatically in thesubsequent operation simulation procedures. Thesettlement of the system configuration for the coolingsystem can also be performed in a similar manner byproviding the data for the ratio of being in charge ofturbo or absorption type chillers.2. Modelling of DHC system for networkingoperationIn contrast with small cogeneration or CES system, theDHC system is not authorized to sell the electricity tothe customer directly in Korea [6]. As a result, theoperation mode differs from that of cogeneration orCES system, i.e. the facilities are operating dependingon the heat loads, and CHP facilities stop operatingduring summer to reduce waste heat production.Instead, the hot water load during the summer seasonis usually supplied from incinerators nearby, or heatonly boilers (HOB). However, the operation schemes ofDHC system for stand-alone operation are bound to bemodified to some extent by networking operation withCHP system on-site and the appropriate modelling forsuch an effect of networking operation on DHC systemis a key element for a reliable prediction of theoperation behaviours due to thermal network operation.In this study, the changes of operation schemes andcorresponding variations for physical or mechanicalaspects on existing DHC system side have beenrealized by employing mathematical correlations for thesake of simplicity. The mathematical correlations forenergy productions as a function of energyconsumption are developed based on the annualoperation data of a branch of Korea District HeatingCorporation (KDHC). By applying a simple, but credibleempirical correlations instead of performing anadditional cycle simulation for the existing DHC system,the calculation load and the complexity from thestandpoint of simulation are considerably alleviated.The procedure to obtain the correlations for energyproduction in terms of energy consumption are givenas follows,Fig. 5. Parametric entry of option for CHP system productThe user is to select the most desirable one among thelist of options by referring to the technical specificationfor each option such as the unit capacity of the product,159The required data for the establishment of themathematical correlation is given by,– Annual, heat and electricity production and the salesper day according to the facilities of heat production(CHP, HOB, Incinerator)– Annual, fuel consumption per day according to thefacilities of heat production (CHP, HOB)

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe functional form of the mathematical correlation isgiven as follows,F f ( H,P)(1)Where,F: Fuel consumptionH: Heating loadP: Electricity loadcorrelations is a certain time, not a specific time duringthe year as in the original data. For example, if theDHC system is requested to produce more heataccording to the request from CHP system to theamount of Q , the heat load of DHC system can beregarded to be changed from Q 1 to Q 2 , i.e. Q 2 =Q 1 + Q .Then, the operation behaviour for DHC system at themoment can be estimated simply from themathematical correlations by simply referring the valueof F 2 *, corresponding to Q 2 * and P* corresponding toF 2 *. It means that one can reconstruct the operationbehaviour of the DHC system as a function ofsequential time reflecting the effects of thermal energynetworks. The correlations for the heat and electricityproduction vs. fuel consumption are shown in Fig. 7.(a) Time vs. events(a) Electricity production vs. fuel consumption(b) Events vs. eventsFig. 6. Illustrative diagram for the correlation betweenenergy production and fuel consumptionFig. 6 shows the illustrating diagram for themathematical correlation between energy productionand consumptions. For any time t 1 , an optimizedoperation scenario already exists and correspondingheat and electricity production, and fuel consumptionhas been fixed according to the operation scenario andfor any time t 2 , it is the same as above. On the basis ofthe operation data for a year, the behaviour of systemoperation can also be described between dependentvariables (e.g. F: Fuel consumption, H: Heatproduction, P: Electricity production). In the correlationsbetween dependent variables, the time t is reflectedwith implicit manner and the meaning of time t in the(b) Fuel consumption vs. Heat productionFig. 7. Developed correlations for the energy productionsvs. fuel consumptionSIMULATION OF THE THERMAL NETWORKINGOPERATION1. Operation Conditions and SchemesThe operation of the overall system should be carriedout by the order of priority of operation for the variousheat sources. In this study, the basic schemes in orderof priority for supplying the energy demands in newlydeveloped area are established as shown in Fig. 8,160

1. CHP system operation in A2. Thermal networking operation using CHP in B3. HOB operation in A4. Thermal networking using HOB in BThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia(a) Case A(b) Case BFig. 10. Comparison of energy load prediction: elec. LoadFig. 8. Schemes of the networking operation2. Test Case & Energy Loads PredictionOn the basis of the operation schemes for variousavailable heat sources as described above, theanalysis for the operation behaviour of networkoperation of both systems with those of respectivesystem is performed for two distinct test cases ofresidential buildings only, and a group of nonresidentialones.(a) Case A(b) Case BFig.11. Comparison of energy load prediction: cooling load(Case B)For case A, the area is only comprised of residentialpurpose buildings, i.e. apartments, whereas for case Bit is comprised of non-residential purpose buildingssuch as commercial buildings, offices, hotels, andhospitals. The annual hourly energy load data isestimated by using the energy load prediction module.The comparison of predicted energy loads, in the formof the annual distribution and the cumulative curve, aregiven as shown in Fig. 9 to Fig. 11.3. Operation Simulation ResultsFor the test case comprised of only residentialbuildings, the cooling load is reflected on the electricityload by assuming that it is covered by the airconditioner or electric fan in individual houses.Consequently, the aspect of efficient utilization of therecovered waste heat during the summer is supposedto be a decisive factor in the optimization of the CHPsystem.Fig. 9. Comparison of energy load prediction: heating load161

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia(a) Gas engine(a) Gas engine(b) Gas turbineFig. 12. Typical pattern of heating load and recoveredwaste heat for CHP prime-moversIt is easy to see the typical consumption pattern ofheating and hot water for residential houses in Koreaas shown in Fig. 12. A large variation of heating load isobserved in heat consumption rate and the optimaldesign of CHP system with such a large variation ismore difficult than with a relatively regular consumptionpattern. The typical annual operation results of therespective CHP prime-movers, gas engine and gasturbine, is also shown in Fig. 12. There is a largedifference in the recovered waste heat prediction forrespective CHP arising from the intrinsic feature for gasturbine, i.e. higher heat to power ratio of gas turbineagainst gas engine.(b) Gas turbineFig. 13. Annual thermal energy supply and demandoperating condition for newly developing area(a) Gas engineFigure 13 shows the annual thermal energy supply anddemand operating conditions for two distinct CHPprime-movers. First of all, the quantity of recoveredwaste heat from CHP is not large enough to cover thewhole heat demand in the winter, so that most of heatdemand is covered by HOB operation on-site. It isnoted that the heat supply from DHC network seldomoccurs during the winter. This is mainly due to the factthat it is also short of heat energy in existing developedareas during the winter. Of course, it is a probablescenario to operate the HOB in existing DHC system toproduce the required amount of heat energy for newlydeveloping area. However, it does not actually happenbecause the operation of the HOB on -site has priorityover that of the HOB in existing DHC system accordingto the operation schemes.162(b) Gas turbineFig. 14. Annual electricity supply and demand operatingcondition for newly developing areaThe thermal network operation is observed to takeplace mainly during the intermediate seasons. It showsthat most of heat demand is covered by the thermalnetworking heat supply and it results in bringing down

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniathe rate of operation for HOB on-site considerably.From the view point of system operation efficiency, ithas a very positive impact in that the rate of operationof CHP in DHC system increases to some extent.However, in case of supply of surplus heat to existingDHC system as shown for gas turbine, it is vice versa.It is noted that the heat flow of thermal network can bebi-directional for the gas turbine as shown in Fig. 13.The annual supply and demand operating conditionsfor electricity are shown in Fig. 14. A comparativelygood electricity-tracking operation is observed for bothCHP prime-movers and the supply from the grid tendsto increase during the summer due to the peak of theelectricity demand.Fig. 15 shows the variations of electricity production onexisting DHC system side due to thermal networkingoperation. It is interesting to note that a minor increaseof the electricity production for existing DHC system isobserved during the intermediate seasons. This iscaused by the increased rate of operation of CHP inexisting DHC system due to thermal networkingoperation.The detailed variation of electricity production on theexisting DHC system side is given as shown in Fig. 16.The net increase of electricity production for gasengines is larger than that of gas turbines. This isbecause of the intrinsic feature for gas engine CHPsystem of smaller heat to electricity ratio than that ofgas turbine, which induce that more heat is supplied toon-site by the thermal network and consequentlyincrease the rate of operation of CHP in DHC system.Fig. 15. Variation of electricity production on existing DHCsystem side due to thermal networking operation(a) Gas engine(a) Gas engine(b) Gas turbineFig. 17. Annual LNG consumption rate for newlydeveloping area according to the CHP prime mover(b) Gas turbineFig. 16. Detailed variation of electricity production on theexisting DHC system side according to CHP prime moverThe net amount of LNG consumption for newlydeveloping area is given for different CHP operationsaccording to the heat source facility as shown inFig. 17. It is noted that the composition of LNGconsumption for respective heat source facility variesconsiderably. Since there are various special discountschemes for LNG price in promotion of energy efficientfacilities such as CHP, cooling chillers based oncogeneration system etc. the reliable estimation of LNGconsumption according to their usage is crucial for theassessment of economics for the scenarios.163

In case B, the analysis of operation characteristics fordrawing optimal system configuration becomes muchmore complex due to the existence of cooling load. Thesimulation results for gas turbine with 50% ofabsorption type cooling (i.e. 50% turbo type cooling)are given in the following.The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 18 and Fig. 19 show the annual heat load andoperating conditions of thermal energy supply anddemand for on-site.As shown in Fig. 9, where the heating loads for case Aand B are compared, the heating load for a group ofnon-residential building composition is much smallerthan that of residential building composition. As aresult, the waste heat recovered from gas turbineoperation is sufficient enough to encompass the wholeheat loads in case B as shown in Fig. 18. In terms ofthermal networking operation, there is a great changein the pattern of system operation in that a largeamount of surplus heat energy is available even inwinter not to mention the intermediate seasons. Thismeans that a large amount of heat is flowing toward theexisting DHC system side as shown in Fig. 19, andthere will be serious effects on the operation of existingDHC system.(a) Heat(b) ElectricityFig. 20. Variation of operating conditions due to thermalnetworking operation on the existing DHC system sideThe effects of surplus heat energy on the operationconditions for the existing DHC system side are shownin Fig. 20. First of all, the considerable reduction for therate of CHP system operation during the intermediateseason is observed and it is also expected that the rateof operation for HOB is to be reduced in the winter asmuch as the amount of heat supply from the CHP onsite.Fig. 18. Heating load and recovered waste heat for gasturbine CHPFig. 21. Detailed variation of electricity production due tothermal networking operation on the existing DHC systemsideFig. 19. Annual thermal energy supply and demandoperating condition for newly developing areaConsequently, the heat production on the existing DHCsystem side is reduced to some extent as shown inFig. 20 (a) and it brings about the reduction of LNGconsumption for DHC system. In terms of electricityproduction as shown in Fig. 20 (b), there is a minorvariation for the production of it in winter despite theconsiderable thermal networking operation. It meansthat the CHP system on DHC system side is in fulloperation during winter regardless of thermalnetworking operation and the shortage of heat energy164

is covered by operating HOB. In other words, only theoperation of HOB on the DHC system side is affectedby the thermal networking operation in winter. Asshown in Fig. 20 (b) and Fig. 21, the production ofelectricity on the existing DHC system side during theintermediate seasons is certainly decreasing due to thesupply of surplus heat from CHP on-site, which resultsin the diminution of the rate of operation for the CHP onthe existing DHC system side.The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia(a) Absorption type 80%Fig. 22. Heat balance of operating the absorption chillersThe operating characteristics for cooling load aredescribed in the following with Fig. 22. It shows theheat balance of operating the absorption chillers. Thecooling load exceeding the supply capacity fromrecovered waste heat is modelled to be covered byproviding auxiliary heat for absorption chillers by directgas combustion. The cooling load assigned to turbotype chillers is dealt with as an electricity loadconverted according to the COP of the correspondingproduct of turbo chillers.(b) Absorption type 20%Fig. 24. Heat balance of operating the absorption chillersfor different responsibility by absorption type coolingThe heat balance of absorption chillers for differentratio of responsibility by absorption type cooling isshown in Fig. 24. In case of 80% absorption typecooling, the recovered waste heat is not sufficientenough to handle the assigned cooling load, so anauxiliary heat source, such as direct gas combustion, isneeded to cope with the full absorption cooling load.Whereas, when the 20% absorption type cooling loadis concerned, the required amount of heat for theabsorption chillers can be supplied only by therecovered waste heat as shown in Fig. 24. Theremainder of total cooling load is covered by turbo typecooling system.Fig. 23. Annual LNG consumption rate by use for newlydeveloping area of a grope of non-residential buildingsThe LNG consumption with the cooling load for newlydeveloping area is predicted as shown in Fig. 23. Dueto the lower level of heating loads for non-residentialbuildings, operation of HOB facility is only permissiblein a limited period even in the winter. It is also notedthat a portion of LNG is consumed to provide auxiliaryheat for absorption chillers by direct gas combustion incase of shortage of heat from recovered waste heat.The effects of cooling system configuration on thenetwork operation characteristics are assessed in detailas follows:(a) Absorption type 80%(b) Absorption type 20%Fig. 25. Annual electricity supply and demand operatingcondition for newly developing area for differentresponsibility by absorption type cooling165

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe different operation characteristics in terms ofelectricity demand and supply is given in Fig. 25. It isnoted that the electricity demand during the summerincreases considerably as the ratio of absorption typecooling is decreasing. This peak of electricity during thesummer is due to the consumption of electricity foroperating turbo type chillers. From the view point ofdesign of the CHP system configuration in thesimulation, the cooling load is an important parameterto be considered carefully, because the capacity ofCHP system is given in the form of any percentage ofthe peak value of electricity, i.e. the maximum value ofthe annual electricity consumption rate per hour.Therefore, the criteria for defining the CHP capacity isto be varied depending on the amount of cooling loadassigned to turbo type chillers. The respective LNGconsumption patterns depending on the ratio ofabsorption cooling load are compared in Fig. 26. It isinteresting to note that the composition of fuelconsumed by use is substantially changed according tocooling load treatment during the summer. The resultsconfirm that the effects of various aspects ofconfiguration for CHP and cooling system on theprediction of operational parameters (e.g. fuelconsumption rate by use) are properly realized in thesimulation program.By using the simulation approach as presented in thisstudy, the optimal design of the CHP system innetworking operation with DHC system can be carriedout since one can access the detailed physical dataregarding the whole operation of the network systemsuch as annual rate of fuel consumption for respectivesystems (e.g. CHP, HOB, Chiller etc), annualproduction of electricity, heat, and the amount of heatexchange etc. Along with the appropriate coststructures for fuel, product sales (heat and electricity)and the estimation of capital cost, civil construction,and O&M costs etc, one can also make theassessement for the economic feasibility of variousscenarios. However, the detailed economic analysis forthe test cases and the procedures to determine theoptimized CHP system configuration based on it willnot be described in this paper due to the pageconstraints. These ítems will be discussed in furtherstudies.(a) Absorption type 80%166(b) Absorption type 80%Fig. 26. Annual LNG consumption rate by use for differentresponsibility by absorption type coolingCONCLUSIONA simulation program that predicts the energy loads fora mix of buildings and estimate the operationalcharacteristics for networking operation of existingDHC system with CHP system on-site is developed.The distinctive features of this simulation approach canbe summarized as follows,– The unit energy load models are developed foraccurate prediction of energy consumption by useaccroding to any combiation of building type and scale.– A simple mathematical correlation for reflecting thevariations of the network operation on an existing DHCsystem side is newly proposed for the sake of simplicityand efficient simulation process.– The performance data for the commercial products,operation efficiency in a full and part load condition,has been extensively investigated and the databasehas been realized successfully on the simulationprogram.The operational characteristics of thermal networkingoperation has been assessed in terms of systemconfigurations for the CHP and the cooling facility asfollows.– According to the intrinsic features of the CHP primemovers such as gas engine and gas turbine etc, theaspects for the supply of surplus heat is progressing indifferent manners by and large. For a gas engine, theon-site is almost short of heat so that the predictionresults indicate that the additional operation of CHP onthe exisiting DHC system side is induced in theintermediate seasons. Whereas, surplus of waste heatrecovered from gas turbine CHP is supplied toward theexisting DHC system side. As a result, the amount ofelectricity production is being decreases to someextent.– In case of a group of non-residential buildings, theheating load reduces considerably. Therefore, it isprobable that the heating load can be covered by onlythe recovered waste heat from on-site even in thewinter. Due to the heat flow toward the DHC system

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaside in the winter the rate of operation of HOB will bedecreased.– The thermal energy exchanges via the network andthe corresponding changes in operation on both sidesare prevailing in intermediate seasons in case of similarheat consumption patterns on both sides.– The operation of cooling system on the newlydeveloping area is verified not to have much effects interms of thermal networking operation. However, thesignificant changes in the LNG consumption patternsby use are observed according to the ratio ofresponsibility by absorption chillers for the cooling load.The various aspects of system configuration in terms ofCHP system optimization are discussed with thedevelopment of a simulation program in this study. It isverified that the physical and mechanical mechanismsconcerned with the thermal networking operation hasbeen appropriately modeled from the assessment ofoperational behavior for test cases.ACKNOWLEDGEMENTThe author gratefully acknowledges the financial andtechnical supports for the research from the KoreaDistrict Heating Corporation (KDHC).REFERENCES[1] Korea Energy Management Corporation, ―Statisticsfor district heating and cooling enterprise in Korea‖,2009.[2] A. Marbe, S. Harvey, ―Opportunities for integrationof biofuel gasifiers in natural-gas combined heatand-powerplants in district-heating systems‖,Applied Energy, 2006, Vol.83, pp. 723-748.[3] C. Weber, I. Heckl, F. Friedler, F. Marechal, D.Favrat, ―Network synthesis for a district energysystem: a step towards sustainability‖, ComputerAided Chemical engineering, 2006, Vol. 21, pp.1869-1874.[4] H. Lund, F. Hvelplund, I. Kass, E. Dukalskis, D.Blumberga, ―District heating and market economyin Latvia‖, Energy, 1999, Vol. 24, pp. 549-559.[5] H. C. Park, M. Chung, S. H. Kim, ―Development ofsystem simulator for community energy system‖,Report to Ministry of Industry, 2003.[6] Y. H. Im, H. C. Park, M. Chung, ―A study of optimalheating supply systems for the newly developingarea 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 studyfor small size cogeneration systems in themetropolitan areas of Seoul‖, Final Report to SH(Seoul Housing) Corporation, 2008.[8] M. Chung, H. C. Park, ―Development of a energydemand 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 softwarepackage for community energy system assessment– Part I: Building a load estimator‖, Energy, inpress.[10] H. C. Park, S. S. Lee, D. J. Kim, ―Development ofenergy models for department stores‖, KoreanJournal of Air-Conditioning and RefrigerationEngineering, 2003, Vol. 15, pp. 1088-94.[11] H. C. Park, M. Chung, ―Building load models forhotels in Korea‖, Journal of the Korean SolarEnergy Society, 2009, Vol. 29, pp. 48-57.[12] H. C. Park, ―Development of weighting factors forvariables associated with hourly energyconsumption pattern for hotels in Korea‖, SAREK(Soc. Air-conditioning, Ref., Engineers of Korea)Winter Annual meeting, 2002, pp. 76-82[13] H. C. Park, ―Analysis of energy loads for hospitalbuildings‖, SAREK journal, 2002, pp. 1088-93.167

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaIMPROVED PRIMARY ENERGY EFFICIENCY OF DISTRICT HEATING NETWORKS BYINTEGRATION OF COMMUNAL BIOMASS-FIRED COMBINED HEAT AND POWERPLANTS WITH BIOMASS PYROLYSIST. Kohl 1 , N.A. Pambudi 2 , T. Laukkanen 1 and C.-J. Fogelholm 11 Aalto University, Dept. of Energy Technology, Espoo, Finland1 Corresponding Author: Thomas Kohl, e-mail: thomas.kohl@tkk.fi2 Semarang State University, Semarang, IndonesiaAbstractABSTRACTThis paper investigates the influence of the integrationof communal biomass-fired combined heat and powerplants with wood-pyrolysis on the plant‘s energybalance and product distribution. Further the proposedintegration concept‘s influence on the environmentalperformance of the connected district heating network ispointed out. The environmental performance isevaluated by means of the primary energy factor andthe CO 2 emission coefficient. For this evaluation, theEuropean standards EN 15603 and EN 15613-4-5 areapplied and modified.The concept comprises the integration of a simplepyrolysis model and of a steam dryer with a base casecombined heat and power plant. The yearly plant outputis calculated by applying a multiperiod model of the heatduration curve. The work shows that, by co-generationof valuable pyrolysis product, operation hours andelectricity production can be considerably improved.The integration also clearly improves the district heatingnetwork‘s primary energy efficiency and lowers itscarbon dioxide emissions significantly.INTRODUCTIONThe European Union‘s carbon dioxide mitigation goalsand plans to reduce energy import dependency requireaction towards a more sustainable energy supply that isbased on renewable energy sources available in themember states. Biomass is discussed controversiallydue to its wide range of upgrade possibilities frompower, heat, cooling to chemicals and transportationfuels. Among others, EU directives 2001/77/EC(―…promotion of electricity produced from renewableenergy…‖), 2004/8/EC (―…promotion ofcogeneration…‖) and 2003/30/EC (―… promotion of theuse of biofuels…‖) state that the use of biomass forenergy purposes should be expanded on a sustainablebase.However, the increased use of biomass is expected toraise prices for biomass which will negatively influence,among others, the economy of communal biomass-firedcombined heat and power (CHP) plants – a technologythat is currently competitive to fossil energy production.Furthermore the scarcity of the biomass availabledemands most efficient use of this resource.As shown in a previous study [1] it looks promising tointegrate biorefinery processes, that are linked totransportation fuel production, with CHP plants, sinceCHP plants can provide both a source for hightemperature heat needed for thermal conversion ofbiomass as well as the district heating network (DHN)as a sink for sensible heat that would usually berejected in stand-alone biofuel refineries. It has beenfurther worked out that the integrated production ofinterstage products, such as liquid fast pyrolysis product(often referred to as woodoil) and wood pellets, haveseveral advantages: Firstly, the products areindependent from the transportation fuel marketdevelopments since they can be seen as a universalinput for different upgrading processes to e.g. biodiesel,ethanol, methanol, hydrogen or other chemicalsproduction but they can also be directly combusted forpower and heat generation. Secondly, they increase thebiomass‘ energy density making it more sustainable fortransportation to central plants required for economicfuel production. Thirdly, technologies applied for suchpre-processing are relatively simple and robust, thuskeeping investment cost and system complexity on areasonable level and making it therefore also interestingfor local small-scale solutions.In this paper, outgoing from a base case, we simulatethe retrofit integration of wood fast pyrolysis with anexisting wood-fired CHP plant. The aim is highestpossible pyrolysis product generation using the freeboiler capacity in part loads under the condition that thedistrict heat (DH) demand is still fulfilled.With help of a multiperiod model of the DHN‘s heatduration curve, the work shows the influence of theintegration on plant operating hours, electricityproduction and biomass throughput. In addition theeffects on the DHN‘s primary energy factor and CO 2emission coefficients are studied as well. The primaryenergy factor and the CO 2 emission coefficient arecalculated according to European standards EN 15603[2] and 15316-4-5 [3], applying a modified power bonusmethod. However, no cost estimation is given, since thefocus of the work was to find out if this integration168

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaconcept is possible within the operational limits of theCHP plant.In the following, first the used multiperiod load model isdescribed. Further a brief introduction to biomass fastpyrolysis is given and it is shown how the process hasbeen simulated and integrated. Then the modification ofthe European standards is explained, results arepresented and finally restrictions of the work andoptions for further improvement are discussed.District Heat Load [MW]252015105Heat Duration Curve - Multiperiod ModelIntegrated Case - Lower LoadsDISTRICT HEATING LOADThe CHP plant chosen has been integrated into a virtualDHN. Therefore yearly data of a real DHN has beenscaled so that the CHP plant provides 60% of the hourlypeak demand of the DHN when on full load. The CHPplant is assumed to be shut off at 50% load whichcorresponds with 30% demand in the DHN. As stated in[4], those are common operating parameters forcommunal solid fuel-fired CHP plants.In order to represent the yearly production of the basecase plant a multiperiod load model was developed.One full load and five part load levels have been chosento represent the heat duration curve. The pyrolysisintegrated CHP plant is represented by 7 part loadlevels since lower DH loads can be supplied, asexplained later. Operating time periods per part loadlevel are set of equal length and -together with the fullload period- match the total operation hours and yearlyDH generation of 94.5 GWh as shown in figures 1a and1b. For each load level, fuel input and pyrolysis yield arethen iterated matching the required DH output. DHdemand not provided by the CHP plant is assumed tobe generated in oil-fired heat-only boilers with a thermalefficiency of 0.85District Heat Load [MW]252015105Heat Duration Curve - Multiperiod ModelBase Case00 50 100 150 200 250 300 350District Heat LoadMultiperiod Model DH LoadTime [d]Real CHP DH LoadFig. 1a: DH Load Multiperiod Model - Base Case00 50 100 150 200 250 300 350District Heat LoadMultiperiod Model DH LoadTime [d]Real CHP DH LoadFig. 1b: DH Load Multiperiod Model – Integrated CaseCHP PLANT INTEGRATED WITH WOOD PYROLYISWood Pyrolysis ModelBiomass fast pyrolysis is the thermal conversion ofbiomass in the absence of oxygen at temperatures ofapproximately 500 °C and pressures close toatmospheric [5]. The basic idea of the pyrolysis unit isderived from the bioliq® process developed by theForschungszentrum Karlsruhe (FZK). There, fastpyrolysis is applied in order to yield a high share ofliquid pyrolysis product. Biomass is indirectly heatedand pyrolysed with sand in an inert atmosphere at atemperature of about 500 °C. Subsequently, thepyrolysis gases are condensed and the liquid fraction(also referred to as wood oil) is mixed with the coke andforms the so-called bioslurry which leaves the plant asthe final product. In this work we use data published byFZK [6] and hence assume that 90% of the biomass‘energy is converted into bioslurry whereas 10% accruesin gaseous form. The pyrolysis gas is thought to be cofiredin the boiler and hence its energy is subtractedfrom the fuel input into the boiler.As pyrolysis requires a low fuel moisture content ofapproximately 10% [5] a dryer must be integrated aswell. Indirect steam drying is applied, since this alsoallows the regulation of the DH load. As explained later,regulation is necessary since the enthalpy of the steamflow after the modification exceeds the demand of theDHN and hence must be adjusted.The wood pyrolysis process is modelled as follows: Theheat of pyrolysis of wood is set to 1.87 MJ/kg (moisturecontent 10%) using data for pine derived from [7].Therewith the pyrolysis yield is calculated from the heatextracted from the flue gases.CHP Plant – Base CaseA base case CHP plant with a bubbling fluidized bedboiler (shown in fig. 1) has been simulated in full and169

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniapart load using the thermal power plant simulatorProsim. Performance data of the CHP plant and fuelinput specification is derived from [4] and [8],respectively and given in table 1. The higher heatingvalue (HHV) is calculated by the simulation software.Table 1: Base Plant – Input Specification and Performanceat Design LoadSimulation model input data - design loadWood FuelUltimate analysis C 50.64 O 42.22H 6.10 N 0.16Ash 0.8 S 0.08Moisture 50 % HHV 18.8 MJ/kgFuel input26 MWSteam CycleHigh pressure steam 60 bars Condenser pressure 0.69 bar510°CPlant PerformanceDistrict heat output 16.5 MW Electrical efficiency η el 0.243Power output 6.3 MW Power to heat ration α 0.381Fig. 1: CHP Plant – Base CaseCHP Plant – Integrated Wood PyrolysisThe modified CHP plant is illustrated in Fig. 2. In orderto provide heat for the pyrolysis process, the heat mustbe extracted from the flue gases leaving the fluidizedbed reactor (numbered 3 in Fig.2) boiler at 850 °C. Therequired amount of flue gas is split off (18) after thefluidized bed reactor. As in the FZK process, those fluegases are thought to heat up sand to 550 °C (whichwould provide the heat for the pyrolysis process bycooling down to 450 °C) [6]. The flue gas thereby isestimated to cool down to 480 °C. The flue gas is thenmixed back (20) into the main flow before theeconomizer. The heat extraction needed for biomassfast pyrolysis process is modelled by help of anadditional evaporator (19). 90% of the biomass energyon a lower heating value base will form pyrolysis slurrywhereas 10% accrues as pyrolysis gas. The energycarried by the pyrolysis gas reduces the biomass fuelinput as explained above (―Wood Pyrolysis Model‖).CHP plant – Integrated Steam DryingThe dryer is modelled as a steam tube dryer. Payingattention to the retrofit situation, live steam is extracted,throttled to 10 bars and further cooled to 190 °C byspraying in the saturated water leaving the dryer. Dryingof biomass to low moisture contents requirestemperatures far above the saturation temperature at agiven pressure due to the hygroscopic properties ofbiomass. Heat consumption of the dryer has beenestimated to 2750 kJ/kg water evaporated [9]. Woodand hot flue gases are led in the dryer (24). If heat isavailable from the flue gases, those are cooled down to120 °C and together with the fully condensing steamprovide the heat needed for the drying process. Driedwood leaves the dryer at wet bulb temperature. For thedrying process live steam is extracted (21), throttled(22) to 10 bars and further cooled to 190 °C by sprayingin condensate (23) leaving the dryer. The dryercondensate is throttled to 2 bars (26) and send to thefeedwater tank (15). Flue gas temperatures of 120 °C170

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaare considered not to cause sulphur corrosionproblems, especially not with low-sulphur wood fuels.Certainly this design specification must be reconsideredin case of changed fuel properties.The maximum pyrolysis production for each load pointis restricted by the maximum steam extraction rate andby the boiler‘s maximum burning power. The maximumpossible pyrolysis yield logically requires highestpossible fuel input since heat must be provided both, fordrying and pyrolysis. Conversely, this means that thesteam enthalpy exceeds the demand of the DHN. Thisis because the boiler temperature is controlled bymeans of the evaporator- and superheater tubes in theboiler walls. If now, the heat input in the boiler is kept ona higher level as usual the water amount needed todissipate the heat from the boiler walls is onlydecreasing to a certain amount (resulting from areduced temperature after the economizer).Consequently, in order to match the DH load, this heatmust now be ―dissipated‖ in the pyrolysis heatexchanger (19) or in the dryer (24). By iteration the DHload is matched by adjusting dryer load, correlated splitoffto the pyrolysis heat exchanger and fuel input. In allcases the boiler load (characterised by the fuel heatinput) is restricted to 100%. So, the overload back-upcapacity of the boiler is maintained. With this setup thepyrolysis yield constantly increases with the decrease ofthe DH levels down to 60%. The maximum flow off thedryer (and thus its capacity) is be restricted by thepressure prevailing in the feedwater tank, which in turnis given by the extraction pressure after the turbinestage (11). The pressure decreases with falling livesteam parameters and steam massflow. Hence, there isa pressure dependant maximum enthalpy flow that canbe fed into the feedwater tank until saturation state isreached for the mixture of the condensates from the DHexchanger (13) and the dryer (24). In order to overcomethis restriction the feedwater tank pressure has beenincreased load-dependently to a maximum of 2 barsmatching its design pressure. However, due to thereason mentioned above, for loads below 60% the heatthat would need to be ―dissipated‖ in the dryer (in orderto match the DH load) would result in such a high dryercondensate heat flow which again would bring thefeedwater beyond saturation state. Hence for thosecases the boiler load is gradually decreased, resulting inlower pyrolysis yields. The lowest DH load level that canbe 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 forstable combustion conditions in the boiler-, theintegrated process offers possibilities to increase theoperating hours of the CHP plant considerably.Fig. 2: CHP Plant with integrated pyrolysis and steam drying171

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAPPLICATION OF THE PRIMARY ENERGYCONCEPT ACCORDING TO EN 15603Primary Energy ConceptThe EU standard EN 15603 [2] handles the energyperformance of a building as a whole and givesguidelines how energy use and production of a buildingshall be calculated. In order to aggregate the differentforms of energy produced and used within the building,primary energy (PE) and CO 2 emissions areaccumulated and expressed by means of primaryenergy factors (PEF) and CO 2 emission coefficients,respectively. PE is energy that has not been subjectedto any conversion or transformation process [2]; it ishence not yet extracted from the source. In the PEapproach described in EN 15603, all energy carriersinvolved in the generation process are retraced to theirsources and all energy needed to deliver the finalenergy product are aggregated to the total PEconsumption and CO 2 emissions. Thus the PEapproach applies the holistic principles of life cycleassessment to an energy rating procedure. By retracingenergy consumption to the source, the systemboundaries automatically include the whole world, andthus depict the real impact of the system concerningenergy consumption and CO 2 emissions.In the power bonus method f El. is defined as the PEF ofthe electricity that is thought to be replaced by thepower generated in the CHP plant (for instance, in thisstudy the average power generation efficiency inFinland is used). This allocation pays attention to thefact that the co-generated electricity is more sustainabledue the CHP process‘ high overall efficiency. The PEFof the DHN can thus be determined according to;fDHifF , i EF P QDHfEl .As production of products other than electricity is notdefined in EN 15316-4-5 the power bonus method hasbeen extended by regarding the produced pyrolysisslurry as a ―bonus‖ as well. The PEF of the pyrolysisintegrated CHP plant is thus calculated as:fDH fF , i EF, i P fEl .iQDH EPyro fPyroPrimary Energy FactorThe total primary energy factor is the sum of all PE inputto the energy system divided by the useful energydelivered at the system border. It thus describes howmuch PE input is needed in order to obtain one unit ofenergy used and can hence be seen as an invertedefficiency.In standard EN 15316-4-5 [3] more detailed guidelinesfor the calculation of PEFs of DH systems are defined.According to EN 15316-4-5 PEFs can be calculated fora certain part of the energy system. In this study thesystem boundary comprises the power plants and theDHN.The PEF of the DHN has been calculated applying thepower bonus method. If yearly demand data of the DHNand the generation data are known, the PEF of the DHNcan be calculated by applying the so-called powerbonus method. The power-bonus method is derivedfrom the energy balance of the building which can bewritten as: fF , i EF fDHQDH PfEl .,iwhere E F , Q DH and P are the heat of the fuels used, DHand power co-generated respectively. f F,i , f DH and f El. arethe PEFs of the fuels used, the DHN and of the cogeneratedpower.172In this study PEFs as shown in table 2 have been used:Table 2: Primary energy factors and CO 2 emissioncoefficients for fuels and productsf BM2f Oil2f El.1f Pyro11.09 c CO2/BM21.35 c CO2/Oil23.11 c CO2/El.11.28 c CO2/Pyro1kg/MWh143302701 : value is calculated, 2 : value is taken from EN 15603, Annex EFuels assumed to be used are wood logs for the CHPplant and fuel oil for the heat-only/backup boiler(s) andtheir PEFs are taken form annex E of EN 15603. ThePEF of electricity production in Finland has beenderived from [10]. The PEF of pyrolysis slurry in astand-alone unit has been calculated assuming a fluegas dryer (which is considered as the drying technologymost likely to be applied) with an energy consumption of3300 kJ/kg water evaporated [9] and a heat of pyrolysisof 1.87 kJ/kg [7]. Although the standard asks for moredetailed analysis of the energy chain as e.g.consideration of transport, transmission and otherprocessing should be included, this has not beenimplemented into this study since those factors areassumed not to differ between integrated and separatedproduction of pyrolysis oil.14

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaCO 2 Emission CoefficientThe CO 2 rating is done by calculating CO 2 emissioncoefficients (c CO2 ) that quantify the total amount of fossilfuel derived CO 2 , emitted to the atmosphere, per unitdelivered energy. As for the primary energy factor thesystem boundary comprises of power plants and DHN.Also the power bonus method is applied for calculatingthe DHN‘s specific CO 2 emissions. For the sake ofcompleteness it must be mentioned that CO 2 -equivalentemissions of other greenhouse gases can optionally beincluded. However this has not been implemented intothis study, due to a lack of data. Similarly to the PEF theCO 2 emission coefficients for the base case arecalculated as:cCO2, DHiEF , icCO2, F , iQDHAnd for the modified plant as:cCO2, DHiEF , i cCO2, F , i P c P cQCO2, El .CO2, El .DH. EPyro cCO2, PyroE F,i , E Pyro , P and Q DH represent heat in fuels, heat inpyrolysis slurry and co-generated electricity and DHrespectively. Accordingly, c CO2,F,i , c CO2,Pyro , c CO2,El. andc CO2,DH are the related CO 2 emission coefficients. Thecorresponding values are given in table 2.RESULTSIn table 3, three simulation cases are presented: thebase case (case 1), pyrolysis integration with the sameoperation hours (case 2) and the maximum pyrolysisslurry production (case 3) with prolonged operation.hours and a DH load as low as 30% (matching 18% ofthe total DH load). It can be seen from the table that forall cases the DH output is the same for the 100-50%operating points. This results, in the first two cases, inan identical total DH output of 70.85 GWh. Thiscorresponds with 75% of the total yearly DH load. Dueto steam extracted to the dryer, the enthalpy flowthrough the turbine in part load is decreased, whichresults in a lower electricity production in part load forthe cases 2 and 3. Already for the second casepyrolysis slurry with an energy content in the samerange as the DH load can be produced. Fuel input,which is defined as wood burned in the boiler and woodentering the dryer for subsequent pyrolysis, increaseswith falling load for load levels 60% and higher. In thosecases the boiler combustion power is 100%, but it isdecreased for lower load levels as explained above. Ifoperation hours are extended by supplying lower DHloads with the CHP plant (case 3), total pyrolysis slurryproduction can be increased by approximately 55%,electricity production by 7.8% compared to the basecase. Further DH production is increased byapproximately 14.7%, covering now 86% of the total DHdemand. This directly decreases the fossil fuelledbackup power as shown in table 4. The needed backupheat is almost cut in half. Together with the additionallyproduced electricity this substantially improves theprimary energy factor to 0.68 which certainly will have apositive influence on the PEF of the buildings connectedto the DHN. For case 2 the improvement is marginal.The CO 2 emission coefficient changes somewhatcontroversially by increasing in the 2 nd case. This isbecause the loss in electricity bonus cannot becompensated by the produced pyrolysis slurry, since theCO 2 emission coefficients differ widely. However forcase 3 specific CO 2 emissions become even negative.The negative value is very unlikely to reach and can beexplained with the not fully accounted fuel productionchain. Nevertheless, it is obvious that the DHN‘s CO 2emission factors can be considerably reduced with thepresented integration concept.173

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 3: Results – Multiperiod ModelBase Case - Case 1CHP DH Load [%] 100 90 80 70 60 50 40 30 totalTime [h] 2440 530 530 530 530 530 - - 212 daysFuel Input [MW] 25.90 23.19 20.39 17.47 14.58 11.91 - - 109.56 GWhPower [MW] 6.29 5.64 4.91 4.06 3.22 2.54 - - 26.13 GWhDistrict Heat [MW] 16.50 14.85 13.20 11.55 9.90 8.25 - - 70.85 GWhCHP + Pyrolysis - Case 2CHP DH Load [%] 100 90 80 70 60 50 40 30 totalTime [h] 2440 530 530 530 530 530 - - 212 daysFuel Input [MW] 25.90 36.49 44.24 52.42 60.21 53.88 - - 194.13 GWhPower [MW] 6.29 5.54 4.69 3.71 2.88 2.25 - - 25.45 GWhDistrict Heat [MW] 16.50 14.85 13.20 11.55 9.90 8.25 - - 70.85 GWhPyrolysis Slurry [MW] 12.21 21.16 30.60 39.58 36.76 - - 74.31 GWhCHP + Pyrolysis - Prolonged Operation Hours - Case 3CHP DH Load [%] 100 90 80 70 60 50 40 30 totalTime [h] 2266 633 633 633 633 633 633 633 279 daysFuel Input [MW] 25.90 36.49 44.24 52.42 60.21 50.97 39.56 28.88 256.62 GWhPower [MW] 6.29 5.54 4.69 3.71 2.88 2.27 1.71 1.18 28.17 GWhDistrict Heat [MW] 16.50 14.85 13.20 11.55 9.90 8.25 6.60 4.95 81.26 GWhPyrolysis Slurry [MW] 0.00 12.21 21.16 30.60 39.58 33.79 26.11 19.00 115.46 GWhTable 4: Results - PEF and CO2 CoefficientBase CaseCase 1CHP + PyrolysisCase2CHP + Pyrolysis -Prolonged OperationCase 3Required Backup Power MWh 27.8 27.8 15.5Total PEF [-] 0.80 0.79 0.68CO2 Coefficient kg/MWh 38.6 42.1 -5.3CONCLUSION AND DISCUSSIONThe work shows that by integration of a CHP plant withwood pyrolysis operation hours can be increases by30%, a valuable product can be co-produced and PEEas well as the CO 2 emission coefficient of the DHN canbe substantially improved. As next steps morecomprehensive data of the fuel supply chain should beimplemented to get more realistic values that willapprove the trend shown with this work. The processcan be further improved by integrating heat that is setfree during the condensation of the pyrolysis liquid andgaseous product. The heat is available in atemperature range from approximately 500 °C to 25 °Cand could hence be used for steam superheating,feedwater preheating, but also for DH generation. Thisintegration is not a simple task since many plantparameters influence each other. The heat integrationmust be carried out together with a pinch analysis toassure an energy efficient integration.174Another open question is the influence of the realpyrolysis gas on the combustion temperature and fluegas properties. In order to gather more details of thepyrolysis process a simple pyrolysis model is currentlyunder development. Together with the power plantmodel the integration can be further optimised aimingfor highest PEE along with low CO 2 emissioncoefficients.Further an economic analysis should be carried out inorder to show potential economic benefits. Theintegration itself seems to be viable – a statement thatis supported by a press release from June 2009 whereboiler manufacturer Metso and forestry company UPMannounced the development of a new viable fastpyrolysis process benefitting from the integration with aCHP plant [11].Concerning the European standards used forevaluation, it can be said that the power bonus methodcan be easily adapted to a polygeneration concept

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniayielding heat, electricity and pyrolysis slurry. It verylikely can also be extended to other possible biorefineryproducts as long as those are ―energy‖ products. Thisexpansion option could be implemented into thestandard.However the most difficult question remains how thePEF of other, less common co-products should bedetermined. In the case of pyrolysis slurry it is notpossible to find good average production efficienciessince the technology is not yet on the market. But, howis co-generation of cooling evaluated?In general it can be said that the implementation of theprocess will be strongly dependant on investment costand on the market value of the product. The productvalue is currently difficult to predict and also its futureprice development will be strongly dependant on theuse of biomass in the future.Summarising it can be said that even though manyquestions still need to be answered, this works showsclearly that the integration of communal CHP plantswith wood pyrolysis is beneficial concerning theconnected DHN‘s PEF and CO 2 emission coefficient.Vice versa it also shows that CHP plants can play animportant role in the sustainable bio-refineries of thefuture.ACKNOWLEDGEMENTSThis work is part of the Primary Energy Efficiencyproject of Nordic Energy Research.The funding of the Graduate School of Energy Scienceand Technology (EST) is gratefully acknowledged.REFERENCES[1] Kohl, T., Järvinen, M., Fogelholm, C.J.,―Gasification and biorefinery in combined heat andpower plants‖, Proceedings of the 11th<strong>International</strong> <strong>Symposium</strong> on District Heating andCooling, Reykjavik, Iceland, 2008[2] EN 15603:2008, ―Energy performance of buildings.Overall energy use and definition of energyratings‖, European Committee for Standardization,CEN, Brussels.[3] EN 15316-4-5:2007, ―Heating systems in buildings.Method for calculation of system energyrequirements and system efficiencies. Part 4-5:Space heating systems, the performance andquality of district heating and large volumeSystems‖, European Committee forStandardization, CEN, Brussels, 2007[4] Savola, T., ―Modelling biomass-fuelled small-scaleCHP plants for process synthesis optimisation‖,Doctoral Dissertation, Helsinki University ofTechnology, Espoo 2007[5] Bridgwater, A.V., 2000, Fast pyrolysis processesfor biomass, Renewable and Sustainable EnergyRev., 4(1), pp. 1-73.[6] Henrich, E., 2007, ―The status of the FZK conceptof biomass gasification‖, 2nd European SummerSchool on Renewable Motor Fuels, Warsaw.[7] Daugaard, D., Brown, R., ―Enthalpy for Pyrolysisfor Several Types of Biomass‖, Energy & Fuels2003, 17, 934-939[8] http://www.ecn.nl/phyllis: PHYLLIS is a serviceprovided by the Energy Research Centre of theNetherlands – ECN, 17.9.2009 [selectedsubgroups: untreated wood birch andfir/pine/spruce].[9] Brammer, J., Bridgwater, A., ―Drying Technologiesfor an integrated gasification bio-energy plant‖,Renewable and Sustainable Energy Reviews 3(1999) 243-289[10] Dones, R. et al, 2004, Life Cycle Inventories ofEnergy Systems: Results for Current Systems inSwitzerland and other UCTE Countries, ecoinventreport No. 5, Paul Scherrer Institute Villigen, SwissCentre for Life Cycle Inventories, Dübendorf, CH,p.170.[11] N.N., press release onhttp://www.metso.com/news/newsdocuments.nsf/web3newsdoc/C89A8AC3F77ABD29C22575CF003111F5?OpenDocument&ch=ChMetsoWebEng[12] H. Lund, F. Hvelplund, I. Kass, E. Dukalskis, D.Blumberga, ―District heating and market economyin Latvia‖, Energy, 1999, Vol. 24, pp. 549-559.[13] H. C. Park, M. Chung, S. H. Kim, ―Development ofsystem simulator for community energy system‖,Report to Ministry of Industry, 2003.[14] Y. H. Im, H. C. Park, M. Chung, ―A study of optimalheating supply systems for the newly developingarea in the vicinity of DHC system supplying area‖,Report to Korea District Heating Corporation, 2006[15] Y. H. Im, M. Chung, H. C. Park, ―Feasibility studyfor small size cogeneration systems in themetropolitan areas of Seoul‖, Final Report to SH(Seoul Housing) Corporation, 2008.[16] M. Chung, H. C. Park, ―Development of a energydemand estimator for community energy systems‖,Journal of the Korean Solar Energy Society, 2009,Vol 29, pp. 37-44.175

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia[17] M. Chung, H. C. Park, ―Development of a softwarepackage for community energy system assessment– Part I: Building a load estimator‖, Energy, inpress.[18] H. C. Park, S. S. Lee, D. J. Kim, ―Development ofenergy models for department stores‖, KoreanJournal of Air-Conditioning and RefrigerationEngineering, 2003, Vol. 15, pp. 1088-94.[20] H. C. Park, ―Development of weighting factors forvariables associated with hourly energyconsumption pattern for hotels in Korea‖, SAREK(Soc. Air-conditioning, Ref., Engineers of Korea)Winter Annual meeting, 2002, pp. 76-82[21] H. C. Park, ―Analysis of energy loads for hospitalbuildings‖, SAREK journal, 2002, pp. 1088-93.[19] H. C. Park, M. Chung, ―Building load models forhotels in Korea‖, Journal of the Korean SolarEnergy Society, 2009, Vol. 29, pp. 48-57.176

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaCHP OR POWER STATION? – QUESTION FOR LATVIAD. Blumberga 1 , G. Kuplais 1 , F. Romagnoli 1 and E. Vigants 11 Riga Technical University – Institute of Energy Systems and EnvironmentKronvalda boulv. 1, LV-1010 – Riga, LatviaABSTRACTThis paper presents aspects and problems of theLatvian energy-system connected to the choice of theCHP and/or power stations for the future nationalenergy strategies. In the light of the last EU directive inthe subjects of Renewable Energy Sources (RES) theshare of electricity produced from RES at the momentis attested on the value of 42.4 % but should beincrease to 49.3%. In the same time the share ofrenewable energy resources in the final energyconsumption for 2020 should reach the level of 40%from 30%.Dependence on imported energy sources, growth ofelectricity prices, the need to support local producersare the main reasons for the use of new renewableenergy technologies in the Latvian energy sector toimplemented in refurbished energy supply system.Several methods fro the evaluation of the best strategyare explained.This apaper summarizes the application of the EnergyIndicators for Sustainable Development (EISD) as goodtool for analyzing trends, setting energy policy goalsand monitoring progress. The results from theapplication of a multi-objective optimization regardingthe implementation of the landfill biogas in the biogastreatement plant ―Daibe‖ are reported.1. INTRODUCTIONThe structure of energy user in Latvia is characterizedby high energy consumption in households, public andservice sectors, comparing with relatively lowconsumption in rural and industrial sector. In light ofthis situation, for the power sector development,special tasks are required in connection to the choiceof the more adequate energy resources in order toensure the best energy production and supply.Consequently question on which direction address themain efforts for the energetic national improvement isstill actual: CHP or Power station?If fuel, which is used to produce heat and electricalenergy in Latvia, is taken into account, the dominantone is gas [1] and consequently appear evident howthe Latvian dependence on foreigner energy supplies(mainly from Russian) is not only a weak point inconnection to the energetic sustainability but can serveas a convenient way of exerting economic pressure [2].In CHP station this dominance is almost total and veryhigh in district heat supply boiler houses. As it wellknown Latvia is a great consumer of imported fossilfrom one side but in the same the share of renewableenergy resources is one of the highest of Europe.The use of specific energy resource depends onenergy supply policy, and total consumption of energyresources depends on development of every type ofenergy resources in regions. Now there is unjustifiedhigh proportion of fossil fuel in state energy balancewhich is possible to reduce by a beginning of activeuse of local fuel in regions. The EU directive alsorequires that Latvia in the year 2015 would generate49% of electric power from renewable resources(currently it is 45%) [3]. This is supportable, but thepower supply of Latvia cannot be let out of the sightand this issue is already problematic.Latvia has some electricity production from cogenerationplants and some from hydro-power plants.However, the production of electricity from the hydropowerplants fluctuates a great deal from year-to-year.The rest of the electricity for consumption is importedfrom the neighbouring countries.In order to understand the role played by CHP andpower plant it is fundamental to understand the actualsituation in Latvia for thermal energy where more thana half of Latvia district thermal energy is distributed andconsumed mainly in Riga.Latvian heating primarily is performed on a centralizedbasis and after the used of wood energy the naturalgas imported from Russia is the main source (seeFig. 1).Fig. 1. Main resources used for local and individual heatsupply [3]177

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThis type of heating supply scheme means thatconsumers are grouped and heating is performed fromheat source which is intended for the consumer group.About 70 % out of this thermal energy volume isproduced in the cogeneration cycle (only in Latvenergoowned CHP, and Rigas siltums) and around 30% ofcentrally supplied heat energy is produced in Riga CHPplants and boiler houses. Of course as main fuel inRiga natural gas is used approximately 98% of thermalenergy is produced from natural gas (CHP plants andboiler houses together) [3].As for heat supply outside of Riga, the dominantthermal energy is produced in boiler houses withrelatively high proportion of local fuel usage (as shownin Fig.1). Outside of Riga CHP heat production ratedoes not exceed 5%.2. EXISTING ENERGY SITUATION IN LATVIA:SHORT OVERVIEWDuring the recent past central (large) power plants inLatvia supplied roughly 65% of the total annual powerdemand - distributed energy resources (DERs) covered3–6%, but the rest were received as import suppliesfrom Estonia, Lithuania and Russia (mainly) [4].Regarding fuel sources Latvia has no real fossil-fuels ofits own and the consumption must be imported.However Latvia uses the domestic renewable-energyresources hydro-power and biomass.Table I. – primary energy-consumption in Latvia in theyear 2007 [1]PJ %Natural gas 56.92 27.8biogas 0.32 0.16Biodiesel 0.07 0.03Oil products 73.33 35.8Fuelwood 48.47 23.7Hydroenergy 9.84 4.8Import of Electricity 10.80 5.3of the gross energy-consumption from renewableenergysources in Latvia in the year 2007.The most important domestic renewable-energyresource in Latvia is biomass in the form of fuelwood:in fact approximately 45% of Latvia is covered withwoods and this substantial area makes wood asignificant potential as a resource for energy supplies.Even though the share of renewable is one of the mostlarge Europe the EU directive fixes the target of 40%share of renewable energy resources in the finalconsumption in 2020.This means that the increase is not feasible without theneed of refurbishment and/or construction of energeticinfrastructures.The fact that Latvia has domestic renewable-energyresources makes it interesting because the utilisation ofthe domestic fuels would be sustainable both from anenvironmental and an economic point of view.Latvia has comparatively well developed power, naturalgas supply and district heating systems, and as aconsequence the electricity is basically produced byhydro power plants and by cogeneration plants, whichare operated according to district heating demand, andpart of electricity is imported (fig. 2). Consequently themain objectives of the Latvian energy policy now are toensure sustainable accessibility to necessary energyresources and security of supply in order to favouritethe economic growth and improve quality of life, toensure environmental quality retention and meet theobjectives set in the Kyoto protocol of UN FCCC andLatvian Climate Change Program.Electricity amount, billion kWh87654321Electricity Supply in Latvia02000 2001 2002 2003 2004 2005Yearimported electicitywind generatorssmall HPSsmall CHPCHPHPPImport Coal and coke 4.36 2.1Wind 0.19 0.09Biodiesel 0.07 0.03Total 204.6The use of primary energy for the gross energyconsumptionin Latvia can be seen in Table 1. Theshare of renewable energy in the gross energyconsumptionis made up of fuelwood and hydro energy.That means that there is approximately a total of 30%178Fig. 2. Electricity supply in Latvia (Source: state JSCLatvenergo, Ministry of Economics, Central StatisticalBureau)2.1 Lack of energy sources for electricityThe main domestic electricity capacity consists of 1517MW of hydro and 520 MW [5] of thermal (CHP units inRiga) all of which is controlled by the state company,Latvenergo. The generating potential mainly consists ofthree hydro power plants (HPP) on the Daugava River,hence directly dependent on the river‘s water flow. Dueto small reservoirs, utilization rates are low and the

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaproduction is quite seasonal following the water flows.The amount of power produced by the Daugava riverHPP cascade is average 2.6–2.8 TWh [1] annually,reaching in the years, rich by spring floods and raineven 4.5 TWh [ [5].equipment with this capacity heat production bycogeneration would be a maximum.More in detail the three HPPs, located on the river ofDaugava, form a sort of cascade with the relativecapacity of: Plavinas 870 MW, Kegums 263 MW andRiga 402 MW.Almost two thirds of hydro electricity is produced in thespring month of March, April and May. In this period thesupplies are from the hydro plants. In the high demandwinter season amount of electricity generated by hydroplants is relatively low.Looking the electricity supply statistics [1] the nationalproduction of electricity is around 10.0 PJ where the9.8 are produced using hydro energy and 0.2 PJproduced by wind energy. The net electricity import(including the amount of energy exported) is around10.8 PJ approximately the 50% of the national supply.These figures shows the lacks of energy sources in thenational system and seems reasonable to foreseen amore large fraction of other energy sources for theproduction of electricity, the main question is on whichmethodology base this strategy .2.2 Well organized and developed DH systemLatvian heating primarily is performed on a centralizedbasis consequently consumers are grouped and theheat is supply from heat source which is established fora certain consumer group. The heat source power,depending on type of consumer group, varies from therange of kW to several hundred of MW. In generallower power can correspond to building groups,individual houses or even apartments heating.Residential and separate heating of individual housesbelongs mainly from the decentralized heating. One ofthe benefits of district heating is centralization of heatload, which gives a possibility to increase the heatsource power and to form basis for the development ofcogeneration power. For large heat consumers inLatvia (mainly heating systems in large cities like Riga)large cogeneration plants are installed. The customerswho are not connected to a district heating cannot beprovided from this system. In the other regions far fromthe big cities the heat supply system is mainly basedon district heating, consequently it means that thatthere possibility for a CHP development.CHP plants cover only a part of the total heat load. Therest of the load is covered by the peak load boilers.This means that following the total heat capacity of thesource, the potential heat capacity of cogenerationshould be assessed quantitatively. Heat capacity ofcogeneration plant has to be selected so that operating179Fig. 3. Heating energy distribution by cities in LatviaIf we are looking at the district heating division of Latviaa huge difference can be seen in quantity of heatsupply in Riga and the rest of Latvia (see fig. 3)Two large CHP plants, Riga TPP-1 with an installedelectric capacity of 144 MW and Riga TPP-2 (390 MW),are located in Riga [5]. CHP plants are the main heatgeneratingsources of heating networks of Latviancapital. Power is produced mainly in cogenerationmode, according to the heat–load curve.During the heating season, when there is a substantialdemand for heating and hot water, Riga CHP plantsproduce approximately 80% of the total annualproduction volume, while during summer the volume ofproduction reduces [5].Nowadays Riga CHP plants cover about 20% of thetotal annual power demand of Latvia [5] .The main fuel used in Latvia biggest cities is naturalgas and the rates of thermal energy are 75% - 85% [3].In Riga and other cities where most part of the heat isproduced in cogeneration cycle, the increase of rateswas not so high and currently (in the autumn of 2009)heat rates are lower that in the cities where wood chipsare used.From the thermal energy point of view seventy percentof the heat in Latvia is supplied from district-heatingsystems either from boiler houses or co-generation:37% of the district heating in Latvia was produced bymeans of co-generation plants [6]. This means that63% of the district heating is produced in boiler houses[6]. This means that there is potential to replace someof the heat plants with co-generation units (Eightypercent of the district heating in Denmark is suppliedfrom CHP [6]).As for heat supply outside of Riga, the dominantthermal energy is produced in boiler houses with

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniarelatively high proportion of local fuel usage. Outside ofRiga CHP heat production rate does not exceed 5% [3](combined heat and power plant up to 4 MW of poweroperating in Bauska, Valmiera, Ogre, Vangazi,Daugavpils, Jelgava, Dobele, Grobiņa, Saldus,Ventspils, Ozolnieki, Ādaži, Lielvārde and Cesis).3. METHOD FOR EVALUATIONIn connection to achieving sustainable development onglobal scale the correct and judicious use of resources,technology, appropriate economic incentives andstrategic planning at the local and national levels isrequired. Therefore, choosing energy fuels andassociated technologies for the production, deliveryand use of energy services, it is essential to take intoaccount economic, social and environmentalconsequences. The research on criteria and/orindicators in order to understand the best energeticchoice for Latvia is the first step for a correct energyplanning.gas), constructive parameters of cogeneration plant,parameters of heat energy consumers, heat loadduration curve, duration of heat energy consumptionlevels, behaviour of energy end users, installedcapacity, energy efficiency of technologies,development of demand side management factor, andother factors.3.2. Methodologies: EISD method and MOO methodIn the following paragraph the algorithm of ISED coreset tool, included in the conceptual framework used byUnited Nations Commission on sustainabledevelopment (CED), is shown. After is also shortlyreported the MOO methodologyThe EISD is an analytical tool developed which canhelp energy decision and policymakers at all levels toincorporate the concept of sustainable developmentinto energy policy. EISD core set is organized followingthe conceptual framework used by United NationsCommission on sustainable development (CSD).There are several methodologies that can be chosen toidentify the most suitable indicators, and in the sametime the choice is related and strictly connected onwhat the planning and consequently analysis is basedon.3.1 Criteria and indicatorsThe methodologies can be chosen using severalmethodological tools and approach such us: multicriteriaor multi-objective optimization (MOO) [7],energy indicators for sustainable development (EISD)[8], Life Cycle assessment (LCA) [9, 10]. Each of thesemethodology start from different point of views andbases: MOO methodology is connected to bestoptimization choice of a certain number of variablesthat optimize certain objectives, EISD methodologyaims to evaluate (and consequently increase) theconcept of sustainability based on social, economicaland environmental indicators, LCA aims to figure outthe global environmental load of a process and/orproduct taking into account the entire outflows andinflows connected (in terms of energy, substances andemissions), in this last case the indicators changedepending on type of Life cycle assessment methodschoosen.A summary of the factors that can influence CHPdevelopment in Latvia has been proposed in previouspapers. A. Volkova et al. [2] identify four main factors:political, geographical-climatological, legislative andtechnological.In general the total amount of electricity produced in acogeneration regime and condensing mode dependson constructive solutions (e.g. technical solution for thebiogas‘ collectors), availability of source used (mainlyFig. 4. set of core EISD [8]There are 30 indicators, classified into threedimensions (social, economic and environmental) andgrouped in 7 big themes. There are four socialdimension indicators: three of them represent equity(accessibility, affordability, disparities) and one healththeme (safety). The set of energy indicators ofeconomic dimension consists of 16 indicators. Thereare nine environmental dimension indicators in theEISD core list. The scheme of core EISD indicators ispresented in Fig. 4. The priority areas for energy sectoranalysis in Latvia can be were selected based on themain EU energy policy directions. These priority areasare as follows:Energy use.Energy intensities.End-use intensities of economic branches.Energy security.Environmental energy impacts.The next Fig. 5 shows the linkages among theindicators selected for energy policy analysis in BalticStates. Relevant policy actions based on analysis180

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaconducted in the previous sections are defined basedon targeted indicators.of produced biogas. The main barriers for improvedbiogas injection are the high costs of improvement andgrid connection. Grid injection is limited by location ofbiogas production and improvement sites, which haveto be close to natural gas grid [12].Problems are connected with biogas utilization incogeneration plants (CHP) since there are nopossibilities to find heat energy consumers, which inturn resulted with low efficiency landfill power plantsalmost all over Latvia.Due to high electricity feed-in tariff there is aneconomical motivation for power plant operation withlow efficiency. For electricity produced in renewableenergy power plants with nominal capacity of up to4MW high feed in tariff has been transposed in Latvia‘slegislative acts.Fig. 5. Linkages between indicators and relevant policyactions based on the targeted indicators [8]Multi-objective optimization (MOO), also known asmulti-criteria optimization, particularly outsideengineering, refers to finding values of decisionvariables which correspond to and provide the optimumof more than one objective. Unlike in single objectiveoptimization (SOO), which gives a unique solution,there will be many optimal solutions for a multiobjectiveproblem. Multi-objective optimization involvesspecial methods for considering more than oneobjective and analyzing the results obtained [7].Often, the various objective functions conflict with eachother (i.e., optimizing one of them usually tends tomove another towards undesirable values), for solvingsuch models one needs to know how many units of onefunction can be sacrificed to gain one unit of another,but this trade-off information is not available. In otherwords, one is forced to determine the best compromisethat can be achieved.In the following paragraph an example of MOO appliedto the evaluation of possibilities to utilize landfill biogasfor electricity production in one of Latvia‘s landfills.4. TESTING OF LANDFILL GAS PRODUCTIONThe improved biogas is one of the cleanest fuels with alittle impact on the environment and human health [11].One of the advantages of biogas injection into naturalgas grid is the fact that natural gas grid connects aplace of biogas production (usually in rural areas) withdensely populated areas. It allows new consumers touse gas. In this way it is possible to increase the biogasproduction in remote areas not being worried about useThe development of Latvia‘s landfill sites is at thecrossroads. On one hand it is economically feasible tooperate CHP just for electricity production, but on theother – it is important to use natural resources on fullvalue by producing from biogas the maximum amountof heat energy. In first case it means that there is noneed for waste sorting in landfills, but in the other it isimportant to sort both – before waste collection and inlandfills.Utilization of landfill biogas in Latvia is based on energyproduction in power station placed close to landfill fordifferent reasons. One of the most important reasons isfinancial state support of small scale power stations (4MWe) from renewable energy resources. Such kind ofsupport prevents both, development of waste sortingand utilization of refuse derived fuel in cementproduction, and biogas improvement to cover needstransportation sector or to connect to natural gas grid.In the following is shortly reported the methodologyregarding the optimization model of biogas use inlandfills in Latvia in connection to the data collectedfrom landfill ―Daibe‖. After the analysis only two of theindependent parameters have been chosen: quality ofbiogas (characterized by heat value), and technologicalequipment (characterized by electrical capacity).This optimization model for biogas utilization in landfillsincludes four modules and is based on technological,climate and economical sub models.Results of economical optimization show that in case oflow biogas quality (4 kWh/m 3 ) the optimal installedcapacity is 2.2MW. In case of biogas quality of 5kWh/m 3 , optimal installed capacity is 2.8MW, and3.4 MW – in case of high biogas quality (6 kWh/m3).181

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia5. COMMENT AND DISCUSSIONThe use of CHP instead of conventional plant willalways improve energy efficiency and will reduce CO 2emissions significantly, in Latvia there is potential toreplace some of the heat plants with co-generationunits (comparing with Denmark where approximately80% of the district heating in Denmark is supplied fromCHP [6].Fig. 6. Diagram of the economical optimization[12]Results of technological optimization show that, thehigher the installed capacity, the shorter the operationtime of equipment. If assumed that operation time ofthe equipment could be 5 up to 10 years, then theinstalled capacity can be 0,5MW and higher.Hence promotion of high-efficiency cogeneration (CHP)based on a useful heat demand is a priority with regardto saving primary energy, avoiding network losses andreducing emissions, in particular of greenhouse gases[2].Of course the choice of the fuel is fundamental in orderto reach the target required from the last EU directive interms Renewable Energy Sources (RES).More use of energy from biomass in terms of woodfuel,biogas, landfilled gas and biofuels seems to be a gooddirection in order to displace the part of energy sourcesgiven by the imported natural gas.Fig. 7. Equipment operation time vs installed capacity [12]Results of the climate sub model show that the higheris installed capacity, the greater the reduction ofgreenhouse gas emissions. Besides that, it is notpossible to reach extremis by using two objectivefunctions (heat value of biogas and installed capacity),which have been used in case of economical andtechnological sub models, and it is necessary tointroduce another objective functions.The use of wood in the energy sector (through theproduction of heating and electricity) must become notonly an objective for the development of the energysupply system, but it must also become part ofstrategies for economic development and for theimprovement of the import/export balance of thecountry. These measures can succeed in not onlydeveloping local production and job creation, but if canalso stimulate and increase the potential export.It is particularly important to conduct engineer-technicaland economic analysis of the various technologicalsolutions possible to implement wood use in thecogeneration plants of the larger cities (including RigaTEC 1 and TEC 2) [13]. Any possible choice and/orscenario cannot be complete if it not references to aLife cycle assessment (LCA) that it a good tool in orderto understand the environmental load of a certainprocess strategy and in order to give a comparablecommon base.Based on the targeted indicators for Latvia the beststrategy can be identify in:Fig. 8. Diagram of the environmental optimization [12]Model of power production in landfill shows that feed-intariff stated as financial support today in Latvia allowsto reach economically feasible projects even in case ifcogeneration unit is operated in power station regime(generates only electricity). Results show that statepolicy needs corrections to improve energy efficiency ofbiogas utilization for energy production.182– Enhance the diversity and variety of the energymix.– Improve maintenance of existing energyinfrastructure.– Eliminate constraints and investment in newfacilities.– Increase the efficiency of energy supply inelectricity generation.– Increase the share of electricity produced bycombined heat and power (CHP) plants.– Increase the share of renewable and domesticenergy sources in the energy mix.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia6. CONCLUSIONS1. In the paper has been presented aspects andproblems of the Latvian energy-system connected tothe choice of the CHP and/or power stations for thefuture national energy strategies in the light of the lastEU directive in the subjects of RES. The dependenceon imported energy sources, the growth of electricityprices, and the need to support to local producers arethe main reasons for use of new renewable energytechnologies in the Latvian energy sector.2. In this paper has been summarized the results fromthe application of the Energy Indicators for SustainableDevelopment (EISD), a good tool for analyzing trends,setting energy policy goals and monitoring progress inorder to indentify good policy indicators. Also a testingof landfill gas using MOO method has been reportedwhere only two of the independent parameters havebeen chosen: quality of biogas (characterized by heatvalue), and technological equipment (characterized byelectrical capacity). Model of power production inlandfill shows that feed-in tariff stated as financialsupport today in Latvia allows to reach economicallyfeasible projects even in case if cogeneration unit isoperated in power station regime (generates onlyelectricity), but if is feasible from an economical pointview is not the same if reference to environmentalimpact.3. In the paper has been discussed how LCA can be agood approach that enables the energy requirements,GHG balance and other environmental impacts ofbioenergy production chains to be accounted andaccurately compared. Hence LCA is good tool in orderto give the possibility to compare different RES usagestrategies.4. Due to high electricity feed in there is an economicalmotivation for power plant operation with low efficiency.For electricity produced in renewable energy powerplants with nominal capacity of up to 4MW high feed intariff has been transposed in Latvia‘s legislative acts.OF course this is not good from environmental point ofview.5. The use of CHP instead of conventional plant willalways improve energy efficiency and will reduce CO2emissions significantly, in Latvia there is potential toreplace some of the heat plants with co-generationunits.6. Only crucial measures such as the reconstruction ofenergy sources in the larger cities (including RigaTEC 1 and Riga TEC 2) adjusting the use of fossil fuelsto biomass and conversion to non-natural gas sources,will produce results. Biogas and landfill gas favorite theenvironmental impact displacing usage of natural gas,the possibility of the feasibility solution for connectedCHP in out-of-city region to heat consumer must beevaluated.7. REFERENCES[1] Construction, Energy and Housing State AgencyEnergy Department, Latvian energy in figures,Riga, 2008.[2] A. Volkova, E.Latõšev, A. Siirde, Small-scale CHPpotential in Latvia and Estonia, Scientific Journal ofRTU Environmental and climate technologies, Ser.13, n. 2, Riga, 2009.[3] Latvia‘s district heating association , Heat supply inLatvia,http://www.lsua.lv/en/index.php?option=com_content&task=view&id=4&Itemid=5.[4] D. Streimikiene, I. Roos, J. Rekis, External cost ofelectricity generation in Baltic States, Renewableand Sustainable Energy Reviews n. 13, 2009, pp.863–870.[5] D. Streimikiene, I. Roos, J. Rekis, External cost ofelectricity generation in Baltic States, Renewableand Sustainable Energy Reviews n. 13, 2009, pp.863–870.[6] L.H. Rasmussen, A sustainable energy-system inLatvia, Applied Energy n. 76, 2003, pp. 1–8.[7] G.P. Rangaiah, Multi-Objective Optimization:Techniques and Applications in ChemicalEngineering, World Scientific, 2008, p. 454.[8] D. Streimikiene, R. Ciegis, D. Grundey, Energyindicators for sustainable development in BalticStates, Renewable and Sustainable EnergyReviews, 2007, Vol. 11, pp. 877–893.[9] G. Rebitzera et al., Life cycle assessment - Part 1:Framework, goal and scope definition, inventoryanalysis, and applications, Environment<strong>International</strong> n. 30, 2004, pp. 701– 720.[10] D.W. Pennington et al., Life cycle assessmentPart 2: Current impact assessment practice,Environment <strong>International</strong> n. 30, 2004, pp. 721–739.[11] D.Blumberga, Ģ. Kuplais, I. Veidenbergs, E.Dace,The benchmarking method for an evaluation ofbiogas improvement methods, Scientific Journal ofRTU 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 inLatvia, 7th <strong>International</strong> conference ORBIT2010:Organic resources in the carbon economy, June29-July 3, 2010, Heraklion, Greece.[13] A. Blumberga et al., Assessment on the use ofrenewable energy resources in Latvia until 2020:report, LVAF, December 2008, Riga.183

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaLCA OF COMBINED HEAT AND POWER PRODUCTION AT HELLISHEIÐIGEOTHERMAL POWER PLANT WITH FOCUS ON PRIMARY ENERGY EFFICIENCYMarta Ros Karlsdottir, Olafur Petur Palsson, Halldor PalssonUniversity of Iceland, Faculty of Industrial Engineering, Mechanical Engineering and Computer Sciencemrk1@hi.isABSTRACTThe aim of the study is to calculate primary energyfactors, f p , stating the primary energy efficiency as wellas factors for CO 2 emission, K, for geothermalcombined heat and power production at the HellisheidiCHP plant in South-West Iceland. These factors statehow much primary energy consumption and CO 2emissions result from the production of 1 MWh of heatand electricity due to geothermal utilization. Methods oflife cycle assessment (LCA) are used to calculate thesefactors by taking into account all energy and materialstreams to and from the CHP plant during constructionand operation. The results show that producing heatand electricity in a combined heat and power plantminimizes the primary energy factor for the electricitygeneration and produces a relatively low primaryenergy factor and CO2 production factor for the heatgeneration process. From the results, it can also beseen that life cycle assessment is a useful method toevaluate the total impacts of the geothermal energyconversion process, especially for the emission ofgreenhouse gasses during the lifetime of theproduction facilities. The experience in this study alsodemonstrates that the method can equally be used forprocesses as it is commonly used for the analysis oftotal impact of products.INTRODUCTIONThe calculation of primary energy and CO 2 productionfactors for geothermal power production has had littleattention while factors for some other types of energytechnologies such as hydropower, nuclear and coalfired power plants have been developed during therecent years. The importance of these factors is statedmainly in the new recast of Directive 2002/91/EC of theEuropean Parliament and of the Council on the energyperformance of buildings [1]. There it is stated thatbefore the end of year 2010, all new building occupiedby public authority should be issued energyperformance certificates showing these factors, basedon the energy mix used by the building and thebuildings‘ energy performance.At present time, geothermal power plants are situatedin 24 countries [2] and a total of 78 countries havereported direct use of geothermal energy [3]. Withincreasing fossil fuel prices and focus on renewableenergy sources, these power plants producing ―green184energy‖ become more viable in various locationsaround the world. It is thus important to investigatetheir primary energy efficiency and environmentalimpact for comparison with other energy conversiontechnologies. These energy performance indicatorscan be used to help decision making of futuredevelopments, policy making and energy rating ofbuildings.Countries that have access to geothermal areas andproduce power by geothermal utilization within theEuropean Union (EU) are: Austria, France, Germany,Greece, Hungary, Italy, Netherlands, Portugal,Romania, Slovakia and Spain. Other Europeancountries such as Iceland and Turkey, which are notcurrent member states of the EU, also utilizegeothermal energy extensively [2]. Also, 32 Europeancountries use geothermal energy directly for variouspurposes such as district heating [3]. Thus, electricityand heat based on geothermal energy are a part ofEurope‘s energy mix. For countries using geothermalbased power and/or heat and complying to EUlegislation, it is therefore important to have easy accessto standardized factors accounting for the primaryenergy efficiency and CO 2 emissions from geothermalbased heat and power.The aim of this study is to produce standardized factorsfor primary energy efficiency (f p ) and CO 2 emission (K)for geothermal heat and power production.ENERGY PERFORMANCE INDICATORS FORPRIMARY ENERGY CONSUMPTION AND CO 2EMISSIONSThe primary energy factor is defined as the ratiobetween the total primary energy inputs involvingenergy production to the actual energy delivered to theconsumer. According to [4], it should always accountfor the extraction of the energy carrier and its transportto the utilization site, as well as for processing, storage,generation, transmission, distribution and delivery.There are two primary energy factors defined: Total primary energy factor, accounting forprimary energy use of both renewable energysources and non-renewable sources. Non-renewable primary energy factor,accounting only for the primary energyconsumption of non-renewable energy sources.This factor is used when expressing only the use of

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniafossil or other non-renewable or polluting energysources in the energy conversion system.The CO 2 production coefficient, K, shall include all CO 2 -emissions associated with the primary energy used.Furthermore, equivalent emissions of other greenhousegases, e.g., methane, may be included [4].According to Directive 2002/91/EC, indicators on theenergy performance of buildings shall include theconsumption of primary energy and the CO 2 emissionsresulting from the buildings energy usage. Factors forprimary energy consumption and CO 2 emissions havebeen calculated for various energy chains producingelectricity, and values for these factors are given inAnnex E of the standard EN15603 on the energyperformance of buildings. An overview of these factorsis given in Table 1.Table 1: Energy performance indicators for varioussources of electricity [4]Source ofelectricityHydraulicpowerNuclearpowerPrimary energy factors f p[MWh primary energy /MWh delivered energy]Non-RenewableTotal0.50 1.10 72.80 2.80 16CO 2productioncoeff. K[Kg/MWh]Coal power 4.05 4.05 1340ElectricitymixUCPTE3.14 3.31 617As seen in the standard EN15603:2008 [4] andTable 1, no indicators are given for geothermal power.The directive is under reconstruction and a recast hasbeen released, as mentioned before. Also, the tabledoes not give factors for sources of thermal energyused by buildings for space heating. Thus, there isclearly a need to calculate these factors for energychains that involve geothermal energy, since theyproduce both electricity and heat which is delivered tobuildings within the European Union and in countriesfollowing EU legislation.GEOTHERMAL HEAT AND POWER PRODUCTIONAT HELLISHEIDI CHP PLANTHellisheidi geothermal CHP plant is situated at theHengill geothermal area close to Reykjavik, the capitalof Iceland. A 90 MW electricity production started in2006 after several years of construction and research.In 2007, a low pressure turbine was added, increasingthe power generation to 120 MW. A year later, another90 MW were added, resulting in a power generationcapacity of about 210 MW (213 MW in February 2009).Further developments of the power plant includeadding heat production in 2010 for district heating andalso increasing the power production if possible.Estimated production capacity for the completedHellisheidi Plant is 300 MW electricity and 400 MWthermal energy [5].The plant today is a double flash power plant with highandlow-pressure turbines and separators as seen inFigure 1. The heat production facilities are currentlyunder construction with a planned 133 MW thermalcapacity at the end of year 2010. The technicalcomplexity is moderate and the plant makes a goodbasis for a LCA study to evaluate the primary energyefficiency and CO 2 emission of this type of geothermalpower plant. Since it is fairly newly constructed, accessto detailed background data for the inventory modellingis possible, making the study more reliable andaccurate. Environmental assessment for theproduction is available as well as measurements ofvarious environmental impacts of the power plant,providing data for the impact assessment of the LCAstudy.In this study, a steady production of 213,6 MWelectricity and 121 MW heat is used as a basis for theLCA model. The reason for this choice is that thenewest inventory data on the construction phase andmass extraction are built on these productioncapacities, and that the base thermal load is estimatedto be 121 MW and not the full capacity of 133 MW.PRIMARY ENERGY OF VARIOUS ENERGYSOURCESThere is a matter of inconsistency in primary energycalculations of various energy sources as manydifferent methods are in use and accepted by differentenergy authorities [6]. As an example, the primaryenergy factors for power produced from renewableenergy sources such as hydro power, wind energy andsolar energy are sometimes calculated by assumingthat the primary energy factor for the energyconversion system is one, which is the same asassuming that the energy conversion process is 100%efficient. The reason for this assumption is that theprimary energy is defined as the first usable stage ofthe energy flow, which in the case of wind, solar andhydro is the electricity itself produced from theseprimary sources [7]. For electricity production fromheat sources, the first usable stage of the energystream is defined as the steam input into the turbine,according to an energy statistics manual from the<strong>International</strong> Energy Agency (IEA) [8]. The methods185

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaWHot watertankTVHPSLPSHPTHPCLPTLPCGGCTCPHX1IWCold watertankHX2Figure 1 – A simple schematic of the Hellisheidi geothermal CHP plantW: Geothermal production wellG: GeneratorHPS:HPT:HPC:LPS:LPT:LPC:CT:CP:HX1:HX2:High pressure steam separatorHigh pressure steam turbineCondenser for high pressureturbineLow pressure steam separatorLow pressure steam turbineCondenser for low pressureturbineCooling towerCooling water pumpHeat exchanger 1 for DH systemHeat exchanger 2 for DH systemIW: Reinjection wellused to calculate primary energy demand of powerproduction from renewable energy sources tends tounderestimate the primary energy input from theoriginal energy sources into the energy conversionsystem compared to the assumptions made for theheat conversion processes such as coal, oil and alsogeothermal.PRIMARY ENERGY OF VARIOUS ENERGYSOURCESThere is a matter of inconsistency in primary energycalculations of various energy sources as manydifferent methods are in use and accepted by differentenergy authorities [6]. As an example, the primaryenergy factors for power produced from renewableenergy sources such as hydro power, wind energy andsolar energy are sometimes calculated by assumingthat the primary energy factor for the energyconversion system is one, which is the same asassuming that the energy conversion process is 100%efficient. The reason for this assumption is that theprimary energy is defined as the first usable stage ofthe energy flow, which in the case of wind, solar andhydro is the electricity itself produced from theseprimary sources [7]. For electricity production fromheat sources, the first usable stage of the energystream is defined as the steam input into the turbine,according to an energy statistics manual from the<strong>International</strong> Energy Agency (IEA) [8]. The methodsused to calculate primary energy demand of powerproduction from renewable energy sources tends tounderestimate the primary energy input from theoriginal energy sources into the energy conversion186system compared to the assumptions made for theheat conversion processes such as coal, oil and alsogeothermal.Definition of Primary Energy of Geothermal FluidThere is no clear definition of primary energy fromgeothermal energy sources. Published methods ofdetermining the primary energy consumption ingeothermal power plants are the following [6]:Working Group III (WG III) of theIntergovernmental Panel on Climate Change(IPPC) records electricity from geothermal on a 1:1basis. This results in a f p factor of 1.The Engineering Information Administration (EIA)uses a factor of 6.16 units of primary geothermalenergy for each unit of geothermal electricity.<strong>International</strong> Energy Agency (IEA) records a f pvalue of 10 by assuming 10% conversion efficiencyof geothermal power plants.In this LCA study, where the main goal is to calculatean accurate f p factor for a specific conversiontechnology, the main issue is the primary energycontent of the geothermal fluid extracted from theproduction wells. The primary energy content of thegeothermal fluid can be based on differentassumptions. The first one is the energy content of thegeothermal fluid based on its enthalpy in kJ/kg.Second, the exergy content of the fluid can be used asa basis. However, in this study, the primary energycontent of the geothermal fluid taken from theproduction wells (and utilized for both electricity andheat production) is chosen to be the enthalpy above

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia15 °C, an <strong>International</strong> Standard Atmosphere (ISA)reference temperature [9], and calculated in thefollowing manner:Where(1)is the specific primary energy content inkJ/kg, is the enthalpy of the fluid and isthe saturated liquid enthalpy of the fluid at standardreference temperature of 15 °C.LIFE CYCLE ASSESSMENTThe Directive 2002/91/EC defines the concept ofprimary energy as energy that has not undergone anyenergy conversion process [1]. The primary energyfactor must thus represent all the primary energyconsumed in order to provide one unit of heat or powerto the consumer. Primary energy consumption ofenergy chains is not only based on the consumption offuel (or other energy source) in the power or heatgeneration process, but also all the primary energyneeded for the construction, operation and possiblydemolition of the production facilities. Also, someprimary energy is needed for the distribution of theproduct. To calculate such accumulated primaryenergy, the method of life cycle assessment is wellsuited. LCA is a method that has been developingsince the earliest performance of such a study in 1969and standards on the methodology where issued in thelate 1990s [10].LCA has been considered a good tool to achieve aholistic approach on evaluating the environmentalimpact of products. Today, it is widely used toinvestigate all kinds of production systems and hasgiven valuable insight on the total impact of productsand systems on the environment by not only focusingon the operational aspect [11]. Many interesting resultshave been achieved by using this methodology andthose results form a basis for evaluating and comparingdifferent solutions for production of various products,such as vehicles for transport, soft drink containers andpower conversion technologies. On the other hand,LCA in the process industry has had much lessattention than for manufacturing products, andresearch is needed before complete methods forprocesses are readily available [11]. The application ofLCA on geothermal energy utilization can be valuablefor LCA developers working on further improvementsand adjustments on the LCA methodology for theprocess industry.Using LCA to calculate the total primary energyconsumption and CO 2 emission for heat and powerproduction based on geothermal energy will helpidentify how much effect the construction, collection ofgeothermal fluid and even the demolition phase of thepower plant and the distribution system have on thetotal primary energy consumption. It can identify theimpact of the drilling of wells, manufacturing of powerplant components and piping, construction of buildingsand roads associated with the power plant, operation ofthe power plant itself and the primary energy extractedfrom the geothermal reservoir and even the impacts ofconstructing and operating the distribution facilities.The different phases of performing LCA will bedescribed in the following sections. The main phases ofLCA include:Defining the goal and scope of the studyPerforming inventory analysisPerforming impact assessmentGoal and Scope of the StudyThe main goal of this LCA study is to analyze the twoenergy performance indicators presenting the primaryenergy efficiency and the CO 2 emissions for both theelectricity and heat production at Hellisheidi powerplant. The LCA calculations and impact assessmentwhere done by using the LCA software SimaPro 7 [12]and using different databases such as the Ecoinventdatabase [13] for the inventory information on variousraw materials and processes used in the geothermalpower plant.There are numerous geothermal power plantsworldwide using similar technology as the Hellisheidipower plant to produce electricity (double flash powerplants produced 23% of the electrical power fromgeothermal resources in 2007 [14]), so the results forthe energy performance indicators for the powerproduction at Hellisheidi could be used to representthese power plants. Other types of geothermal energyconversion systems, such as single flash and binarysystems, should be treated individually whencalculating energy performance indicators for theelectricity production.Geothermal combined heat and power plants are notcommon worldwide, but regarding Europe they can befound in Iceland as well as Austria and Germany. Byproducing heat as well as electricity in geothermalapplications, the utilization of the heat taken from thegeothermal reservoir in the form of geothermal fluid ismaximized. The heat produced has a variety of usefulapplications, such as for district heating, agriculture,fisheries, swimming pools, snow melting and heatingup greenhouses [3]. The calculations of the primaryenergy factor of the heat production at Hellisheidigeothermal CHP plant will emphasize this increase inthermal efficiency of the power plant.187

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe scope of this LCA study includes making thefollowing choices [11]:Functional unitSystem boundariesChoice of impact categoriesMethod for impact assessmentPrinciples for allocationData quality requirements1 kWh electricity1 kWh heatElectricity,geothermal, atHellisheidi CHPplantGeothermalfluid, fromelectricityproductionHeat, fromcondenserHeat,geothermal, atHellisheidi CHPplantGeothermalpower plant unitGeothermalfluid, at powerplantCollectionpipelinesGeothermalheat productionunitPower plantequipmentPower plantstructuresGeothermalfluid, in groundDrilling ofgeothermalwellsHeating stationStructureHeating stationequipmentFigure 2: Flow model for the life cycle assessment of the Hellisheidi CHP plantFunctional UnitThe primary energy and CO 2 factors are defined asprimary energy usage and CO 2 emission per MWh andthus, the functional unit of the study is chosen to beMWh of electricity or heat produced in the Hellisheidigeothermal CHP plant. The functional unit is thereference flow to which all other modelled flows of thesystem are related.System BoundariesThe processes included in this LCA study are mainlythe operation and construction of the power plant. Thedemolition or end-of-life phase is disregarded due toinsufficient information at this time. Also, the energyand material flows due to maintenance in theoperational phase of the power plant are disregardedbut both the demolition and the maintenance will beincluded in further studies. The time horizon in thisstudy is chosen to be 30 years, which is the technicallifetime of the power plant capital goods.A flow model of the CHP plant as modelled in the LCAstudy is shown in Figure 2. The two outputs of theproduction system are 1 MWh of electricity and 1 MWhof heat. The main material and energy inputs into theenergy conversion system are the geothermal power188plant unit, the geothermal heat production unit and thegeothermal fluid. The geothermal power plant isconstructed from the power plant structures andequipment while the fluid is transported in collectionpipelines from geothermal wells that need to be drilledfor the production. The heat production unit consists ofthe heating station structure and equipment. Theenergy input into the heating process is waste heatfrom the power production process in form of heattaken from the steam in the condenser for preheatingof district heating water, and the waste geothermal fluidfrom steam separators used for final heating of thedistrict heating water. Inventory data on all thesedifferent components in the flow model was collectedand used for the LCA study of the Hellisheidi CHPplant.Impact Categories and Methods for ImpactAssessmentTo calculate the two energy performance indicators,the two main impact categories to be used are theprimary energy demand of the production process inMWh and Global Warming Potential (GWP) given inCO 2 equivalents.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTwo different methods of impact assessment had to beused in the impact assessment calculations. For theprimary energy factor, the Cumulative Energy Demand(CED) method [15] was used which is based on amethod published by Ecoinvent 1.01 and available inSimaPro 7 impact assessment methods. For thecalculation of the CO 2 emission factor, the IPCC 2007GWP 100a V1.01 [16] was used to get the CO 2equivalent total global warming potential for the chosenfunctional unit of 1 MWh electricity produced.Principles for AllocationTo allocate the impacts of the different products,electricity and heat, produced at Hellisheidi CHP plant,several methods can be used. The method usedshould reflect the physical relation between the twoproducts, such as how the different inputs and outputsof the process are dependent on the two differentproducts. Simple methods of allocation for an energyconversion process can be:Based on energy content of the productsBased on exergy content of the productsBased on the monetary value of the productsThe abovementioned methods can be used when thephysical relation between the two products is unclear.In the case of the Hellisheidi CHP plant, the physicalrelation between the two outputs (electricity and heat)is mainly the use of waste heat from condensers andthe geothermal fluid from the production wells, asshown in Figure 2. The impacts of construction caneasily be divided between the electricity and heatproduction with the detail of inventory data provided.Also, the geothermal fluid used in the heat productionis taken from steam separators in the electricitygeneration process and would otherwise be reinjectedback into the geothermal reservoir via reinjection wells.The disposed heat in the condenser is utilized topreheat the district heating water by using it as coolingwater. The condenser pressure determines thetemperature of the steam output from the turbines andthus, also the final temperature of preheating of thedistrict heating water. If the heat demand is high, thecondenser pressure must be higher than the optimumfor power production in order to supply high enoughtemperatures to the district heating water. This limitsthe electrical power production and requires that moregeothermal wells have to be drilled in order to sustainthe electrical production under high thermal loads ofthe district heating system. These limitations on theelectrical production imply that the allocation of impactsfrom the drilling of wells should be related to thenumber of wells that have to be drilled to sustain boththe electricity production and the highest thermal loaddesigned for the district heating system.Data QualityTo calculate the energy performance indicators bymethods of LCA, reliable inventory information isneeded on material and energy flows to and from thegeothermal power production facilities during theirlifetime.. The inventory in this study is constructedfrom data provided by Reykjavik Energy, the powercompany in ownership of the Hellisheidi plant. Thedata on the construction phase is retrieved from theconditions and specifications in a tender for theconstruction of the power plant, where quantitativeinformation is collected on all major material flowsrequired for the constructions and machinery. Theinventory information for the fluid collection and drillingis retrieved from a report done by Reykjavik Energy,including the power and performance of the geothermalwells drilled for the power and heat production [17].For a LCA study, the following data quality indicatorsmust be presented:Time periodRegionType of technology and representativenessAllocationSystem boundariesIn this study, the time period of the data is from 2005 to2009 and the region is Western Europe. The type oftechnology is modern and the representativeness isdata from a specific company. The allocation, asmentioned before, is by physical connections betweenthe two outputs. The system boundaries are describedby three different criteria. First, the cut-off criteria is ingeneral set to be less than 5% which means that allinventory data that does not contribute more than 5%to the overall impacts of the two products isdisregarded. Also, the system boundary is chosen tobe of the first order, only to account for the materialsused in the construction and operation of the CHP plantbut not the processing and transportation of thesematerials. The third system boundary criterion is thesystem boundary with nature, which in this study isdescribed as unspecified at this stage of the LCAstudy.RESULTS FOR THE ENERGY PERFORMANCEINDICATORSEnergy Performance Indicators for ElectricityProductionThe results for the impact assessment of the electricityproduction alone, focusing on the two energyperformance indicators, is shown in Table 2. Thehighest value of fp 6.33 MWh primary energy/MWhproduced energy, is obtained when no heat productionis present at the power plant and the effects ofreinjection of waste streams is not taken into account.189

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe value of 5.33 for fp is obtained in the two lattercases, where the waste heat is either reinjected backinto the reservoir or used for heating of DH water. Inthose cases, the primary energy content of the wastestream can be subtracted from the primary energycontent of the geothermal fluid used for the electricityproduction, resulting in lower fp values. The share ofnon-renewable primary energy sources such as oil andgas used in the construction phase or in themanufacturing of various power plant components, onlyaccount for about 0.01 of the total fp value in all cases.The factor K for the CO 2 emissions is the same for allthree cases of electricity production as reinjection andutilization of waste stream does not have significanteffects on the total emissions due to the power forproduction. The origins of the CO 2 emissions can beseen in Figure 4. The largest contributor to the CO 2emission from the electricity generation over 30 yearsof production is the geothermal fluid, responsible 88%of the CO 2 emissions per kWh of electricity production.Table 2 – Results for the primary energy factor and CO 2 emission factor for electricity based on geothermal energySource of electricityPrimary energy factors f p[MWh primary energy / MWh producedenergy]Non-RenewableTotalCO 2 productioncoeff. K[Kg/MWh]Electricity from Hellisheidi geothermal powerplantElectricity from Hellisheidi geothermal powerplant, with reinjection0.01 6.33 290.01 5.33 29Electricity from Hellisheidi CHP plant 0.01 5.33 29A small share of 8% originates from the drilling ofgeothermal wells while the construction of the powerplant, along with the manufacturing of its maincomponents, is responsible for 4% of the CO 2emissions.GWP 100a for Electricity Productionin kg CO2 eq4%8%0.5%87.5%Geothermal fluid(87.5%)Power plant andcomponents (4%)Geothermal welldrilling (8%)Collection lines(0.5%)value reduces to 0.69. In both cases, the share ofprimary energy from non-renewable energy sources isless than 0.01. In both cases, the CO 2 productioncoefficient is 0.98 kg CO 2 equivalents per producedMWh.The origins of the CO 2 emission from the heatgeneration process can be seen in Figure 4. Thelargest contributor to the total emissions is the drillingof the geothermal production wells that were needed tosustain the electricity production while the heatproduction is at maximum load of 133 MWth. Themanufacturing of the district heating pipeline from theproduction area to the rural area of Reykjavík citycontributes to 15% of the total emission resulted fromthe heat generation process.Figure 3 – Origins of CO 2 emissions from the differentprocesses of the power generationEnergy Performance Indicators for ThermalProductionThe energy performance indicators for the productionof heat for district heating are given in Table 3. Twocases are presented for the heat production; heatproduction process with or without the effects ofreinjection of waste geothermal fluid. The highest valuefor f p is obtained in the case where reinjection is nottaken into account, with the value of 1.78 MWh primaryenergy/MWh produced energy. With reinjection, theFigure 4 – Origins of CO 2 emissions from the differentprocesses of the heat generation190

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 3 – Results for the primary energy factor and CO 2emission factor for heat from a geothermal CHP plantSource ofheatHeat,HellisheidiCHP plantHeat,HellisheidiCHP plant,reinjectionDISCUSSIONPrimary energy factors f p[MWh primary energy /MWh produced energy]Non-RenewableTotal>0.01 1.78 0.98>0.01 0.69 0.98CO 2productioncoeff. K[Kg/MWh]The following discussion highlights the most significantresults from this study:1) By comparing the energy performance indicatorscalculated in this study and shown in Table 3 to theindicators given in Table 3 it can be seen that electricityfrom geothermal power plants has the highest total f pfactor while the share of non-renewable energysources is the lowest. The main reason for the high f pfactor is the low conversion efficiency of geothermalpower plants due to low working temperatures andpressures. The CO 2 production coefficient is relativelylow compared to the other energy conversiontechnologies and could be lowered even further ifmeasures are taken to control the emissions from thepower plant. The results for the Hellisheidi geothermalCHP plant cannot be used to represent all geothermalpower plants producing either electricity alone or with acombined production of electricity and heat. Furtherstudies are needed on different types of geothermalpower plants, such as single flash and organic Rankinecycles, to be able to produce specific or averagefactors representing geothermal utilization.2) The results for the heat production at theHellisheidi geothermal CHP plant, given in Table 3,show that the energy performance indicators arerelatively low and, in the case of reinjection, belowunity. This is because the primary energy needed topreheat the DH water is not accounted for in the heatproduction but rather assigned to the electricityproduction. This is due to the fact that the preheating ofthe DH water from 5 °C to 41 °C is done in thecondenser for the high pressure steam turbine as seenin Figure 1 and is a necessary step in the electricityproduction, but a beneficial step in the heat productionfor the DH system.3) Values for the indicators for both electricity andheat are calculated with and without reinjection of thecooled geothermal brine from the energy conversionprocess. Reinjection of geothermal brine is recognizedto improve heat mining and stabilize the productioncapacity of geothermal fields, if successfully carriedout. It can also counteract pressure draw-down in thereservoir by providing an artificial water recharge [18].In this study, reinjection of the waste stream ismodelled, which decreases the use of primary energyin the energy conversion process, since a part of theprimary energy from the geothermal fluid it is returnedback to the reservoir. Reinjection is present at theHellisheidi geothermal CHP plant so the values of theenergy performance indicators with reinjection are validfor the power plant.4) Life cycle assessment is especially useful toevaluate the total impact of geothermal power plantswith respect to their emission of greenhouse gasses.Figure 3 and Figure 4 show how the different phases inthe life cycle of the power plant significantly contributeto the overall emission in CO 2 equivalents. If LCA hadnot been carried out for the process, 12% of the CO 2emissions resulting from the electricity generationwould not have been accounted for and no emissionswould have been found for the heat production, sincethe emissions from drilling, construction of buildings,and manufacture of components had not beenaccounted for.ACKNOWLEDGMENTSSpecial thanks are given to the following partners:Nordic Energy Research (NER) for funding the studyand the Energy Research Fund of Landsvirkjun for theirsupport. To Orkuveita Reykjavíkur for providing data forHellisheidi Power plant, to Mannvit engineering fordiscussion and data provision and to Ragnar Gylfasonfor his contribution in the data gathering phase.REFERENCES[1] EU. (2003, January 4). Directive 2002/91/EC ofthe European Parliament and of the Council of 16December 2002 on the energy performance ofbuildings. Official Journal of the EuropeanCommunities .[2] Bertani, R. (2010). Geothermal Power Generationin 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 2010Worldwide Review. Proceedings World GeothermalCongress 2010, (April), 25-29.[4] EN 15603:2008. Energy performance of buildings.Overall energy use and definition of energy ratings.191

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaGeneva: <strong>International</strong> Organisation forStandardisation (ISO).[5] Helisheidi Geothermal Plant. (2009). RetrievedMay 2009, from http://www.or.is/English/Projects/HellisheidiGeothermalPlant/[6] H. Douglas Lightfoot. (2007). Understand the threedifferent scales for measuring primary energy andavoid errors. Energy, 32, 1478-1483.[7] Segers, R. (2008). Three options to calculate thepercentage renewable energy: An example for aEU policy debate. Energy Policy , 36 (9), 3243-3248.[8] IEA. (2004). Energy Statistics Manual.<strong>International</strong> Energy Agency (IEA). Paris:OECD/IEA.[9] ISO 2533:1975. Standard atmosphere.<strong>International</strong> Organization for Standardization,Geneva, Switzerland.[10] Russell, A., Ekvall, T., & Baumann, H. (2005). Lifecycle assessment - introduction and overview.Journal of Cleaner Production , 13 (13-14), 1207-1210.[11] Baumann, H., & Tillman, A.-M. (2004). The HitchHiker's Guide to LCA. Lund, Sweden:Studentlitteratur AB.[12] PRéConsultants. (2009, September 6). SimaProLCA software. Retrieved October 14, 2009, fromSimaPro LCA software: http://www.ecoinvent.ch/[13] Ecoinvent. (2009, August 13). Home. RetrievedOctober 14, 2009, from Home:http://www.ecoinvent.ch/[14] DiPippo, R. (2008). Geothermal Power Plants –Principles, Applications, Case Studies andEnvironmental Impact (2nd edition ed.). Oxford:Butterworth-Heinemann.[15] Klöpffer, W. (1997). In defense of the cumulativeenergy demand (editorial). <strong>International</strong> Journal ofLife Cycle Assessment , 2, 61.[16] PRéConsultants. (2009, September 6). Methods.Retrieved October 14, 2009, from SimaPro LCAsoftware:http://www.pre.nl/simapro/impact_assessment_methods.htm#CML2[17] Gunnlaugsson, E., & Oddsdóttir, A. L. (2009).Helisheidi - Gufuborholur 2008 (Hellisheidi - Steamwells 2008). Reykjavík: Orkuveita Reykjavíkur.[18] Stefansson, V. Geothermal reinjection experience.Geothermics, 26, (1997), 99–130.192

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFLEXIBILITY FROM DISTRICT HEATINGTO DECREASE WIND POWER INTEGRATION COSTSJ. Kiviluoma 1 and P. Meibom 21VTT Technical Research Centre of Finland2 Risø DTUABSTRACTVariable power sources (e.g. wind, photovoltaics)increase the value of flexibility in the power system.This paper investigates the benefits of combiningelectric heat boilers, heat pumps, CHP plants and heatstorages in a district heating network when the share ofvariable power increases considerably. The results arebased on scenarios made with a generation planningmodel Balmorel [1]. Balmorel optimises investmentsand operation of heat and power plants, including heatstorages. It uses hourly resolution and enforcestemporal continuity in the use of the heat storages.Scenarios with high amount of wind power wereinvestigated and the paper describes how the increasein variability changes the profitability and operation ofdifferent district heating options in more detail than wasdescribed in the article by Kiviluoma and Meibom [2].Results show that district heating systems could offersignificant and cost-effective flexibility to facilitate theintegration of variable power. Furthermore, thecombination of different technologies offers the largestadvantage. The results imply that, if the share ofvariable power becomes large, heat storages shouldbecome an important part of district heating networks.NOMENCLATUREIndicesi, I Unit, set of unitsI HeatStoHeat storage unitst, T Time steps, set of time stepsa, A Area, set of areasVariablesCPQZParametersc Invc Fixc OperationwhNew capacityPower generationHeat generationCharging of heat storageAnnualized investment costFixed operation and maintenance costsOperation cost function of the unitWeight of time periodHeat demandINTRODUCTIONWind power is projected to be a large contributor tofulfil electricity demand in several countries. This couldtake place due to relatively low cost of wind powerelectricity or policy mechanisms promoting renewableenergy. In any case, power systems with a largefraction of power coming from a variable power sourcewill need to be flexible. Flexibility is used to cope withthe increased variation in residual load (electricitydemand minus variable power production) and with theincreased forecast uncertainty in the residual load. Onthe other hand, lack of flexibility will cause larger costsfrom increased variability and forecast errors.Therefore, it is prudent to investigate the cost optimalconfigurations for the combined power and heatgeneration portfolios.Heat generation could offer significant possibilities forincreasing the flexibility of the power system. Currently,part of the inflexibility of the power system comes fromCHP plants that are operated to serve the heat loadwhile electricity is a side product. Installation of electricresistance heaters next to the CHP units or elsewherein the heat network could break this forced connection.During periods of low power prices, which will becomemore common with high share of wind power, CHPplants could be shut down and heat would be producedwith electricity. The dynamics can be made moreeconomic with the use of heat storages. Further optionis to have heat pumps in the DH network, but they willrequire large amount of full load hours to be profitableand will compete with CHP plants for the operatingspace.In most countries heat demand is in the same order ofmagnitude as electricity demand. For example, in UKthe demand for primary energy due to heat is around40% of total primary energy demand [3]. About 25% ofthe primary energy demand is due to space and nonindustrialwater heating. In the US all kind of heat useaccounts for about 30% of the primary energyconsumption [estimated from 4].Heat is inexpensive to store compared to electricity.Electricity storage has been seriously considered toalleviate the variability of wind power [5-6]. Therefore, itis apparent that the use of heat storages should alsoreceive serious consideration in the current context.Some work has been done [7-9], but not considering193

optimal investments in new power plants and heatstorages.The study has been restricted to residential andindustrial district heating systems. Buildings notconnected to district heating systems were notconsidered, although these also require heat. Coolingdemand could also offer similar possibilities, but theproblem was not addressed here. Industrial heatdemand and water heating do not usually have strongseasonal variation and can therefore be more valuabletowards the integration of variable power.METHODS AND DATAThe model and assumptions used for the analysis aredescribed in more detail in [2]. For convenience, mostimportant sections are referenced below. The heatsector of the model is described more thoroughly here.The Balmorel model is a linear optimization model of apower system including district heating systems. Itcalculates investments in storage, production andtransmission capacity and the operation of the units inthe system while satisfying the demand for power anddistrict heating in every time period. Investments andoperation will be optimal under the input dataassumptions covering e.g. fuel prices, CO2 emissionpermit prices, electricity and district heating demand,technology costs and technical characteristics (eq. 1).The model was developed by (Ravn et al. [1]) and hasbeen extended in several projects, e.g. (Jensen &Meibom [10], Karlsson & Meibom [11], Kiviluoma &Meibom [2]).min iIExOperation Ci CiwtciPi, t, QitInvFixciCi ci,iItTiIThe optimization period in the model is one yeardivided into time periods. This work uses 26 selectedweeks, each divided into 168 hours. The yearlyoptimization period implies that an investment is carriedout if it reduces system costs including the annualizedinvestment cost of the unit.The geographical resolution is countries divided intoregions that are in turn subdivided into areas. Eachcountry is divided into several regions to represent itsmain transmission grid constraints. Each region hastime series of electricity demand and wind powerproduction. The transmission grid within a region isonly represented as an average transmission anddistribution loss. Areas are used to represent districtheating grids, with each area having a time series ofheat demand. There is no exchange of heat betweenareas. In this article, Finland is used as the source formost of the input data.The hourly heat demand has to be fulfilled with the heatgeneration units, including heat storages (eq. 2).The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia(1)194Loading of heat storage adds to the heat demand. Lossduring the heat storage process is not considered. Thedynamics of heat networks were not taken intoaccount.iIQi t hrt Z t Ta A, ,i,tHeatStoiIa ; (2)Analysis is done for the year 2035. By this time, largeportion of the existing power plants are retired. Threedistrict heating areas were considered. These have arather different existing heat generation portfolio by2035. This helps to uncover some interesting dynamicsin the results section.In this paper, scenarios without new nuclear power arecompared (scenarios ‗Base NoNuc‘ and ‗OnlyHeatNoNuc‘ in article [2]). This meant that wind power had avery high share of electricity production. Accordingly,there was more demand for flexibility in the system.‗Urban‘ area presents the heat demand in the capitalregion of Finland. The existing power plants in 2035cover over half of the required heat capacity. Largestshare comes from natural gas, which is a relativelyexpensive fuel in these model runs. The annual heatdemand is smallest of the considered areas: 6.2 TWh.‗Industry‘ area aggregates the known industrial districtheating demand from several different locations. This isa necessary simplification, since Finland has overhundred separate DH areas and the model would notbe able to optimise all of these simultaneously. Theindustrial heat demand in Finland is driven by paperand pulp industry, which produces waste that can beused as energy input. This capacity is assumed to beavailable in 2035 and as a consequence the modeldoes not need more industrial heat capacity. Theannual heat demand is 46.8 TWh.‗Rural‘ area aggregates non-industrial heat demandexcluding the capital region considered in ‗Urban‘. Thisis probably the most interesting example, as theexisting capacity covers only 20% of the heat capacitydemand. Therefore, the model has to optimise almostthe whole heat generation portfolio. There are woodresources (limited amount of forest residues and moreexpensive solid wood) available unlike in the urbanarea. The annual heat demand is 21.0 TWh.RESULTSFigures 1–3 give an example how heat productionmeets heat demand in the different areas during thesame 4.5 days in January. Negative productionindicates charging of heat storage. Electricity price ison separate axis together with the cumulative contentof heat storage. When electricity price is low, storage isloaded with electricity using heat boilers and heatpumps. When electricity price is high, CHP units

produce heat and electricity. Fluctuations in electricityprice are mainly driven by changes in wind powerproduction, since these are larger than changes inelectricity demand (Fig. 4).Heat production (MW)Heat production (MW)60005000400030002000The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia1200450NG_EX_UR1000375NG_BP_UR800MW_HB_UR600300MW_BP_UR400225EL_HP200EL_HB0150Storage use-20075Stor. content-400Elec. price-6000Fig. 1. Example of operation in ‗Urban‘ heat area. Negative production indicates charging of heat storage.Heat production (MW)1000150Storage use0Stor. content-100075Elec. price-20000Fig. 3. Example of operation in ‗Industrial‘ heat area. Negative production indicates charging of heat storage.Electricity production (MW)4000MW_HB_RU3000375WR_EXWW_EX2000300NG_BP_RUNG_CC_EX1000225PE_BP_RUWO_BP_RU0150EL_HPEL_HB-100075Storage useStor. content-20000Elec. priceFig. 2. Example of operation in ‗Rural‘ heat area. Negative production indicates charging of heat storage.17500150001250010000Wood waste (WW)Fig. 7500 1. Example of operation in ‗Urban‘ heat area. Negative production indicates charging of heat storage.Peat5000Solid wood (WO)25000-2500195450375300225Electricity price (€/MWh)Heat storage content (%)PE_BP_INWR_BP_INWW_BP_INEL_HB-5000Fig. 4. Electricity production. Negative production indicates the use of electric heat boilers and/or heat pumps.450Electricity price (€/MWh)Heat storage content (%)Electricity price (€/MWh)Heat storage content (%)WindNatural gas (NG)HydroForest residues (WR)Municipal waste (MW)NuclearElectricity to heat

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaEffects of heat measures in the three heat areasIn 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 isthat the existing heat production capacity fromindustrial wood waste and the associated no-costHeat capacity (MW)Heat production (GWh)2500020000150001000050000BaseOnlyHeatCap.BaseOnlyHeatProd.HEATSTOREL_HBWW_BP_INWR_BP_INPE_BP_INNG_BP_INFO_BP_INFig. 5. Heat capacity and production 8 in the ‗Industrial‘heat area.waste wood were not easily replaced. However, therewere some high wind situations with low power priceswhere it was beneficial to use electric heat boilers toproduce heat and decrease heat production from woodwaste in the ‗Industry‘ area. There was an annualresource limit on wood waste on the country level andthe wood waste use was transferred to the ‗Rural‘ heatarea. It was also profitable to install some heat storagecapacity. This enabled the full shut down of woodwaste back pressure power plants for the duration oflow electricity prices. This decreased electricityproduction and gave more room for the upsurge in windpower production.In the ‗Urban‘ heat area heat measures enabled thereplacement of CHP coal units with production fromheat pumps and to smaller extent from electric heatboilers (Fig. 6). Also wood based heat boilers werereplaced. Investment in heat storage was relativelysmaller. However, they were cycled more due to fastercharging rate.Heat capacity (MW)Heat production (GWh)3500300025002000150010005000BaseOnlyHeatCap.BaseOnlyHeatProd.HEATSTOREL_HPEL_HBNG_HBCO_EXWO_HBNG_EX_URNG_BP_URMW_HB_URMW_BP_URFig. 6. Heat capacity and production 1 in the ‗Urban‘ heatarea.The combined utilization of the heat measures wasused to shut down existing natural gas based CHPpower plants during hours of average or lowerelectricity prices. During low electricity prices electricheat boilers were used to charge heat storage.Accordingly, during average electricity prices heat wasused from heat storage to prevent the use of electricheat boilers. During the highest electricity priceselectric heat pumps were also shut down with the helpof heat from the heat storages.The most important difference between ‗Urban‘ and‗Rural‘ heat areas is the availability of wood residues inthe ‗Rural‘ heat area (Fig. 7). For the most part thisresource was able to outcompete heat pumps asmeans to produce heat. Heat measures still helped toreplace coal CHP. The combination of electric heatboilers and heat storages was again a large source ofadditional flexibility to the system.Heat capacity (MW)Heat production (GWh)120001000080006000400020000BaseOnlyHeatCap.BaseOnlyHeatProd.HEATSTOREL_HPEL_HBNG_HBNG_CC_EXCO_EXWR_EXWW_EXWO_HBWO_BP_RUPE_BP_RUNG_BP_RUMW_HB_RUFig. 7. Heat capacity and production 1 in the ‗Rural‘ heatarea.8 Heat production is from the modelled 26 weeks and should bemultiplied by 2 to get an estimate on annual production.196

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDynamics of heat storageMost of the daily fluctuation in heat demand wassmoothed with heat storages and electric heat boilersin all heat areas. If CHP units were operated, they wereusually operated at maximum heat output.The investment cost for heat storage was assumed tobe 1840 €/kWh. With the assumed ratio of 12 betweenstorage capacity and heat capacity this translates to153 €/kW. In comparison the capacity cost of electricheat boilers was assumed to be 40 €/kW and 50 €/kWfor natural gas heat boiler. This means that investmentinto heat storage capacity was not driven by need fornew capacity since heat boilers were cheaper. Therehad to be operational benefits from the use of heatstorage to cover the additional investment costs.Heat storages create operational benefits by movingconsumption from more expensive sources of heat toless expensive by shifting demand in time. In allheating areas whole operating ranges of heat storageswere extensively utilized. During most 168 hour periodsheat storage reached both the minimum and maximumstorage capacities. In the ‗Rural‘ area heat storage was2.1% of the time either full or empty. With a largerstorage capacity this could have been reduced, but itwas not worth the investment.The size of the heat storage in ‗Industry‘ area waslarger than in other areas in relation to daily heatdemand (Fig. 8). In ‗Industry‘ area charging of heatstorages took place over several days during higherpower prices, when wood waste CHP units wereproducing extra electricity. Storing the extra heatrequired larger heat storage capacity. On the contrary,in ‗Rural‘ and ‗Urban‘ charging and discharging wasmore balanced and smaller heat storage was enough.Heat (GWh)180160140120100806040200Heat storage sizeMax daily heatMin daily heatAverageRural Urban IndustryFig. 8. Heat storage size compared to maximum, minimumand average daily heat demands.In the ‗Rural‘ area during winter time, charging of heatstorages is mostly based on the use of electric heatboilers. They create large amount of heat in relativelyshort time during periods of low power prices. Duringsummer time, heat storages are charged by turning onwood waste and forest residue CHP units. Duringspring and fall CHP units operate more often, since theheat load is larger, but still the heat storage helps toshut them down for periods of some hours.‗Urban‘ area has similar dynamics, but during summertime the adjustment is made by heat pumps instead ofCHP. In the winter during high power prices old naturalgas CHP units are less expensive to operate than theheat pumps.CONCLUSIONSDistrict heating systems offer good possibilities forincreasing the flexibility of the power system, if thepenetration of variable power like wind power increasesgreatly in the future. According to the results, mainvessels to increase flexibility are the use of heatstorages, electric heat boilers and flexible operation ofCHP units.Investment in electric heat boilers in district heatingsystems is driven mainly by periods of very high windpower production. The resulting cheap electricity isconverted to heat and to some extent stored in heatstorages for later use. Investments in heat storage inturn are driven by the same mechanisms, but also tocreate flexibility in the electricity production when pricesare higher. To enable this, the operation of CHP unitsand heat pumps is altered with the help of heatstorages. Heat pumps mainly compete against CHP asa 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 hasa considerably penalty due to CO 2 cost. Heat pumpsare not very important as a source of flexibility, sincethey require lot of full load hours due to theirinvestment cost.While the research has been conducted on districtheating, similar dynamics could be achieved inhousehold heating not connected to district heatingnetworks. However, the costs are likely to be largerunless there is an existing hot water tank. Flexibilitycould also be gained from district cooling or airconditioningunits with the addition of a cold storage.Further research should also address some of theshortcomings of current study. Sensitivity analysiswould be important, especially concerning the costestimates of the analysed heat measures. Heat storagemodel 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.197

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaHeat pumps were assumed to work at constant COPand this is a crude approximation even if the heatsource is groundwater or sea water.REFERENCES[1] H. Ravn et al. Balmorel: A Model for Analyses ofthe Electricity and CHP Markets in the Baltic SeaRegion. Balmorel Project 2001. See also:http://www.balmorel.com/Doc/B-MainReport0301.pdf[2] J Kiviluoma and P. Meibom, ―Influence of windpower, plug-in electric vehicles, and heat storageson power system investments‖, Energy, Volume35, Issue 3, March 2010, pp. 1244-1255. Elsevier.doi:10.1016/j.energy.2009.11.004[3] Energy consumption in the UK: overall data tables,2009 update. Department of Energy and ClimateChange - secondary analysis of data from theDigest of UK Energy Statistics, Office of NationalStatistics and the Building ResearchEstablishment.[4] Annual Energy Review 2008. U.S. EnergyInformation Administration.[5] H. Ibrahim, A. Ilinca and J. Perron, ―Energy storagesystems—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[6] J.K. Kaldellis and D. Zafirakis, ―Optimum energystorage techniques for the improvement ofrenewable energy sources-based electricitygeneration economic efficiency‖, Energy, Vol. 32,pp. 2295–2305. Elsevier.[7] H. Lund and E. Münster, ―Modelling of energysystems with a high percentage of CHP and windpower‖, Renewable Energy, Vol. 28, 2003, pp.2179-2193. Elsevier. doi:10.1016/S0960-1481(03)00125-3[8] H. Lund, ―Large-scale integration of wind powerinto different energy systems‖, Energy, Volume 30,Issue 13, October 2005, pp. 2402-2412. Elsevier.doi:10.1016/j.energy.2004.11.001[9] H. Lund, B. Möller, B.V. Mathiesen and A.Dyrelund, ―The role of district heating in futurerenewable energy systems‖, Energy, Vol. 35, 2010,pp. 1381-1390. doi:10.1016/j.energy.2009.11.023[10] K. Karlsson and P. Meibom, ―Optimal investmentpaths for future renewable based energy systems –Using the optimisation model Balmorel‖,<strong>International</strong> Journal of Hydrogen Energy Vol. 33,2008, pp. 1777-1787.[11] S.G. Jensen and P. Meibom, ―Investments inliberalised power markets. Gas turbine investmentopportunities in the Nordic power system‖, Int. J.Electr. Power Energy Syst. Vol. 30, 2008,pp. 113–124.198

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDAILY HEAT LOAD VARIATION IN SWEDISH DISTRICT HEATING SYSTEMSH. Gadd and S. WernerSchool of Business and Engineering, Halmstad UniversitySE-301 18 Halmstad, Phone: +46 35 167757henrik.gadd@hh.se, sven.werner@hh.se, www.hh.seABSTRACTIf daily heat load variations could be eliminated indistrict heating-systems, it would make the operation ofthe district heating system less costly and morecompetitive . There would be several advantages in theoperation such as:Less use of expensive peak load power whereoften expensive fuels are used.Less need for peak load power capacity.Easier to optimize the operation that leads tohigher conversion efficiencies.Less need for maintenance because of moresmooth operation of the plantsThere are a number of ways to handle the dailyvariations of the heat load. Two often used are largeheat storages or using the district heating network astemporary storage. If it would be possible to centrallycontrol the customer substations, it would also bepossible to use heavy buildings connected to thedistrict heating system as heat storages.To be able to find the best way to reduce or eveneliminate the daily heat load variations, you need tounderstand the characteristics of the daily variations.This paper will describe a way of characterizing dailyheat load variations in some Swedish district heatingsystems.INTRODUCTIONFor all heat generation/distribution systems, heat loadvariations leads to inefficiencies. You need to designyour system for the peak load even though you onlyneed the top capacity for a very short period of time ofthe year. This is of cause expensive. The solution tothis problem is heat storage. There are a number ofpossibilities to store heat in DH systems:Large heat storages at the heat generation plantsHeat storage in district heating networksHeat storage in heavy buildings in by allowingsmall variation in indoor temperatures[1].If it would be possible to extinguish daily variations itwould lead to several profitable advantages such as:Less use of expensive peak load power whereoften expensive fuels are used.Less need for peak load power capacity.Easier to optimize the operation that leads tohigher conversion efficiencies.Less need for maintenance because of moresmooth operation of the plantsTo do this some questions need to be answered:What input and output capacity to/from the heatstorage is needed?What size of the heat storage is needed?Are the daily heat variations in the specific systemlarge or small during a year?METHODNomenclatureP h = Present hour value [MWh/h]P d = Mean hour value for the present day [MWh/h]P a = Mean hour value for the whole year [MWh/h]S h = Energy transfer capacity [MWh/h]S d = Size of heat storage [MWh/day]S a = Total annual daily heat load variationh= Momentary daily variation [h/h]d= Total daily variation [h/day]a= Total annual relative daily variation [h/year]VariablesMeasured data has been collected from some districtheating systems in Sweden. The collected data is theheat power that is generated and fed into the districtheating network. It is hour mean power that is used, i.e.8 760 data points per year. Only whole years is usedfrom 1 of January to 31 of December. To describe thedaily variation three variables is defined.1. Momentary daily variation ( h)2. Total daily variation. ( d)3. Total annual relative daily variation. ( a)Three system examples are presented in this paper toexemplify the method to characterize district heatingdaily heat load variation:System A: From a city in South of Sweden with anannual heat generation of 200 GWh.199

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSystem B: From a city in Southwest of Sweden with anannual heat generation of 64 GWh.System C: From a city in the middle of Sweden with anannual heat generation of 1550 GWh.Momentary daily variation ( )The momentary daily variation is proportional to theamount of heat that needs to be fed in or out to the DHnetwork to extinguish the daily variation. This variabledescribe the heat power capacity needed for in and output from and to the heat storage. For each districtheating systems you will get 8 760 (8 784 during leapyears) values per system and year.hd1224h1P PThe total daily variation is presented in Fig. 2 for thethree example systems. The figure verifies that thevariations are more pronounced in the two smallersystems compared to the larger system. Anotherimplication is that the highest day values are very few,giving an incentive to construct heat storagessomewhat smaller than the peaks in the figure. Hence,the investment costs will be reduced more the lostbenefits from the storage, giving a more optimised heatstorage.PhadThe momentary daily variation is defined as thedifference of each hourly measured value and themean value of heat per hour of the same day dividedby the mean heat per hour of the year.54,54Total daily variationSysten ASystem BSystem CPh PdhPaThe momentary daily variation is presented in Fig. 1 forthe three example systems. The figure shows that thevariations are more pronounced in the two smallersystems compared to the larger system.Total daily variation, τd[h/day]3,532,521,510,5Momentary daily variation, τh[h/h]0,70,60,50,40,30,20,10-0,1-0,2-0,3-0,4-0,5-0,6-0,7Momentary daily variation0 1000 2000 3000 4000 5000 6000 7000 8000Hour of the yearSysten ASystem BSystem CFig. 1 Momentary daily variation sorted by size hour byhour for the three different district heating systems.Total daily variation ( d )Total daily variation is defined for each day and is avariable that is proportional to the amount of heat thatdivert from the daily mean heat load. If you want toextinguish the daily variation in a system this variabledescribe the size of the heat storage. For each DHsystems you will get 365 (366 during leap years) valuesper system and year.The total daily variation is defined as the sum over theday of the difference of each hourly measuring valueand the mean value of energy per hour of the sameday divided by two times the mean energy per hour ofthe year.0- 50 100 150 200 250 300 350Days of the yearFig. 2 Total daily variation sorted by size day by day forthe three different district heating systems.Total annual relative daily variation ( )Total annual daily variation is a variable that isproportional to the total amount of energy that at dailybasis divert from the mean value accumulated for aperiod of one year. It is used to compare differentsystems between themselves. For each DH systemsyou will get 1 value per system and year.Total annual daily variation is defined as the sum overthe year of the difference between each hourlymeasuring value and the mean value of energy perhour of the same day divided by two times the meanenergy per hour of the year. a128760,365hh1,d 1PP PThe annual daily variation is presented in Fig. 3 for 10different Swedish district heating systems. Since theannual daily load variation has a magnitude of250–500 h, only 3–6% of the annual heat load isgenerated above the daily average heat loads. Hence,it is the seasonal variations that dominate the heat loadvariations in the Swedish district heating systems.ada200

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTotal annual daily variation, τa[h/year]50045040035030025020015010050Total annual daily variationSYSTEM B01 10 100 1 000 10 000Annual heat supply [GWh]SYSTEM ASYSTEM CFig. 3 Total annual daily variation for 10 different districtheating systems in Sweden.RESULTSTo characterize daily heat load variations in districtheating systems three variables have been defined.h= Momentary daily variationd= Total daily variationa= Total annual daily variationTogether with the mean annual heat per hour (P a ) andthe energy transfer capacity in and out of the heatstorage, size of storage to extinguish the systems dailyvariation and the total daily variation and can bedetermined according to the expressions below.Energy transfer capacity:S h = ·P a [MWh/h]hSize of heat storage:S d = ·P a [MWh/day]dCONCLUSIONSAn expected conclusion would be that large districtheating systems have smaller relative daily variations ) than small district heating systems. There are two(areasons for that:1. In a large district heating system, the use of heatpower is spread on different distances from the heatplant, i e the chilled water in the return pipe return backto the heat generation at different time compared towhen the return water left ach substation (geographicaldiversity)2. In large district heating networks, you would expectthat the operators have more active operation of theheat distribution network with respect to temporary heatstorage.But as can be observed in the Fig. 3 there does notseem to be such a trend. One explanation could be thatthe heat users differ in different systems. e.g. in thesystem in Fig 3 with an annual heat supply of 9 GWh,mostly single and multi family houses are connectedand very few industry or office buildings are connected.Since there is a large diversity among the annual dailyvariation more data need to be collected to be able tomake any further conclusions.REFERENCES[1] Olsson L, Werner S: ―Building mass used as shortterm heat storage‖, The 11th <strong>International</strong><strong>Symposium</strong> on District Heating and CoolingReykjavik 2008.Total annual daily heat load variation:S a = ·P a [MWh/year]a201

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDISTRICT HEATING AS PART OF THE ENERGY SYSTEM:AN ENVIRONMENTAL PERSPECTIVE ON ‘PASSIVE HOUSES’AND HEAT REPLACING ELECTRICITY USEMorgan Fröling 1,2 and Ingrid Nyström 31Engineering and Sustainable Development, Mid Sweden University, Östersund, Sweden2 Chemical Environmental Science, Chalmers University of Technology, Göteborg, Sweden3 CIT Industriell Energianalys, Göteborg, SwedenABSTRACTEnergy use for space heating, hot tap water and otherheat use at comparatively low temperature levelsrepresent a substantial part of the total energy use inSweden and countries with similar climate. It is thus ofimportance to meet this demand in a way generating assmall environmental impact as possible. However, it ispossible to create a system with higher environmentalimpacts with energy efficient buildings compared toless energy efficient buildings through choice of lessgood energy carriers. It is not enough that theindividual parts of a system are good and efficient togive a low environmental impact; the parts must beconnected into the system in a good way.From environmental perspective energy efficientbuildings and district heating don‘t oppose each other– good parts connected in a good system will give anoptimal. The results from the study of the three items ofhousehold equipment show possibilities for districtheating to be an alternative with good environmentalperformance, but not under all heat generationregimes.INTRODUCTIONIt is of importance to meet for space heating, hot tapwater and other heat use at comparatively lowtemperature levels in a way generating as smallenvironmental impact as possible. This can be done byincreasing the efficiency in the use phase and in theheating systems of buildings as well as through heatgeneration systems with low environmental impact.During recent years there has been a focus on houseswith low need of space heating, low energy houses or―passive houses‖. In such buildings the heat from theincoming sun radiation together with body heat frompeople living in the houses and different householdequipment will cover the whole or at least substantialparts of the space heating need over a year (extraheating might be needed during the coldest days of ayear). Hot tap water still need to be heated. For parts ofthe year this can be achieved by solar panels, but thereis a need for extra heating during winter. This mightresult in the extra heating demand being covered byelectricity, directly or indirectly.Increased energy efficiency is in itself a desirable goalfor a society – it increases the robustness of the energysystem and the possibilities for a resource efficient andmore sustainable energy system in the long run.However, it is possible to create a system with higherenvironmental impacts with energy efficient buildingscompared to less energy efficient buildings throughchoice of less good energy carriers. It is not enoughthat the individual parts of a system are good andefficient to give a low environmental impact; the partsmust be connected into the system in a good way.Thus it is important to identify system solutions thatavoids sub optimization and gives us energy efficientbuildings and an efficient energy system with goodenvironmental performance.In a synthesis studies within the framework ofChalmers Energy Center [1] the role of district heatingin a future society with more energy efficient buildingshave been investigated. Here we report on generalfindings of this study with a special focus on theenvironmental performance of the possibility to convertsome household electricity use into district heating - forthe use in dish washers, washing machines and tumbledriers [2]. The environmental performance is studiedusing life cycle assessment methodology and differentassumptions regarding electricity and district heatinggeneration.DISTRICT HEATING – DEMAND SIDEThere are today several drivers in the direction of lowertotal heat market for district heating in future [1].Among possible such drivers in Sweden are:Warmer climate (due to climate change)Higher energy pricesIncreased environmental awarenessIncreased energy efficiency of existing buildingstockLimited amounts of new housingNew housing more energy efficientHowever, there are also possible drivers for a largerheat market in future, e.g.:Increased wealth giving larger living space perperson and higher demands on comfort202

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaElectricity prices might increase faster than heatprices might lead to interest in heat instead ofelectricity for ―new‖ applications (washer,dishwasher, et.c.)Heat for comfort coolingIncreased use of heat for other purposes – e.g.drying of biofuels et c.With strategic planning the resulting effect for districtheating might be a lower total but at the same timemore even demand of heat (Fig. 1).DISTRICT HEATING – SUPPLY SIDEA strategic role of district heating in the energy systemis the ability to utilize and deliver resources thatotherwise would have been lost. Among possiblesystem drivers on the supply side in Sweden are [1]:Increased utilization of industrial surplus heatRemaining large potential of waste incinerationIncrease of CHP power productionFig. 2 Focus on the use of biomass e.g. for making optimalamounts of high qualitative energy carriers with heat as aresidue (it could also e.g. be biomaterials production).At the same time we can also expect:Increased competition for bio fuel resourcesHigher prices on high quality energy carriers(electricity and fuels) might drive towards smallerfraction as heat.Increased energy efficiency in industrial processes.With strategic planning district heating might utilizeresidual heat from processes producing combinationsof high quality energy carriers (or bio based materialproduction). The focus can probably not be on heatproduction. Even combined heat and power productionfrom bio fuels might not be efficient enough forcompetitive district heating (Fig 2).Fig. 1 Possible change for district heating demand infuture – decreasing demand but more even over the year.a) b)c) d)Fig. 3 Illustration of the need for a systemic perspective in planning the details of the energy system; a): A CHP plant and apotential energy customer (building); b): A CHP plant delivering district heat and electricity to a customer; c): A power plantdelivering 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 andelectricity to a customer with passive house standard (less total primary energy demand than in the b case).203

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTHE OVERALL ENERGY SYSTEMThe energy system of a country is complex, and it isimportant to understand how changes in sub systemsmay affect the whole system. Sub optimizations mighteasily occur. A simplified example of a situation wherea more energy efficient building through suboptimization of the total system gives a larger overallprimary energy need is illustrated in Fig. 3. Obviously itis possible to create a system with higherenvironmental impacts with energy efficient buildingscompared to a system with less energy efficientbuildings. It is not enough that the individual parts of asystem are good and efficient to give a lowenvironmental impact; the parts must be connected intothe system in a good way.Thus it is important to identify system solutions thatavoids sub optimization and gives us energy efficientbuildings and an efficient energy system with a goodenvironmental performance.IMPLICATIONS OF NEW TYPES HEAT LOADTo better understand implications of different new typesof heat load (as illustrated in the right hand side ofFigure 1) a life cycle assessment (LCA) has beenperformed regarding the use of heat instead ofelectricity for the three examples of house holdappliances: dish washer, washing machine and tumbledrier. Basic data regarding the appliances areexemplified with those in the ―district heating villa‖ inGöteborg, Sweden. The LCA model includes energyproduction (electricity or/and heat) for an average useof each machine and the materials needed to produceit. Different types of energy mixes for electricity anddistrict heat generation were studied. Details of thesystem boundaries and data can be found in the fullreport of the study [2].The results indicate that the total energy systeminfluences the results greatly. If we consider electricityproduction with large environmental impacts, to utilizedistrict heating is a good alternative, even in caseswhere the district heating generation in itself is notoptimally environmentally friendly. This is exemplified inFig. 4 where we consider Swedish average districtheating fuel mix (bio and residue heat, but also fossilfuels and some peat [5]) and European averageelectricity generation. If we for the long termdevelopment consider electricity generation that ismuch less fossil carbon intensive and compare it withdistrict heating based on forest bio fuels the results aremuch more narrow, and it become important whatenvironmental impact category is considered. In Fig. 5this is exemplified with climate impact and acidificationimpact.If district heating should continue to be seen in generalas an environmentally preferable option it is importantthat district heating companies continue to developdistrict heating production in a favourable direction.Heat for district heating should originate from resourcesthat are otherwise wasted. In the long term that willmean that bio fuelled district heating is not enough, butheat from other primary production like bio energy orbiomaterial combines producing transport fuels and/orbio based materials.CONCLUSIONSFrom environmental perspective energy efficientbuildings and district heating don‘t oppose each other– good parts connected in a good system will give anoptimal. It is not enough that the individual parts of asystem are good and efficient to give a lowenvironmental impact; the parts must be connected intothe system in a good way. The results from the study ofthe three items of household equipment showpossibilities for district heating to be an alternative withgood environmental performance, but not under allheat generation regimes. Heat generation mustcontinuously be considered.Fig. 4 Environmental impact from using district heat fordishwasher, drier and washer. Case: Swedish av. districtheating and European av. electricity.204

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaREFERENCES[1] Ingrid Nyström, Martin Eliasson, TorbjörnLindholm, Morgan Fröling, Jan-Olof Dahlenbäck,Erik Ahlgren and Elsa Fahlén (2009): Energieffektivbebyggelse och fjärrvärme i framtiden (in Swedish:Energy efficient built environment and districtheating in future). Swedish District HeatingAssociation, Stockholm, Sweden. Available as pdffrom www.svenskfjarrvarme.se[2] Morgan Fröling and Ingrid Nyström (2009):Miljöpåverkan från energieffektiva hus ochalternativ värme- eller elanvändning (in Swedish:Environmental impacts from energy efficientbuildings and alternative heat or electricity use).Published in [2].[3] Morgan Fröling; Charlotte Reidhav; Jan-OlofDalenbäck and Sven Werner (2008): Is there a rolefor district heating in future cities with low energybuildings? 11th <strong>International</strong> <strong>Symposium</strong> on DistrictHeating and Cooling, August 31 to September 2,2008, Reykjavik, ICELANDFig. 5 Environmental impact from using district heat fordishwasher, drier and washer. Case: bio based districtheating production and Swedish av. electricity.ACKNOWLEDGEMENTFinancial support from the Knut and Alice Wallenbergfoundation and the Swedish District HeatingAssociation is gratefully acknowledged.[4] Göteborg Energi. Fjärrvärmehuset (published inSwedish; ―The district heating house‖). Brochure.Göteborg Energi AB.[5] Morgan Fröling (2004): Environmental limitationsfor the use of district heating when expandingdistribution into areas with low heat density. 9th<strong>International</strong> <strong>Symposium</strong> on District Heating andCooling, August 30-31, 2004, Espoo, Finland.205

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaADAPTIVE CONTROL OF RADIATOR SYSTEMS FOR A LOWEST POSSIBLERETURN TEMPERATUREP. Lauenburg and J. WollerstrandLund University, Faculty of Engineering, Department of Energy Science, Sweden,patrick.lauenburg@energy.lth.seABSTRACTThe present paper describes how the control of aradiator system connected to a district heating networkvia a heat exchanger can be optimised to provide thelowest possible district heating return temperature. Thiscan be achieved for each operating point by employingan optimal combination of radiator circuit supplytemperature and circulation flow rate. The controlalgorithm gradually creates a modified control curve forthe radiator circuit, enabling it to consistently providean optimal cooling of the district heating water. Sincethe heat exchanger is dimensioned for very low outdoortemperatures, it is oversized for all other heat loads. Inaddition, radiator systems are often oversized due tosafety margins. Such facts render it possible to reducethe district heating return temperature.The objective of the present study was to develop acontrol algorithm and to test it in practice. A descriptionis here given of the algorithm, as well as of field teststhat were carried out to practically verify it. The controlmethod could be implemented in any modern controllogics for adaptive control of a radiator circuit, and theobtained results indicated that one can expect alowering of the return temperature in line with previoustheoretical calculations.oversized for all other heat loads. In addition, radiatorsystems are generally also oversized for safetyreasons, as presented in both Swedish studies [3], [12]and international ones [5], [8] and [10], thus providingfurther potential to reduce the return temperature.ObjectiveThe objective of the study was to develop a controlalgorithm for determining the optimal choice of supplytemperature and flow in an arbitrary radiator system forevery heat load in order to minimise the primary returntemperature.LimitationThe present investigation has dealt with DHsubstations that were indirectly connected to theDH network, i.e., hydraulically separated by HEXs.OPTIMISED HEATING SYSTEM TEMPERATURESThere exist various ways to control the heat output in aheating system. Here, we have dealt with the prevailingcontrol method used in Sweden; an outdoortemperature-compensated supply temperature,ensuring that an adequate amount of heat is suppliedto the building at each outdoor temperature.INTRODUCTIONThe present paper demonstrates how the control of aradiator system connected to a district heating (DH)network via a heat exchanger (HEX) can be optimisedto provide the lowest possible DH return temperature.This is done by always choosing the optimal radiatorsupply temperature and flow rate.Relevance of the topicLow return temperatures are beneficial for theproduction as well as the distribution of DH. A specificadvantage of the control method demonstrated in thispaper, as opposed to, for example, conventional lowflow balancing, is its robustness, enabling the lowestpossible return temperatures to be consistentlyobtained. This is the case independently of the currentoutdoor temperature and heat load, even if the DHsupply temperature changes, the HEX becomes fouled,or the house heating requirements change. The idea isalso to utilise the fact that, since a HEX is dimensionedfor an extremely low outdoor temperature, it is in fact206The benefits with regard to the primary returntemperature from adjusting the flow according to theheat load are known. The idea of using an optimalcombination of flow and supply temperature wasconceived by Frederiksen and Wollerstrand [2], andthis theory has been further studied [13] [11]. Theguidelines from Euroheat & Power [1] state that thelowest return temperature is obtained by varying theflow according to the consumption. If such a variableflow is used, it is controlled by thermostatic radiatorvalves (TRV) either in combination with a constantsupply temperature or with an outdoor temperaturecompensatedsupply temperature. Langendries [4]suggests a central control of the flow rate through thepump‘s rotating speed, but claims that it appears to bea rather difficult and expensive system. Petitjean [9]proposes a lowering of the pump speed at low heatloads, when the TRVs are almost fully open, but finds itproblematic to determine which parameter to use forcontrolling the pump speed.It should be possible to implement the control algorithmpresented in this paper in any modern, state-of-the-art

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniacontrol logics for building automation, which are todayoften used for controlling DH substations. The controlmethod suggests how the flow can be determined foreach heat load. The flow is regulated by adjusting thepump‘s rotating speed. Speed-controlled pumps arecommonly used nowadays and they provide a superiorcontrollability [1], [10].Table 1: A summary of flow-weighted mean primary returntemperatures (bold) and resulting reduction for varioustemperature programmes.Let us first study an example of an optimal controlcurve for a 100 % oversized system. Such a curve ispresented in Fig. 1, which also shows the relativemagnitude of the varying radiator flow in relation to therequired flow. The blue dashed line in the diagramcorresponds to the primary return temperature. For thesake of comparison, the primary return temperature fora 55/45 °C system is also shown (gray dashed line).Temperature [ C]Rel. flow [%]10090807060504030755025Primary return temperature reduction0-15 -10 -5 0 5 10 15Outdoor temperature [C]T p,sT p,r,optT s,s,optT s,r,optT p,r,55/45Fig. 1 Temperatures with an optimised temperature curveand a variable flow in a 100% oversized system. Theprimary return temperature from a 55/45 °C programme isshown for comparison.Flow-weighted, yearly mean primary returntemperatures from the radiator HEX have beencalculated with regard to the outdoor temperatureduration. Above the dashed line in Table 1, results areshown for a correctly dimensioned system, with an80/60°C programme as well as with an optimisedprogramme. The gain is estimated to just under twodegrees C. The last column shows how the primaryreturn temperature is affected when the length of theHEX is doubled. This comparison can be justified bythe fact that the primary return temperature issignificantly influenced by the lower secondary flow thatthe optimisation entails, while the pressure drop andheat transfer rate in the HEX can remain at amagnitude close to the original ones.m sUnder the dashed line, results are shown for a systemthat is oversized by 100 %. The first three temperatureprogrammes are 55/45, 60/40 and 80/30 °C, whereasthe last two are optimised ones with variable flow.The following conclusions could be drawn from thetable: The oversizing of a radiator system leads, in itself,to a significant reduction of the primary returntemperature, provided that some kind ofcompensation has been made in order for thesystem to work properly, i.e., that an accurateindoor temperature has been provided. By optimising the system (through the use of avariable secondary flow), the primary returntemperature can be further reduced, especially ifthe system is oversized. By extending the radiator HEX, the returntemperature can be further reduced with thetemperature programmes that employ a relativelylow flow. Regardless of the degree of oversizing, acombination of an optimised temperatureprogramme and an extended HEX provides asubstantially reduced primary return temperature.The values presented in the table have been calculatedonly for the radiator HEX. When considering thesubstation‘s total return temperature, it can be said tobe smoothed by the DHW consumption. Calculationscorresponding to those in Table 1 for a parallel and a 2-stage substation for 20 flats (based on the SwedishDistrict Heating Association‘s recommendations forsizing) result in reductions in the return temperaturethat are approximately 20 % lower than the valuesshown in the table. The difference between the paralleland the 2-stage connection is negligible when thereturn temperature from the radiator HEX is low ormoderate, a fact that has been previously207

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniademonstrated [6] [3]. Euroheat and Power recommendthat a 2-stage connection be used only in large multiresidentialbuildings if the primary radiator returntemperature is high. However, it should not beemployed if a low-flow heating system providing lowreturn temperatures is used [1].The advantage of extending the HEX when thesecondary flow is low actually demonstrates theoptimisation problem: When the secondary flow isreduced, the secondary return temperature willdecrease. In the radiator HEX, the situation is different.As the secondary flow decreases, the differencebetween the primary and the secondary returntemperatures, increases as a result of the heat transfercoefficient in the HEX being strongly flow dependent.Fig. 2 shows how the secondary return temperature islowered with a decreasing secondary flow while thedifference between primary and secondary returntemperatures increases. This results in a primary returntemperature that, at first, decreases and then increaseswhen the secondary flow is further reduced. The valuesin the figure have been taken from one of the testobjects. For this heat load, the lowest primary returntemperature was achieved for a secondary flow ofapproximately 30 % of the original flow.Return temperature, °C4039383736353433Tp,r,radTs,rGrädigkeitOptimum, lowest T p,r15 20 25 30 35 40 45 50Flow, %Fig. 2 Primary and secondary return temperatures, as wellas the difference between them, as functions of theradiator flow.Another reason for including the impact of an extendedHEX in the comparison in Table 1 is the opportunity ofconnecting to new installations. Large parts of thehousing stock in Sweden, built under strong politicalincentives during the 1960s and 1970s, are facingsubstantial renovation needs. The results of this projectcan be considered consistent even if fewer radiatorsystems be oversized in the future, whetherincorporated in older, renovated, or new buildings. Thesmaller potential for return temperature reductionsresulting from less oversized radiator systems may becompensated by the ability to install a HEX that isdimensioned for of an optimised radiator programme,76543210Grädigkeit, °Ci.e., a longer HEX. Furthermore, with optimised control,there exists a preparedness for future changes insystem temperatures in the DH network. Should theDH supply temperature be changed, an adaptivecontrol will ensure that the lowest possible returntemperature is always achieved.In order to operate according to Fig. 1, the algorithmmust combine a control of the radiator supplytemperature with a control of the radiator flow as afunction of the heat load and the DH supplytemperature. In previous work [7], we have shown thatit is possible to manually determine the optimal radiatorsupply temperature and flow. A natural continuation isto develop a method for automatic adjustment ofparameter values for the optimal control algorithm.THE TEST OBJECTSThe tests have been carried out in four multi-residentialbuildings in the city of Karlshamn, Sweden. The houseswere built in 1967-1968: three of them had three storiesand a basement, and one had six stories and abasement. The number of flats varied between 20 and30 per house.The radiators in all houses were fitted with TRVs, butthese were at least ten years old. It was thus uncertainwhether they functioned properly. The circulation flowwas found not to vary significantly in any of the radiatorcircuits, which may have been an indication that manyof the TRVs were not working. However, it should benoted that the presented control algorithm isindependent of the use of TRVs in a system. Whatevercombination of optimal supply temperature and flowthat is identified for a given outdoor temperature, theheat supply will be the same. The main task for TRVsis to limit the heat supply in a room where additionalheat supply (solar radiation, bodily warmth or electricalequipment) would result in an overheating of the room.The substations were of the 2-stage type and equippedwith control logics of the brand IQ Heat (Alfa Laval AB).The equipment for the building automation wasmanufactured by Siemens and furnished with aseparate communications module that could also beused for executing minor computer programmes. Therewas also an internet connection, rendering it possibleto communicate in a number of ways, such as via thesoftware Saphir ScopeMeter© (Siemens), or FTP. Aftera reconfiguration, the pump speed could be controlled,since all pumps were equipped with communicationmodules.In order to monitor the circulation flow in the radiatorcircuits during the tests, clamp-on ultrasonic flowmeterswere utilised. However, the objective was todevelop a control algorithm based on modern, state-ofthe-artequipment without using additional installations.208

To assure that the temperatures measured in thesubstation corresponded to the average temperaturelevels in the various risers in the radiator circuits,temperature sensors were installed in two of thehouses. This enabled measurement errors ordisturbances in the radiator circuit to be identified. Theindoor temperature could be monitored thanks to sixwireless sensors installed in each house in the area.Modifications in the substationsAfter some initial tests, the circulation pumps werefound to be generally oversized to such an extent thatthe flow rate could not be decreased as much asdesired. There exists a predetermined minimumrotational speed for this type of pump, implying that thespeed could be reduced by 60–70%. Discussions withthe manufacturer revealed that the lowest pump speedcould not be changed in this model, for which reasonthe decision was made to throttle the flow with anexisting shut-off valve located after the pump, whichshifted the pump‘s operating range. The throttling wasconducted in order for the pump to give half the flowrate at 100% rotational speed. The control curve wasmodified accordingly, leading to the temperature dropin the radiator circuit becoming doubled and the heatsupply remaining unaltered.We were unable to receive a comprehensive reply fromthe pump manufacturer with respect to the possiblemeasures regarding the regulation of the pump. Adiscussion with another manufacturer implied that therewere no technical limitations for how far down thepump speed could be controlled. However, such anextension of the manoeuvrable range has so far notbeen requested. After a simple modification of thepump‘s frequency converter, the working range couldbe extended from today‘s 30–100% to, in an extremecase, 2–100%.Existing control of the radiator circuitsAlthough the radiator circuits within the area weredesigned by the same consultant, there is today a largespread in the choice of control curve and resultanttemperature drop (10–30 °C). It is likely that the curveshave been gradually adapted to the circuits‘ hydraulicproperties and balancing, and one can assume that thisis a common situation.When older houses are renovated and their radiatorcircuits are modernised, there are no guarantees thatoversizing is taken into consideration. For example, theradiator HEX in a substation that was installed in 2005in one of the houses was dimensioned for 185 kW heatoutput at DOT with temperatures corresponding to80/60 °C at a flow of 2.25 l/s. However, whenexamining data for this substation, it turned out that thesubstation delivered less than 40 kW at an outdoorThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia209temperature around 0 °C, which corresponded to aload of approximately 50%. The actual flow rate wasabout 1.1 l/s and the temperatures corresponded to60/40 °C, thus representing an oversizing around100%.ADAPTIVE OPTIMISATION - METHODIn the theoretical example, the system was assumed tobe 100 % oversized, while in an arbitrary system onecannot be sure of the degree of oversizing. It is alsodesirable to have a robust and adaptive controlalgorithm. The method found to function the best isdescribed below. This approach consists in graduallymodifying, by automatically performed tests, the controlcurve and determining the associated flow rate.Online testingBy locking the control valve (CV), one can assume tohave approximately the same primary flow through theradiator HEX, and since the variations in the cooling ofprimary water is relatively small, the heat supply is alsoapproximately constant. If the secondary flow isreduced while the CV is maintained locked, thetemperature of the secondary flow leaving the HEX willrise. When a new flow and its associated supplytemperature are tested, the current level of the primaryreturn temperature is compared to the level before theexperiment. In this way, the new combination of flowand supply temperature can be either accepted orrejected. This method renders it possible to implementthe adaptive algorithm in any arbitrary system, leadingto the control curve becoming gradually modified. Thismethod we suggested in [7].One problem associated with this kind of optimisation isthat the method is sensitive to disturbances. If theprimary supply temperature, primary differentialpressure or the outdoor temperature changes duringthe test, one cannot be sure that the heat supply isconstant. In that case, a reduced return temperaturecould be the result of a heat supply that is too low.Such tests have to be rejected.In order to render the tests less sensitive todisturbances, the CV is locked only briefly, in order forthe HEX to stabilise. Subsequently, we return toautomatic control, but instead of using the controlcurve, the control aims at maintaining a constanttemperature drop in the radiator system. If this issuccessful, the heat supply is also kept constant. Onecan assume that the secondary flow is relativelyconstant: as long as tests are conducted at night, nosolar radiation is present and internally generated heatis likely to be at a relatively steady level. If, for instance,the primary supply temperature or differential pressurerises during the course of a test, the CV will closesomewhat causing the secondary supply temperature

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniato decrease, and thereby also the temperature dropand heat supply, to be detained at the same level.A test is started by keeping the CV locked for tenminutes. This leaves enough time for the HEX tostabilise. The new level of the difference between theprimary and secondary return temperatures becamestable already after about two minutes in the testedobjects. The CV was maintained locked for tenminutes, which should be sufficient even for very lowflows and most types of HEXs. Subsequently, thecontrol was resumed in order to ensure a constanttemperature drop on the secondary side.The temperature drop was controlled by verifying thecurrent temperature drop, e.g., every five minutes, andcomparing it with the desired temperature drop, i.e., thetemperature that was observed when the CV waslocked. If the difference exceeded a certain value,0.2 °C has been used so far, the set-point for thesupply temperature was updated according toT setpoint = T s,r + T setpoint .Fig. 3 displays a performed test: At 1:00 a.m., the CVwas locked and the radiator flow rate was reduced from0.59 to 0.36 l/s with the result that the secondarysupply temperature rose from 40 to 44 °C. After tenminutes, the temperature drop in the radiator circuitwas automatically controlled (in this case, thetemperature drop was stable and it took more than 15minutes before the CV opening degree requiredadjustment). After ninety minutes, the second flowreduction was carried out, to 0.24 l/s, and thesecondary supply temperature increased to about48 °C.The total primary return temperature varied to arelatively large extent, partly because of tappings ofdomestic hot water (DHW), but also due to the DHWcontrol in this substation being very unstable when notappings were made. However, the return temperaturefrom the radiator HEX was of interest for the tests. Inthis object, the difference between the primary andsecondary return temperatures was very small, andeven for a low radiator flow, the grädigkeit was belowone degree. One can see from the figure that the returntemperature had fallen from just under 32 °C to slightlyover 28 °C during the test. This resulted in, for acurrent outdoor temperature of 8 °C, the set-point forthe secondary supply temperature being changed from40 to 48 °C while the flow should be reduced from 0.59to 0.24 l/s.Temperature [ C]Flow [l/s]%Heat supply [kW]8070605040302010000:30 01:00 01:30 02:00 02:30 03:00 03:30 04:00Time2010010.750.50.250604020CV,heatCV,DHW000:30 01:00 01:30 02:00 02:30 03:00 03:30 04:00TimeT p,sT s,rT p,r,radT p,r,totT s,rT sT oT o,dampFig. 3 Results from a test. The flow was reduced at 1:00and 2:30. The top graph shows temperatures in thesubstation, the next graph presents the valve position forheat 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 primarysupply temperature fluctuated a lot. Since thesecondary temperature drop was kept constant, it hadno impact on the outcome of the test. One can see thatthe CV generally demonstrated a lower opening degreelater in the night, as opposed to before 1:00, when theprimary supply temperature increased. Without the Tcontrol, the heat supply would have been too highduring the last part of the test.The radiator flow was altered by changing the set-pointfor the pump speed, expressed as a percentage of themaximum speed. It has been found that two flowalterations of ninety minutes each are suitable per test,as this would allow the secondary return temperature tostabilise even at very low flows. The first test for anyoutdoor temperature, as was the case in Fig. 3, meansthat starting conditions include the original controlcurve and flow rate. It is then desirable to perform twofairly large flow reductions since, according to thetheoretical calculations, one can expect to find anoptimum at a relatively low flow. If, however, the flow isalready on a low level, it is reasonable to attempt oneslightly higher and one slightly lower flow rate. Thealgorithm for the adaptive control is illustrated by theflow chart in Fig. 4.m pQ pm sQ sGr210

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaYes11:50 PM < time < 0:00 AMWait 60 min withconstant pump speedStart timerWait 5 min< 80 min?NoSave currentvaluesSet controlvalve to AutoNo,> 80 minGet T o , usemodified controlcurveSet pump speedfor testSet TPump speed 1?YesWait 10 minTest pumpspeed 1Wait 10 minTest pumpspeed 2Set controlvalve to Manualpoint, one could expect a stable secondary returntemperature, e.g., during the last five minutes. Inaddition to the secondary supply temperature, also theprimary supply temperature is recorded. However, thedampened outdoor temperature, i.e., the input signal tothe controller, is recorded when the CV is locked for thefirst time. The reason for this is that the heat supply issubsequently kept constant at a level matching theoutdoor temperature (and heat load) at the time beforethe test was started.NoYes No, pump speed 2T – T set-p > 0.2°C Yes T set-p = T s,r + T set-pReject testresultUpdatecurvesNot okOkCheck maximumdeviation for Q(e.g., 5%) and To(e.g., 2°C)Test doneDetermine T p,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 aboutto start, the control should be interrupted and the pumpspeed kept constant for an hour prior to the test. Thisway, one avoids the risk of the flow changing (due toalterations in the outdoor temperature) too close to thetest, which could result in unstable radiator systemtemperatures.The supply and return temperatures were measured onfour of the most remote risers from the substation,during the tests. A continuous matching againstmeasurements on risers gives a good indication thatthe flow distribution in the system was not impaired bythe optimisation. The temperature profile was closelymatched to the profile at the substation. Both flowreductions resulted in increased temperature drops.Updating the control curvesAfter the completion of a test, the obtained informationneeds to be evaluated. The influence of the variation ofthe outdoor temperature is not entirely obvious; itsinfluence decreases with an increasing time constantfor the building. Variations on the primary side normallyhave is compensated for since the heat supply is keptconstant. As a result, it is sufficient to verify that theheat supply was maintained at a steady level during thetest, avoiding any disruptions.If a test result is accepted, the primary returntemperatures for each tested flow are compared inorder to verify which flow resulted in the lowest returntemperature. This flow also gave rise to a secondarysupply temperature. It is however not obvious how toread this temperature, given that it was regulated bythe controller and changed continuously. The mostlogical choice is to read the mean value at the end ofthe test period, before the pump speed changes. At thisThe next step consists in using the information attainedfrom the test to modify the control curves. Initially, theoriginal curve was used and the pump was, in ourcase, controlled to give a constant differential pressure.If the result of a test is that a lower primary returntemperature is obtained at a lower secondary flow rate,the control curve is updated for that outdoortemperature. A reasonable resolution is 1 °C. Theoriginal control curve, generally based on 5–8 points,was therefore initially extended to comprise values foreach outdoor temperature.If the experiment, as in Fig. 3 above, was performed at8 °C, this point on the curve would be updated. Alongwith the new supply temperature there followed a newradiator flow, which in our case was expressed as anew set-point for the pump speed.The adaptive control continues in this manner nightafter night, and the control curves are continuouslyupdated. Outside the test periods of approximatelythree hours each night, the modified control curves areused for controlling the heating system.Fig. 5 shows an example of the gradual development ofthe modified control curve. The first graph shows a newpoint at 0 °C (used for 0 ± 0.5 °C). In the second(upper) graph, a point for 3 °C has been added, whilethe range 0 to 3 °C is complete in the third. The fourthgraph shows a much more complete control curve(-5 to 10 °C). Temperature curves corresponding toconstant flow systems with lower flows than the originalsystem have been included as thinner lines. The valuefor 10 °C coincides with the curves of a system with alow flow, while the value of -5 °C coincides with thecurves of a system with a moderately reduced flow(normal flow). The last graph clearly demonstrates thatthe modified curves are based on a variable flow, i.e.,they coincide with various constant flow curves atdifferent points.As shown in the second graph of Fig. 5, the modifiedcurve could emerge in sections that subsequently arecombined. One way to speed up the modification of thecontrol curves is to interpolate intermediate valuesrather than wait for a flow optimisation at the missingoutdoor temperature. Even the return temperaturescould be interpolated, since it is possible to determine211

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniathe required radiator flow for a known temperature drop(and heat supply).where T s,r,n is determined in analogy with T s,s,n ,according to:7070Temperature605040OriginalcontrolcurveModifiedcontrol curveConstant,decreased flowConstant, furtherdecreased flowTemperature605040T Ts,, test s,r,n1T, ,rs r n(3)23030Temperature20-10 -5 0 5 10 15Outdoor temperature7060504030Temperature20-10 -5 0 5 10 15Outdoor temperature7060504030To ensure that the heat supply is kept constant, therequired flow for the new temperature drop iscalculated. Since the flow is inversely proportional tothe temperature drop, it can be determined from thelast used flow and temperature drop, together with thenew temperature drop, according to:20-10 -5 0 5 10 15Outdoor temperature20-10 -5 0 5 10 15Outdoor temperatureFig. 5 A stepwise modification of the control curve. Thesupply temperatures are drawn in solid lines while thereturns are dashed.For the first test to be carried out at a specific outdoortemperature, it is logical to let the results of this testfully replace the original points on the curve. As moretests are performed for the same outdoor temperature,one can proceed in several ways. Since the controlshould be adaptive and thus able to take into accountchanging circumstances both in the DH network and inthe building, the results of new tests should beemployed. However, one may expect that testsperformed close to one another in time, at equivalentoutdoor temperatures, still provide slightly differingresults for varying reasons. A solution would thereforebe to use a forgetting factor, i.e., to gradually ―forget‖old values when the supply temperature curve isupdated with new data. A possible approach for doingso consists in calculating the new supply temperature,T s,s,n , as a mean value of the obtained, T s,s,test , and thelast used, T s,s,n-1 , supply temperature according to:( m T)s n1s,n(4)Ts, nmAs mentioned earlier, the flow rate is set by changingthe set-point for the pump speed. According to theaffinity laws for fluid machines, the flow is proportionalto the rotational speed. The process of letting the lastmodified supply temperature and the result of a newtest form a new modified supply temperature isillustrated in Fig. 6.Temperature60504030OriginalcurvesT s,s,testT s,s,nModifiedcurvesT Ts,, test s,, n1T, ,sss s n(1)220T s,r,testT s,r,n2 4 6 8Outdoor temperatureWhen a new test is performed at the same outdoortemperature, a new mean value is calculated, whichmeans that older values will have less and lessinfluence. To determine the secondary flow associatedwith the new supply temperature, i.e., the one providingthe correct heat supply at the current outdoortemperature, the expected temperature drop iscalculated as:Ts, nTs, s,n Ts, r,n (2)Fig. 6. An approach for modifying the control curve basedon new test results.The proposed method for updating the control curvesindicates that if for instance the DH utility demonstratesa long-term change in the supply temperature in thenetwork, the control system gradually adapts to thenew temperature. However, there are always variationsin the primary supply temperature. This may includeboth unintended and intended variations which may bethe result of, for example, a charging of the network ifthe outdoor temperature is expected to fall. Since theprimary supply temperature affects the primary return212

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniatemperature, it is desirable for the adaptive control toalso compensate for such short-term variations. Oneway of doing so is to develop a number of parallelcontrol curves for various intervals of the primarysupply temperature. If the temperature is greater than acertain level, an alternative control curve is employed,whereas if it is below a certain level, one utilisesanother. This method has yet to be tested and there isno basis for assessing how much impact one canexpect from normal variations in the supplytemperature or what would constitute reasonableintervals for parallel control curves in this case. Anothervariant could be to perform a linear adjustment for thesecondary supply temperature depending on theprimary supply temperature, according to:Ts, s s,s,0( p,s,0p,s T 1a(T T))(5)where a is a constant that can be determined fromtests.Regarding the measurement of temperatures and flowsRegarding the temperature measurement in thesubstation, supply and return temperatures on both theprimary and the secondary sides are required. Oneshould keep in mind that, on the primary side, thereturn temperature from the radiator HEX is neededsince the total return temperature is affected by theDHW system. This temperature is normally available inmodern substation control equipment.It is desirable to avoid installation of a flow-meter in thesecondary circuit. On the primary side, where theenergy-meter is located, the total primary flow and thetotal temperature drop in the substation are measuredand the energy required for DHW provision is thusincluded. Since the tests are performed at night, DHWtappings can be avoided to a large extent. By closingthe DHW CV for a short time, the primary flow passesexclusively through the radiator HEX. By comparing theaverage level of heat supply with a closed valve to thelevel prior to closing the valve, the flow required forDHW re-circulation can be estimated.In the test objects, indoor temperature measurementswere used to verify that the adaptive control was ableto give the correct indoor temperature. However, onecan in fact be sure that the correct amount of energy istransferred to the system for each operating point,regardless of whether the original control curve or theoptimised curve is used. A possibility is that there is animbalance in the system. For example, the most distantriser may not receive the required flow because of atoo low differential pressure when the pump speed isdecreased. It is, however, more likely that a betterbalance in the system is achieved when the differentialpressure is lowered this since the pressure losses inthe system decreases and all risers receive a moresimilar differential pressure. However, one must be onthe look-out for errors (e.g., short circuits) in thesystems, a problem that is often emphasised inconnection with low-flow systems, as these tend to bemore sensitive to hydraulic imperfections [12].Reduction of the primary return temperatureTo estimate a yearly mean return temperaturereduction (as presented in Table 1) achieved by theadaptive control, an entire, or a major part of the,heating season needs to be evaluated. The controlmethod presented in this paper was developed duringthe winter and spring of 2009, and only a limitednumber of tests were performed during the spring.However, Fig. 7 shows the obtained primary returntemperature that was attained for the tests that wereperformed in one of the houses. Note that these resultswere ―first runs‖ for each outdoor temperature (i.e., theflow was reduced to approximately 40%), signifyingthat no further optimisations were undertaken. Thecurve displaying the original return temperatures wasbased on the average return temperatures from theradiator system prior to any of the modifications (i.e.,for the tests or the constant flow rate change, asdescribed in section 3.1).Primary return temperature50454035302520Tp,r,rad,origTp,r,rad,opt-10 -5 0 5 10 15Outdoor temperatureFig. 7. Primary return temperatures in the radiator systemwhen the flow is reduced (dots), compared to the originalreturn temperatures (curve).CONCLUSIONS AND DISCUSSIONAn adaptive control algorithm was developed in orderto minimise the DH return temperature. The controlalgorithm can be implemented in any modern controllogics for building automation. Some refinement maybe done by compensating for short-term temperaturevariations in the DH network. During the field studies,limitations in the speed control of the circulation pumpshave presented a complication. A modification of thefrequency converter could increase the working range.213

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThere was not enough time to develop completelymodified control curves for the test objects during thepresent heating season. On the other hand, a controlcurve with an adaptive controller is never definitive;rather it increases as more operational points (differentoutdoor temperatures) are added and is then graduallymodified if outer conditions change. In order to receivevalues for the primary return temperature on a yearlybasis using the adaptive control algorithm, the newcontrol curve needs to be modified for the entiretemperature range. During the performed field studies,the reduction of the primary return temperature wasabout 3 °C. Even though the test period limited thenumber of tests, the temperature range was still ratherwide, including temperatures from -2 to 14 °C.It is plausible that certain circuits are more suitable fora variable flow rate, e.g., depending on hydraulicbalancing. It would also be possible to map out underwhich circumstances other heat emitters than radiators,such as fan coil heaters, can be included in a radiatorcircuit where the flow varies.ACKNOWLEDGEMENTThe Swedish District Heating Association, the SwedishEnergy Agency and Nordic Energy Research aregratefully acknowledged for financing this work.REFERENCES[1] Euroheat & Power, Guidelines for District HeatingSubstations, Downloaded from:http://www.euroheat.org/documents/Guidelines%20District%20Heating%20Substations.pdf,20081117.[2] Frederiksen, S., Wollerstrand, J., Performance ofdistrict heating house station in altered operationalmodes, 23rd UNICHAL-Congress, Berlin, 1987.[3] Gummérus, P., Petersson, S., Robust Fjärrvärmecentral(Robust District Heating Substation), ReportA 99-223, Dept. of Energy and Environment,Chalmers Univ. of Technology, Gothenburg, 1999.[5] Liao, Z., Swainson, M., Dexter, A.L., On the controlof heating systems in the UK, Building andEnvironment 40 (2005) 343-351.[6] Lindkvist, H., Walletun, H., Teknisk utvärdering avgamla och nya fjärrvärmecentraler i Slagsta(Technical evaluation of old and new districtheating 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 theaim of lowering the primary return temperature,Proceedings from 11th <strong>International</strong> <strong>Symposium</strong> onDistrict Heating and Cooling, Reykjavik, 2008.[8] Peeters, L., Van der Veken, J., Hens, H., Helsen,L., D‘haeseleer, W., Control of heating systems inresidential buildings: Current practice, Energy andBuildings 40 (2008) 1446-1455.[9] Petitjean, R., Total hydronic balancing, Tour &Andersson Hydronics AB, Ljung, Sweden, 1995.[10] Skagestad, B., Mildenstein, P., District Heating andCooling Connection Handbook, published by the<strong>International</strong> Energy Agency (R & D Programme onDistrict Heating and Cooling), 2002.[11] Snoek, C., Yang, L., Frederiksen, S., Korsman, H.,Optimization of District Heating Systems byMaximizing Building Heating System TemperatureDifferences, Report 2002:S2, <strong>International</strong> EnergyAgency (R & D Programme on District Heating andCooling) & NOVEM, Sittard, 2002.[12] Trüschel, A., Hydronic Heating Systems – TheEffect Of Design On System Sensitivity, DoctoralThesis, Chalmers University of Technology,Gothenburg, Sweden, 2002.[13] Volla, R., Ulseth, R., Stang, J., Frederiksen, S.,Johnson, A., Besant, R., Efficient substations andinstallations, Report 1996:N5, <strong>International</strong> EnergyAgency (R & D Programme on DHC) & NOVEM,Sittard, The Netherlands, 1996.[4] Langendries, R., Low Return Temperature (LRT) inDistrict Heating, Energy and Buildings, 12 (1988)191-200.214

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaPOLICIES AND BARRIERS FOR DISTRICT HEATING AND COOLINGOUTSIDE EU COUNTRIESA. Nuorkivi 1 and B. Kalkum 21 Energy-AN Consulting2 Energy & Utility ConsultingABSTRACTThe policies and barriers faced by DHC in the countriesoutside the EU will be investigated during 2010–2011as a part of the Annex IX of the IEA ImplementingAgreement on District Heating and Cooling (DHC),including the integration of CHP.The countries to be covered are China, USA, Canada,South Korea, Russia and some other selectedEuropean countries outside the EU. The work is basedon both interviews of the key officers and specialistsand the existing laws, regulations and policies of eachselected country. The project will also provideexamples of best practices useful for sustainabledevelopment of DHC as well as offer recommendationsto the countries to improve the institutional set up of theDHC.Regarding each country, the project will review, forinstance, the tariff setting, DHC related legislation,taxation rules, price regulation, customer definition andpoints of delivery; ownership of fixed assets; allocationof CHP costs and environmental fees; socialconsiderations; municipal heat planning; and, heatmetering and control.The project here is a twin project to EcoHeat4EU that isa thorough analysis of the barriers and opportunities ofDHC as well but in the selected EU member countries.INTRODUCTIONThere is no reliable statistics of DHC in most of thesubject countries. The countries are in different stagesof 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 fromall countries, including the EU that might be useful toboost DHC development in the particular subjectcountry. Nevertheless, the lessons learned andrecommendations will be developed in fall 2010, afterthe <strong>Symposium</strong>, and the final and complete study willbe available in May 2011. Therefore, all informationpresented in the paper regarding four countries,Canada, China, Ukraine and USA is based on thepreliminary survey that will be finalized byOctober 2010.PRELIMINARY COUNTRY SPECIFIC SURVEYS1. Canada1.1. Status of DHCThe old DH systems before 1985 are predominantlywith steam, whereas water systems have been builtsince 1985. Both domestic hot water (DHW) and spaceheating (SH) have been included. Based onwater/steam carrier, various combinations of heatingand cooling are available in Canada.Historically, Canada has had the highest per capitaenergy use of the developed countries, as a result ofthe harsh climate and relatively low-cost, abundantenergy. So the benefits of DHC would be particularlywelcome to save energy. In Canada, there are recordsof some 120-160 DHC systems in the country, andalmost a half of them located in Ontario Province alone.About 27 Mm 2 of residential, industrial and institutionalfloor area are connected to the DHC systems. Thisrepresents about 1,3% of all floor space in Canada.The largest DHC system is in Toronto with 522 MWthermal capacity.[1]Natural gas distribution has spread everywhere, whichis a challenge for DHC expansion. Moreover, atrelatively low electricity prices, there is a little marketfor CHP. No economic market for CHP exists inCanada unless the feed-in tariff is in place or theelectricity is used in-house of producer. Power and gasutilities have not been co-operating so far, becausethere has not been any incentive to such co-operation.Because of the structure of the provincial utilities andlow electricity prices, only a few CHP based DHCsystems are in operation.The utilities are empowered to provide the people withgas and electricity at the lowest costs possible.Economic drivers support the selection of the propertechnologies, and the provincial regulators ensure thatthe system availability and safety are maintained at alltimes. Provincial governments provide some directionsto the energy industry, but limit themselves to settingoverall goals only. The selection of the technologies isleft to the utilities. Natural gas is widely availablethroughout the country, which is a challenge for otherheating modes to enter the market. Serious lack of gasreserves is expected in the future, which meansalternative energy sources to become increasingly215

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniarealistic. To substitute natural gas, DHC based onbiomass and possibly with CHP is a superior option.For DH, two-tier tariffs are used in which energy fee ispass-through of energy costs, and the fixed fee coversthe profit, the connection costs and all other costexcept energy. The fixed fee can be adjustedannually/biannually with CPI (Consumer PriceIndex).The customer contracts are made for a longperiod, say 10-20 years, during which the capital costhave been discounted to the fixed fee. Municipalcompanies operate as non-profit but private companieswith reasonable profit.1.2. Market DriversIn Canada, the federal government is committed toreducing GHG emissions by 17% below 2005 levelsby 2020, being the main driver of DHC. The DHCmarket is expanding smoothly to start creating adifferent infrastructure to substitute depleting resourcesof natural gas.As mental drivers, there is strong interest inmunicipalities to consider DHC introduction and furtherexpansion very much based on European practise.Many municipalities have set voluntarily targets to thereduced GHG emissions. DHC systems are widelyrecognized as a potential measure to achieve thetargets. The DHC is considered a tool for the urbanplanners but not an energy issue per se.As an example of investment support, Ontario PowerAuthority (OPA) subsidizes investments in electricitysavings by paying up to $800/kW of the saved electriccapacity. The subsidy used to be 400/kW, but wasdoubled at the end of 2009. Customers can use thatmoney as the partial payment of the connection costsof DHC, thus DHC companies indirectly benefittingfrom the subsidy system as well.1.3. Main BarriersThere is no formal DHC strategy or policy supportingDHC and CHP development in Canada. TheGovernment does neither have the tradition nor thewillingness to take strong position in DHCdevelopment. The private sector that could bringinvestments and entrepreneurship cannot be muchinterested, because starting the DHC is risky: long paybacktimes ranging beyond 10 years, limited access tomunicipal property, challenging contracting ofresidential, municipal and federal buildings, overallbilling and collection of different types of customers.Nevertheless, the municipalities are rather weak,because the municipal taxation only covers propertyand tourism taxes but no corporate or income taxes.Moreover, municipalities have no mandate on energy.The federal government hesitates to take a strong rolewhile fearing of intervening the private sector drivenheating market.1.4. Current ActivitiesThe Integrated Community Energy Solutions (ICES)Roundtables have been established to accelerateprogress toward reducing GHG emissions by bringingtogether senior-level stakeholders to exchange viewson the best way forward from here. The Roundtablesbuild upon ICES. The Roadmap for Action, which wasreleased by the Canadian Council of Energy Ministersat its annual meeting in September 2009, describes therole that Canada's federal, provincial and territorialgovernments can play in advancing ICES and it setsout a broad strategy for action. It also includes a varietyof options from which the governments can choose,according to their priorities, to advance communityenergy performance and complement existing energyefficiency activities in different sectors.The ongoing collaboration of key energy actors andenablers across Canada from the private and publicsectors through the Quality Urban Energy Systems ofTomorrow (QUEST) collaborative also informed theRoundtable discussion. In particular, preliminary resultsfrom a QUEST-led study suggest that ICES couldreduce GHG emissions at the community level by asmuch as 40% to 50%, resulting in reduction of 65 Mtby 2020, which is about 20% of Canada's official2020 target reductions. These results are verypromising and highlight how ICES could contributesignificantly to improving Canada‘s energy and GHGperformance.2. P.R. China2.1. Status of DHCIn China, the DH development has been very strong,more than 10% annually during the past decade onaverage. By the end of 2005, DH supply (includingsteam and hot water) was over 2 100 PJ; of which CHPaccounted for 47% and boilers accounted for 51%.Inthe supply of steam and hot water, steam supply is 715PJ, of which CHP accounts for 81% and boilersaccount for 17%; the total hot water heating supply is1395 PJ, of which CHP accounts for 29% and boilersaccount for 69%. The heating supplied by CHP unitsand boilers are respectively 992 PJ and 1086 PJ.Apart from Europe, only SH is supplied by the DHsystems, and the DHW by individual systems: solarcollectors, propane, electricity, etc. [2,3]During the few years to come, China will become thelargest DH country in the world.216

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia2.2. Market DriversThe rational of strong DH development in China isbased on eliminating the small and polluting coal firedboilers in the northern, western and central provincesand to provide feasible living conditions to thepopulation massively moving in to the cities.DH has been encouraged by the Chinese governmentfor several decades. China's DH heating area hasincreased from over 276 Mm 2 in 1991 to over1100 Mm 2 in 2000, and exceeded 2500 Mm 2 in 2005,with an annual growth rate of 17%. The growth in DHmainly came from the northern and the northeastregions. In China, residential buildings account forabout 70% of the total DH area and commercialbuildings the balance of about 30%.The urban communities are very densely built, whicheffectively supports centralized heating and coolingsolutions. The new buildings comprise about half of theDH connections, whereas the balance for existingbuildings, the latter previously having had been heatedby small coal boilers.2.3. Main BarriersThe DHC sector is expanding fast but there are stillsome barriers regarding economy, policy, financing andtechnology as summarized below.ECONOMIC AND PRICING BARRIERSIn order to become cost-effective and an attractiveinvestment, power and heating reform policies will needto be undertaken. Some of the key issues include:Energy price policy reform is a priority. At present,in China, the coal price is based on the market,which has grown rapidly in recent years. However,electricity and heating prices are still controlled bythe government, and have only slightly increased.While the government has provided limitedsubsidies to DH companies, most CHP enterprisesand DH companies are currently not making aprofit as a result of the lack of energy price reform.In addition, heating reform needs to be furtherdeveloped. Currently, in most cases, heat tariffsare based on the building area, rather than on theactual heat consumption, which has a negativeinfluence on improving the energy efficiency indistrict heat facilities and buildings.Power sector reform is also needed. At present,the electricity produced by most DHC (and someCHP) projects cannot interconnect with the powergrid, which has strongly reduced development. Thetechnical issues of grid connection can likely beaddressed. However, there are also administrativeinterconnection issues, such as added-capacitycharges and power grid balancing that need to beaddressed. At present, the State Power Grid Groupis responsible for the power grid operation. Assuch, more communication and coordinationactivities could be conducted between the DHCindustries and the State Power Grid Group.Centralized DHW would benefit CHP. MissingDHW load hampers economic development ofCHP schemes. Without DHW, the CHP plants canoperate all year round only if there is industrialsteam load existing nearby.POLICY BARRIERSThere also exist barriers in the area of economicsupport and administrative policies related toCHP/DHC, including:There is a lack of monitoring and enforcement ofthe government‟s policies related to the efficientoperation of CHP projects. Currently, it appearsthat some newly- built CHP projects are operatingonly in thermal generation mode after they havebeen approved, thereby reducing their energyefficiency.There is a lack of targeted policy for smaller CHPunits. In order to fulfill the energy conservationtarget, China is attempting to increase the numberof more efficient large power generation plants andto close down smaller, older units. While it isimportant that the smaller, more inefficient units beclosed down, some small CHP units with highefficiency are also being targeted for phase-out.Based on the goal of increasing energy supplyefficiency, a different policy should be adopted. Forexample, in regions with low heating loads, smallCHP units could provide most of their energyneeds at a fraction of the cost of larger units.FINANCING BARRIERSThere are promising energy conservation projects– particularly in the DH sector – that could be realized ifthere were sufficient funds or other means available toaddress the gap in investment capital. In particular:Some planned CHP/DHC projects are not operatedefficiently because they lack sufficient resources toinvest in expanded heat pipeline infrastructure.Further, at many existing DHC projects, the heatloss in pipelines is high, reducing the overallefficiency of the heating system. Additionalfinancing is needed to invest in cost-effective heatpipeline retrofit projects, which will generatesizeable energy efficiency benefits and GHGreductions.217

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaWhile energy service companies are expanding inthe commercial building energy conservationarena, they have not yet entered the CHP/DHCarea. There is some room for these types of thirdpartyplayers to come up with innovative means tofinance projects.TECHNICAL BARRIERSWhile CHP/DHC are proven, existing technologies thatdo not require major research and development, thereare some advanced technologies that could beintroduced from IEA Member Countries to improveefficiency and operational benefits. In addition, there iscurrently some debate about the relative merits of DCtechnology. China-specific research studies could beconducted to confirm the primary energy conservationperformance of these technologies.ORGANIZATIONAL BARRIERSThere are some organizational barriers for optimaldevelopment as well.Scattered organizations with several heat suppliersand distributors prevail in one city. In the same DHsystem, the heat supplier is responsible foroperation and maintenance until the groupsubstations that serve several buildings throughthe secondary network, and the distributors beingresponsible from the substations to the indoorheating elements. Therefore, the holisticoptimization can be often compromised by partialoptimizations.The DHC companies are operation andmaintenance companies only, whereas investmentdecisions and financing depends on the municipaland provincial budgets. This is one more reason forthat there is little business minded atmosphere inthe extensively staffed DHC companies.2.4. Current ActivitiesThe DH systems are expanding fast in China,simultaneously restricting coal consumption andreducing overall GHG emissions of the heatingservices.The Ministry of Construction has issued the Housingand Building Reform on Energy Efficiency (HRBEE),which requires more efficient buildings to be built aswell as introduction of heat metering and consumptionbased billing. The first consumption based billing pilotwas initiated in Tianjin a few years ago with a two-tierheating tariff. Such billing systems are slowlyexpanding to other regions.3. Ukraine3.1. Status of DHCUkraine is one of the largest DH countries in Europe.Currently almost 80% of urban housing is supplied withDH 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 gasboilers or even remain without heating, thus enjoyingon the heat losses penetrating to them through wallsfree of charge from their heated neighbours. Even themunicipalities are offering investment subsidies to theapartment owners to purchase apartment level gasboilers while disconnecting the DH services.Such practices have led to extremely poor quality ofDH services: low water and room temperatures a wellas periodical heating are used to minimize fuel costs.There are coal (and anthracite) mines in Ukraine, butlittle used for providing fuel for DH: Most DH is basedon natural gas imported from Russia. The costs of gascomprise 50–70% of the DH, which explains why theDH is vulnerable to gas price changes.Ukrainian heat generating facilities are ineffective formany reasons. The most important reasons are asfollows:technology used for heat generation is outdatedand inefficient;key assets are heavily deteriorated;equipment is being used in a switching mode onunspecified fuel;delays and failures to carry out regular repairs.According to the Ministry of Fuel and Energy, morethan 90% of energy units have worked out theirprojected 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 oftariffs for production, transportation and supply.Tariffs for heat that is produced by CHPs, cogenerationor alternative/renewable energy sources areset by the National Energy Regulatory Commission(NERC) but they should not be higher than heatproduced by other sources.Tariffs for heat production, transportation and supplyother than CHPs, co-generation or alternative/renewable energy sources are approved by localgovernments. Due to that the tariffs differ much acrossthe territory of Ukraine.According to the Law of Ukraine ―On Heat Supply‖,heat tariffs should cover all the economically soundexpenses for heat production, transportation andsupply. Tariffs should include full costs of heat218

production and provide for marginal profitability levelthat is not lower than the level defined by the Cabinetof Ministers on the base of calculations by the centralbody of executive power in heat supply.If heat tariffs do not cover the cost of heat and marginalprofitability level, the body that has set the tariff shouldprovide for the compensation according to effectivelegislation. That is, if the tariffs for heat from thermalpower station and boilers that are approved by the localgovernment on the basis of heat producer calculation,and they are lower than economically sound costincluding marginal profitability level, the localgovernments must compensate the losses from thelocal budgets.Meanwhile, the Ministry of Economy elaborated thedraft that specifies binding of the household servicestariffs to energy prices. First of all, it means heat, hotwater and gas supply to households. The currentsystem of tariff setting reduces the competitiveness ofUkrainian industry, since industry is forced tocompensate for low households tariffs.The procedure to raise the heat tariffs is rathercomplicated and time consuming, as follows:1) The district heat supply company receives officialnotification from NERC on gas price increase. Onlyafter that the company may start developing theproposal on the heat tariffs increase.2) The new heat tariffs have to be approved by thefollowing authorities: Commissions of the MunicipalCouncil (mis‘krada) and regional council (oblrada). Thetariff proposal has to be reviewed by several instancesas listed below:Trade unionsAntimonopoly CommitteeDepartment for Price AdministrationDepartment for the Protection of Consumer RightsPublic hearings3) Municipal Executive Committee (misk‘vykonkom)has to approve the new heat tariffs as well.4) The tariff changes shall be publicized via officialmass media of Municipal or Regional Council. If duringa month there are no official protests from the Office ofPublic Prosecutor, the company is entitled to apply thenew tariffs.The above steps clearly show how cumbersome anytariff increase can be in practise.3.2. Market DriversArticle 54 of the state budget of Ukraine for 2006 andthe Cabinet of Minister‘s Decree No.207 of 9 March,2006 stipulate for subsidies from the state budget tolocal budgets. No less than 75% of the subsidy must beThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia219directed to energy saving in heat, water supply andsewerage. But according to monitoring results, thefunds are allocated to other purposes. Only fourregions used the funds for energy saving. Otherregions used from 7% to 40% instead of the required75% to energy saving.Other measures of energy saving that would beappropriate include:replacement or reconstruction of steam and gasboilers with efficiency that is lower than 89%;improvement of heat pipes insulation to decreaselosses in transmission pipelines;installation of heat meters; and,Installation of co-generation equipment.Another stimulus for companies to introduce energysaving technologies is outlined in the Law of Ukraine―On Heat Supply‖. According to the Article 8, ―in caseheat supply or heat transportation companies introduceenergy saving measures that result in saving of heatlosses, the body of executive power, that is entitled toregulate heat tariffs according to the Law, may leavethe tariffs unchanged for the three consecutiveyears‖.[4]3.3. Main BarriersIn general, there are a number of decent laws andregulations that would support DHC development, butthey are not implemented properly, as mentionedabove already.Therefore, there is little if any incentives to businessoriented development of the heating services, but thesystems are run at minimum investments and reducedtechnical performance. The DH companies are solelyoperation organizations, mainly departments of themunicipality. The municipalities take care of billing andcollecting based on subsidized lump sum tariffs, and oninvestment decisions.There are many privileged customer categories thatenjoy reduced costs of DH services. In Odessa, forinstance, 25% of the customers in year 2006 enjoyedsuch privileged heating prices. Their billings weredecreased by 20, 30, 50, 75 or even 100%, whicheffectively destroys the business opportunities of DH.Individual and autonomous heating in every apartmentseems the most favourable option for consumers. Insuch a case they do not pay for heat and hot water butonly for gas and cold water. In addition, they canregulate temperature in their apartments and do notsuffer from overheating in spring and insufficientheating in winter. But sometimes it is impossible toinstall autonomous boilers in every apartment, becausethere is not enough space for heating equipment andthe vertical ventilation ducts are not designed for fluegases. Therefore, it would be appropriate to install one

oiler for the whole building (several apartments) orseveral buildings. Another problem for individual andautonomous heating is that in case of gas supplyinterruption there is no reserve fuel resources tocontinue heating. Reserve fuel can be provided only forcentralized DH.Frequent failures in the heating systems as a result ofoutdated equipment and poor funding are still commonthroughout the country. Some service breaks in coldestwinter times have caused serious impacts on humanlife already.Legally, local authorities that establish tariffs forpopulation lower than the cost coverage level have tocompensate the difference to energy‐generatingcompanies. In practice the compensation is not alwayspaid in full which leads to arrears accumulation andaggravates financial state of heat‐generators. Theprocedure of heat tariffs increase is rather complicated,as well as time consuming.According to the Law of Ukraine adopted in April 2006,heat producers such as CHPs and renewable sourcespower plants are not allowed to cross‐subsidy heatproduction to cover losses from heat production at thecost of electricity production or other activity.Nevertheless, official sources say that due to low heattariffs for CHPs heat production is subsidized by thecost of electricity production. But the unofficial sourcesassert that CHPs may charge heat tariffs that are evenhigher than heat production cost to cover losses fromelectricity production, because electricity tariffs are setonly by NERC while heat tariffs are set by heatproduction companies with the approval of local bodiesof power.3.4. Current ActivitiesThe DH strategy is under preparation in Ukraine as amulti-ministerial approach and it should be ready in fall2010. CHP development is in the focus of the strategy.There has also been comprehensive frameworksupport initiated by USAID, EBRD and EU toreformulate the national energy policy, including DHCand CHP. It is uncertain now how much the politicalelection of April 2010 will influence availability of suchforeign technical assistance in the years to come.4. U.S.A.4.1. Status of DHCThe total DHC industry base comprises approximately2 500 systems, in which the number of customerbuildings served by a typical DHC system may rangefrom as few as 3 or 4 in the early stages of new systemdevelopment to the largest system served byConsolidated Edison in Manhattan. The downtownThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia220DHC/CHP system in New York City is the World‘slargest steam system with 1850+ customers.DHC (primarily DH currently) delivers about 3,5 % ofthe total final energy demand in the industrial,residential, public, and commercial sectors. In the pasttwo decades, some 47 Mm 2 has been connected to theDHC systems, but the total customer base volumenumber is not available.The DHC systems are predominantly (80%) withsteam, the consumption being a mixture of steamheating, cooling and DHW depending on the particularcase. There is little residential heat load but themajority is public: offices, malls, universities andmilitary bases.Countrywide, the DH and DC markets are expanding at3-4%/a and up to 10%/a, respectively, but almost solelyon campuses, hospitals, military bases and in thedowntown commercial and public buildings.[5]In general, however, DHC together with CHP has beentragically underutilized as a tool to combat climatechange, to reduce life-cycle costs of energy supply andto defend energy independence in U.S.A.4.2. Market DriversThe U.S. Congress has acknowledged the benefitsDHC/CHP by stating that: approximately 30% of the total quantity of energyconsumed in the United States is used to providethermal energy – heating and cooling buildingspace, DHW and industrial processes; thermal energy is an essential, but oftenoverlooked segment of the national energy mix; DHC systems provide sustainable thermal energyinfrastructure by producing and distributing thermalenergy from CHP, sources of industrial ormunicipal surplus heat and from renewablesources such as biomass, geothermal, and solar; DHC systems provide advantages that supportsecure, affordable, renewable, and sustainableenergy for the U.S., including use of local fuels orwaste heat sources that keep jobs and energydollars in local economies, stable, predictableenergy costs for businesses and industry,reduction in reliance on fossil fuels, reduction inemissions of GHG, and flexibility to modify fuelsources in response to future changes in fuelavailabilities and prices and development of newtechnologies; DHC helps cut peak power demand and reducepower transmission and distribution systemconstraints; and,CHP systems increase energy efficiency of powerplants by capturing thermal energy and using thethermal energy to provide heating and cooling, more

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniathan doubling the efficiency of conventional powerplants.The Department of Energy has estimated thatincreasing CHP from its current 9% share of U.S.electric power to 20% by 2030 would avoid 60% of theprojected increase in U.S. carbon dioxide emissions(equivalent to taking half of all U.S. passenger vehiclesoff the road); and, generate $234 billion in newinvestments.DHC would be a critical component of this CHP growth.The local electric distribution companies (LDCs) areinterested in DHC as a means to reduce the summerpeak and to release transmission and distributioncapacity to other electric applications that have moreeven consumption during the year.The developers of the building sector are interested inDHC as well, because it would leave more room spacein the building for sale.At the municipal level, the market driver for DHC is thereduction of the GHG emissions. Many municipalitieshave set voluntarily targets to the reduced GHGemissions.4.3. Main BarriersIn general, the barriers are very much the same asalready discussed in Canada. Private sector asinvestor cannot be much interested, because startingthe DHC is risky: long pay-back times ranging beyond10 years, limited access to municipal property,challenging contracting of residential, municipal andfederal buildings, overall billing and collection ofdifferent types of customers.Only little expansion on residential sector isrecognized, and that is because there is voting neededamong the condominium owners. The centralizedenergy systems, that the condo owners are not familiarwith and perhaps difficult for them to understand thebenefits, have not been adopted on the residentialsector in a considerable scale so far.4.4. Current ActivitiesThere are several laws and regulations that areexpected to support DHC development in theU.S.A.[6,7]Rising interest on development and extension ofrenewable energy sources as well as improving overallenergy efficiency is to be converted to legislation at themoment. Unfortunately, DHC has not been successfulin the legislation process so far, but both theDepartment of Energy as well as the DHC and CHPassociations such as IDEA and USCHPA are workingon it.The definition of CHP is rather complicated. TheInternal Revenue Code 26 USC and its § 48 defineCHP as producer of:at least 20 % of its total useful energy in the formof thermal energy which is not used to produceelectrical or mechanical power (or combinationthereof), andat least 20% of its total useful energy in the form ofelectrical or mechanical power (or combinationthereof), andthe energy efficiency percentage of which exceeds60%.The Thermal Energy Efficiency Act of 2009 establishesthe Thermal Energy Efficiency Fund that would awardgrants for DHC, CHP, and recoverable waste energyprojects. It includes biomass facilities. Under a federalGHG emissions regulation program, 2% of emissionallowances established for each calendar year from2012–2050 would be allocated to the Fund.This legislation would dedicate 2% of revenues fromclimate change legislation to fund CHP, waste energyrecovery, and DHC projects. Based on variousestimates, this could mean roughly between $1 billionand $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 publicor non-profit hospitals, local and state governments,school districts and higher education facilities, tribalgovernments, municipal utilities, or their designees),40% for commercial and industrial entities, and 20% tobe used in the discretion of the Secretary of Energy tofund institutional entity projects, commercial andindustrial projects, or federal facility projects. A matchis 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 is75% for construction of infrastructure, 15% forplanning, engineering, and feasibility studies, and theremaining 10% to be used at the discretion of theSecretary for either infrastructure or planning,depending on the need.In competition with grid power plants receivinggenerous allowances in ACES, CHP systems could beshut down. Unless allowances are allocated to theDHC CHP system, it will have to purchase allowancesfor all gas consumed in the facility, resulting in anadditional cost equal to 15% of the average 2007wholesale power price ($57 per MWh) at the $16 permetric ton allowance price projected by EnvironmentalProtection Agency (EPA) for the year 2020. In contrast,the merchant coal plant will only have a GHGallowance cost of only 5% of the average 2007wholesale power price, because allowances will beallocated for nearly all (83%) of its emissions.221

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFaced with this huge competitive disadvantage in themarginal cost of power generation, some existing CHPfacilities will shut down and construction of new CHPplants will be choked off.In the ACES, DHC systems are not directly coveredentities unless they qualify as ‗electricity sources‘.District systems would be covered indirectly throughthe costs of allowances built into the prices ofpurchased fuel oil, or natural gas if purchased from thegas LDC. However, gas purchased on the wholesalemarket or coal users not qualifying as an electricitysource would not be required to submit allowances.This is a fundamentally good framework with theexception of the concerns about CHP systems to becovered or not. However, if the upcoming legislativeprocess results in modifications that make many DHCsystems covered entities, it is critical that changes inallowance allocations be made as discussed below.For example, if the final climate change bill regulates allsources with emissions greater than 25 000 metric tonsCO2e (the threshold generally used in the ACES aswell as a number of past bills), over 70% of DHCsystems and over 95% of DHC output would becapped. In such a way, more efficient systems will havecompetitive advantage, because the quantity ofallowances needed per unit of energy will be lower.American Clean Energy Leadership Act (ACELA) inJune 2009 and Federal Renewable/Energy EfficiencyStandard establishes a Renewable Electricity Standardwhich includes provision for energy efficiency credits aswell as renewable energy credits that can benefit DHCas well.Renewable Electricity and Energy Efficiency Standardestablished by ACELA is applicable with the electricutilities selling >4 TWh a year. The utilities are requiredto supply 20% of demand from combination ofrenewable sources and increased energy efficiencymeaning 15% renewable together with 5% efficiencyincrease. If the state determines that it cannot meet therenewable requirement, then the portion of renewablesources may fall lower to 12% but with efficiencyincrease equal or higher than 8%. These requirementsprovide important leverage for DHC/CHP development.CONCLUSIONThe survey work is still underway, and therefore, thelessons learned and recommendations will be issued inthe final report in spring 2011.ACKNOWLEDGEMENTThe authors express their gratitude to the interviewedspecialists, Mr. B. Gilmour (Canadian Urban Institute),Mr. K. Church (Natural Resources Canada), Mr. M.Wiggin (Public Works and Government ServicesCanada), Mr. R. Thornton (<strong>International</strong> District EnergyAssociation – IDEA), Mr. D. Kaempf and Ms. P.Garland (U.S. Department of Energy), Mr. B. Hedman(ICF <strong>International</strong>) and Messrs. G. Draugelis and P.Salminen in the World Bank.REERENCES[1] National DHC Survey, Canadian DHC Association(CDEA), 2009.[2] Ministry of Construction, China City ConstructionStatistic Annual. The DH data does not includeindustrial steam and hot water.[3] T. Kerr, IEA Collateral, Sustainable Energy inChina: The Role of CHP and DistrictHeating/Cooling, 2008.[4] A. Tsarenko, Overview of Heating Sector, CASEUkraine, 2007.[5] IDEA Report, The DHC Industry, 2005.[6] DHC Services, Commercial Data Analysis for EIA‘sNational Energy Modeling System, Energy andEnvironmental Analysis, Inc. and <strong>International</strong> DHCAssociation, 2007.[7] M. Spurr, Climate Change Legislation in Dollarsand Cents, presentation in IEA DHC inGustavelund, Finland, in Aug 2009.Energy and Water Development Appropriations Act of2010 will provide $15 M for DHC feasibility studies.222

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaBARRIERS TO DISTRICT HEATING DEVELOPMENTIN SOME EUROPEAN COUNTRIESDag Henning 1 and Olle Mårdsjö 21 Optensys Energianalys, Örng 8c, SE-582 39 Linköping, Sweden, phone +46 70 536 59 22,e-mail dag.henning@optensys.se, www.optensys.se2 Manergy, P.O. Box 271, SE-581 02 Linköping, www.manergy.seABSTRACTDistrict heating (DH) offers low primary energydemand, high security of supply and small CO 2emissions. Barriers to DH in the UK, Ireland, France,Romania and the Czech Republic have been compiledthrough publications and interviews.DH systems require large investments, have negativeinitial cash flow and long payback time, which obstructsfinancing. One actor should control DH from source toconsumption. If the value chain is fragmented,contracts are required between the links. It increasesrisks and financing costs, like in the UK and Ireland,where DH is not established. There are few multi-familyhouses with central heating and it is expensive to buildDH networks in built areas.Most French DH systems are operated according tolong-term concessions by companies that sell electricityand gas. No strong actor provides unbiased DHsupport. In the Czech Republic, gas offers DH severecompetition. Much DH is produced at the expense ofelectricity that is considered more valuable, and wasteincineration is not popular. In Romania, DHconsumption was reduced by one-half. Distributionlosses are enormous. New less polluting plants areneeded.Consortia from established DH countries could offer DHsystems from fuel to customer if local policies facilitateDH development.are difficult to use for individual buildings, such asunrefined biomass fuels, heat from waste incinerationand industrial surplus heat. The latter may, forexample, be a by-product from production ofautomotive biofuel. District heating can provide cheapenergy to consumers by using low-cost energysources, such as wood, waste and surplus heat. Manyof these resources can be of local origin and promotelocal business and industry.The main advantages with district heating are highsecurity of supply through utilisation of domesticrenewable energy resources, if available, low primaryenergy demand due to high conversion efficiency, aswell as small CO 2 emissions thanks to low fossil fueluse and the high energy efficiency. Incineration ofwaste with heat recovery to district heating may beused at very low cost. District heating also givesopportunity for cogeneration of power and heat withvery high efficiency. District heating enables profitableheat supply with outstanding environmentalperformance but there are in many places variousbarriers to a prosperous DH development.Barriers to district heating in the United Kingdom (UK),Ireland, France, Romania and the Czech Republic, aswell as barriers to export of Swedish district heatingknowledge and products to these countries have beencompiled from publications and through personalcommunication with people in public and private energybodies and companies in Sweden and abroad [1].INTRODUCTIONThis paper describes barriers to district heating (DH) invarious parts of Europe and to Swedish involvement indistrict-heating business abroad. The paper is basedon a report called ―District Heating in Europe: Barriersto overcome for Swedish export‖ [1], which wasprepared for The Swedish District Heating Association.The losses by energy conversion in Europe are of thesame magnitude as the European heat demand andconsist mainly of heat that is wasted by electricitygeneration [2]. District heating is a means to utilisesuch surplus heat to cover heat demand.District heating can utilise the heat from electricitygeneration in combined heat and power (CHP) plants.District heating can also use other heat sources that223In the studied countries, there are large potentials fordistrict-heating development and for Swedish sales ofDH related goods and services. But for district heatingand export to succeed, there are several barriers toovercome in Sweden as well as in the other countries.It should be emphasised that this paper focusesbarriers and does not give the full picture of theconditions for district heating, which also includes manypossibilities.BARRIERS IN WELL-DEVELOPED DH COUNTRIESIn many countries with well-developed district-heatingindustry, such as Sweden, much DH competenceresides in municipally owned energy companies. Theyhave system knowledge, which could be applicable inother countries. District-heating companies owned by

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSwedish municipalities must, for judicial reasons, limittheir business abroad to sales of services, and to avery limited extent goods. For municipal district heatingcompanies, domestic judicial restrictions are the firstbarriers to overcome before operations in othercountries can commence.Only certain components for production and distributionof district heating are manufactured in a single country,which calls for international cooperation. The SwedishGovernment provides certain but limited support topromotion of district heating business abroad. Forexample, Swedish district heating consultants workabroad but it is seldom followed by goods export.FINANCING THE DH VALUE CHAINFinancing is a large barrier to district heatingdevelopment. DH systems require large investmentsand may have long payback times. The cash flow isnegative for a long time during the establishment of anew DH system. Time horizons are distant, whichstresses financers in our present situation of rapidlychanging conditions. Private companies often focus onshort-term profit and public involvement may benecessary for the deployment, modernisation and longtermdevelopment of district heating systems.District heating is a comprehensive concept for heatfrom source to consumption. Its strength lies inmaintaining the value chain (Fig. 1). This may fit badlyin an exaggerated market context where every little linkof the value chain is organised separately with aninterface of costs and revenues to other links. Afragmented value chain increases interface costs andtotal risk. EU regulations have a tendency to promotesuch fragmentation. Between the links of a fragmentedsupply value chain, many complicated agreements arerequired, which all include risks. It means a larger totalfinancing risk, which raises interest rates and shortensamortisation periods for loans. This implies a mismatchwith the depreciation in the balance sheet due to thelong economical lifetime of district heating versus theshort amortisation time.TWO GENERAL DH BARRIERSTwo general district-heating barriers are related to CO 2emissions and the attempts to reduce these through,for example, reduced energy use. Global warming andbetter insulated houses reduce heating demand and,hence, the advantages of district heating becauseinvestment costs must be carried by less supplied heat.Another general barrier to district heating is the EUemission trading scheme, which favours individualheating because individual CO 2 emissions do not needallowances.TYPES OF DH BARRIERSIn the countries analysed in this project, the barriersare of very diverse nature. The obstacles aredominated by difficulties for district heating itself ratherthan for foreign companies‘ operations in the countries.In the British Isles, it is largely a question ofestablishing district heating as a natural element insociety. In France, it is about large domestic companiesthat may offer superior competition to foreign firms. Inthe Czech Republic, French and other companies fromabroad dominate the DH business but the technicaldesign of district-heating production may hamper DHdevelopment. In Romania, there are several problemswith facilities in bad shape and public bodies that havenot addressed the issues properly.Table I is an attempt to assess how large the variousbarriers are in the studied countries. The table startswith some general conditions. Ownership andorganisation considers if district-heating companies areowned, or DH operations are organised, in ways thatmake it more difficult for Swedish companies to dobusiness. Corruption may be a problem through, forexample, indirect bribes by procurement. National andlocal control encompasses national laws and policyinstruments that are disadvantageous for districtheating, DH price regulations, as well as municipalitiesnot facilitating district heating by planning of newdevelopments. But rules complicating combined heatand power production are included in the CHP line inTable I.Financing is one of the largest barriers to districtheating, primarily because DH schemes give a low rateof return. A fragmented value chain cause contractrisks at several instances. Entrance barriers for foreigncompanies in Table I consider additional difficulties forforeign firms besides the other parameters and thegeneral disadvantage of not being familiar with thedomestic business culture.Some parameters in Table I are related to districtheatingsales. DH competitiveness includes theavailability and price of other forms of heating, primarilynatural gas. Customer relations concern customerattitudes toward district heating, customers‘ andsuppliers‘ perceived insecurity whether they canFig. 1. District heating value chain with heat production, distribution and sales in focus [1]224

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaestablish and maintain relations, as well as ifdisconnections have occurred or may occur. Builtenvironment relates to how common multi-familybuildings are and if these have a central heatingsystem for the whole house. Table I ends with districtheatingproduction and distribution issues. Biomassconsiders domestic biomass supplies andinfrastructure for biomass fuel supply. Waste includescurrent waste management and attitudes towardwaste incineration. CHP concerns regulationshampering CHP production as well as problems inexisting plants. Finally, district heating distribution inTable I encompasses difficulties with building networksand deficiencies in existing distribution.The assessments in Table I were primarily made withineach country and secondly countries were compared butmostly the ranking of countries for a parameter isappropriate. However, every grade has a certain ‖width‖and two countries with the same digit may differ. As anexample, district heating is assessed to be somewhatless competitive in Romania than in the Czech Republic.It follows a description of barriers in the individualcountries emphasising the largest barriers.Table I. – Height of DH barriers in analysed countries [1]BARRIER UK IRELAND FRANCE CZECH REPUBLIC ROMANIAOwnership and organisation 1 0 4 2 3Corruption 0 0 0 2 3National and local control 3 2 1 2Financing 4 3 2 3 3Fragmented value chain 4 3 1 2 1Entrance barrier for foreign companies 1 1 4 2 2DH competitiveness 2 1 3 4 4Customer relations 2 2 1 4Built environment 3 4 2 0 0Biomass 3 3 1 3 1Waste 1 1 3 4 2CHP 3 3 2 4 4DH distribution 4 4 1 4THE BRITISH ISLESIn the United Kingdom (UK), and even more in Ireland,district heating is not really an establishedphenomenon. Figure 2 shows that residences mostlyare heated with gas in the UK, often through a gasboiler for the individual household. Oil is the mostcommon fuel in Irish homes but gas is expanding.The largest problem is district heating distribution(Table I). It is expensive and complicated to build DHnetworks in already built areas and, at least in the UK,it is not straightforward to obtain a licence for puttingdistrict heating pipes into streets. The financingdifficulties in the British Isles are primarily due to afragmented value chain with many contract issues thatneed to be solved before a larger district heatingscheme can be deployed. British thinking is based oncompetition and individual choices. A collective largescale solution, such as district heating, may conflictwith principles and tradition. Another large barrier is thebuilt environment. Few people live in multi-familyhouses in the UK and even fewer in Ireland [3], andeven these buildings often lack central heating, butindividual heating of apartments is common. Biomass225is rated as a rather large barrier in Table I becausesupplies are limited in the British Isles and fuel supplysystems are less developed.UK Government and municipalities have hitherto notfacilitated district-heating development sufficiently andstrong incentives for deploying district heating systemsare lacking. Heating is generally not regarded as apublic concern, but as a concern for each individual.National and local control is therefore indicated as arather large barrier in Table I. In Ireland, the situationseems to be slightly better but in both countries certainregulations, designed with electricity and gas in mind,are disadvantageous for district heating. CHP suffersespecially from rules on how produced heat and powermay be supplied.Customer relations are complicated because districtheating is a rather unknown energy form and there is acertain resistance against collective solutions [3]. Thereis a lack of standardised terms of contract forconnection to and delivery of district heating. Potentialheat suppliers and customers feel insecure concerninghow many users that will connect to a DH grid, for howlong they will stay and if heat supply may be

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniainterrupted. The competitiveness of district heatingcompared to gas concerning availability and price isconsidered as a medium severe barrier in the BritishIsles (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%90%80%70%60%50%40%30%20%10%0%UK Ireland France France Czech Czech Republic Republic Romania RomaniaDistrict District heating heating Gas Gas Biomass Biomass Peat Peat Electricity Electricity Oil Oil Coal CoalFig. 2. Heating of residences [1], [4]–[6]FRANCETable I shows that one of the largest barriers in Franceconcerns the organisation of district-heating operations.Most DH systems are managed by private Frenchcompanies according to long-term concessions [7]. Thecompanies have successfully applied this DHmanagement model in several other countries. By sucharrangements, it is important that operators haveincentives to make investments even if these havepayback times longer than the concession period [8]. Itis unclear if the French DH management model isdisadvantageous for district heating development but itshould anyway be a large barrier for foreign companieswanting to enter the French market. In general,domestic solutions are preferred. There is no strongactor who provides unbiased support for districtheating. The dominating DH operators also sellelectricity and gas, which both cover a large fraction ofthe heat demand (Fig. 2) and offer district heatingsevere competition. Only ten percent of the apartmentsand four percent of all residences have district heatingtoday, and DH expansion is slow [6].Fig. 3 shows that one-half of the district heating inFrance is produced with natural gas, mostly in CHPplants. The main part of the renewable energy used fordistrict heating production is waste, which is used to aslowly growing extent [7]. But French wasteincineration plants are mostly built far away from towns,which makes it difficult to utilise the heat [6].CoalRenewablesMiscellaneousOilNaturalgas CHPNaturalgas heatFig. 3. District heating production in France [9]Financing is considered to be a smaller problem inFrance. The market domination by a few actors maypresent an indirect financial barrier. Quite a few peoplelive in apartments but most multi-family houses lackcentral heating. The large French nuclear powerproduction is one reason for worse CHP conditions,which is assessed as a medium-grade barrier (Table I).THE CZECH REPUBLICFig. 2 shows that district heating covers a substantialpart of residential heating in the Czech Republic, butelectricity is used to the same extent and gas is themost common heat source. District heating covers onehalfof the apartments and 60% of urban heating [7].A large barrier in the Czech Republic is, according toTable I, the competitiveness of district heating.226

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaGas prices make it difficult for gas-based districtheating to compete with individual gas heating [7].There are some disconnections from DH systems.use. Many district heating users switched to gas due tolow gas prices and heavy, government-regulated DHprice increases [7], whereas households and districtheating plants had the same gas price.MiscellaneousNaturalgasHardcoalCoalNaturalgasLigniteOilFig. 4. District heating production in the Czech Republic[10]Domestic coal dominates Czech district heatingproduction (Fig. 4). Most of the district heating isproduced in CHP plants. The problem concerning CHP(Table I) is that a large share of Czech district heatingcomes from coal-fired power plants with extractionturbines where the heat is produced at the expense ofelectricity [7], which is considered more valuable. Thebenefit of this CHP production is not allocated to theheat [8]. Some biomass is used to produce districtheating, but biomass use is complicated due todeficient fuel supply systems [7] and governmentscepticism toward renewable energy. There is alsomuch resistance to waste incineration from the publicas well as from politicians.Financing may be a rather large barrier, partly due to acertain district heating disconnection tendency. Themany private foreign district-heating companies in theCzech Republic [7] may be a difficult target for Swedishand other district heating companies from abroad thatare not established in the country. There may also besome reluctance toward foreign enterprises. A certainbarrier is the common corruption by public procurement(Table I). The value chain is sometimes fragmentedinto production and distribution run by different actors.ROMANIAIn Romania, biomass covers the largest fraction ofresidential heat demand among the countries understudy (Fig. 2). Individual boilers and stoves for woodand gas cover more than one-half of the heat use inhouseholds. Gas is the most widely used heatingsource for residences and it is expanding at theexpense of district heating [7].Table I shows that district heating has large problemswith competitiveness and customer relations. Today,the DH consumption is just one-half of the previous227Fig. 5. District heating production in Romania in 2005 [7]Fig. 5 shows that Romanian district heating productionis completely based on fossil fuels. One-half of the heatis produced in, normally coal-fired, CHP plants. Largeinvestments are required in the Romanian districtheating systems. CHP plants and heat-only boilersmust be replaced for environmental reasons.Distribution losses are enormous [7].Organisation is a rather large obstacle for districtheating in Romania (Table I). The municipalities arenow mostly in charge of the district heating systems [7]but much lobbying is required to achieve improvementsand it takes time to reach an investment decision.Corruption is common. Some politicians andemployees try to make their own profit on DH business.Financing difficulties largely concern insecurity whethercustomers will remain because many havedisconnected from district heating. National and localcontrol is a certain barrier because DH companiespartly get heat production costs covered by centralGovernment and City Councils [7]. Besides thementioned problems, the entrance barrier for foreigncompanies should be rather low. Waste collection andsorting are now deficient but, on the other hand, newpossibilities should emerge when Romania wants tointroduce waste incineration, and waste is thereforeconsidered to be a medium-size barrier in Table I.HOW TO OVERCOME BARRIERSThis paper focuses barriers and omits more positivecircumstances for district heating. It may be depressingbut the message is not that district heating has noprospects. The report should rather be understood as arealistic guide to DH development in the studiedcountries.To have a chance to overcome the outlined barriers toany significant extent, powerful initiatives are requiredfrom countries with established district-heating

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaindustries, such as Sweden. Initiatives should comprisemany different players, for example, district-heatingcompanies, equipment manufacturers, consultants andgovernmental bodies. Such consortia could offerdistrict-heating systems from fuel supply, via heatproduction plants and DH networks to customercontracts. Now, many foreign groups visit municipaldistrict-heating systems in Sweden but theseopportunities are seldom utilised to sell acomprehensive DH solution.Municipally owned district heating companies havesystem knowledge that can be applicable in othercountries. A competence transfer may be realisedthrough deeper involvement that might includeownership of plants in other countries. Businessmodels should be developed, which allow utilisation ofmunicipal knowledge abroad and give municipalitiesreasonable returns.For a successful transfer of district-heating solutionsfrom established to emerging markets, private andpublic companies must focus marketing on thecountries, places, projects and forms of involvementthat have the greatest expectations to succeed. At thesame time, national and local policies should reduceand remove described barriers and facilitate districtheating development as a means for increasedefficiency of energy utilisation, higher security of supplyand decreased environmental impact.CONCLUSIONThere are several barriers to district heatingdevelopment in the countries under study. In the UK,there are not many district heating systems. There arefew multi-family buildings with central heating inIreland. The long-term operating concessions of Frenchdistrict heating systems might hamper theirdevelopment. In the Czech Republic, much districtheating is produced in extraction turbines at theexpense of more valuable electricity. Romanian districtheating use was reduced by one-half by cheap gas.In general, it should be advantageous that one actorcontrols the whole district-heating value chain fromsource to consumption in order to utilise synergies andto avoid economic risks with contracts between theseparate entities of a fragmented value chain. Like forother long-term large-scale infrastructure investments,public involvement may be necessary for districtheating development.Through cooperation among various well-establishedplayers in the district heating industry, knowledge,products and services can be transferred to evolvingdistrict heating markets, which promotes industrialprosperity for all parties and helps building sustainableenergy systems in Europe.ACKNOWLEDGEMENTThe Swedish District Heating Association and TheSwedish Energy Agency are gratefully acknowledgedfor financing this study through the Fjärrsynprogramme. We would also like to thank everybodywho has contributed to the study with facts andviewpoints.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[2] S. Werner, Ecoheatcool work package 4:Possibilities with more district heating in Europe,Euroheat, Brussels (2006)www.euroheat.org/ecoheatcool[3] WS Atkins Consultants Ltd, Assessment of theBarriers and Opportunities Facing the Deployment ofDistrict Heating in Ireland, Sustainable Energy Ireland,Dublin (2002)www.sei.ie/uploadedfiles/InfoCentre/DistrictHeatingReportatk.pdf[4] S. Werner, Ecoheatcool work package 1: TheEuropean heat market, Euroheat, Brussels (2006)www.euroheat.org/ecoheatcool[5] SEI, Energy in Ireland: Key Statistics, SustainableEnergy Ireland, Dublin (2008)www.sei.ie/Publications/Statistics_Publications/EPSSU_Publications/Energy_in_Ireland_Key_Statistics/Energy_in_Ireland_Key_Statistics_Final.pdf[6] P. Cousinat, District Heating: A Tool for RationalHeat Management, Master thesis 2006:21, Departmentof Civil and Environmental Engineering, Chalmers,Gothenburg (2006).[7] Euroheat, District heating and Cooling country bycountry 2007 survey, Euroheat, Brussels (2007).[8] J. Zeman and S. Werner, District Heating SystemOwnership Guide, DHCAN project, BRE, Watford(2004) http://projects.bre.co.uk/DHCAN/guides.html[9] SNCU, Les réseaux de chaleur et de froid: l‘énergiecitoyenne, 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 EHPmeeting, Budapest, 11 September 2008,www.lsta.lt/files/seminarai/080911_Budapestas/CZ.pdf228

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaINTRODUCTIONIMPACT OF THE PRICE OF CO2 CERTIFICATES ON CHP ANDDISTRICT HEAT IN THE EU27Markus Blesl 11Institute of Energy Economics and the Rational Use of Energy (IER)In the current energy and climate policy debate, one ofthe key points is the discussion about emissionreduction targets and how they are spread amongdifferent world regions or countries and also amongdifferent sectors. To find a cost optimal burden sharingof an emission reduction target, the different reductionpotentials of the particular sectors or technologies haveto be known. To reach a reduction target, emissioncertificates in a country or region (like EU-27) areallocated among the different sectors or betweendifferent types of heat and power generationtechnologies. This allocation (for example, auctioning)of emission certificates is an important issue tonegotiate since the costs of buying certificates could bean important factor in technology choices forinvestment.The significant advantage of this approach is that theanalysis of the different competing pathways to achieveemission reductions also assesses how they influenceeach other. In the context of efficiency improvement inindustrial CHP and district heating and cooling, the useof waste heat becomes an interest field. Efficiencyimprovements in the residential or commercial sector isexamined in the topic of energy saving. Withoutanalysing the entire energy system the possibleadvantages of CHP and district heating and coolingcouldn‘t be taken into account. This shows thedifference to a standard cost potential curve approach,which has a fixed order of measures depending ontheir avoidance cost.This analysis will evaluate the reduction potential ofCHP plants or in general the production of districtheating and cooling in the EU-27 using the energysystem model, TIMES PanEU /Blesl et al 2008; Blesl2008; Blesl et al 2008b, Kuder Blesl 2009; Blesl 2009/.TIMES PAN-EU MODELThe energy system model, TIMES (The IntegratedMarkal Efom System), is a further development of thetwo model generators, MARKAL and EFOM-ENV,written in GAMS. TIMES was developed in recentyears within the „Energy Technology Systems AnalysisProgramme―(ETSAP) from the IEA with contributionfrom the IER. It is classified in one category with themodels MARKAL, EFOM or MESSAGE. The modelgenerator, TIMES, was developed in the generalUniversity of Stuttgart229modelling language of GAMS due to reasons of beingbetter transferable. TIMES is a multi-periodic linearoptimization model based on a technical approach atwhich single plants are aggregated. The purpose is theevaluation of the economically optimal energy supplystructure at a given need of end use energy and energyservices and also at given energy and climate policyrequirements. For this, the discounted system costs areminimized, whereas the single players (industry,supply, households) could have different economicconsiderations. The main objective of the modeldevelopment of TIMES is the flexible structure toensure a simple mathematic adjustment to therespective problem.The pan European TIMES energy system model(abbreviated as TIMES PanEU) is a model of 30regions which contains all the countries of EU-27 aswell as Switzerland, Norway and Iceland. The objectivefunction of the model is a minimization of the totaldiscounted system costs over the time horizon from2000 to 2050. A perfect competition among differenttechnologies and paths of energy conversion isassumed in the model. The TIMES PanEU modelcovers on a country level all sectors connected toenergy supply and demand such as the supply ofresources, the public and industrial generation ofelectricity and heat and the industrial, commercial,household and transport sectors. Both greenhouse gasemissions (CO2, CH4, N2O) and pollutant emissions(CO, NOx, SO2, NMVOC, PM10, PM2.5) are coveredby TIMES PanEU.The transport sector is disaggregated into four areas:road transport, rail traffic, inland shipping and.aviation.The road traffic includes five demand categories forpassenger transportation (car short distance, car longdistance, bus, coach, motor bikes) and one for freightservice (trucks). The rail traffic includes threecategories: rail passenger transportation (divided intoshort and long distance) and rail freight transportation.The transport modes of inland shipping and aviationare represented by a non-specified general processwhere the development of the transport demand isembodied by the final energy demand.The household sector contains eleven demandcategories (space heating, cooling, hot water, cooking,refrigeration, lighting, washing machines, laundry dryer,dishwasher, other electrics, other energy use), whereof

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniathe first three correlate to specific building types (singlefamily houses in urban and rural areas and multi-familyhouses each described as existing stock and newbuild). The commercial sector is represented by asimilar reference energy system (RES) and consists ofnine demand categories (space heating, cooling, hotwater, cooking, refrigeration, lighting, public streetlighting, other electrics, other energy use). The firstthree of them are subdivided according to differentbuilding types (large/small).The agricultural sector is described by a generalprocess with a mix of several energy carriers as inputand an aggregated demand of end use energy asoutput.The industrial sector is subdivided into severalbranches (for example, iron and steel, cement, lime…)and into energy intensive and non-intensive branches.While the intensive ones are modelled by a processorientated approach, the other industries have a similarstructure but with five energy services (process heat,steam, machinery drive, electrochemical, others)..The generation of electricity and heat in power plants,CHPs and heating plants is differentiated into publicand industrial production. The model contains threedifferent voltage levels of electricity (high voltage,medium voltage, low voltage) and two independentheat grids (district heat, local heat).In the supply sector, all primary energy resources(crude oil, natural gas, hard coal, lignite) are modelledby supply curves with several cost steps. Threecategories can be differentiated: discovered reserves(or developed sources), growth of reserves (orsecondary and tertiary extraction) and new discoveries.Additionally, seven bio energy carriers aredifferentiated: matured forest, bio gas, householdwaste, industrial waste, as well as energy plantscontaining sugary, starchy and lignocelluloses..Due to its regional resolution, TIMES PanEU allows theconsideration of country specific features, for exampledifferent structures of the stock of power plants,different extension potentials for renewables as well aspotentials for storing CO 2 . An interregional electricitytrade is implemented in the model, so that exports andimports of electricity according to the existing bordercapacities could be calculated endogenously in themodel.The role of CHP and district heating will be influencedin the future by the heating demand for the heat, spaceheating and cooling processes. The following chaptersdescribe the status and the assumed development forEurope.Industrial heat demand by temperature and subsectorin the EU27The particular sub-sectors of the industrial sector usedifferent chemical and physical conversion processes.Therefore, they need heat on different temperaturelevels (Figure 1). Processes with a need for very hightemperatures (> 1400 °C) are e.g. blast furnaces(iron/steel industry) or kilns (cement or lime industry).Processes with lower temperature levels occur in thefood/tobacco (sugar production, dairy) industry, otherindustries or in general for the supply of space heatingand hot water. Also, the pulp/paper industry has a highneed for heat at a lower temperature level (< 100 °C).Most of the heat is produced by the combustion offuels. Other heat is generated by the use of electricity.Key processes using electricity for high temperatureheat are chlorine electrolysis, aluminium electrolysis,electric arc processes (iron/steel) and copperelectrolysis.Final energy consumption [PJ]2,5002,0001,5001,00050000-6060-100100-120120-180180-240240-300300-360360-420420-480480-540Figure 1: Final energy consumption for industrial heatproduction by temperature and sub-sector in the EU-27in 2005540-600600-700700-800800-900900-10001000-11001100-12001200-13001300-14001400-1500> 1500OthersFood/TabaccoPulp/PaperOth. non-metallicmineralsGlass flatGlass hollowLimeCementOth. chemicalsChlorineAmmoniaOth. non-ferrousmetalsCopperAluminiumIron/SteelOn the country level, the role of the different memberstates concerning a particular temperature leveldepends on the structure of the industrial sector in thatcountry. In general, the final energy consumption forheat production at a specific temperature level isdominated by the bigger member states and membersof the EU-15 like Germany, Italy, UK, France andSpain. However, new member states like Poland,Czech Republic or Romania also play an significantrole. Some countries only play a key role at single subsectorsand thus only for some temperature levels.The lower temperature levels are dominated by theindustrial sub-sectors pulp/paper, food/tobacco andothers. Due to high activities in those areas, theheating demand is clearly influenced by France (strongfor food/tobacco), Sweden and Finland (strong forpulp/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 highamount of cement production. In the Netherlands, the230

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniachemical and food & tobacco industries are the mostimportant ones. Each country is clearly specialised indiffering industrial sub-sectors.Space heating and cooling demand in Europetoday and in futureThe demand for space heating and cooling differsamong the countries in Europe due to the differences inclimatic conditions and in living standards (e.g. squaremeters per capita) and building standards. This isespecially applicable to the assessment of current andnear future energy demand.Heat demand in [PJ]Cooiling demand in PJ14000120001000080006000400020000120010008006004002002005 2010 2015 2020 2025 2030 2035 2040 2045 20500Commercial Urban MFH Urban SFH Rual New HousesResidential SFHResidential MFHCommercial2000 2005 2010 2015 2020 2025 2030 2040 2050yearFigure 2: Demand for space heating/hot water and forcooling in the EU27In the 2000, the useful demand for cooling was lessthan 5% lower than the useful demand for spaceheating and hot water. In the long term, the coolingdemand will be dominated by the commercial sector.The increase of cooling demand in the EU27 up to2050 will reach approx. 1120 PJ in the residential andcommercial sectors.occuring at this price level are analysed according tothe role of the different reduction possibilities.The foundation for the CO 2 price variation is set basedon the CO 2 price outcomes from two scenario runs witha reduction target of 15% [scenario: 15% reduction(2020)] and 40% [scenario: 40% reduction (2020)] in2020 compared to the Kyoto base year (Table 1). In thelong run (2050), both of these restricting scenarioshave the same target which equals a 450ppm goal(-71% in 2050 compared to 1990).2020 2025 2030 2035 2040 2045 205040% reduction (2020) -40% -45% -50% -55% -61% -66% -71%15% reduction (2020) -15% -20% -25% -37% -48% -60% -71%Table 1: CO2 reduction pathways for the two restrictingscenariosThe resulting CO 2 prices of these two restrictionscenarios build the framework for the price variations.Within the range of the resulting CO 2 prices, the carbonprice varies between 10 €/tCO 2 and 110 €/tCO 2 in 2020in increments of 10 €. In 2030, the price varies between27 €/ tCO 2 and 123 €/ tCO 2 . The price increases until itreaches the level of a 450 ppm scenario in 2050(Figure 3). The emission reductions are evaluatedusing the results from the different scenarios incomparison to the case of the lowest CO 2 prices (10 €/tin 2020, 27 €/t in 2030). First, the total reductions overall sectors are presented and afterwards the focus willbe on the industrial sector. The drivers of the reductionare shown separately. Looking at the industrial sector,the reasons for the emission reductions could be splitup into more efficient production processes, moreefficient heat supply, fuel switch in heat generatingunits or CCS technologies in production processes andenergy supply.CO2 price [€ 2000]4504003503002502001501005002020 2025 2030 2035 2040 2045 205015% reduction(2020)40% reduction(2020)CO2_10CO2_20CO2_30CO2_40CO2_50CO2_60CO2_70CO2_80CO2_90CO2_100CO2_110Figure 3: CO2 prices of the different scenariosSCENARIO DEFINITIONA parameter variation is used to evaluate the reductionpotential and the role of CHP and district heat in theenergy system of the EU27. By varying the CO 2 price,the reduction potential curves are constructed.Therefore, different scenarios with different CO 2 prices(one common price for ETS and Non-ETS sectors) arecalculated with TIMES PanEU and the reductions231OVERVIEW OF THE DYNAMIC DEVELOPMENT OFTHE ENERGY SYSTEM OVER TIMEIn the following analysis, the two scenarios with thelowest (10 €/t CO 2 in 2020, scenario CO2_010) and thehighest (110 €/t CO 2 in 2020, scenario CO2_110)prices are displayed to show the range in which theresults of the price variation occur. Therefore, thedevelopment over the whole modelling horizon(2000–2050) is presented to rank the more detailed

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaresults from a point of time within these more generalresults over a period of time. Since the CO 2 reductiontarget 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 moredifferent in these periods (Figure 3), the energy systemshows the most variations during this time.To show the development over the modelled timeperiod, first of all the net electricity generation of EU-27is displayed (Figure 4). The overall electricitygeneration remains almost constant at 2010 levels(about 3 200 TWh) until 2030. In later periods, there isa clear increase in electricity generation up to4 255 TWh (2050, scenario CO2_110). The increase inthe later periods is driven by stronger emissionreduction targets. To fulfil the restrictions, moreelectricity with low specific emissions and high end useefficiency in the demand sectors is used.According to the given CO 2 prices of the two scenarios(CO2_010 and CO2_110), the main differences occurin the mid term periods. While the total electricitydemand in 2020 is lower in the scenario with higheremission certificate prices (-22 TWh in 2020 betweenCO2_110 and CO2_010), the demand is higher by86 TWh in 2030. The increase is due to the use ofmore efficient technologies in the end use sectorsresulting in lower electricity demand in 2020, while by2030 the switch to electricity based technologies to fulfilthe emission restrictions has already taken place.Net electricity generation [TWh]450040003500300025002000150010005000StatisticCO2_010CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_1102000 2010 2015 2020 2025 2030 2035 2040 2045 2050Figure 4: Net electricity generation in the EU-27Others / Wastenon-ren.OtherRenewablesBiomass /Waste ren.SolarWindHydroNuclearNatural gasOilLigniteAside from the changes in the total electricity demand,there is also a change in the structure of the electricitygeneration. At higher CO 2 prices, less coal (-120 TWhfrom coal fired power plants in 2030) and more gas(+44 TWh) and nuclear (+30 TWh) are used and moreelectricity from renewable energy sources (+35 TWhfrom wind, +56 TWh from biomass and renewablewaste) is generated. Furthermore, CCS is used morewidely under the conditions of the CO2_110 scenario in2030 compared to CO2_010.CoalThe electricity generation from CHP plants in the EU27increases by 79% from about 380 TWh in the year2000 to 640 TWh by the year 2020 (see Figure 4). Theextension of the electricity generation from CHP plantsis essentially supported by gas-fired and biomassbased CHP plants. Additionally, existing public CHPplants with an extraction condensing turbine aresubstituted by CHP plants with a higher power-to-heatratio and there is also an extension of industrial CHPplants, which are often used in cooperation withcommunal facilities. The intermediate growth of CHPplants in the commercial sector between the years2015 and 2035 are based on efficiency advantages ofCHP plants with a medium sized internal combustiongas engine. In the long term, the limited possibilities ofusing CO 2 free fuels in commercial CHPs will result inthese phasing out in the commercial sector. Until theyear 2050 the electricity production by CHP plants inthe scenarios further increases up to a level of 1055 to1100 TWh. CHP plants based on biomass as well asCCS CHP are an important option in the year 2050.Net electricity generation CHP in [TWh]1200.001000.00800.00600.00400.00200.000.00CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_1102000 2005.0 2010 2015 2020 2025 2030 2035 2040 2045 2050Figure 5: Net electricity generation CHP by sector in theEU-27In addition to the net electricity generation, the primary(Figure 6) and final energy (Figure 7) consumption ofthe EU-27 are also analysed over the whole timeperiod. Overall, the primary energy consumption (PEC)does not show clear changes and remains at a level ofabout 75 000 PJ. The lowest total PEC occurs in themid-term periods. The total consumption is influencedby an increasing efficiency till 2030 and later on by ahigher share of renewables and also CCS which bothlead to a higher consumption due to the lower thermalefficiency in the combustion processes.Looking at the impact of the single energy carriers,there is a distinct change between the two scenariosthan in the total sum of the PEC. In 2030 at a higherCO 2 price, less coal (-1 675 PJ) and petroleumproducts (-881 PJ) and more Hydro, wind, solar(+338 PJ) and other renewables +4856 PJ) (mainlybiomass) are used.PublicComercialIndustry232

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaPrimary energy consumption [PJ]9000080000700006000050000400003000020000100000StatisticCO2_010CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_1102000 2010 2015 2020 2025 2030 2035 2040 2045 2050Figure 6: Primary energy consumption in the EU-27ElectricityimportWaste (nonrenewable)OtherrenewablesHydro, wind,solarNuclearNatural gasOilLigniteThe final energy consumption (FEC) showscomparable results (Figure 7). The use of petroleumproducts declines over time in both scenarios(-9 052 PJ in scenario CO2_010 between 2000 and2050). The use of gas increases at lower CO 2 prices inthe mid-term periods (up to more than 13 500 PJ in2020 at scenario CO2_010), but declines in bothscenarios at the very end. This shows that one earlyand cost-effective measure for emission reduction isthe fossil fuel switch from petroleum products and coalto gas in the end use sectors.As already shown with electricity generation, the use ofelectricity also increases in the end use sectors.Especially in the long run at higher carbon prices, thereis a clear rise. The use of renewable energy sourcesalso increases constantly in both scenarios. In 2020and 2030, clearly more renewables are used in theCO2_110 scenario due to the higher CO 2 prices(+3900 PJ in 2030).In contrast to the PEC, the total FEC decreases slightlyin the long run. The reason for this differentdevelopment is that the higher conversion lossesarising from a higher electricity demand and theextended use of renewables and CCS at the publicelectricity generation are balanced at PEC and do notinfluence the FEC.Even though more renewables (mainly biomass) areused, due to the higher use of electricity with its highend use efficiency and other efficiency improvements,the total FEC declines to 49 482 PJ (in 2050 atscenario CO2_110). This efficiency improvementoccurs in the industrial sector mainly at industrialproduction processes, but is also clearly driven byefficiency improvements ain the residential andtransport sectors.99For a detailed discussion of the effects in the different enduse sectors and its impact on the total final energyconsumption see /Blesl et al. (2010)/Coal233Total final energy consumption [PJ]6000050000400003000020000100000StatisticCO2_010CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CO2_010CO2_1002000 2010 2015 2020 2025 2030 2035 2040 2045 2050Figure 7: Final energy consumption in the EU-27Others (Methanol,Hydrogen)WasteRenewablesHeatElectricityGasPetroleumproductsIn contrast to the year 2000, the distribution of localand district heat to the household, commercial andindustrial sectors changes by the year 2050 with anadditional approx. 1000 PJ district consumed in theyear 2050 (see Figure 8).Final energy consumption heat in [PJ]3500300025002000150010005000CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CO2_010CO2_100SupplyResidentialIndustryComercialAgricultureCO2_010CO2_100CO2_010CO2_100CO2_010CO2_100CoalCO2_010CO2_100CO2_010CO2_1002000 2005 2010 2015 2020 2025 2030 2035 2040 2045 2050Figure 8: Final energy consumption district heat in theEU27In the long term, the CO 2 contents of the heat supply forthe end use sectors will be reduced from 130 kgCO2/MWh to 122 kg CO2/MWh in 2020 and from 113kg CO2/MWh to 36 kg CO2/MWh in the year 2050,which is one explanation for achieving the CO2reduction targets in this area. On the other hand, thepossibility to use renewable energy or to install CCS,increasingly influences the penetration of CHP. By2050, fossil heat plants will also be substituted withlarge heat pumps and solar thermal heat plants incombination with storages, biomass heat plants fuelledwith wood or woody crops and biogas.The overall emissions decrease is based on theemission reductions of the single sectors leading todifferent CO 2 abatement costs (Figure 9). The totalemissions correspond to the emission pathway of thetwo restricting scenarios (scenario ―15% reduction(2020)‖ and scenario ―40% reduction (2020)‖, seeTable 1. The earliest and strongest reductions takeplace in the conversion/production sector. Theindustrial sector and the residential/commercial sectoralso show clear reductions. The transport sector tends

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaonly to reduce its emissions with very strict reductiontargets connected to high carbon prices.900800TransportCommercialEmissions of CO 2 [Mt]450040003500300025002000150010005000StatisticCO2_010CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_110CO2_010CO2_1102000 2010 2015 2020 2025 2030 2035 2040 2045 2050TransportHouseholds,commercial,AGRIndustryConversion,productionAdditional emission reduction [Mt]700600500400300200100036 46 56 65 75 85 94 104 114 123Carbon price [€ 2000/tCO2]ResidentialIndustryConversion/ProductionFigure 10: Additional CO2 reduction in the EU-27 in 2030by sector compared to the scenario with the lowest CO2price of 27 €/tFigure 9: CO2 emissions in the EU-27ANALYSIS AT A SPECIFIC POINT OF TIME WITHFOCUS ON 2030After the general effects are described and thescenarios with the lowest and highest CO2 prices areanalysed over the whole period of time, a more detailedanalysis shows the effects in the industrial sectorduring the mid-term periods with a particular focus on2030.Firstly, the reduction potential of the different sectorsshould be analysed (Figure 10). Both conversion/production and the other end use sectors are takeninto account. As in the results of the emission reductionfrom 2000 to 2050 (Figure 9), the industrial reductionpotential plays the key role next to theconversion/production sector. Looking at the year 2030and comparing the additional CO 2 reductions when theCO 2 price is increased from 27 €/t to 123 €/t, thestrongest additional reduction occurs at the conversionsector (+351 Mt at a price of 123 €/t compared to27 €/t). An additional 301 Mt of CO 2 are reduced by theindustrial sector.Especially at higher prices above 94 €/t, the reductionpotential of the industrial sector becomes more andmore important. Its share of the total additionalreduction increases from 33% (36 €/t compared to27 €/t) to 37% (123 €/t to 27 €/t). The lowest reductionoccurs in the transport sector. Till a price of 85 €/t, onlyan additional 6.3 Mt are reduced, while at a price of123 €/t an additional 18.9 Mt are reduced. In theresidential and commercial sector, some reductionpossibilities are cost-effective even without a price onCO 2 . The energy savings outweigh the additionalinvestment costs. Those reduction measures areespecially connected to the building/heating sector.One reason for the CO2 reduction in the residential,commercial and industrial sectors is the increase infinal energy demand from district heat (see figure 11).The overall increase of the district heat demandinfluenced by the different CO2 prices is 14%. Thebiggest growth can be seen in the commercial sector,where the total district heat demand for district heatinggrows by over 30% between the min the minimum andmaximum CO2 certificate price.Final energy demand heat in [PJ]200018001600140012001000800600400200027 36 46 56 65 75 85 94 104 114 123Carbon price [€/T CO2]Figure 11: Final energy demand of district heat in theEU-27 in 2030 by sectorResidentialIndustryComercialAgricultureHowever, the generation of district heat from the use ofrenewable sources and CCS will be one reason for thegrowth of the reduction potential in the conversionsector (Figure 12). The share of the use of renewables,especially biomass, will rise from 29% to 60%. Morethan 1300 PJ of additional biomass will be needed.Due to this increase, the average heat to power ratio ofall CHPs will fall from 0.9 to 0.66. In the cases whereCO 2 prices exceed 56 €/tCO 2 , the district heatgeneration in CCS CHP plants grow more rapidly. Thespecific emissions of the district heat generationdecrease from approx. 380 kg / MWh to 84 kg / MWh.234

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia2000420030%District heat generation in [PJ]18001600140012001000800600400200027 36 46 56 65 75 85 94 104 114 123carbon price in [€/t CO2]Heat Plant RESHeat PlantCHP RESCHP FCCHP CCSFigure 12: District heat generation in the EU-27 in 2030 bytechnology groupCHPUse of Electricity [PJ]418041604140412041004080406040404020400027 36 46 56 65 75 85 94 104 114 123Carbon price [€ 2000/tCO2]25%20%15%10%5%0%-5%-10%Change of Electricity supply by technologycompared to CO 2 price of 27 €/tFigure 13: Use of electricity in the EU-27 in 2030 bytechnologySUMpublicgenerationCondensingindustrialCHPindustrialIn the industrial sector, the share of CHP will grow. Theadditional emission reductions by the industrial sectorof 301 Mt in the year 2030 could be split into industrialsupply and industrial production processes. The supplyside covers the industrial generation of energycommodities or energy services. These are electricityfrom industrial condensing power plants and CHPs,heat and steam from CHPs and boilers, space heatingand heat for hot water as well as cooling. The supplyactivities play an important role in the industrial subsectorswith a high share of space heating (such asfood & tobacco or other industries) or low temperatureprocess heat (such as pulp & paper or food & tobacco).In total, from the additional reduced emissions, 147 Mtare reduced by industrial supply processes and 154 Mtby production processes in 2030. While at lower a CO 2price more emissions are reduced on the supply side(66% of the additional reduction based on supplyprocesses at 46 €/t), at higher prices more and morereductions take place on the production side (49 %based on supply processes at 123 €/t).The additional electricity needed at high CO 2 prices ismainly generated by industrial autoproducers. Withinthis industrial production, the additional electricitymainly comes from CHP power plants. The use ofelectricity in the industrial sector from public generationremains relatively constant even when the CO 2 priceincreases. Accordingly, one key way to reduce theemissions on the supply side is through the extendeduse of CHP plants for industrial power generation. Thishigher amount of electricity from industrialautoproducers (Figure 13) leads to higher conversionlosses in total when the fuel use is considered. Asdescribed above, that is one reason for the differencebetween final energy consumption and fuelconsumption. Another reason is the lower efficiency ofelectricity generation due to the higher use of CCS.The other part of the industrial supply processes is theindustrial heat generation. The drivers for the emissionreduction in industrial heat production are a switch tobiomass (from coal and clearly from gas) and the useof CCS in industrial CHPs (Figure 14). Between a CO 2price of 36 and 56 €/t of CO 2 in 2030, there is a clearincrease in the use of renewables in boilers. The shareof renewables in the total fuel use in industrial boilersincreases from 33% to 51%. As a result, the thermalefficiency of boilers has an overall decrease.In industrial CHPs, there is also a slight increase in theuse of s. This switch takes place between CO 2 pricesof 27 €/t to 65 €/t. However, the main changeconcerning CHPs is the increasing use of CCS. At aCO 2 price above 94 €/t, there is a clear rise in the useof this technology. These CCS CHPs are mainly gasfired 10 . This is why the share of renewables used inindustrial CHPs declines at a price over 75 €/t again.Like biomass, the extended CCS use also leads tolower efficiencies resulting in both the efficiency ofboilers and CHPs to decline over time. Accordingly, thekey driver is not efficiency improvements, but the useof renewables and CCS. The effects of renewables andCCS compensate the trend to lower energy intensitywithin one technology. Gas boilers become moreefficient and as do biomass boilers. However, the moreefficient biomass boilers still use more fuel than the gasboilers.Looking at the heat output by technology, there is alsoa shift (Figure 14). At lower emission prices, the heatoutput from industrial boilers stays almost constant.Within this range, the share of renewables usedincreases (as illustrated in Figure 13). Afterwards, at aprice above 65 €/t, boilers are substituted with heatfrom CHPs and district heat. Both heat commodities23510For a detailed analysis of the CCS potentials, costsand the modelling of CCS in TIMES PanEU see/Kober, Blesl (2010a), Kober, Blesl (2010b), Kober,Blesl (2009)/

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaare generated in combination with an increasing shareof renewables, a higher CO 2 price and from CCS.Efficiency and share [%]90%80%70%60%50%40%30%20%10%0%27 36 46 56 65 75 85 94 104 114 123Carbon price [€ 2000/tCO2]CHP industrial η(total)CHP industrialshare RESBoiler industrialηBoiler industrialshare RESCCS CHPindustrial (shareFuel input)Figure 14: Efficiency of heat supply technologies andshare of CCS at industrial CHP in the EU-27 in 2030150District Heatthat the climate conditions within Europe differsubstantially.Within the energy system of the EU-27, there aredifferent emission reduction pathways. The emissionscould be reduced by a fuel switch in more efficient (orbetter, less carbon intensive) energy supply or by achange in production processes. Key driversconcerning the emission reduction in productionprocesses and in heat demand side are efficiencyimprovements due to new technologies andtechnological improvements. The key driver concerningthe supply side of electricity and heat generation is theincreased use of renewables, mainly biomass, for heatgeneration. The CCS technology also plays animportant role in the reduction of emissions. Due to theincreased use of renewables in CHP and heat plantsand the use of CCS, the efficiency in the supplyprocesses decreases at higher CO 2 prices.change in heat output [PJ]100500-50-100CHP industrialBoilerIn the long run to a CO 2 -free world, the possibility togenerate district heat with renewable energy and theuse of CCS make the decarbonisation of the energyconsumption in the end use sectors possible.In general, the progression of district heat dependscrucially on the possibility of generating CO 2 emissionfree district heat and electricity.-15036 46 56 65 75 85 94 104 114 123Carbon price [€ 2000/tCO2]Figure 15: Heat supply by technology in the industrialsector in the EU-27 in 2030 compared to the scenario withthe lowest CO 2 price of 27 €/tIn total, all these described effects concerning theindustrial supply processes lead to the additionalemission reduction in 2030 of 147 Mt at a price ofbetween 123 €/t and 27 €/t. In general, more emissionsare reduced in boilers than in CHPs. The reasons arethe fuel switch from coal and mainly gas to renewablesat lower CO 2 prices and later on the substitution ofboilers with CHPs (less boilers are used and therewithproduce less emissions).Due to a higher use of CHPs, there is no clear increasein emissions during the mid-term ranges. When theoutput of heat stays constant and a higher share ofCCS is used, then clear emission reductions fromCHPs (additional 48.7 Mt in 2030 at 123 €/t comparedto 27 €/t) occur.CONCLUSION AND OUTLOOKDistrict heating generation offers an economic potentialfor expansion in the future. Depending on the regionsor countries, the development will be different becausethe starting point is economic growth and the existingnational laws or cross-subsidies for competitor‘s energycarriers. In addition, it is necessary to take into accountREFERENCES[1] Blesl, M.; Kober, T.; Bruchof, D.; Kuder, R.: Effectsof climate and energy policy related measures andtargets on the future structure of the Europeanenergy system in 2020 and beyond, Energy Policy,2010 (forthcoming)[2] Blesl, M.; Kober, T.; Bruchof, D.; Kuder, R.: Beitragvon technologischen und strukturellenVeränderungen im Energiesystem der EU 27 zurErreichung ambitionierter Klimaschutzziele,Zeitschrift für Energiewirtschaft 04/2008[3] Blesl, M.: CHP and district heat in the Europeunder an emission reduction regime, in: 11th<strong>International</strong> <strong>Symposium</strong> on District Heating andCooling in Reykjavik, Island[4] Blesl, M., Cosmi, C. ,Kypreos, S. , Salvia, M.:Technical paper n° Technical Report n° T3.18 –RS 2a ―Summary report of Pan European modelresults – BAU scenario‖ EU Integrated ProjectNEEDS ―New Energy Externalities Developmentsfor Sustainability‖ October, 2008[5] DEHSt (2010): DeutscheEmissionshandelsstelle, Kohlendioxidemissionender emissionshandelspflichtigen Anlagen im Jahr2009 in Deutschland, Mai 2010236

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia[6] EEA (2010): European Environment Agency,European Union emission trading scheme (ETS)data viewer, 2010[7] Kober, Blesl (2010a): Analysis of potentialsand costs of storage of CO2 in the Utsira aquifer inthe North Sea; report work package 4: Regionalanalysis at North Sea level, 2010, www.fencoera.net[8] Kober, Blesl (2010b): Perspectives of CCSin Europe considering technical and economicpower plant uncertainties; in PLANETS workpackage 6 deliverable No. 15 ―Report onProbabilistic Scenarios‖, 2010, www.feemproject.net/planets[9] Kuder, Blesl (2009): Kuder, R.; Blesl, M.: Effects ofa white certificate trading scheme on the energysystem of the EU-27, Fullpaper 10th IAEEEuropean Conference in Vienna, Austria, 2009[10] UNFCCC (2009): GHG inventory reports for thesingle member states of the EU-27, submission2009 situation / problems / wishes, Energy PolicyEHP meeting, Budapest, 11 September 2008,www.lsta.lt/files/seminarai/080911_Budapestas/CZ.pdf237

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaCONSIDERATIONS AND CALCULATIONS ON SYSTEM EFFICIENCIES OF HEATINGSYSTEMS IN BUILDINGS CONNECTED TO DISTRICT HEATINGMaria Justo Alonso 1 , Rolf Ulseth 2 and Jacob Stang 11SINTEF Energy Research, Department of Energy Processes2NTNU, Faculty of Engineering Science and Technology,Department of Energy and Process EngineeringABSTRACTIn order to harmonize the implementation of the ECDirective on the energy performance of buildings(EPBD) [1], and to provide guidelines and commoncalculation tools, several technical standards havebeen worked out by CEN in accordance with amandate from the EC. This paper focuses oncalculating system efficiencies of hydronic heatingsystems by using the standards EN 15316-x-x [4], [5],[6].The paper has been written in order to ease anddiminish the time consuming process of interpretingdetails in the standards such as the numbered EN15316-x-x, and with the goal to enlighten main parts ofthese standards.To exemplify some results, an apartment building of1000 m 2 floor area located in a climate like Oslo ischosen. In the base case, the design distributiontemperatures in the building are 80/60. The differentefficiency figures applying for this case are calculatedefficiency values for the production of the heat, for itsdistribution through the building and its emission in theroom. The room efficiency is the one that has thebigger influence on the total system efficiency.INTRODUCTIONThe Directive on the energy performance of buildings iscarried out in order to be used together with a numberof EN-standards. The main goal of the Directive is topromote the improvement of the energy performance ofbuildings within the Community, taken into accountoutdoor climatic and local conditions as well as indoorclimate requirements and cost-effectiveness. The mainfocus is on reducing the primary energy use and theassociated CO 2 emission of buildings.Figure 1 shows how the Primary energy use iscalculated based on all the steps where the energy ischanging its nature from the source to the end use. Inthe current case, the energy calculations are performedfor the systems within the building to be able tocalculate the delivered energy to the building. Thismeans that the building substation with the heatexchangers and tap water storage are included.In the current scenario, all the losses before the heat isdelivered to the building are included in the primaryenergy factor (PEF) for the delivered heat. In case ofconsidering the complete scenario, the boundaries forthe energy performance indicators are the wholeenergy chain from the source to the end use. In thiscase, if a CHP plant is represented, the ―power bonusmethod‖ (EN 15316-4-5) should be used. This methodis giving the produced district heat a bonus for theelectricity produced assuming that this electricityreplaces electricity production with a high PEF-value.According to the implementation of EPBD, it is crucialthat the system borders are clearly defined so that thedelivered energy is doubtlessly defined.Calculation of Primary Energy use according to EPBD and mandated EN-standardsPrimary energy use = DEdh • PEFdh (f (x,y,z)) +DEel • PEFel (f (x,y,z)) = (Weighted delivered energy indicator (kWh/m 2 )) • A CDelivered Energy (DE)dh +el(PEF might be PEF R or PEF T depending on purpose)( A C = conditioned floor Area )Net energy demandPrimary energy use calculated by PEF(x,y,z)ElectricityHot tap waterAir and room +heating systemDH substationEnd usedemand !HeatingsystemsDistribution andTransmission(el)Delivered energy (el)Delivered energy (dh)Distribution andTransmission (dh)•CHP-plantHeat boilersStorageGenerationTransformationEnergy carrier (z)Energy carrier (y)Energy carrier (x)TransportationWaste coll.LoggingExtractionProcessingStorage2010/ 04/RUCalculating end use and losses by ENstandards worked out according to mandate from the EU CommissionHeating systems efficiencyCalculating directionSystem border for the energy performance indicators is the whole energy chain from the source to the end useFigure 1.- Sketch of the calculation of Primary Energy use according to EPBD and mandate EN-standardsFigure 1 Sketch of the calculation of Primary Energy Use according to EPBD and mandate EN-standards238

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaMETHODDefinition of building and system build upIn this exemplified case, the main chosen building is anapartment building. This building category seems to bethe most representative concerning heat use amongthe building categories defined in the EPBD [1].In the shown example, the size of the building ischosen to be 1000 m² floor area since this size shouldbe rather representative and be a good compromisebetween the previous and the proposed new recast ofthe EPBD. [2]A building of these features corresponds to a threestoreys squared building with four flats of about 80 m²per storey.In this setting, the total heating system efficiency in thebuilding is built up based on the differentiation betweenthe three main parts of the system. It must be definedwhere the substation is located in the building, i.e.where the heat is exchanged from the distributionnetwork – DH stage in Figure 2. The heat supply to theheating system within the building from the districtheating system is assumed to be provided by two heatexchangers and hot water storage defined as thebuilding substation part of the system.Finally; once distributed, the heat is emitted accordingto the demand. For the present case, the heat isdelivered either by radiators (80/60 ˚C) in the basecase, floor heating (35/28 ˚C), or domestic hot water at60 ˚C. Figure 2 gives a further visual explanation. Forthe present paper, the supply of heat is just done by ahydronic heating system. The possible heat loss fromthe distributed air is neglected since the temperature ofthe air is assumed to be slightly lower than thetemperatures in the rooms.Categories of buildingThe presented analysis shows results for five kinds ofbuildings described in the EPBD which are: singlefamily house and apartment block, office buildings,hotel and restaurants, educational buildings andhospital buildings. When it comes to heat consumptionfor these buildings, the measurements performed inLinda Pedersen‘s PhD thesis [3] show that theconsumption of the apartment buildings is about116 kWh/m², while hospitals use 150 and officebuildings use 100 kWh/m². These measured valuesinclude the domestic hot water (DHW) and the spaceheating (SH) consumption.The calculated efficiency for the system will depend onthe size of the building as well. The present apartmentbuilding shows a higher efficiency value than a singlefamily house with the same consumption. This is due tohigher relative losses in the substation.Climate influenceThe calculations in the present paper are based on aclimate like in Oslo, Norway. This climate is defined tohave approximately 5100 degree days with 20 ˚C asthe internal reference temperature and an externaldesign temperature of -20 ˚C [8].Ventilationair+●•2010/05/RU+ +DHCWIn practice, the outdoor climate can vary widely fromplace to place. Owing to this, the outdoor climateaffects not only the heat consumption but also therelative losses. In general the relative losses areincreasing with an increased ratio between the degreedays and ΔT between the dimensioning internal andexternal temperature.Hot tapwaterdistrib.Roomheatingdistrib.SubstationsystemborderFigure 2 Sketch of the system elements for production,distribution and conditioning of the roomsFrom the substation, the hot water is distributed eitherfor air and space heating or as domestic hot water.Both uses are provided by their own heat exchangerand the necessary pipelines will now be referred to asdistribution pipelines.239The average outside temperature affects the heatconsumption and the temperature variations affect theregulation of the heat emitters. This means that duringcold periods, the temperature of the supply water tendsto be increased imposing an increase in the lossesrelated to the transport of water with highertemperatures. The design temperature for the radiatorsin the base case in this paper is 80/60, and in warmerperiods, this temperature is decreased in order toreduce losses and adapt the supply temperature to theoutside temperature. This affects the efficiencies in apositive way.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFinally, the outdoor climate affects the length of theheating season. Usually, the lower the average outsidetemperature, the longer the heating season. Thishowever, does not affect the DHW since this is more orless steady all the year along.Positioning of substationFigure 2 shows the positioning of the substation. Theheat is delivered from the district heating pipelinesthrough two separate heat exchangers, one for heatingthe water in the storage tank for DHW by a circulatingloop, and the other for the air heating and spaceheating system. The main reason for having two heatexchangers is due to the different needs of temperaturelevels. In the calculations dealing with the production,the used heat demand used is the total demand, whilefor the distribution the heat is divided into heatdistribution for SH and for DHW.In order to calculate losses related to the storage tank,it is assumed that the tank is of a common type with acommon value for the stand by heat loss. The systemdesign consists of the coupling in series of 289 litresstorage tanks. There are considered two tanks for theapartment building of 1000 m 2 but one more tank if thecase is dealing with hospitals, educational and hotelbuildings. For other sizes the number of tanks isadjusted according to the demand.temperature is here constantly at the designed point of60 o C.When dealing with distribution of SH, the losses areconsidered dependent on the kind of insulation materialand the ambient and the mean water temperature inthe supply and return pipes.The heat emission to the room from the DHW draw-offtap discharge cocks is considered to be negligible incomparison to the total heat consumption.Dealing with space heating a distinction is done withrespect to the kind of emission. Two major groups areconsidered: the emission by floor heating and byradiators. The first has a low temperature distribution of35/28 ˚C. As for the radiator system, the analyzed basecase is 80/60 ˚C for supply/return design values.Besides the temperature level, the placing in the roomaffects the stratification efficiency and the loss throughthe outside wall. Furthermore, another point related tothe temperature is the regulation of the roomtemperature, which in our case, is assumed to be aPI-regulator, even if in a lot of apartments thisregulation is quite often done by on/off regulation.RESULTSProduction efficiencyEfficiencies to be studiedFor the present paper, as written previously, thesystem is divided in three smaller system parts whichare independent. For every component, the efficiencyis calculated following different standards: Production; according to EN 15316-4-5:2007[4]Distribution; according to EN 15316-2-3:2007 andEN 15316-3-2:2007[5] Room emission; according to EN 15316-2-1:2007[6]The efficiency of the production includes the lossesdepending on the thickness of the insulation material,the insulation material itself, the storage tank, thecomplete local piping system of the substation systemand the temperature difference between the two mediaand the ambient. It takes into account the thermal lossof the total substation. For this case the substation isconsidered to be in an unheated part and therefore thelosses are considered as unrecoverable.In case of the distribution, the efficiency depends onthe use of the heated water. In case of being a part of aDHW system; the energy used for heating the waterwhich is not drawn-off and which slowly gets cold in thepipelines, has to be considered as loss. Moreover, heatis used to heat up the pipes and fittings. Since thebuilding is large enough to need a circulation loop thisloop is considered to be a source of loss? The water240The production efficiency is shown in Figure 3. By usingthe losses on the heat demand and the temperaturedifference as basis for calculation, the values inTable 1 are obtained.System efficiencyProduction efficiencies for the different types of buildings according to theEN 15316-4-5 :2007 (Oslo climate) with a distribution temperature 80/600,9900,9800,9700,9600,9500 250 500 750 1000 1250 1500 1750 2000 2250Building floor area [m 2 ]Apartment blockOffice buildingHotel and restaurantbuildingEducational buildingHospital buildingSingle familyFigure 3 Production distribution of the 80/60 ˚C districtheating for different buildingsAs Figure 3 shows, the bigger the building, the higherthe efficiency. This effect is due to the reduction of therelative losses when the size of the substation (kW)increases. The curve profile is decreased slightly from2000 m 2 and downwards, and then decreasing rapidlyfrom about 1000 m 2 down to 500 m 2 .

In addition, it can also be observed that among alltypes of buildings, apartment houses representsomehow the highest efficiencies which justify the mainfocus in this study. The displayed case applies for thevalues where the design temperature level is 80/60 ˚C.It can be concluded from other calculations that thehigher the design distribution temperature level, thelower the production efficiency. This conclusion is whatcould be expected considering the difference betweenthe average temperature and the ambient temperature;the larger this difference, the larger the losses.As shown in Table 1, the efficiency varies only between0.9784 and 0.9673. It can be concluded, compared tothe distribution loss values that the productionefficiency is not changing significantly even if thetemperature level is changed. As a conclusion it can besaid that the losses in the production are relatively lowfor bigger houses but increasing quite rapidly forsmaller buildings.Table 1 Efficiency of DH production system for differentdesign temperature levels.Kind ofbuilding 80/60 70/55 55/45 35/28Apartmentblocks 0,9776 0,9778 0,9780 0,9784Office building 0,9729 0,9732 0,9735 0,9740Hotel andrestaurantbuilding 0,9701 0,9703 0,9706 0,9709Educationalbuilding 0,9676 0,9678 0,9681 0,9685Hospitalbuilding 0,9773 0,9775 0,9777 0,9780The quality of the insulation of the storage tank will alsoinfluence the production efficiency. Manufacturesshould follow the standard pr EN5044:2005 [7] in orderto calculate these losses. Losses from storage tanksshould be considered closely in practice, and tankswith relative high losses should be considered forreplacement or to be replaced by direct heatexchangers for DHW.Distribution efficiencyFirst the system for the distribution of tap water isanalyzed. In this case, the building includes acirculation loop (in small dots Figure 2) which goesfrom the storage tank, to the third floor and thedistribution branches (in bigger dots in Figure 2) whichdeliver DHW from the central loop to the consumer.The water temperature in the circulation loop isassumed to be at 60 ˚C throughout the whole year. Thebiggest share of the losses come from the circulationThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia241loop ranging up to 35 % of the total losses from thetotal distribution system. This loss is related to the factthat the water that remains in the distribution pipelinesrepresents 5% of the total losses per flat.For this calculation it is assumed that the pipelineshave insulation which a loss of 0.3 W/m∙K (thepipelines are considered to be according to thecategory ―installed after 1995‖ in [5]). In this calculationthe losses due refilling the pipes with hot water areincluded. This heat could be considered as recoverableloss for space heating during the heating season but inlack of a special national annex all the losses related tothe distribution of DHW should be considered as ―nonrecoverable‖. These losses are not related to thedemand for heat and will consequently be lost or resultin increased room temperatures.When it comes to SH, the losses are related to thetemperature difference in the non-heated areas wherethe water goes through. These losses are relatively lowcompared with tap water since most of these lossesare considered to be recoverable. The values used aretabulated in the EN standard [5].The percentage of recoverable losses is the cause ofthe higher efficiency for distribution of space heatingwhich ranges 0.99, whilst the efficiency for distributionof DHW is in thee range of 0.60.Emission efficiency in the roomsIn this case, domestic hot water is not considered tocontribute to the room heating since the losses fromthe discharge cocks are considered negligible.In case of space heating a difference has to be madebetween floor heating and radiator heating when itcomes to the efficiency calculations.Floor heating is by its nature emitted at lowertemperature, which has an effect on the stratificationefficiency since the lower the temperature level, thehigher this η str. By definition in [6] the stratificationefficiency of floor heating is 1 whilst this parameter forradiators goes down to 0.91 on the 80/60 distributionsystem. This value is combined with the efficiencyvalue of 95 % due to the positioning of the radiators ona normal external wall. Together these values make atotal room efficiency of 0.93.However, a regulation with PI controllers for theradiators delivers an efficiency of 0.97 while the samecontroller remains at 0.95 for floor heating.In case of the embedded floor heating efficiency theefficiency is 0.93. Since it is considered to be normalinsulation layer according to EN 1264, it results in aη emb of 0.95, the combination results in η emb of 0.94.Due to these three parameters, floor heating all in allhas a room efficiency of 0.90 and radiators of 0.88.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTotal system efficiencyThe total system efficiency is in this paper calculatedas the product of the efficiencies of the different pastsof the heating system in the building.Figure 4 shows the room efficiency and the totalsystem efficiency related to the temperature levelsupply/return. The delivering water temperature to theradiators is the parameter which affects the efficiencythe most.The total system efficiency follows the pattern of theroom efficiency since this parameter has far the largestinfluence.Another conclusion from Figure 4 is that the efficiencyof the complete system varies significantly with thedesign temperature level of the heating system. This isdue to the temperature difference between the heatingsystem components and the ambient. With a lowerdistribution temperature the losses will be smaller.Efficiency0.940.920.90.880.860.840.82System efficiency for an hydronic heating system inan apartment building of 1000m 2 floor areaΔθ=80/60 Δθ=70/50 Δθ=55/45Temperature distributionRoomEfficiencyTotal systemefficiencyFigure 4 Room efficiency and total system efficiency forthe different design distribution temperatures according toEN 15136-2-1:2007 (Oslo climate). Radiators withthermostatic valves mounted on normal external walls andwith heat supply from district heatingTable 2 Total efficiencies of the systems for spaceheating with floor heating and hot tap water systemCONCLUSIONSThe design temperature level for the system is themost important factor when referring to the efficiency ofa hydronic heating system in buildings supplied bydistrict heating. . Therefore the possibility of loweringthe design temperature level of the heating systemshould be considered closely. This increases theemission efficiency in the room and reduces the lossesfrom the distribution pipelines. It saves energy andincrease the cooling of the district heating waterthrough the substation. Changing the positioning of theradiator from the external wall to the internal wallactually decreases the room emission efficiency.The introduction of an energy performance certificatefor buildings according to EPBD requires a transparentcalculation model according to the standards in the EN15316 series. This paper gives a picture of systemefficiencies for hydronic heating systems and also anidea of the time consuming process that has to beperformed in order to calculate the efficiency of asystem in detail. Therefore, it is concluded that someuser-friendly guiding material should be desirable inorder to enlighten and facilitate the calculation process.In the present paper the potential heat losses from theventilation system are neglected due to the fact that itis assumed that the air temperature is distributed attemperatures slightly below the room temperatures.ACKNOWLEDGEMENTTotal system efficiencyFloor heating 0.87Hot tap water 0.59Calculated efficiencies for a floor heating system andthe hot tap water system are presented in Table 2.Space heating with floor heating has a slightly lowerefficiency than radiators due to the lower efficiency forthe emission of the heat in the room.Tap water systems have a lower efficiency since thesystem is by its nature losing a considerable amount ofheat when leaving the hot water in the pipes betweenthe tapping cycles. This water is cooled down insidethe pipelines and is then being tapped without beinguseful. In the present case the distance from thesubstation to the furthermost apartment forces aninstallation of a circulation loop in order to reduce thewaiting time for hot tap water at the tapping cocks. Thisis a stand by source of loss. These two factors causethe rather low efficiency of the hot tap water system.242This work has been supported by SINTEF, NTNU andhas been related to the project ―Systemvirkninsgrader‖(System efficiencies) which was initiated by StandardNorge and paid by the Norwegian Water Resourcesand Energy Directorate. It has also been supported bythe Primary Energy Efficiency project which is paid byNordic Energy Research and companies in the heatingfield in Norway.REFERENCES[1] European Parliament and Council on energyefficiency of buildings, ―Directive 2002/91/EC onthe energy performance of buildings‖ (EPBD)[2] Proposal on a recast of Directive 2002/91/EC onthe energy performance of buildings, 2009-11-25

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia[3] PEDERSEN, L. (2007) ―Load Modelling ofBuildings in Mixed Energy Distribution Systems‖,Department of Energy and Process Engineering,NTNU,(Norwegian University of Science andTechnology), Trondheim[4] EN 15316 ―Heating systems in buildings – Methodfor calculation of system energy requirements andsystem efficiencies – Part 4-5: Space heatinggeneration systems, the performance and quality ofdistrict heating and large volumes”, 2007[6] EN 15316 ―Heating systems in buildings - Methodfor calculation of system energy requirements andsystem efficiencies – Part 2-1: Space heatingemission systems.”, 2007[7] CEN: “Efficiency of domestic electrical storagewater-heater – German version pr EN 50440”,2005[8] “VVS-tekniske klimadata for Norge”, Norgesbyggforskningsinstitutt, Håndbok 33[5] EN 15316 ―Heating systems in buildings – Methodfor calculation of system energy requirements andsystem efficiencies – Part 2-3: Space heatingdistribution systems.”, 2007243

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaHEAT LOAD REDUCTIONS AND THEIR EFFECT ON ENERGY CONSUMPTIONChristian Johansson 1 and Fredrik Wernstedt 21Blekinge Institute of Technology, PO Box 520, SE-372 25, Ronneby, Sweden, chj@bth.se2 NODA Intelligent Systems AB, Drottninggatan 5, SE-374 35, Karlshamn, Sweden, fw@noda.seABSTRACTIn this paper we investigate the consequences of usingtemporary heat load reductions on consumersubstations, from the perspective of the individualconsumer as well as the district heating company. Thereason for using such reductions are normally to saveenergy at the consumer side, but the ability to controlthe heat load also lie at the core of more complexcontrol processes such as Demand Side Management(DMS) and Load Control (LC) within district heatingsystems. The purpose of this paper is to study the waydifferent types of heat load reductions impact on theenergy usage as well as on the indoor climate in theindividual buildings. We have performed a series ofexperiments in which we have equipped multiapartmentbuildings with wireless indoor temperaturesensors and a novel type of load control equipment,which gives us the ability to perform remotelysupervised and coordinated heat load reductionsamong these buildings. The results show that asubstantial lowering of the heat load and energy usageduring periods of reductions is possible withoutjeopardizing the indoor climate, although we show thatthere are differences in the implications whenconsidering different types of heat load reductions.INTRODUCTIONThe main purpose of this paper is to investigate theconsequences of using temporary heat load reductionson consumer substations within a district heatingnetwork. The most common way to perform temporaryheat load reductions is to use night time set-back, i.e.to lower the wanted indoor temperature during nighttime while social activity is expected to be low.Emerging technologies like Demand Side Management(DMS) and Load Control (LC) also use temporary heatload reductions in order to accomplish system widecontrol strategies, although the characteristic of thesehead load reductions differ significantly from night timeset-back.In the context of this study we regard a heat loadreduction to be the whole process from the initialchange of heat load, through the return to normal heatload, and until no evidence of the heat load reductioncan be noticed in the dynamics of the building energy244balance. This definition is based on the fact that theheat load reduction will continue to exert an influenceon the buildings thermal buffer for some time even afterthe heat load reduction in itself is ended. The length ofthis interval is specific to each building and is related tothe thermal inertia of the building in question.In this paper we study the consequences of usingdifferent types of heat load reductions, and try toanalyse the way the thermal buffer of the building isaffected along with the actual heat load and energyusage from both a local and a global perspective. Westudy the performance of both long low-intensity heatload reductions (e.g. night time set-back) as well asshort high-intensity reductions (e.g. those frequentlyused in DMS schemes). The use of night time set-backhas received some attention in previous works, e.g [1],and the possibilities to use the building as a heat bufferhas been evaluated [11], but heat load reductions suchas those used in DSM and LC have to the knowledgeof the authors not been thoroughly investigated.Night Time Set-backNight time set-back means to lower the wanted indoortemperature during night time, with the purpose ofsaving energy through reduced heat losses due todecreased difference between indoor and outdoortemperature. This is the most common way to performtemporary heat load reductions, and many commercialcontrol systems support this feature. This is normallydone by a parallel displacement of the heat controlcurve during night hours. During night time set-back thewanted 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 setbackactually gives an energy saving or not [4], andmost practical implementations of night time set-backsuffer from morning peak loads when the controlsystem returns to the original operational level. Still,almost all control equipment companies sell equipmentthat facilitates the use of night time set-back, and theuse of this technique is widespread.Demand Side Management and Load ControlWhile night time set-backs are a solely local energysaving technique, DMS and LC are usually performedwith a system wide perspective in mind. A building

owner is normally only interested in lowering theenergy consumption, while the district heating companyis more interested in being able to optimize the wholeproduction and distribution process. Optimizing theproduction normally translates to avoiding expensiveand, more often than not, environmentally unsoundpeak load boilers or trying to move heat load demandin time in order to maximize utility during combinedheat and power generation. Basically, from theperspective of the district heating company it is aquestion of finding a balance between loweringexpensive heat load demand while still selling as muchenergy as possible. Implementing this on a systemwide scale requires complex coordination controlstrategies that dynamically adapt to the state of thedistrict heating system [2]. On the local building levelthis is implemented by performing temporary heat loadreductions. On a local level these reductions arenormally very short, i.e. one or a few hours, but theycan be of high intensity, even sometimes completelyshutting of the heat load during shorter periods of time.This behaviour requires the control system to be highlyadaptive in relation to the dynamics of the buildingsthermal inertia in order to avoid jeopardizing the indoorclimate. By coordinating such local heat loadreductions among a large group of buildings it ispossible to achieve system wide DMS and LC.Previous workMost previous work regarding temporary heat loadreductions deals with night time set-back. This is atechnique that has been around for a long time, and isbased on the general idea that if you decrease thedifference between the outdoor and indoor temperaturein a building you will save energy. One of the firstlarge-scale evaluations of night time set-back wasperformed in 1983 when buildings in Sweden, USA,Belgium and Denmark were evaluated. Thisexperiment concluded that night time set-back did notsave as much energy as was expected, at most a fewpercent for multi-apartment buildings [3]. In hindsight itis possible to see that these meagre results were aconsequence of several interacting factors. First of allthe control systems of the time were not capable ofproperly handling the transition from night time setbackto the original operation mode, which causes aconsiderable over-compensation of heat load when thesystems tries to find the new control level. This extraboost in heat load during the mornings counteractslarge portions of the energy saving done during thenight. The theoretical part of the experiment also had afew draw-backs, e.g. assuming optimally adjustedradiator systems and linear relations between indoortemperature and energy savings. Other articles showthat there is indeed a substantial level of energy savingto be found by controlling the local heat load [5].The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia245Most of the previous work done on the subject is basedon simulated results. This is expected since thedynamic thermal processes within a building areextremely complex and it is not surprising thatcomparisons between measurements and calculationssometimes show large discrepancies. It is noted thatmost calculations are dependent on variables thatcannot be measured and verified, and that the buildingtime constant is really not a constant [6].EXPERIMENTAL METHODIn order to study the effects of temporary heat loadreductions we equipped a building with several wirelesstemperature sensors in order to measure thefluctuations in indoor temperature. The building inquestions is an office building with semi-light thermalcharacteristics (light construct with concrete slab) and atime constant of about 150 hours [7]. The indoortemperature sensors were placed on different locationswithin the building in order to get a good overview ofthe thermal behaviour of the indoor climate. In additionto the existing outdoor temperature sensor an extrawireless sensor was also placed on the outside of thebuilding. Unlike the existing outdoor temperaturesensor the wireless one was placed in a position wereit was fully exposed to any possible sunshine. Thisgave us an extra indication of the impact of free heatingthrough window areas, even though we did not haveany ability to measure the actual solar irradiance.In order to control the district heating consumer stationwe connected a load control platform for system wideLC and DSM [8]. This platform is based on a novelform of hardware and software which enables us tomanage the heat load of the substation without anymajor alterations or any damage on the existinghardware. The software system is based on the opensource Linux operating system and is equipped with anapplication programming interface (API) for I/O. Thismakes it easy to apply additional sensors, e.g. formeasuring the forward and return temperatures of theradiator system. The platform also featuresconnections to a database system which enables realtimelogging and analyse of sensor data. The actualheat load reductions are implemented by supplying theexisting control system with adjusted outdoortemperatures, which gives us the ability to manage thebehaviour of the heat load without exchanging anyexisting hardware. This adjusted outdoor temperaturecan be managed with a resolution of at most 60seconds. The computer platform uses either Ethernetor GPRS modems to communicate with the database.In our case we used the existing Internet access in thebuilding. In addition to this primary experimentalbuilding we also collected and analysed data frompreviously installed buildings using the same basiccomputer platform.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaEnergy and heat load usage was primarily evaluated bystudying the dynamic differences between the forwardand return temperature of the radiator system inrelation to the flow. These readings were then verifiedby specifications from the district heating providerregarding energy consumption and momentary heatload usage.Using this set-up we scheduled different types oftemporary heat load reductions and studied theireffects on the measured data. During this study westudied three primary types of temporary heat loadreductions:system performs a controlled heat load recovery inorder to avoid unwanted heat load peaks after thereduction.The same values are shown for a long heat loadreduction in Figure 2. The heat load reduction startsslightly before the 600 minute mark and continues forseveral hours until about the 900 minute mark. Afterthat the control system performs a controlled recoveryin order to return to the original operational state.Long – Four to eight hours of continuous heatload reduction with different intensityShort – Up to one hour long heat loadreductions with different intensityRecurring – Several short subsequent heatload reductions with short pauses in betweenWhen we studied the different types of heat loadreductions we took care in allowing the buildingsthermal process to return to its original state betweeneach reduction so that the reductions would notinfluence each other. This was done in between eachreduction except in those cases when then purposewas to explicitly study the interaction betweensubsequent heat load reductions.EXPERIMENTAL METHODFigure 2: dT in radiator circuit with long heat loadreductionFigure 3 shows the same values for a series ofrecurring heat loads.Figure 1 shows the temperature difference between theforward and return temperature in the radiator circuitduring a short heat load reduction.Figure 1: dT in radiator circuit with short heat loadreductionThe heat load reduction starts at about 60 minutes andcontinues until the 120 minute mark. Between the 120minute mark and about the 160 mark the control246Figure 3: dT in radiator circuit with recurring heat loadreductionEach of the heat load reductions in Figure 3 is one hourlong intersected by one hour long recovery periods.The first reduction starts at the 60 minute mark andcontinues until the 120 minute mark.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFigure 4 shows the energy consumption in relation tothe outdoor temperature during week long periods withand without heat load reductions implemented as LC.The squares are from periods without LC and thetriangles are from periods with LC. LC in this regardmeans that temporary heat load reductions are beingperformed in recurring sets throughout the week aslong as the thermal inertia of the building allows it, i.ewithout jeopardizing the indoor climate. In this examplethe energy usage is about 8.2% lower during periods ofheat load reductions.Figure 6 shows recurring heat load reductions insteadof single long ones. It is clear that the building is able torespond to the control scheme in this example also.The largest heat load reduction during the recurringscheme is about 25%.Figure 6: Heat load reductions shown 24 hours withoutreductions (black), 24 hours with reductions (dark grey)and control scheme for reductions (light grey)Figure 4: Energy usage in relation to outdoor temperature.The squares are values during periods without LC, andtriangles show periods with LCFigure 7 shows a range of indoor temperature readingsduring periods with heat load reduction (triangles) andduring periods without (squares). The averagedeviation during heat load reduction is about 0.29 whilethe average deviation during periods without reductionsis about 0.19.Figure 5 shows the heat load (kW) during 24 hourswhen using reductions compared to not usingreductions. The control scheme is also added to thefigure in order to show when the reduction wasperformed.Figure 5: Heat load showing 24 hours without reductions(black), 24 hours with reductions (dark grey) and controlscheme for reductions (light grey)Figure 5 clearly shows that the reduction in heat loadclosely follows the control scheme. The largest heatload reduction is about 30% in this example.247Figure 7: Indoor temperature during periods with heatload reductions (squares) and during periods withoutheat load reductions (hourglass)Figure 8 shows readings from two different outdoortemperature sensors during a time period of two days.The graph shows the outdoor temperature sensorwhich is connected to the actual consumer sub-stationin the building (black line). Normally these sensors areplaced somewhat in the shadow to avoid largefluctuations due to solar radiation. We added anothertemperature sensor (grey line) in order to estimate the

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaimpact of this solar radiation. Hence this sensor wasplaced in full view of the sun. The first day was sunnyduring most of the morning until midday, while thesecond day was cloudier.will lower the need of additional heating from theradiator system, by coordinating the thermal inertia ofthe building with freely available heat, e.g. heat fromsunlight or electrical appliances, to balance the heatingneed. This notion is supported by our results as wehave shown that the thermal inertia of even a small ormedium sized multi-apartment building is considerable.How people perceive the indoor climate is dependantnot only on the actual indoor temperature itself but alsoon other factors like air quality, individual metabolismand behaviour, radiation temperature and airmovement. In relation to this it can be noted thatprevious work have shown that about five percent ofany group of people will always be unsatisfied by theindoor climate [9], and that it is not possible to create aperfect climate that will make everyone happy.Figure 8: Outdoor temperature sensors placed in theshade (black line) and in full view of the sun (grey line)DISCUSSIONWhen dealing with temporary heat load reductions it isimportant to include the whole process of the reduction.This also includes what happens after the actual heatload reduction has been performed. For example, whenjust restoring the wanted control level after a longreduction, e.g night time set-back, the forward flowtemperature in the radiator system will rise much fasterthan the return flow temperature. This causes asubstantial, although temporary, heat load increase inthe radiator system which negates large portions of theenergy saving done during the actual reduction. Apartfrom decreasing the local net energy saving thisbehaviour is also less than desired from a system wideperspective, since it causes massive heat load peaks ifdone in many buildings simultaneously, e.g.contributing to morning peak loads. In order to avoidthis it is important to factor in the whole process of thereduction, and make sure that the control systemproperly handles the transition from the reduction levelto the original level. The inability among mostcommercially available control systems to properlyhandle this over-compensation is most likelycontributing a great deal to the lingering controversywhether night time set-back actually gives an energysaving or not.It is important to realize that the definition of anacceptable indoor temperature is not about having theindoor temperature at a certain precise level at all time,but rather to have it within a certain, sociallyacceptable, temperature interval at all time. This hasbeen discussed at great length in previous work [6].The general idea is that a greater temperature interval248CONCLUSIONSThere is an ongoing debate whether night time setbackslead to an energy reduction or not. Results fromthis study clearly show an energy saving in relation toheat load reductions, although this assumes that thecontrol system is able to smoothly handle the transitionfrom reduction to normal operation. The resultsshowing energy saving is evaluated in relation to thetotal energy usage which also includes tap-waterusage. Normally this is estimated to about 30% of thetotal energy use in a multi-apartment building.In prior studies of temporary heat load reductions thefocus has been on the fluctuations in the indoortemperature as a way of evaluating the energy saving[3]. This idea is based on the widespread notion thatany energy saving is linearly proportional to thetemperature difference between the indoor and outdoortemperature. This model might be true in a steady statesimulation where the temperature difference isassumed to have had time to permeate the air mass aswell as the entire building structure, but it is obviouslyinadequate in a dynamic situation. We have insteadfocused on the heat load and energy usage directly, i.e.the difference between forward and return temperaturein relation to the flow within the radiator circuit. In mostof the buildings evaluated there has been aconsiderable reduction of energy consumption withoutany noticeable change in indoor temperature. Thereason that there does not need to be a measurablechange of the indoor temperature is due to thedynamics of the thermal inertia of the building, e.g. thetime constant of a building is not a constant [6]. Thisaspect comes into play when using very short heat loadreductions, at most one or a few hours long. During thisfirst part of the reduction it is mainly the actual air massthat is influencing the indoor temperature drop sincethis body has a low resistance to change, i.e. the shorttime constant [10]. If the heat load reduction isprolonged, like during a night time set-back, the

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniabuilding mass will start to interact with the air mass andthus stabilizing the continuing temperature drop, i.e. thelong time constant [10].The influence of external and internal free heat is largeenough that when these heat sources interact withother parts of the thermal process it hides shorter heatload reductions in the ambient temperature. This canbe seen in Figure 7 where it is shown that although theaverage indoor temperature is not noticeably affectedthere is still a somewhat larger deviation in the indoortemperature which implies that there is indeed a higherlevel of temperature flux within the air mass and thatthis is triggered by the heat load reductions. Thecontrol policies used during this work obviously set ahigh bar for the control system to handle, but as theaverage hardware develops it should be possible toimplement such techniques on a larger scale.Figure 8 gives another clear indication of just howsubstantial such sources of free energy can be. Thisextra heating due to solar radiation through thewindows directly interacts with the mass of air insidethe building, thus raising the temperature.In addition to being able to help save energy usage in abuilding temporary heat load reductions also form thebackbone of DSM and LC, in which the goal is tomanage the heat load (kW) rather than the energyusage (kWh).FUTURE WORKIn the future we plan to further develop models in orderto dynamically estimate the temperature flux withinbuildings and develop theoretical and practicalinterfaces for incorporating this data dynamically intothe control systems.ACKNOWLEDGEMENTThis work has been financed by Blekinge Institute ofTechnology and NODA Intelligent Systems AB.REFERENCES[1] N. Björsell, Control strategies for heating systems,University-College of Gävle-Sandviken.[2] F. Wernstedt, P. Davdisson and C. Johansson,―Demand Side Management in District HeatingSystems‖, in Proc. Of Sixth <strong>International</strong>Conference on Autonomous Agents and MultiagentSystems, 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 niofjärrvärmecentraler i Slagsta‖ Report ZW 04/05,ZW Energiteknik, 2004 (In Swedish)[5] F.B. Morris, J.E. Braun and S.J. Treado ―Experimentaland simulated performance of optimalcontrol of building thermal storage‖, ASHRAETransactions, Vol. 100, No. 1, 1994[6] E. Isfält and G. Bröms. ―Effekt- och energibesparinggenom förenklad styrning och drift avinstallationssystem I byggnader‖, ISRN KTH/IT/M--22--E. Institutionen för Installationsteknik. KungligaTekniska Högskolan, 1992. (In Swedish)[7] S. Ruud. ―Energimyndighetens program förpassivhus och lågenergihus‖ Remissversion 2009-03-10. Forum för Energieffektiva byggnader, 2009.(In Swedish)[8] F. Wernstedt and C. Johansson. ―Demonstrationsprojektinom effekt och laststyrning‖. ISBN 978-91-7381-041-8, The Swedish District HeatingAssociation, 2009. (In Swedish)[9] J. Skoog, ―PM avseende komfort‖, ÅF-InfrastrukturAB, 2005. (In Swedish)[10] C. Norberg. ―Direktverkande elradiatorers regleringoch konstruktion‖ Vattenfall Utveckling AB, Rapportnr F-90:5, Älvkarleby, 1990. (In Swedish)[11] L. Olsson Ingvarsson, S. Werner. ―Building massused as short term heat storage‖ in Proceedings ofThe 11th <strong>International</strong> <strong>Symposium</strong> on DistrictHeating and Cooling. Reykjavik, Iceland, 2008.249

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaVERIFICATION OF HEAT LOSS MEASUREMENTSJ.T. van Wijnkoop 1 , E. van der Ven 21Liandon B.V, 2 Thermaflex <strong>International</strong> Holding B.V.ABSTRACTHeat loss tests are performed on different samples ofthe Thermaflex Flexalen 600 series and oneST-PUR-PE sample at the Thermaflex heatlossequipment and two German test facilities. At thesefacilities two different testing methods are used. Thesemethods are both described in the European standard[1] but show significant differences in the results. In thispaper the different methods of testing are described.Furthermore the Thermaflex heat loss equipment isverified with the test institute that uses the same testingmethod.INTRODUCTIONLast year Liandon developed a test-rig for Thermaflexto measure heat loss of insulated plastic pipingsystems. With this test-rig it is possible for Thermaflexto test the in house produced pre-insulated, semiflexible pipes in various diameters.To verify the test results, the results of the Thermaflexheat loss equipment are compared with the test resultsof two acknowledged institutions. For this paper twoGerman institutions are chosen, since they bothmeasure in compliance with the European standard EN15632 [1], however with different methods described inthis paper. In order to give an appropriate comparison,knowledge of the testing methods of both systems isrequired. In this paper the testing methods of all threesystems is covered, together with the comparison ofthe test-results. Since the testing facilities use twodifferent methods described in the standard, thecomparison refers to the test methods and the testresults.The objective of this paper is to compare the testmethods and test results of the two different testinstitutes with the Thermaflex heat loss equipment andverify the outcome. As in “Heat loss of flexible plasticpipe systems analysis and optimization”(E. Van der Ven et Al.) [4] and “Performance of preinsulated pipes” (I. Smits et Al.) [6] these results areused to compare different sizes of the Flexalen 600series and competitive products.conditions, as defined by the European standard [1].The ability to conduct equally based heat lossmeasurement result in an objective comparison ofdifferent types of (semi) flexible piping systems,providing the opportunity to highlight strengths andweaknesses of (competitive) piping systems.Furthermore, in contradiction to most heat loss tests,the test time in the Thermaflex test-rig is only a fewhours so the test can be performed during production.This provides the opportunity to optimize the productionprocess real-time and measure the heat loss of theproduct several times during a production run. Thisguarantees the quality of the produced batch.In addition the handling of the equipment is made easy,so no specially trained staff is needed for testing,making it possibly for operators to carry out the tests.EUROPEAN STANDARD METHOD DESCRIPTIONThe European standard EN 15632 [1] allows twodifferent methods of heat loss or thermal conductiontesting. These methods both state the same on internalheating of the service pipe but vary on the method ofcompensation for heat loss in axial direction.The first method, the guarded end method, states noaxial heat transfer is permitted. This should beaccomplished by the use of end guards, an extra pairof heating elements at both ends of the service pipe asshown in Fig 1. By heating the ends separately to thesame temperature as the middle test section no heattransfer will take place to the ends of the service pipe.In this case a theoretical compensation is not requiredsince the test section only has losses in radial direction.This method is used in the Thermaflex heat lossequipment and at one of the institutes.NOVELTY AND MAIN CONTRIBUTIONThe Thermaflex test-rig is newly developed for theresearch of heat loss of pre-insulated pipes. Thenovelty of this system is its ability to measure theoverall heat loss of different samples under similar250Fig 1, Longitudinal section guarded end heating probeThe second method described for compensating foraxial heat loss is the calibrated or calculated endmethod. The calibrated end method will not be coveredin this paper since it is not used in our comparisons.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe calculated end method states the ends of theservice pipe shall be insulated with a known thermalconductivity as shown in Fig 2.X, L: distance to next measuring point from the middle,sample lengthT 0m ,T 0X ,∆T 0m ,T 2 : pipe temperature at the middle of thetest section, temperature at distance X of the middle,temperature correction, temperature at insulationsurface.Fig 2, Configuration calculated end cap.The service pipe is heated, using a heating elementwith only one section. During the tests a thermal profileis made of the outer casing of the sample, showinglower values at the ends. After testing the heat loss iscompensated for the end loss with the van Rinsum orNukiyama theory. For this investigation only the vanRinsum theory is used and therefore described.According to the van Rinsum theory, the axial heat losscauses a decrease in temperature not only towards theends of the service pipe, but in the test section as well.With the use of the equations (1), (2), (3) thistemperature decrease in the test section can becalculated and added to the measured value,compensating the end loss. This corrected temperatureis used in equation (4) to calculate the overall thermalconductivity. This method is used by one of theGerman institutes. calcc D 2 lnD 0 2LT 0m T 2 2 calc DA 1 1 A 2 2 2 lnD 0 T 0m T 0XT 0m cosh Xc 2 D 2 lnD 0 LT 0m T 0m T 2 λ calc : approximate value of thermal conductivity(1)(2)(3)(4)D 2 /D 0 : outer/inner diameters of casing and service pipeA 1 , A 2 : areas of the heating probe, inner service pipeλ 1 , λ 2 , λ: thermal conductivity of heating probe, thermalconductivity of medium in the service pipe, thermalconductivity total sample.251VERIFICATION OF SAMPLESTo verify the outcome of the Thermaflex heat lossequipment and the laboratory tests, three samples ofthe Flexalen 600 piping system are tested on theiroverall heat loss. These samples consist of 2 or 3 m ofthe pre-insulated piping system. More informationabout the Flexalen 600 system can be found in “Heatloss of flexible plastic pipe systems analysis andoptimization” (E. van der Ven et Al.) [4]. Furthermore,method comparison tests are performed on competitivepre-insulated piping systems, a comparison of theproducts themselves is given in “Performance of preinsulated pipes” (I. Smits et Al.) [6].The tests on the Flexalen 600 products are performedby Thermaflex and by one of the acknowledgedinstitutes, using the different methods. To ensure theeffect of ageing in the Flexalen 600 system is the sameduring all tests, the Flexalen 600 samples are testedsimultaneously. To exclude effects of the productionprocess both tested samples are half of a 6 meter stick.An alternative method is used for Flexalen 50A25 andcompetitive products. Here the same sample is testedat the different test facilities.The comparison of the results is based on the outcomeof heat loss per meter, calculated as described in theEuropean standard [1]. This loss per meter is onlyconclusive on a very small part of the entire system.Therefore the complete Flexalen 600 system will becovered in paper “Heat loss system optimisation” (J.Korsman et Al.) [3] and ‗‟New economical connectionsolutions for flexible piping systems” (C. Engel et Al.)[5].In this report the following diameters of the Flexalen600 piping systems are used for comparison of themeasurements: Flexalen 600: 50A25, two guarded end tests* and calculated endtest. 160A90, one guarded end test* and calculated endtest. 200A110, one guarded end test* and calculatedend 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 notyet available.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTHERMAFLEX HEAT LOSS EQUIPMENT 600The Thermaflex heat loss equipment is speciallydesigned for the Thermaflex Flexalen 600 series. Oneof the major design goals was to develop a fast andeasy to use test rig with the precision of a laboratorialtest. These goals have resulted in a test rig that is ableto measure heat loss in a few hours, allowing directoptimization during the production process, and isoperable by the production staff without the loss ofaccuracyPhysical test facilityThe physical part of the Thermaflex heat lossequipment consists of three segments.The first is the water cooled compartment in which alltests are performed. This compartment is kept at aconstant temperature, (23 °C), during eachmeasurement.The second is a heat source, for which heating probesare used. These heating probes are custom made byequipping a two meter Thermaflex piping segment, ofall available diameters, with three heating coils.The third part of the heat loss equipment is the controlunit. Here the heating probe is powered and all thermalreadings are done. By applying custom made softwareall desired readings can be done. The final output is theactual heat loss in W/m through the entire pre-insulatedFlexalen pipe, consisting of the service pipe, insulationand outer casing.section. With this method it is possible to measure theheat loss by measuring the power needed to maintain aconstant temperature of the test sample. Incontradiction to the measurements at the test institutes,the Thermaflex heating probes temperature isregulated by PID controlled power supplies. In the testresults Graph 2 the power consumption versus testtime is shown. This variable power supply makes itpossible to pre-heat the probes in a short period oftime, shortening waiting times considerably.Furthermore the use of the actual pipe material as aheating probe increases the accuracy. Moreover iteliminates all additional heat loss by convection thatwill be present with the use of smaller, not inner servicepipe connecting heating probes.For testing competitive products with differentdiameters these advantages are lost. However by theuse of thermal compartments in the service pipe thetest results can be guaranteed.Thermaflex method of testingFor testing, the heating probe with the appropriatediameter is inserted in the insulation covered with outercasing, and inserted in the cooled test section. Afterconnecting the probe to the control unit themeasurement can be started. Different testingconditions can be entered at this point such as theinner pipe temperature, representing the internalmedium. When the test is started the heating coils heatthe inner side of the probe until the desiredtemperature is reached. When the inner temperature isconsidered constant and uniform throughout the threeheating coils, the actual measurement is started. Toensure a constant temperature in the probe, a waitingtime is built in the software that will reset themeasurement if temperature exceeds presettemperature values.Fig 3, Thermaflex heat loss equipment 600.Measurement principle ThermaflexThe Thermaflex test rig is designed in compliance withthe European standard [1] and also the tests arecarried out according to ISO 8497 and EN 15632. Inthe design of the heating probes the most realisticmethod, the guarded end method, is used. Accordingto this method the heating probes are equipped withthree heating coils with separate power supply. Asshown in Figure 2, two 400 mm heating coils located ateach end of the 1000 mm test section. These twosections provide a thermal insulation at both ends ofthe test section since all three are kept under uniformtemperature, eliminating axial heat loss of the test252The heat loss measurement is done by measuring theenergy required to keep the probe at a constanttemperature, by measuring the current at constantvoltage in the heating coils, and calculating the powerconsumption. Since the middle/testing coil is exactlyone meter in length the required energy represents theexact heat loss through one meter of piping andinsulation in W/m. Since the actual piping material isused during the measurement, there are no otherlosses, nor advantages, than there will be in practise,ensuring an objective measurement. Furthermore arealistic fit of the insulation material is guaranteed. Asstated in the foregoing paragraph these advantagesare lost for divergent diameters. However during thisinvestigation the probes have proven suitable fortesting, as both testing institutes also use smallerheating probes.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTest results Thermaflex heat loss equipmentIn this paragraph the test results are presented for thetests carried out with the Thermaflex heat lossequipment. For this study four different types of theFlexalen 600 series were tested. The tests for theFlexalen 600 series took place at three differenttemperatures, 60, 70 and 80 °C. The values at lowertemperatures are calculated using the linearizationmethod described in the European standard EN 15632[1]. In the following tables and graphs the test results ofthe Thermaflex heat loss equipment are presented.The power usage during the testing cycle is shown inGraph 2. In this graph the first 40 minutes represent theheating and stabilisation time for the heating probe andinsulation, whereas the last 30 minutes is the actualtest time. Since, as the figure shows, the temperatureis constant, the power usage equals the heat lossthrough the piping system in radial direction during thelast 30 minutes. The results, as given in Table 1, arecalculated by using the mean of the powerconsumption during the last 30 minutes of the heat losstest. The results in Table 1, are also displayed inGraph 1 for the three tested samples.Table 1, Results heat loss equipment for the Flexalen 600productsHeat loss of the Flexalen 600 series in W/m tested onthe Thermaflex heat loss equipmentProduct 40 °C 50 °C 60 °C 70 °C 80 °C50A25 3.6 6.4 9.3 12.0 15.0160A90 6.2 10.1 14.0 17.9 21.8200A110 6.5 12.0 17.5 23.0 28.5Graph 2, Power and temperature of the Thermaflexheating probe.Outcome Competitive products for comparison oftesting method:For the comparison with test institute two, two samplesof competitive products are tested. As these samplesare ST-PUR-PE system, a correction has been madefor using the PB heating probe using the Wallentén [2]method. First the thermal conductivity of the insulationis determent by the use of equation (5), hereafter theheat loss is recalculated without the heating probe,using the temperature of the inner service pipe inequation (6). The results are presented in Table 2 andGraph 3. iT p T c2 probe d 3ln d 2 1 dln2 1 d 6 ln 1 d ln4 st d 1 p d 5 c d 3 Wm K (5)Heat loss [W/m]302010Heat loss results thermaflex heat loss equipment 2 T st T c corrected1 d 2ln 1 dln3 1 d 4 ln st d 1 i d 2 c d 3Where: Wm (6)T p ,T c , T st =Probe, Casing and Steel pipe temperatured 1 to d 6 = inner/outer diameters of service pipe, casingand heatingprobe040 50 60 70 80Temp erature inner service pipe [°C]Flexalen 50A25Flexalen 160A90Flexalen 200A110λ st , λ i , λ c , λ p = heat coefficient of service pipe, insulation,casing and probeΦ probe , Φ corrected =probe power and corrected heat loss.Graph 1, Results heat loss equipment Flexalen 600products.253

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 2, Results heat loss equipment for the competitiveproductsHeat loss of competitive products in W/m tested onthe Thermaflex heat loss equipmentSample 40 °C 50 °C 60 °C 70 °C 80 °CSample 1 2.5 4.5 6.4 8.4 10.3Sample 2 6.5 8.8 11.2 13.5 15.9made of metal. Furthermore no heat guards are used.This means the outer ends of the piping system areinsulated and the heat loss is corrected with acalculated value. In the paragraph ―European standardmethod description‖ a more detailed description isgiven. As can be seen in Fig 4 the heat distribution inthis case is not uniform along the test specimen,proving the need for the van Rinsum correction.Sample 320Heat loss results thermaflex heat loss equipmentHeat loss [W/m]15105040 50 60 70 80Temp erature inner service pipe [°C]Competitive samp le 1Competitive samp le 2Fig 4, Thermal image of the sample at institute oneIn contradiction to the Thermaflex test rig, no integratedcomputer controlled power supply system is used. Thepower for the heating probe is first theoreticallycalculated and manually set to this value. For thetemperature measurement thermocouples and a datalogger with computer link are used.Graph 3, Results heat loss equipment for the competitiveproductsTEST INSTITUTE ONEThis institute is specialized in measuring heat loss indifferent types of insulation. The test facility used forthe Flexalen 600 system is specially designed formeasuring the heat loss of (pre-) insulated pipingsystems. This means the facility is designed tomeasure all different types and diameters.Measurement principle institute oneThe measurements are all based on the calculated endapparatus, using the van Rinsum theory as correction,as described in the paragraph European standard [1]method description of this paper.Physical test facilityThe physical part of the test facility is similar to theThermaflex test rig and also consists out of the threeelements: A temperature controlled compartmentwhere the tests are carried out at a constanttemperature of 23 °C. FIW also uses heating probes asa heat source but, since it is not specially designed forthe Flexalen 600 system, they are made to fit allsystems. To ensure the fit of the probes in all differentsystems the diameters are smaller, and for durability254Method of testingThe heating probe is positioned in the centre of the testpipe with positioning foam in three sections of the pipe.On these foam blocks four thermocouples are placed in0, 90, 180 and 270 degrees on the inner surface of theservice pipe. For the outcome of the pipe innertemperature the mean of the four values is used. Tomeasure the temperature on the outside casing of theinsulation, five groups of four thermocouples are usedin the same configuration as the inner pipe. Thedifference being that the thermocouples are placedboth on and in between the corrugations of the casing.The test sample, with the heating probe, is placed inthe conditioned container thereafter the test can bestarted. The power supply of the heater is turned on bysetting the voltage and current of the power unit to afixed value so the electrical power equals thecalculated heat loss.Depending on the diameter of the test sample and thetest temperature the waiting time for the heating of thesample is five to eight hours due to the low, fixed powerinput. After a constant temperature of the outer casingis achieved the actual test cycles start. Each test cycleconsists of a measurement of 30 min in which the outercasing temperature is to be constant. If not the cyclehas to be restarted. In total ten cycles will be performedon each sample. After the test the values are corrected

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniafor axial heat loss and the thermal conductivity, thermalresistance and overall heat loss are calculated.Test results institute oneAs the actual measurement data are not available dueto the correction factor, only the calculated values canbe discussed in this paragraph. As soon as the actualmeasurements become available this section will beupdated. Furthermore the results are not given forexactly 60, 70 and 80 °C due to the fixed power supplywith no temperature set point, the displayed results arecalculated heat loss values at the set temperatures tomake the data more interpretive. For this calculationthe linearization method described in the Europeanstandard [1] is used. In Graph 4 the data from Table 3is presented as a graph.Table 3, Results test institute one for the Flexalen 600productsHeat loss of the Flexalen 600 series in W/m tested attest institute oneProduct 40 °C 50 °C 60 °C 70 °C 80 °C50A25 5.6 8,8 11.9 15.1 18.3160A90 9.1 15.1 21.1 27.1 33.0200A110 9.8 15.1 20.5 25.8 31.2Heat loss [W/m]40302010Heat loss results test institute oneAn update to this paper will be made as soon as theFlexalen 600 results will become available.Testing method institute twoThe method used by this institute is generally the sameas the method used by Thermaflex; however the testfacility itself is different.Physical test facilityThe testing facility at institute two consists of atemperature controlled room, kept at the prescribed23 °C. As a heat source a heating probe, consisting ofa 2 m test section and two 50 cm end guards is used.At the time of writing no further information on the testfacility was available. This paragraph will be updatedwhen this information becomes available.Method of testingPrior to testing, the sample is prepared by placingthermocouples in various locations on the inner servicepipe and outer casing. Subsequently the sample isplaced in the temperature controlled room and theheating probe is inserted. By setting the power supplyto a calculated value for all three heating coils theheating process of the sample is started. Because ofthe low fixed value of the power supply, this heating willtake approximately 5 to 8 hours. After the desiredtemperature is reached at the test section as well as atthe guarded ends, the actual test is performed. The testconsists of a power reading during a 30 min cyclewhere het temperature of the test section and guardedends may not exceed the limit of an yet unknownbandwidth.Test results test institute twoThe test results of institute two are given in Table 4 andGraph 5. As not all data was available during writingthere tables and graphs will be updated.040 50 60 70 80Temperature inner service pipe [°C]Flexalen 50A25Flexalen 160A90Flexalen 200A110Graph 4, Results test institute one for the Flexalen 600productsTEST INSTITUTE TWOTable 4, Results test institute two for the competitiveproductsHeat loss of competitive products in W/mtested at test institute twoSample 40 °C 50 °C 60 °C 70 °C 80 °C90DN25 1.95 4.00 6.02 8.06 10.09For the second test institute in this research, aninstitute using the same guarded end method ischosen. This makes it possible to provide a correctcomparison between the test results and not only thetesting method. The tests carried out by test institutetwo at the time of writing are of competitive productsonly as the facility was already running on full capacity.255

Heat loss [W/m]15105Heat loss results test institute two040 50 60 70 80Temperature inner service pipe [°C]Competitive sample 1Competitive sample 2The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniacompetitive products show consistency with testinstitute two as shown in Graph 7. The heat lossequipment values are just a little higher, which can beexplained by the need to cut the sample in order toplace the heating probe with thermocouples in the rightposition. The difference between the outcome of thetest on sample 1 are 0.39 and 0.22 W/m at an innerservice pipe temperature of 60 and 80 °C respectively.These values are within the combined accuracy rangeof both facilities. This comparison, although only basedon one test, proves the worthiness of the Thermaflexheat loss equipment and will be updated as more datacomes available.15Heat loss comparison Thermaflex and test institute twoGraph 5, Results test institute two for the competitiveproductsCOMPARISON OF THE TEST RESULTSComparison of the Thermaflex flexalen 600 series:Although both methods, guarded end and calculatedend, are approved and described in the Europeanstandard [1], the difference between the results issubstantial as displayed in Graph 6. Moreover allresults vary more as the temperature differenceincreases. This can be explained by the use of thecalculated end caps that conduct more energy athigher temperature differences. As these end caplosses increase, the corrected thermal conduction forthe sample also increases, resulting in a highercalculated heat-loss.Heat loss [W/m]3530252015105Heat loss comparison Thermaflex and test institute one040 50 60 70 80Temperature inner service pipe [°C]Flexalen 50A25 Thermaflex resultFlexalen 50A25 Institute one resultFlexalen 160A90 Thermaflex resultFlexalen 160A90 Institute one resultFlexalen 160A90 Thermaflex resultFlexalen 160A90 Institute one resultGraph 6, Comparison results of the heat loss equipmentand test institute oneComparison of competitive products:Although the Thermaflex heat loss equipment wasdesigned for Flexalen series, test results onHeat loss [W/m]105040 50 60 70 80Temperature inner service pipe [°C]Competitive sample 1Competitive sample 1Competitive sample 2Competitive sample 2Graph 7, Comparison results of the heat loss equipmentand test institute twoCONCLUSIONDuring this research it has become clear that theEuropean standard [1] tolerates differences in heat lossvalues by allowing different testing methods. Theoutcome of the tests indicate that the result of theguarded end cap method varies from the result of thecalculated end cap method, however no assumptionscan be made based on only one comparingmeasurement. Further study that is being conducted atthis moment will provide more comparison data. Thiswill be updated with this data as soon as becomesavailable. This new data could point out that the vanRinsum theory is not suitable for accurate heat lossmeasurement of plastic piping systems.The comparison of the guarded end method resultsfrom test institute two and the Thermaflex heat lossequipment conclude that the results of the heat lossequipment are correct and comply with the Europeanstandard [1]. This validation makes the results of theThermaflex heat loss equipment valid for not only inhouse testing but also for publication as done in “Heatloss of flexible plastic pipe systems analysis andoptimization” (E. van der Ven et Al.) [4] and “Performanceof pre insulated pipes” (I. Smits et Al.) [6].256

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFURTHER INFORMATIONQuestions concerning the paper can be addressed to:Thermaflex <strong>International</strong> Holding B.V.Veerweg 15145NS WaalwijkThe NetherlandsLiandon B.V.Dijkgraaf 46920AB DuivenThe NetherlandsACKNOWLEDGEMENTAcknowledgments go to both the test institutes for theiropen and honest explanation of their testing methodsand facilities and for even showing the entire facilityand methods.Furthermore acknowledgements go to all involvedemployees of Thermaflex Isolatie B.V. and LiandonB.V. who made this research possible. Specialacknowledgements go to P. Blom and P. van Rijswijkfor their devotion on all the heat loss measurementsthey performed during this research in a short amountof time.REFERENCES[1] NEN-EN 15632 and NEN-EN-ISO 8497[2] P. Wallentén, ―steady-state heat loss frominsulated pipes‖, Lund Institute of Technology,Sweden, 1991[3] J. Korsman and G. Baars, ―Heat loss systemoptimization‖, <strong>12th</strong> ISDHC 2010[4] E. van der Ven and R. van Arendonk, ―Heat lossanalysis and optimization‖, <strong>12th</strong> ISDHC 2010[5] C. Engel and G. Baars, ―New economicalconnection solution for flexible piping systems‖,<strong>12th</strong> ISDHC 2010.[6] I. Smits and E van der Ven, ―Performance of preinsulated pipes‖, <strong>12th</strong> ISDHC 2010.257

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDISTRICT HEATING AND COOLING WITH LARGE CENTRIFUGAL CHILLER-HEATPUMPSUlrich PietruchaFriotherm AG, SwitzerlandABSTRACTWith prices for primary energy resources soaring, therecovery of "waste energy" was getting into the focus ofattention within the last years. Also the global climatechange reminded us to limit the use of primary energyresources to a minimum, thus exploiting "waste energy"potentials wherever feasible. The process of upgradinglow grade waste heat is especially interesting wherelarge amounts of such energy are available at onepoint, e.g. next to sewage water treatment plants,alongside main sewers, in power plants or close toground water sources.Even if the "waste energy" potential is abundant andeasily exploitable, the aspect of overall thermalefficiency is considered crucial for the final decision toinvest in large heat recovery installations.Number of units 2TypeUNITOP® 50 FY/34FYRefrigerantR134aHeat source medium Raw waste waterRaw sewage water inlet 10.0 °C ... ~ 15 °CRaw sewage flow water flow 3800 m3/hHeating water temp. in/out 60 / 90 °CHeating water flow824 m3/hPower at terminal9‘750 kWHeat capacity27‘600 kWCoefficient of performance 2.83 up to >3.0INTRODUCTIONDescribed are five applications of large centrifugal heatpumps-chillers for the use in large districtheating/cooling systems.Application 1: Heat recovery from raw sewage waterand hot water production at 90 °C.Application 2: Combined heating and cooling with araw sewage water heat pump/chiller installation. Thisplant is operated successful since 1989.Application 3: Combined heating and cooling: acombination of cooling with simultaneously heatproduction in summer and heat recovery from cleanedsewage water in winter.Application 4: Heat recovery from wet flue gascleaning processApplication 5: Combined heating and cooling inStockholm1. SKOYEN VEST PLANT IN OSLO: HOT WATERPRODUCTION AT 90 °CThis is the world's largest heat pump plant using rawwaste water as heat source. It is installed in a cavernalongside one of the main waste water channels inOslo. With 2 heat pumps a heating capacity of 27'600kW is generated by recovering heat from raw wastewater.One of the Skoyen heat pumps2. THE SANDVIKA PLANT IN OSLO: COMBINEDHEATING AND COOLING FROM A RAW SEWAGEWATER HEAT PUMPThis is the oldest combined chiller/ heat pumpinstallation in the world, producing simultaneouslycooling, taking out heat from raw sewage water andproducing heating capacity for the district heatingsystem.The heat pumps are in successful operation since 1989and each one has an additional heat exchanger, whichis used either as raw sewage evaporator or as rawsewage water condenser.Each heat pump has an overall operating time of about160'000 hours, means the heat pumps were operatedsince 1989 each year for more than 8'400 hours.A 3rd larger heat pump was taken into operation in2008.258

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniapoints with lower district heating temperatures where aCOP of up to 6.5 can be achieved.2 Friotherm heat pumpsType UNITOP 28CX-71210UFrom districtheating-networkHeat source capacity24.3°C 15.5MWHeat sink capacity19MW50°CFlue-gasCombined chiller / heat pump at Sandvika plantFlue- gascondensingPower consumption3.5MW3. THE KATRI VALA PLANT IN HELSINKI:COMBINED HEATING AND COOLINGThis is the largest combined chiller heat pumpinstallation in the world producing simultaneously60MWth cooling and 90MWth heating, i.e. totalproduced thermal energy is 150MW. The requiredelectrical input is 30MW i.e. a superb COP of 5 can beachieved (150MW / 30MW).During Winter season the required cooling is done bysea water, while heat is produced by using cleanedwaste water as heat source.Flue-gascleaning34.2°CSteam- turbineBoilerGeneratorWaste-to-Energy plant SYSAV Malmö Sweden59.2°C5. NIMROD STOCKHOLM: COMBINED HEATINGAND COOLINGDue to the fact that with every cooling process there isalso waste heat generated, Friotherm AG, which hasworked since many years on chillers with heatrecovery, has worked out a concept which allowsvarious operating modes in order to operate the chiller /heat pumps more efficient over a longer period and,making therefore the investment more attractive:Typical Unitop 50FY heat pump (Q heat 15 to 23MW)There are 4 chiller / heat pumps installed in the Nimrodplant. The centrifugal compressors are switched inparallel for Summer cooling production of 48MW.However during this period heat recovery is notrequired as there is sufficient capacity available fromthe existing heat pumps.4. SYSAV MALMÖ: HEAT RECOVERY FROM WETFLUE GAS CLEANING PROCESSSYSAV Malmö in Sweden has built a new waste-toenergyplant. An important part in this plant was theinstallation of a 19MW heat pump using the flue gascondensation as heat source. The heat pump issupplying hot water with a temperature of up to 70 °Cto the district heating system of the community ofMalmö.The two heat pumps are connected in series on theheat source side and on the heat sink side; thisimproves considerably the COP. There are operating259The same units are producing during Spring, Autumnand Winter a cooling capacity 24MW with a full heatrecovery of 35.6MW at a temperature level of 78 °C.For heat recovery operation mode the centrifugalcompressors are switched in series.Each chiller / heat pump consists of two centrifugalcompressors Type Uniturbo 33CX and 28CX and isable to operate at the following modes, describedbelow:5.1 Cooling only:During Summer with high cooling demand, the wasteheat from the condenser is removed with sea water of

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniamax. 22 °C, therefore the condenser and sub coolerare equipped with Titanium tubes.The two compressors Uniturbo 33CX and 28CX arethen working in parallel, in a single stage mode, with asingle stage expansion, producing a cooling capacity ofup to 7MW plus 5MW = 12MW i.e. with 4 units a total of48MW.Depending on the cooling demand, one or the other, orboth compressors can be put in operation.If needed, the part load of each chiller / heat pump canbe controlled down to 10% of its nominal capacity, witha reasonable high efficiency, with the use of inlet guidevanes. The chilled water temperature outlet is keptconstant to 5 °C5.2 Combination of cooling and heating:During Spring, Autumn and Winter, with moderatecooling demand of up to 24MW, but simultaneous needof heating, the waste heat from the condenser issupplied to the district heating network at atemperature outlet of 78 °C and a maximum heatcapacity of 35.6MW.The two compressors Uniturbo 33CX and 28CX arethen working in series in two stage compression mode,with two stage expansion using an economiser afterthe first stage expansion.The compressor Type Uniturbo 33CX with the largervolume flow is working as 1st stage and the TypeUniturbo 28CX with the smaller volume flow as 2ndstage compressor.The control system is controlling the required coolingcapacity; the surplus heat is supplied fully to the districtheating network at a temperature level of up to 78 °C.I.e. this operation mode delivers heat which can besold in addition to the cooling, with a total COP ofabove 5.The cooling only mode and the combination of heatingand cooling mode are explained in the below P&I‘s:CONCLUSIONReliability of technology, future developments andchallengesAbout more than 140 heat pumps, producing hot waterwith temperatures above 70 °C, are installed worldwide since 1980. The heat pump plants described inthis article are only showing a small part of thenowadays available applications.Almost all of the installed heat pumps plants startingfrom the early 1980's are today still in operation, whichis showing the high reliability of this technology.Nimrod single stage operation2 compressors in parallel288Today developments are the extension of the heatpump operation range in temperature and capacity toexploit new heat sources and to extent the field ofapplications.The adaptation of the centrifugal heat pumps to newrefrigerants with GWP close to zero is already on theway.REFERENCESSea waterCooling capacity 12 MWNimrod two stage operation2 compressors in series28Heating water33Heating capacity 9 MWCooling capacity 6 MWText and pictures from Friotherm AG / Switzerland.260

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaNEW ECONOMICAL CONNECTION SOLUTION FOR FLEXIBLE PIPING SYSTEMSChristian Engel, Gerrit-Jan BaarsThermaflex <strong>International</strong> Holding B.V.ABSTRACTMost Energy Companies are facing the same problem:Connection costs per house shall be cheaper andfaster to install to reach more customers. At the sametime high level of durability and a system free ofmaintenance must be guaranteed.This paper shall give an insight of the practicalexperience with new solutions showing the economicand ecological advantages in projects with severalEnergy companies.INTRODUCTIONDistrict Heating & Cooling networks are a major costfactor for Energy Providers and subject to permanentsearch for cost improvements.Flexible plastic pipe systems have been a major stepfor cost reduction in low temperature networks. Withthe new EN 15632 [1] the necessary basis forcertification of these systems has been laid. This is amilestone in terms of acknowledgement for flexiblesystems as a proven part of future networkdevelopments.As flow temperatures and pressures are reduced, thefield of application for flexible plastic systems isincreasing. Until recently only a small percentage ofDistrict Heating companies have started to use plasticpipes in their networks. These were kind of pioneerswho co-created systems together with the industry.Together with the University of Leoben, the long termdurability of two types of plastic medium pipes, made ofPB and PE-X were investigated. The research made byDipl.Ing. E.Kramer and Univ.Prof.Dr.J. Koppelmann [2]was based on OIT (oxygen induction time), tearstrength, elongation at break and internal pressuretests to determine the lifetime of plastic pipes at 80, 95and 110 °C. The final results were in favour for pipesmade of PB. The calculation of the lifetime for PB pipeswas based on a typical temperature profile used insecondary district heating networks of STEWEAG. Thelifetime expectancy was stated with 36 years.The decision of STEWEAG was made for PB pipes dueto their more homogenous structure, superior flexibilityand allowance for welded joints. More than 250 km ofthis system have been installed since 1981 insecondary networks operated by STEWEAG. See alsoUniv. Prof. Dr. E. Hönninger, STEWEAG [6].WHAT CAN BE SOLVED WITH FLEXIBLE PLASTICSYSTEMSApart from the high and long term investment costs, thefollowing main problems had to be solved as well:Corrosion problems in conventional Systems madeof Steel/PUR/PE or Cu/PUR/PE or Cu/Mineralwool/PEHeat loss due to wet and aged insulationSystem shut downs for maintenance and repairThe first co-development of such a system was startedalready in 1980 by the Austrian Electricity companySTEWEAG. They were looking for a pre-insulatedpiping system as easy to install as an electric cable.Photo no.: 2 corroded steel pipe connectionTHE NEW FLEXIBLE PIPE GENERATIONPhoto no.:1 first Flexalen installation 1981In 2001 the Dutch Energy Provider NUON started adevelopment co-operation with Thermaflex to createeven more flexible and moisture resistance pipingsystems. Target was again to reduce the connectioncosts for new district heating projects. The new system261

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniadeveloped is called FLEXALEN 600, the improvedversion of the system used by STEWEAG. The systemconsists of a PB (Polybutene) medium pipe and PO(Polyolefin) insulation foam welded to a HDPE (highdensity Polyethylene) outer casing.up to d63 – there trenches can be compared to cableducts. Another fact is that connections are most of thetime only necessary at branches, for sticks trenchesneed to be suitable for execution of the welding and theinsulation process.With a new inline production process it was possible toweld the moisture resistant insulation to the outercasing. The targets of a corrosion proof and moistureresistant insulation were met.Photo no.: 4 Steel compared to FlexalenPhoto no.:3 Flexalen 600 longitudinal cutFLEXALEN has been the first system to pass acertification and 3 rd party control by KIWA, which issimilar to the new EN 15632.DECREASING INSTALLATION COSTSThe most obvious advantage is the chance to reducethe installation time with flexible systems supplied incoil lengths of 100m and more. Compared to rigidsystems the following relative costs have been realizedin actual projectsAlthough material cost for plastic pipes are higherespecially for larger dimensions, the total installedsystem costs are lower, especially when using doublepipe systems wherever possible (see Photo nr. 5).BRANCH SOLUTIONSUntil now 2 types of branch solutions have been used.On site welded solution with Half-shells plus insulationto cover (see Photo no.:4). This technology has beenused for smaller networks. Due to homogenouswelding techniques, either with polyfusion or withelectrofuction fittings, the branches are corrosion freeand offer the same inner diameter as the pipes.Table 1Pre-insulatedsteel pipesFLEXALENMaterial costs 100% 90–150%Installation time 100% 20–25%Trenching 100% 50–70%Total 100% 60–85%Material costs are depending on the dimensioning ofthe system in the first place. In case of optimization ofpipe sizes and lengths according to the advantages ofPB pipes and connection systems, as described lateron, the material costs can be reduced for Flexalen.Installation costs are proven in practical experiencessince almost 30 years. Flexalen systems can beinstalled 5 times faster than rigid systems.Lower costs in trenching is related to the fact thatFlexalen systems are supplied in double line systems262Photo no.: 5 Polyfusion welded Flexalen branchPre-insulated Tees are another way to secure morereliable network quality due the reduced number ofjoints to be made on site (see Photo no.: 6). Thestraight connections are insulated with special kits witha robust slide over HDPE tube, which is sealed to theouter casing with heat shrinks.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaPhoto no.: 7 Narrow space for district heating lines underbuildings in NLThe result was the ―Flexalink (Flexalen T-Link)‖ solutiona very small, flexible and 100% watertight system, prefabricatedand pressure tested by Thermaflex NL.Photo no.: 6 Pre-insulated branchThe pressure for even more economic solutions forconnections has led to further innovations in close cooperationbetween the Dutch Energy Provider Enecoand Thermaflex.PRE-FABRICATED NETWORKSCompared to the branch solutions described before,these new solutions take full advantage of flexiblewelded systems, in order to further reduce the numberof joints on sites.A new type of pre-fabricated network has beendeveloped. High flexibility and a minimum ofconnections was the goal.The first application was the district heating networkCapelle a/d Ijssel in the surroundings of Rotterdam forrenovation in difficult circumstances under houses(high ground water level) to replace corroded heatingand sanitary distribution systems.The space under the houses is so small that neitherwelding nor mechanical connections can be carried outin a safe way. Steel welding is even forbidden underthese conditions.Photo no.: 8 Flexalen T-LinkThis solution combines the following advantages: Factory made welding and branch insulation – allwatertight and pressure tested. Connections are made under clean manufacturingcircumstances. No weather influences, no failurecosts. Customer made connections according to the realsituation. Light weight and flexible for easy sliding into thetrench or under the house Fast Installing time (first project experience 10houses/day) Reducing connection costs in new buildingprojects. Less system parts on the building area.This development covered all the wishes andrequirements of the Dutch Energy Provider ENECO.They have already ordered this system for 800 houseconnections in 2010.263

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaPhoto no.: 9 Installation of Flexalen T-LinkPhoto no.:12 Welding of T-Link to the next sectionAll experiences so far have been very positive andhave led to further applications already.RENOVATION IN PURMERENTPhoto no.: 10 Installation of Flexalen T-LinkThese 2 photos (no 8 and 9) demonstrate how easy itwas to slide the connection into the duct under thehouse.Also further connections between the pre-fabricatedsections were made before sliding the entire systemunder the house. Only the last connection had to bemade in the duct. See also photos no.:11 & 12.The situation of the current district heating network inPurmerend (Energy supplier StadsverwarmingPurmerend) is very critical. Due to higher ground waterlevels than expected the current metal pipe systemshave corroded and need to be replaced.As the network has been installed under the basementof the attached housing schemes, the space for theinstaller is very tight and it is not allowed to use anysteel welding process in these circumstances. FlexalenT-Link has been identified to be the ideal solution.Purmerend has ordered this system for 300 houseconnections for 2010 already.For the renovation market in the Netherlands, thissolution has shown big advantages. This solution isnow available for Energy Provider worldwide not justfor renovation, but also for new networks.PRE-FABRICATED NETWORKS FOR NEWPROJECTSPhoto no.: 11 pipes under the buildingThe conditions in new building situations are mucheasier and this solution can help to reduce theconnection costs.Both Dutch Energy Provider, Eneco and Nuon areinvestigating this new solution for new building projects.264

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 1 network scheme suitable for T-LinkEspecially for networks with short distances betweenthe branches and the connection to the houses, a highdegree of pre-fabrication can be offered.One solution is the T-Link as described before. Anotherone can be a main line up to 100 m with factory weldedand insulated Tees. This reduces the work on site tojust 1 welding for the house connection line. Thissolution is interesting for longer distance houseconnections.THE IMPORTANCE OF THE NETWORK DESIGNFlexible PB piping systems offer important advantagescompared to other plastic and steel systems in terms oflayout and design.Compared to steel pipes flexible PB systems can belaid more direct as the system is flexible and fully selfcompensating.Expansion loops and elbows can besaved. The saving in pipe length can be calculated with7–10%.PB systems offer low friction loss and show nocalcification or incrustation during the lifetime. Thepolyfusion welded fittings have at least the same innerdiameter as the pipe and offer the same high abrasionresistance. Taking this into account, some extrasecurity factors used in pipe dimensioning can beeliminated.265PB systems can be operated with much higher flowspeed; hence smaller dimensions can be used for thesame load requirement. See also J. Korsman, I.M.Smits, E. van der Ven [4].With relation to the topic of this paper, the followingadditional savings can be made during the networkdesign: Looking for a new building area, mostly streets withblock of houses there are two possibilities:Installing under the floor or Installing in the streets. For every house connection under the floor onlytwo welds and two insulation sets are necessary.Reduction of the installing time/costs by 50–60%. For every house connection in the street, apre-fabrication e.g. for 8–10 house-connectionsbuilt in into one 100m coil can safe installingtime/costs totally including excavating the trenchesof 70%Taking all these possibilities for savings andoptimization into account, the next most important topicfor Energy Provider, the efficiency of the network inoperation, can be tackled as well. Due to the possiblereduction in network length and pipe diameter, theoverall heat loss can be reduced as well. See alsoresults 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 <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaCONCLUSIONNew connection solutions, which meet therequirements of Energy Companies in terms of lowerinvestment costs, faster installation time and durability,have been co-created with leading Energy Suppliers inAustria and The Netherlands.These solutions are based on flexible and weld-ableplastic systems and have been used successfully up to29 years in secondary networks with maximumoperation temperatures of 95 °C (peak temperature)and maximum pressure of 8 bars.The latest development is going into the direction of ahigher degree of pre-fabrication, by including the entireconnection line to the houses as well as parts of themain line into one piece, made up and fully pressuretested in the factory.The experiences in recent projects are showinginstallation times 5–10 times faster compared toconventional pre-insulated steel. The number ofconnections to be made on site is significantly reduced.Successful projects with Energy Suppliers in TheNetherlands are confirming the advantages of this newconnection solution.ACKNOWLEDGEMENTREFERENCES[1] EN 15632 District heating pipes, Pre-insulatedflexible pipe systems, Requirements and testmethods[2] Dipl. Ing. E. Kramer, Univ.Prof. Dr. J. Koppelmann,„Untersuchung zur Dimensionierung einer flexiblenFernwärmeleitung aus Kunststoff―, UniversityLeoben, Austria; 1984.[3] J.T. van Wijnkoop, E. van der Ven, ―Verification ofheat loss measurements‖, 12 th ISDHC 2010.[4] J. Korsman, I.M. Smits, E.J.H.M. van der Ven―Heat loss analysis and optimization of a flexiblepiping system‖, 12 th ISDHC 2010.[5] I.M. Smits, J. Korsman, J.T. van Wijnkoop andE.J.H.M. van der Ven, ―Comparison of competitive(semi)flexible piping systems by means of heatloss measurement‖, <strong>12th</strong> ISDHC 2010.[6] Univ. Prof. Dr. E. Hönninger, „Sekundärnetzefördern die Fernwärmeanwendung―,STEWEAG,Fernwärme <strong>International</strong> 14/85.[7] C. Engel, „Polybutene – The alternative material forheating and domestic hot & cold water systems―,PLASTIC PIPES IX, Edinburgh 1995.Acknowledgement go to the innovative engineers inEnergy Providers like STEWEAG, NUON and ENECO,who are drivers for co-creation of new solutions for thebenefit of the entire industry.266

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaABSTRACTCOMPETITIVENESS OF COMBINED HEAT AND POWER PLANT TECHNOLOGIESIN ESTONIAN CONDITIONSThe goal of this paper is to evaluate competitiveness ofmarket ready combined heat and power (hereaftercalled as ―CHP‖) technologies for CHP expansionpotential locations in Estonian. The main criteria toindicate preference of CHP technology is a heat priceby which the internal rate of return is equal to investors‘expectations.Calculation results shows, that in spite of theadvantages of gas engines (relatively low investmentcosts and high electrical efficiency) the calculated heatprices are the highest. Heat price for expected 7% IRRis 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 steamturbine/engine technology. Heat prices are lower for1–4 EURO/MWhfuel, depending on heat demands.Heat prices for places with annual heat demand under20 000 MWh are mainly above 45 EURO/MWhfuel(average heat prices for biomass boiler houses inEstonia is between 40–45 EURO/MWh). Developing ofCHP plants in such areas is feasible in the case ofgrant payments for investments. CHP plantdevelopment based on wood chips or peat could befeasible without grant payments in the places whereheat demand exceed 30 000–40 000 MWh annual.Carefully selected CHP technology and capacity canafford higher IRR when keeping competitive heatprices.The most feasible places for CHP expansion in Estoniaare Maardu, Viljandi, Rakvere, Valga, Haapsalu, Võru,Paide and Põlva.INTRODUCTIONThis paper draws on ongoing project ‗Analysis on thetechnical and economic consequences of renewableenergy based CHP systems in new areas with thelowered useful heat demand or after implementation ofenergy conservation measures in the areas with olderbuildings‘ within the project ‗Primary Energy Efficiency‘partly financed by NER, which contributes to the effortof enhancing the primary energy efficiency (PEE) andreducing CO2 emissions in the energy sector.Present-day world energy policy is based on two maindirections: energy efficiency and environmentalprotection. Efficient CHP production is one of theE. Latõšov 1 and A. Siirde 11Department of Thermal EngineeringTallinn University of Technology, Tallinn, ESTONIA267energy consumption effective methods, where CHPproduction from the renewable fuels is preferable. [1]During the last 2 years a few CHP plants working onwoodchips and peat were build in Estonia. A few ofbiomass CHP plants are under active development. Allof them are planed or constructed in major Estoniancities and are based on backpressure steam turbinetechnology. At the same time feed-in tariffs as well aspossibilities to get grants for expanding of CHP andusage of renewable fuels makes CHP expansion moreattractive for locations with a lower heat demands.Steam turbine technology is a classic for CHP plants.But in relatively small-scale boilers and district heatingsystems use of steam turbines is connected toeconomically less efficient operation (commonly higherspecific investment costs, O&M costs and lowerelectrical efficiency) where use of other alternative CHPtechnologies could be preferable.The goal of this paper is to evaluate competitiveness ofmarket ready CHP technologies for CHP expansionpotential locations in Estonian. The main criteria toindicate preference of CHP technology is a heat priceby which the internal rate of return (hereafter called as―IRR‖) is equal to investors‘ expectations.The paper is structured as follows. After an overview ofplaces where construction of CHP plants can bereasonable the paper provides principles for evaluationof CHP technologies competitiveness. Next sectionsprovide an overview of the CHP technologies whichcan be used in CHP plants and descriptions of mainfuel sources for energy production in Estonia. The lastsection provides heat price calculation examples basedon proposed principles for evaluation of CHPtechnologies competitiveness.LOCATIONS OF POTENTIAL BIOFUELED CHPPLANTS IN ESTONIAFig. 1 shows major Estonian cities and municipalitieswhich are distributed by the annual heat demands.Places where CHP plants are already constructed orunder construction, as well as in a state of activedevelopment are marked separately.Fig. 1 reflects well known principles, where theconsumers with higher annual heat consumptions aremore preferable.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 1 Distribution of Estonian cities and municipalities by the annual heat demandsPRINCIPLES FOR EVALUATION OF CHPTECHNOLOGIES COMPETITIVENESSThe revenues of a CHP company are generated fromthe heat and electricity sales. Theoretically they mustcover the operation and maintenance (hereafter calledas ―O&M‖) costs of the CHP plant completely andprovide an expected IRR. Main CHP plant related costsand incomes are shown in Fig. 2.Income from the heat sells depends on amount of soldheat (computable value) and heat price.Knowing investment costs (specified in section CHPtechnologies), and other above mentioned costs andincomes the power plant operation annual net cash flowscan be calculated and IRR defined.The principle for evaluation of CHP technologiescompetitiveness is based on finding such heat pricewhich will cause an expected (proposed) IRR, wherecalculation/estimation rules for the other cash flowscomponents are clearly defined.CHP TECHNOLOGIESThere are numerous CHP technologies that can betheoretically used for small scale CHP systems, but notall of them are economically and technically feasible.The list of main CHP technologies ordered by marketreadiness and common heat outputs are shown in Fig. 3.Fig. 2 CHP plant incomes and costsFuel costs, pollution charges and ash handling costs aremainly depend on used fuel properties and areestimated in section Fuel sources for energy production.CHP technology related fixed operation andmaintenance costs depend on selected CHP technologyand are defined in % from the investment costs annual.They are estimated in section CHP technologies.Electricity sells depends on amount of producedelectricity (computable value) and fuel prices. Fuel pricesare estimated by taken into account feed-in tariffsdescribed in Electricity Market Act [2].It is important to consider the market ready solutions firstof all, such as a steam turbine (hereafter called as ―ST‖),steam engine (hereafter called as ―SE‖), ORCtechnology (hereafter called as ―ORC‖) and gas engine(hereafter called as ―GE‖). Hereafter SE and ST areconsidered jointly, where capacities less than 1 MWelcorrespond to SE by default.For CHP plant economical calculations it is important toknow such CHP plant parameters as efficiencies, priceand O&M costs.Above mentioned parameters are obtained andsystemized on the basis of information regarding CHPplants collected from different information sources suchas [3, 4, 5, 6, 7, 8].268

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFig. 3 Main prime mover CHP technologiesValues for investments as well as nominal electricalcapacities used to calculate fuel prices depend on CHPnominal electrical capacities and CHP technology, asshown in Table 1.Table 1. Values for investments and nominal electricalcapacities for selected CHP technologiesCapacityTechnologySpecificinvestmentcostsElectricalnominalefficiencyMW el MEURO/MW el %0,1 ST/SE 10,3 100,1 GE 1,6 321 ST/SE 5,1 151 ORC 5,8 151 GE 1,0 405 ST/SE 3,2 225 ORC 4,5 165 GE 0,8 4110 ST/SE 2,9 2210 ORC 4,2 1610 GE 0,8 42In this paper investment means all costs before CHPplant commissioning.For the evaluation of CHP competitiveness the efficiencydrop working at partial load is taken into account. It isassumed, that minimal CHP heat load for alltechnologies is 25% from the nominal heat load. It isassumed, that electrical efficiency working at minimalheat load is 35% for steam engine/turbine, 80% for gasengine and 85% for ORC from the nominal electricalefficiency.It is assumed, that CHP technology related fixed O&Mcosts for SE/ST, ORC and GE are relatively 2.5%, 2%and 3.5% from the investment costs annual.FUEL SOURCES FOR ENERGY PRODUCTIONMain fuel sources for under 10 MWel CHP plants inEstonia are natural gas, peat and wood chips.Fuel pricesThe fuel prices taken as basis for heat price calculationsare as follows:Peat price – 11.7 EUR/MWhfuel. Proposed price isbased on average peat price levels obtained fromTootsi Turvas AS, the biggest peat milling andexporting enterprise in Estonia.Wood chips price – 12.8 EUR/MWhfuel. Proposedprice is based on latest data, published by EstonianInstitute of Economic research in their web basedprice information system [9].Natural gas – 35 EUR/MWhfuel. Proposed price isan average price for the latest data published byStatistics Department of Estoni [10].269

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAsh handling costsAsh handling costs calculations are based onassumptions, that:Peat ash content is 5%. Average calorific value is3.3 MWh/t [11];Wood chips ash content is 1%, calorific value2.4 MWh/t;Natural gas combustion does not emit any ash;Regarding to information obtained from differentlandfill owners, an average for year 2012 expectedash removal costs (ash transportation to landfill, andstoring) are 45 EUR/t.The combustion plant is equipped with dry ashremoving system.Taking into account above mentioned information theash handling costs per MWh of fuel energy content forthe peat and wood chips are ~0.19 and 0.72 EUROrespectively.Pollution chargesPollution charges and levels are calculated base on theEnvironmental Charges Act [12], Regulation No 99/2004[13] and No 94/2004 of Estonian Minister of Environment[14].The method described in [13] takes into account differentcombustion technologies, flue gas cleaningtechnologies, control devices as well as capacities todefine emission factors of pollutants.Table 2. Summarised results of the heat price calculations for different CHP expansion scenariosDistrict heating areaCHP plant capacityHeat price, EURO/MWhHeatdemandMaximumheat capacityTechnologyFuelHeatElectricalInvestmentIRR 7% IRR 12%MWh MW MW h MW e MEURO Withoutgrant270WithgrantWithoutgrantWithgrant5000 1,5 ST/SE Peat 0,83 0,13 1,31 55 --- 67 ---5000 1,5 ST/SE Woodchips 0,83 0,13 1,31 53 39 65 445000 1,5 Gas engine Natural gas 0,83 0,63 0,77 61 --- 66 ---10000 3 ST/SE Peat 1,65 0,27 2,54 54 --- 65 ---10000 3 ST/SE Woodchips 1,65 0,27 2,54 51 37 62 4310000 3 ORC Peat 1,65 0,34 2,93 53 --- 66 ---10000 3 ORC Woodchips 1,65 0,34 2,93 49 32 61 3810000 3 Gas engine Natural gas 1,65 1,36 1,29 56 --- 60 ---20000 6 ST/SE Peat 3,30 0,60 4,46 49 --- 59 ---20000 6 ST/SE Woodchips 3,30 0,60 4,46 46 34 56 3920000 6 ORC Peat 3,30 0,70 4,94 46 --- 57 ---20000 6 ORC Woodchips 3,30 0,70 4,94 42 29 53 3420000 6 Gas engine Natural gas 3,30 2,94 2,65 55 --- 59 ---40000 12 ST/SE Peat 6,60 1,44 7,08 41 --- 49 ---40000 12 ST/SE Woodchips 6,60 1,44 7,08 38 29 45 3240000 12 ORC Peat 6,60 1,44 8,10 40 --- 49 ---40000 12 ORC Woodchips 6,60 1,44 8,10 36 25 44 2940000 12 Gas engine Natural gas 6,60 6,40 5,34 53 --- 57 ---80000 24 ST/SE Peat 13,20 3,90 14,57 39 --- 47 ---80000 24 ST/SE Woodchips 13,20 3,90 14,57 35 --- 42 ---80000 24 ORC Peat 13,20 2,96 15,20 38 --- 46 ---80000 24 ORC Woodchips 13,20 2,96 15,20 34 --- 42 ---80000 24 Gas engine Natural gas 13,20 14,00 11,66 53 --- 57 ---To avoid complexity of the analysis to be issued fromdifferent combinations of capacities, combustiontechnologies, fuel gas cleaning and control equipmentit is assumed that:Thermal capacity of combustion plants is below50MW;Selected combustion technology provides lowestemission level than the others in [13] mentionedcombustion technologies;Combustion plant is equipped with the mosteffective control systems mentioned in [13];Combustion plant is equipped with the mosteffective flue gas treatment technology mentionedin [13].Calculated levels for pollution charges for year 2013are:~0.07 EUR/MWhfuel for wood chips;~0.95 EUR/MWhfuel for peat;~0.43 EUR/MWhfuel for natural gas.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaHEAT PRICE CALCULATIONSCalculations of heat prices are provided incorrespondence with principles described in sectionPrinciples for evaluation of CHP technologiescompetitiveness.Heat prices are evaluated for different scenarios.Scenarios include described heat demands, consideredfuels and technologies.Heat prices are calculated for 7% and 12% IRR.Heat price for CHP plant developing scenarios whichsatisfy the requirements described in regulation [X],which define conditions for grant payments to expandrenewable energy production and construction of CHPplants in Estonia, are calculated separately.For calculating heat prices in addition to informationfrom previous paper sections, some other figures haveto be specified: Cash flows are calculated for 20 years; CHP starts energy production in the beginning of2013; Expected rate of inflation is 1.5%; Heat loses in district heating network are 15%; Heat load profile is estimated based on heat loadmodel described in [15] taking as a basis the heatload duration curve shape of Tallinn.The results matrix of heat price calculations is shown inTable 2.CONCLUSIONThe technologies for smaller CHP applications aremore expensive (specific price) and less efficient thanthose for larger CHP plants.At present peat is considered as a good alternative forwood chips. Lower fuel price (11.7 EUR/MWh) smoothover higher than for wood chips ash handling costs andpollution charges. At the same time wood chips aremore preferable because of higher feed-in tariffs forproduced electricity.The advantages of gas engine CHP plants arerelatively low investment costs and high electricalefficiency. But because of high natural gas price(MWhfuel price is 2.5–3 times higher than for woodchips and peat) and relatively high fixed O&M costs thecalculated heat prices are the highest. Heat price forexpected 7% IRR is between 53 and61 EURO/MWhheat depending on heat demand.Under 5 MWel ORC is competitive to SE/STtechnology. Calculated heat prices are lower for1–4 EURO/MWhfuel, where higher fuel price differencecorresponds to places with lower heat demands.Heat prices for places with annual heat demand under20 000 MWh are mainly above 45 EURO/MWhfuelwhere an average heat prices for biomass boilerhouses are between 40–45 EURO/MWh [16].Developing of CHP plants in such heat demand areasis feasible in the case of receiving of grant paymentsfor investments.CHP plant development based on wood chips or peatcould be feasible without grant payments in the placeswhere heat demand exceed 3000-40000 MWh annual.Carefully selected CHP technology and capacity canafford higher IRR when keeping competitive heatprices.The most feasible places for CHP expansion in Estoniaare Maardu, Viljandi, Rakvere, Valga, Haapsalu, Võru,Paide and Põlva.Calculation results are valid for assumed cases only.Other particular cases should be calculatedindividually.REFERENCES[1] C. Dötsch and A. Jentsch, ―District heating (DH) inareas with low heat demand density (HDD):A chance for the integration of renewable energysources (RES)‖, 10th <strong>International</strong> <strong>Symposium</strong> onDistrict Heating and Cooling, 3–5, September2006, p. 2www: http://www.lsta.lt/files/events/20_doetsch.pdf[20.01.2010][2] Electricity Market Act www:https://www.riigiteataja.ee/ert/act.jsp?id=13279771[14.05.2010][3] Schwaiger, H., Jungmeier, G, (2007) Overview ofCHP plants in Europe and Life Cycle Assessment(LCA) of GHG emissions for Biomass and FossilFuel CHP Systems CIBE Conference„Cogénération biomasse dans l'industrie et sur lesréseaux de chaleur opportunités – retoursd'expérience-perspectives―[4] Obernberger, I., Thek, G, Techno-economicevoluation of selected decentralised CHPappications based on biomass combustion in IEApartner countries Graz (2010)[5] Bryson, T., Major, W., Darrow, Ken. Assessment ofOn-Site Power. Opportunities in the IndustrialSector, Carlsbad (2001) www:http://www.uschpa.org/files/public/Assessment%20of%20Onsite%20Power%2001.pdf[14.05.2009][6] Kirjavainen, M., Sipilä, K., Savola, T. Small-scalebiomass CHP technologies. Situation in Finland,Denmark and Sweden, VTT Processes (2004)271

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniawww:http://www.opetchp.net/download/wp2/small_scale_biomass_chp_technologies.pdf [14.05.2010][7] Institute for Thermal Turbomachinery and MachineDynamics, Cogeneration (CHP) TechnologyPortrait, Vienna (2002) www:http://www.energytech.at/pdf/techportrait_kwk_en.pdf[14.05.2010][8] U. S. Environmental Protection Agency CombinedHeat and Power Partnership, Biomass CombinedHeat and Power Catalog of Technologies, (2007)www:http://www.epa.gov/chp/documents/biomass_chp_catalog.pdf [14.05.2010][9] Estonian Institute of Economic research www:http://www.ki.ee [14.05.2010][10] Statistics Estonia www: www.stat.ee [14.05.2010][11] Paappanen, T., Leinonen,A. Fuel peat industry inEU, 2005, p. 134 www:http://turbaliit.ee/index.php?picfile=21 [14.05.2010][12] Environmental Charges Act, [14.05.2010] www:http://www.riigiteataja.ee/ert/act.jsp?id=13316043[14.05.2010][13] Procedure and Methods for Determining Emissionsof Pollutants from Combustion Plants into AmbientAir www:http://www.riigiteataja.ee/ert/act.jsp?id=789462[14.05.2010][14] Välisõhu eralduva süsinikdioksiidi heitkogusemääramismeetod www:http://www.riigiteataja.ee/ert/act.jsp?id=127572 15[14.05.2010][15] Latõšov, E., Siirde, A. (2010). Heat load model forsmall-scale CHP planning. In: Proceedings of<strong>International</strong> Conference on Renewable Energiesand Power Quality: <strong>International</strong> Conference onRenewable Energies and Power Quality(ICREPQ‘10), Granada (Spain), 23-25th March,2010., 2010.[16] Estonian Competition Authorities approved districtheat maximum prices (without VAT) to end-userswww:http://www.konkurentsiamet.ee/file.php?15416[14.05.2010]272

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDISTRIBUTION OF HEAT USE IN SWEDENMargaretha Borgström, Sven Werner1 School of Business and EngineeringHalmstad University, PO box 823, S-301 18 Halmstad SwedenABSTRACTThe current heat use refers normally to the averageheat use in a country or a sector during the course of ayear. But it is also important to be aware of thedistribution of high to low use when estimating thepotential for reducing total heat use.Energy statistical data published in the annual reportfrom Statistics Sweden have been supplemented by adeeper analysis of distribution of heat use andsystematic causes regarding high heat use.The aim of this paper is to explain the variation in heatuse with respect to construction year, degree days andenergy efficiency measures.In the Swedish energy efficiency debate, many voicesrefer 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 individualcauses. The most important implication of the studyresults is that systematic policy measures will have alow impact on the total national energy efficiency.INTRODUCTIONcorresponding to 310 million square metres in multifamilybuildings and service sector premises. Thesurvey sample thus constituted a sizable portion of theentire building stock.This energy statistical data, published in the annualreports from Statistics Sweden, have beensupplemented with a deeper analysis of the distributionof the heat use and the systematic causes regardinghigh heat use. Independent variables for explanation ofvariations were number of degree-days, constructionyear, ventilation system, energy efficiency measure,and co-use of heat supply. High and low users werealso analysed by location, construction year, heatsupply method, ownership, and building size [3]. In thisshort paper, the specific heat use will be presented byits distribution, construction year, degree days andenergy efficiency measures.1. Distribution of heat useThe total distribution of specific heat use as a functionof the percentage of the building area of all multi-familybuildings and service sector premises in Sweden isshown in Fig. 1.Multi-family residential buildings and service sectorpremises constitute 80% of the customer stock in theSwedish district heating systems. The level of futureheat use in these buildings will thus have a stronginfluence on the future district heating economy and thecorresponding investment demand. It is therefore ofinterest to collect information and make analyses of thecostumer heat use and how the heat use will develop inthe future.Specific heat use in multi-family buildings and servicesector premises has decreased considerably since the1970‘s. In 2006, the specific heat use in multi-familybuildings has decreased by 38% compared to the heatuse in 1972. The lower heat use is due to increasingenergy prices and more energy efficient buildings.An extensive study of the current heat use for buildingsin Sweden has been performed. The input informationfor this study was constituted by the anonymousresponses to the annual survey of energy use in multifamilybuildings and service sector premises performedfor 2006 by Statistics Sweden, [1] & [2]. The responsesprovided input data from 11253 buildings having a totalarea of 77.6 million square metres. By using scalingfactors, estimates could be made for the entire country,273Heat usekWh/m 240035030025020015010050Multi-family buildingsPremises00% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%Share of all national building spaceFig. 1 Heat use distribution during 2006 as a function ofthe share of all national building space. The diagram is anestimation for all multi-family and service buildings inSweden.The area under each curve is the total heat used inmulti-family buildings and service sector premisesduring 2006. The figure shows that 13% of the area inmulti-family buildings had a specific heat use of morethan 200 kWh/m 2 , and 12% of the area in servicesector premises had a specific heat use of more than200 kWh/m 2 . This result shows that there are no majordifferences between the percentages of the building

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaarea with high heat use in multi-family buildings andservice sector premises.The results in Fig. 1 also show that 11% of the buildingareas in multi-family buildings, and 31% of the buildingarea in service sector premises have a specific heatuse lower than 100 kWh/m2.Buildings with heat use 200 kWh/m2 or more havebeen further analysed and the results are presented inthe following section considering construction year.2. Construction yearFig. 2 shows specific heat use in multi-family buildingsas a function of construction year. The figure alsoincludes the average value each year, together givingthe total average specific heat use of 152 kWh/m2.There are no major differences in heat use in buildingsconstructed before 1980. After 1980, the heat use wasapproximated 15% lower than the average heat use forall buildings in Sweden.The relationship between construction year and highheat use in buildings has been analysed. The definitionof high heat use is 200 kWh/m2 or more. Fig. 4 showsthe results for multi-family buildings. There were a totalof 179.3 millions square metres in multi-family buildingsin 2006 and 13% of the heated area had heat use of atleast 200 kWh/m2.Of special interest are buildings built during the period1965–74, when a large part of the existing buildings inSweden were built. During this period there were norequirements for low energy use in buildings.In multi-family buildings built during the period1965–74, 30% of the total area had heat use of at least200 kWh/m2 and for buildings built in the period1941–60. 42% of the total building area had heat useof 200 kWh/m2 or more.Heat usekWh/m 24003503002502001501005001930 1940 1950 1960 1970 1980 1990 2000 2010Construction yearFig. 2 Specific heat use as a function of construction yearfor 4285 multi-family buildingsFig. 4 Total square metres where heat use is higher orequal to 200 kWh/m2 in multi-family buildings categorisedby construction year.Heat usekWh/m 24003503002502001501005001930 1940 1950 1960 1970 1980 1990 2000 2010Construction yearFig. 3 Specific heat use as a function of construction yearfor 4061 service sector buildings.The heat use in service sector premises is shown inFig. 3. Also in these buildings, the average heat useafter 1980 is lower (about 10%) than the average heatuse in all service sector premises in Sweden.Fig. 5 Total square metres where heat use is higher orequal to 200 kWh/m2 in service sector premisescategorised by construction year.The relationship between high heat use in servicesector premises and construction year is shown inFig. 5. During 2006, 15.6 million square metres hadheat use of at least 200 kWh/m 2 . Service sectorpremises built between 1965 and 1974 had high heatuse in 3.5 million square metres.274

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe results show that the period 1965–1974 did nothave a dramatically higher heat use in the constructionyear analysis.3. Degree daysThe climate in Sweden varies with a much colderclimate in the northern part compared to the southernpart. Since the statistical data consist of buildings fromdifferent parts of Sweden, the influence of the localclimate on the heat use in buildings can be analysed.This has been done by analysing the correlationbetween the number of degree days for the location ofa building and the corresponding specific heat use.The number of degree days, according to the Swedishdefinition, varies from approximately 3000 in the southup to 7000 in the north of Sweden. Each building in theanalysis was connected to one of 14 climate areas.Heat usekWh/m 2400350300in the theoretical analysis of the optimal wall insulationas a function of degree-days.The results show that the average difference betweenNorthern and Southern Sweden was small, implying asmall climatic impact on heat use. The main conclusionfrom this analysis is that the individual variation in eachclimate area is much higher than the local impact ofclimate. This astonishing conclusion can have severaldifferent explanations:Higher awareness and consequences of lowbuilding heat resistances in Northern SwedenLower regional GDP in Northern Sweden givinghigher incentive to reduce heat costsMore frequent snow cover in Northern Swedengiving extra heat resistance during the winter.4. Energy efficiency measures.The statistical data shows the energy efficiencymeasures during the period 1995–2005. The energyefficiencymeasures were:a. Supplementary insulation250200150100y = 15,63x 0,28b. More energy efficient windowsc. Balancing heating- and ventilation systemsd. Electrical efficiency measures5002000 2500 3000 3500 4000 4500 5000Degree-daysFig. 6 Specific heat use for 5111 multi-family buildings asa function of the number degree days in each climatearea.Heat usekWh/m 240035030025020015010050y = 10,37x 0,3002000 2500 3000 3500 4000 4500 5000Number of degree-daysFig. 7 Specific heat use for 6041 service buildings as afunction of degree days in each climate area.Fig. 6 and Fig. 7 show the specific heat use as afunction of degree days for multi-family buildings andservice sector premises. The figures also show theaverage curve and its equation for specific heat use asa function of degree-days. You should also note thatthe exponent in the fitted equations has only themagnitude of 0.3 instead of the 0.5 exponent obtainede. Heat recovery in ventilation systemsIn multi-family buildings, one or several energyefficiency measures were implemented for anestimated floor area of 57.6 million square metresduring the period 1995–2005. No energy efficiencymeasures had been performed for an estimated floorarea of 92.2 million square metres during the sameperiod.In service sector premises, with an estimated floor areaof 37.2 million square metres, one or several measureshad been taken during the period 1995-2005. Duringthe same period, no measures had been taken for anestimated floor area of 70.3 million square metres.The most common measures in multi-family buildingsand service sector premises were balancing of heatingandventilation systems.In many buildings, a combination of two or severalenergy efficiency measures had been taken in thesame building. In some buildings, up to five measureshave been taken in the same building.The average heat use in multi-family buildings andservice sector premises in relation to measures takenis shown by bars in Fig. 8 and Fig. 9. The horizontallines show the average heat use in buildings, in whichno energy efficiency measure was performed.275

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaCONCLUSIONThe main conclusions from the analysis were:Fig. 8 Average heat use in multi-family buildings inrelation to the measures performed. The measure figurescorrespond to the measures defined in the text.Individual variations dominate compared tosystematic causes regarding the specific heatuse in multi-family and service sector buildings.The district heating companies can help theircustomers by identifying them as high, mediumor low users of heat.On the short term, a significant potential existsfor lower heat use in the Swedish multi-familyand service sector buildings.More efficient heat use in buildings will probablyrepresent the most important competitor todistrict heating supply in the future.Fig. 9 Average heat use in service sector premises inrelation to the measures performed. The measure figurescorrespond to the measures defined in the text.As shown in the figures 8 and 9, there were nosubstantial differences in heat use between buildingswhere energy-saving measures had been taken andthose where they had not. The conclusion from thisanalysis is that the measures taken during these 10years were taken by late-comers rather than by earlyadopters, since heat use after measures were takengenerally corresponds to the average level for allbuildings.In the Swedish energy efficiency debate, manyvoices refer to systematic causes for high heatuse. However, the results from this study do notsupport this opinion, since the distribution ofheat use mostly comes from individual causes.The most important implication of the studyresults is then that systematic policy measureswill have a low impact on total national energyefficiency.REFERENCES[1] Statistics Sweden, Energistatistik förflerbostadshus 2006 (Energy statistics for multifamilyhouses during 2006). StatistiskaMeddelanden EN16SM0702.[2] Statistics Sweden, Energistatistik för lokaler 2006(Energy statistics for premises during 2006).Statistiska Meddelanden EN16SM0703.[3] Andreasson M, Borgström M, WernerS, Värmeanvändning i flerbostadshus och lokaler(Heat use in multi-family buildings and premises2006) Fjärrsyn report 2009:4, Stockholm 2009.Available at www.svenskfjarrvarme.se276

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDAMAGES OF THE TALLINN DISTRICT HEATING NETWORKS AND INDICATIVEPARAMETERS FOR AN ESTIMATION OF THE NETWORKS GENERAL CONDITIONAleksandr Hlebnikov 1 , Anna Volkova 1 , Olga Džuba 2 , Arvi Poobus 1 , Ülo Kask 11 Department of Thermal Engineering, Faculty of Mechanical Engineering,Tallinn University of Technology, Kopli 116, 11712 Tallinn, Estonia2 Tallinna Küte, Punane 36, 13619 Tallinn, Estoniaahleb@staff.ttu.ee, anna.volkova@ttu.eeABSTRACTDistrict heating networks in Estonia are mostly old andin bad condition. The state of the district heatingnetworks of Tallinn is typical for the rest of Estonian DHnetworks. The paper includes analysis of the Tallinndistrict heating networks. Valid data about damages indistrict heating systems received for the last 12 yearswere used for an analysis of the networks damages.Different types of network damages are analysed:external corrosion, internal corrosion, defect ofinstallation, factory defects, defect of construction andother reasons. The number of damages for the differentelements of networks is compared in the paper:armature, compensator, construction and pipes. Mainfactors, which influence damages in district heatingnetworks, are the age of networks, the quality ofconstruction works and the network operationconditions.The damage quantity dependence on the age ofnetworks is also defined and analysed in the paper.The number of damages can be diminished byreducing the average age of networks. This is possibleby replacing old pipelines and other network systemelements. Pipes average age changes for 20 yearsperiod are simulated according different intensities ofrenovation works.INTRODUCTIONDistrict heating (DH) allows centralized heat productionfor an area and hot water transportation to the buildingsthrough a network of pipes. District heating systemsoffer the potential to use energy-efficient andrenewable heat generation technologies, such ascogeneration technologies which implement both fossilfuels, as long as biomass and waste [1]. Districtheating system is traditional in Estonia. It has formedapproximately 70 per cent of all heating in the country.The share of heat produced by combined heat andpower production stations is approximately one third. Atthe same time, the technical situation of the districtheating networks (and production equipment) is poor.[2] Unsatisfactory condition of DH networks andunreliable heat supply can doubt on future of districtheating and the consumers can make a choice towardsa different heat supply alternative. Often thedecentralized heating is not an effective solution forregional heat supply strategy and it decreases potentialof combined heat and power production.[3].Nowadays DH systems operate both in big cities and insmall towns, which means, that there is enough heatload for the installation of new cogeneration equipment.But before new energy sources installation it isimportant to define and analyse the situation with DHnetworks.The purpose of this paper is to define the validcondition of typical old networks in Estonia, to definethe reasons of damage occurrence on the basis ofoperational data and to make forecasts for operation ofa DH network for the next 20 years. The paper includesanalysis of Tallinn district heating networks. The validdata about damages in district heating systemscollected during past 12 years was used for analysis ofnetworks damages.THE PRESENT CONDITION OF TALLINN DISTRICTHEATING SYSTEMDistrict heating networks in Estonia are mostly old andin bad condition. The state of the district heatingnetworks of Tallinn is typical for the rest of Estonian DHnetworks. In Tallinn the heat is transmitted to theconsumers through a 406-kilometres long heatingnetwork including the 93 km of pre-insulated pipes(23%). District heating systems of Tallinn wereconstructed mostly during the 1960-1980 period andtheir average age is 22 years.The AS Tallinna Küte enterprise makes operation ofthe bigger part of district heating networks and boilerhousesof Tallinn.District heating systems of Tallinn consist of fivedistricts of the central heat supply: Kesklinna district(total length ~92 km, length on the balance ofAS 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 <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaMaardu district (total length ~25 km, length on thebalance of AS Tallinna Küte ~14 km). [4]District heating systems of the areas Kesklinna andLasnamäe are connected through the pump stationLaagna. The total length of heating networks is 406 kmfrom which on the balance of AS Tallinna Küte thereare 348 km, or 85,7%.The following CHP stations and boiler-houses supplyheat to the districts of Tallinn: the CHP Iru (natural gas,190 MWel, 748 MWth), the boiler house Ülemiste(natural gas, 232 MWth); the CHP Väo (wood chips, 25MWel, 65 MWth); the boiler house Mustamäe (naturalgas, 390 МWth); the boiler house Kadaka (natural gas,290 MWth).Besides the abovementioned there are some smallscaleboiler houses. In Fig. 1 is displayed the basicscheme of Tallinn heat supply.District heating systems of Tallinn were constructedmostly during the 1960–1980 period and their averageage is 22 years.Mustamäeboiler-house390 MW390 MWKadakaboiler-house290 MWÜlemisteboiler-house232 MW(in reserve)Iru CHP748 MW (190 MW)Laagnapump station200 MWMustamäe network325 MWKesklinnanetwork180 MWLasnamäenetwork268 MWMaardunetworkVäo CHP65 MWThe state of DH networks varies for the differentdistricts of Tallinn.In Lasnamäe the construction of district heatingsystems began in 1970, and the network length is~106 km at present time. Assuming the actual load theheating systems of Lasnamäe district are the mostoverloaded in town.The length of main pipelines DN1000–1200 is ~19 km,the length of pipes DN400-800 is ~4,4 km. The share ofthe main networks is quite big and it is ~22% of totalnetwork length in Tallinn. Thermal isolation is made ofglass wool according to old soviet building norms and itis the reason of big heat losses in the network. Theheat losses in Lasnamäe network in 2008 were 21%from the total produced heat.The interconnected district heating systems of boilerhousesMustamäe, Kadaka and Karjamaa (not inoperation at present time) are related to the Lääne areaFig. 1 The basic scheme of Tallinn district heating system278(districts Mustamäe and Õismäe). Initially there hadbeen two separate networks which were merged lateron as a result of growth. In the area Lääne theconstruction of district heating systems began in 1960.The length of the Lääne area network is ~141 km. Thediameters of the main pipelines are less than those inthe Lasnamäe area.The length of the main pipelines with diameterDN400–900 is ~27,8 km. The heat losses of thenetwork in 2008 were 16% from the total producedheat.The speciality about the heating system of the areaLääne is that in past there was an open system of hotwater supply. The water added to the system had notime to purify sufficiently and oxygen and waterhardness led to an intensive internal corrosion of pipes.In Kesklinn area the network construction began in1959. Initially the heat supply was carried out by the

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTallinna Soojuselektrijaam heat and power station andlater on by the boiler-house Ülemiste. The districtheating system of Kesklinn area is the oldest in Tallinn.The average age of the Kesklinn area network is25 years, the total length is ~76 km.The length of the main pipelines with diameter DN400-900 is ~13,8 km. The share of main pipelines inKesklinn area network is ~18,1%. Relative heat lossesof Kesklinn network are within the limits of 15...18%. Incomparison with other areas the relative heat lossesare less. The reasons for this are: the bigger networkloading, the not oversized pipes and the significantshare of preinsulated pipes.[5]THE ASSESSMENT OF DAMAGESThe analysis of networks damage statistics for Tallinnis made on the basis of valid data collected during thepast 20 years.The distribution of damages of Tallinn district heatingnetwork is shown in Fig. 2 according the periods ofconstruction. It is obvious that the most critical situationis with the sites constructed during the 1980–1985period. It can be explained by the poor quality of bothconstruction works and materials used in construction.During that period the networks were being constructedin a hurry and with lack of proper supervision.damages during1998-200940035030025020015030 years and by today they are already worn out. Theprobability of failures sharply increases. By today the84% of all compensators should be replaced. Someparts of the old locking armature also have to bereplaced. The service life of armature has exceeded25 years. Armature and compensators are partlyrenovated; however some pieces of it are old and alsorequire replacement. [5]3002502001501005001997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008Fig. 3 Places of damage in Tallinn district heatingnetworks elementsarmaturecompensatorconstructionpipesIn Fig. 4 the nature of damages is summarized. Thereare no data about the character of damage for all theareas of Tallinn within past 10 years, that‘s why thedamage allocation by character of damage is shown fora five year period.In Tallinn network the significant part of damages iscaused by external corrosion of pipes. Main reasons ofexternal corrosion are the bad waterproofing ofunderground channels and chambers and thecollapsed drainage. Amongst other reasons are thedefects of pipe supports and the destruction ofconcrete channels.100500… -19651965 -19701970 -19751975 -19801980 -19851985 -1990years of construction1990 -19951995 -20002000 -2005Fig. 2 Damages of Tallinn district heating networksaccording the periods of construction2005 -2008In Fig. 3 the places of damage in the network elementsare shown: armature, compensators of thermallengthening, construction and pipes. The major part ofall damages was the pipes.During the 1997–2003 period there were manyproblems with armature and compensators; after 2003the quantity of damages to these elements hadconsiderably decreased. The oldest thermallengthening compensators work since 1959. Theresource of axial compensators is no more than27990807060504030201002004 2005 2006 2007 2008Fig. 4 Nature of damages in Tallinn district heatingnetworksexternal corrosioninternal corrosiondeffect of constructiondeffect of installationwrong serviceother reasonsThe second main cause of damages is the internalcorrosion. In 2004 many pipes damaged by internal

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniacorrosion were revealed. Internal corrosion is the mostserious problem in Lääne network where an opensystem of hot water supply earlier has been used.Besides the damages caused by defects of installation,defects of construction, factory defects and impropermaintenance, the other reasons have also beenregistered.The main factors, which have an affect on the damagesin district heating networks, are the age of networks,the quality of construction works and the networkoperation conditions. The two latter can be regulatedby control authorities and proper legislation, however,the influence of these factors has been reduced incomparison with the 1970–1990 period. Then quality ofconstruction works was very low, drainage systemswere installed incorrectly or were not installed at all andisolation materials were not qualitative. As regardsdistrict heating operation conditions, theaforementioned open vented hot water supply systemused in some networks has led to intensive internalcorrosion of pipes.One important reason for damages reduction is that inrecent years the networks have significantly reducedpressure. The network works in a stable temperaturemode, the reliability of heat sources is improved andthe quantity of equipment emergency stops forced bysharp fluctuations of the heat-carrier temperature hasdecreased.Other operation condition factor which influencednumber of district heating system damages was higherwater temperatures in networks (up to 130 t °C) thannowadays (up to 110 t °C). Finally we can concludethat such factors as quality of construction works andquality of network operation are close to their optimumat present time in comparison with previous years.Damage quantity also depends on the age of networks.The number of damages can be reduced by reducingthe average age of the networks. This is possible byreplacing the old pipelines and other networks systemselements.Reconstruction and replacement works are made inTallinn, but the intensity of replacement is rather lowand not enough for a stable system operation. It isimportant to define, how intensive the networkreconstruction should be.Data for the three past years were used for defining thedamage dependence (number of damages/km/year) onthe age of networks. Data about damages werecollected 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).Using least squares analysis, a regression equation forthis dependence was defined.2D 0,0096A 1.8985A 1.0496(1),whereD – Number of damages/100 km per yearA – Age of networksBefore using this regression for further calculations, weshould check if this equation is appropriate. One of themain parameters for estimation of regression equationis the correlation coefficient. It is considered, that thecorrelation is good in case when R>0.8. In the case ofdamage dependence on pipes age, R is 0.802.R2=0.643, which means that the equationcharacterizes the 64,3% of damage number changes,but the 35,7% of changes are characterized by anotherfactors. There is still an influence of other factors,which can not be changed, such as construction andinstallation problems in the past.Data about damages allocation by the group andapproximation of these data is shown in Fig. 5.The regression equation can be used for the damageforecasts in future.damages/100 kmper year7060504030201000 5 10 15 20 25 30 35Age of networksFig. 5 Damage number dependence on the age of pipes indistrict heating systems for the 2005–2007 periodAs it has been mentioned before, the age of networksdepends on the intensity of renovation works.In Fig. 6 the length of all repaired sites is shown split byyears.Since 1980 the serial repair of Tallinn district heatingsystem is being carried out.Basically the investments have been directed towardsthe increase of reliability and the reduction of quantityand duration of faults in heat supply. It has beeninvested a lot in the locking armature.For the past 10 years ~35 km of district heatingpipelines have been replaced, which is 10% of totallength of the district heating systems in Tallinn area.280

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe annual replacement of pipes is in average about3,06 km per year, which is less than 1 percent from thelength of Tallinn DH system pipelines.Length, km87654321019851986198719881989199019911992199319941995Fig. 6 Length of replaced pipelines by years in Tallinndistrict heating networkTHE FORECASTS FOR DISTRICT HEATINGSYSTEM AGE1996One of the tasks was to assess, how big the renovationworks should be in order to stop increasing the averageage of pipes. A simulation model, which uses both realdata and also some assumptions, was created for suchestimation.199719981999200020012002200320042005200620072008The average age of pipes for each year was calculatedaccording equation (2)Aavji=b, ifwherejiiabl ( j i)l liiajbiial ( j i) ( j c) ( l ; i=c, ifi jcljl liaAav is average age of pipes in j yearli is length of pipes, constructed in i yearI – year of construction;J – current year;jbl )iiaa – year of construction of the oldest pipes, operating inthe current year.As a result of simulations, seven forecasts for pipesaverage age were calculated according differentintensity of renovation works: for current intensity ofrenovation (3,06 km/year) and for intensities when 1%,1,5%, 2%, 2,5%, 3% and 4% of total DH system lengthwould be annually renovated. The forecasts weresimulated for the 20 year long period.The results of simulation are shown in Fig. 8.(2)Assuming that the length of pipes (360,67 km) will notchange during the forecast period and that the annualscope of renovation works will remain the same duringwhole of the period means that the length of renovatedpipes also will not change. Besides it‘s was assumedthat every year just the oldest pipes would berenovated; however in reality the renovation works arebased on the pipes actual state estimation.Allocation of pipes ages for starting point (2008) isshown on Fig. 7 [5].age, years4035302520151052008200920102011201220132014201520162017201820192020202120222023202420252026202720282029203020312032203320342035203620372038203920401%2%3%4%1,50%2,50%currentlength, km2520151050494643403734312825221916131074age, yearsFig. 7 Length of DH networks by pipes age (in 2008)1Fig. 8 Pipe age forecasts for different intensity of networkrenovation worksAs it can be seen from Fig. 8 in case the renovationstays on the same level, the average age of pipes willgrow till reaching 39 years in 2040. In case the lengthof annually changed pipes is 1% or 1,5% higher, theaverage age will still rise, but in a less steep way.When the 2% of DH system length is annuallyrenovated there will be the minimal changes in ageduring first 5 years, after that the age will start risingand only after 15 years it will begin to decrease.If renovation intensity is 2,5% of the length or higher,the average age will not rise at all or will decrease. Forreducing the damages occurrence probabilityinfluenced by the networks age, the amount of repaired281

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniasites should be at least 9 km/year. This way theprocess of ageing will slow down and also the averageage will stabilize on a certain mark. One of the possiblesolutions is to replace the pipes with higher intensity of3–4% until reaching the 17–20 years average age andthen reduce the length of renovated pipes per year tothe 2–2,5% of the whole length of DH network.CONCLUSIONSDistrict heating networks in Estonia are mostly old andin bad condition. The state of the district heatingnetworks of Tallinn is typical for the rest of Estonian DHnetworks. That‘s why the result of damage analysismade for the DH network of Tallinn can be used for theother networks in Estonia.The AS Tallinna Küte enterprise makes operation of85% from the length of district heating networks inTallinn. Tallinna Küte data about the damages wereused for assessment.Places of damages in the DH system are following:armature, compensator, pipes and construction. Mostof the damages happened in the pipes.As regards the character of damages, the typicaldamages are caused by external corrosion, internalcorrosion, defect of construction, defect of installationand wrong service. The major part of damages iscaused by external corrosion of pipes.The age of networks, the quality of construction worksand the network operation conditions are the mostimportant factors, which influence the damages indistrict heating networks. The number of damages canbe reduced by reducing the average age of thenetworks. This is possible by replacing the old pipelinesand other networks systems elements. The intensity ofreplacement works during last 25 years was less thanone percent from the whole length of pipes.Seven forecasts for pipes average age accordingdifferent intensity of renovation works were simulated:for current intensity of renovation (3,06 km/year) andfor 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 networksaverage age at least at former level, the rate of oldpipelines replacement should exceed the 2,5% of thewhole length of DH system.AKNOWLEDGMENTThis work has been partly supported by the EuropeanSocial Fund within the researcher mobility programmeMOBILITAS (2008–2015), 01140B/2009REFERENCES[1] Cogeneration and district energy sustainableenergy technologies for today…and tomorrow,<strong>International</strong> Energy Agency, 2009.[2] Long-term Public Fuel and Energy SectorDevelopment Plan until 2015, Riigi. Teataja, RT I,23.12.2004, 88, 601[3] Hlebnikov, A.; Siirde, A. The major characteristicparameters of the estonian district heatingnetworks, their problems and development. // The11th <strong>International</strong> <strong>Symposium</strong> on District Heatingand Cooling: University of Iceland, 2008, 141–148.[4] Tallinna küte webpage, www.soojus.ee[5] A. Hlebnikov "The analysis of efficiency andoptimization of district heating networks inEstonia", Doctoral Thesis, Tallinn University ofTechnologies, 2010.282

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaEFFICIENCY OF DISTRICT HEATING WATER PUMPING IN FINLANDAntti Hakulinen 1 , Jarkko Lampinen 1 and Janne Lavanti 11 Pöyry Finland OyABSTRACTThe objective of this study was to determine thesavings potential in district heating pumping in Finland.A measurement method was also developed to quicklyestimate the efficiency of district heating pumping.The work was based on the data gathered from districtheating statistics. The work is divided into two parts.The district heating statistics reveal a number of districtheating networks whose consumption of energyneeded for pumping is exceptionally high. Thesecompanies should clarify the reasons for that.In addition, companies with an exceptionally lowconsumption of pumping energy should check theirmeasurements and data gathering routines.On average the electricity needed for district heatingpumping should not be over 0.5 per cent of the totalenergy supply (=sold+losses). If the density(supply/length of the network) of the district heatingnetwork is less than 3 GWh/km, the energy needed forpumping may rise. In any case the proportionalpumping energy should be lower than 1 per cent oftotal energy supply.90 000 €80 000 €70 000 €60 000 €50 000 €40 000 €30 000 €20 000 €10 000 €The investment costs of a pump0 €0 100 200 300 400 500pow er kWFig. 1. The investment costs of a pump.1.2 Total costs of pumpingTotalPumpFrequency converterTotal costs of pumping include capital, maintenanceand energy used in pumping. The pump lifetime costsare mainly energy costs as we can see from figure 2 onthe next page. The lifetime costs are calculated withthe following assumptions: energy price € 60/MW/h,operating lifetime 15 years, utilization period ofmaximum load 5000 h/a, interest rate 5 per cent andthe O&M 1.2 per cent of the investment.Pump I, pow er: 16 kWMotorThe Finnish potential for saving in district heatingpumping is estimated to be 20 per cent of the currentpumping energy i.e. 30 GWh/a. This is equivalent to ayearly saving of approximately € 2 million.PART 1.INTRODUCTIONIn the Finnish district heating systems no typicalpumping arrangements have been used at heatproduction plants or at booster pump stations. Theways of dimensioning and connecting pumps havevaried a lot. This has led to incorrect dimensioning andconnections of pumps, which in turn has caused higherinvestment costs and greater pumping energy usagethan expected, operational problems and in the worstcase many interruptions in the use of the network. TheFinnish district heating system is based on the variableflow operation (consumer driven scheme)1. COSTS14 %2 %84 %Pump II, pow er: 131 kW7 % 1 %92 %Pump III, pow er: 283 kW5 %1 %CapitalO&MEnergyCapitalO&MEnergyCapitalO&MEnergy1.1 The investment costs of a pump (includingmotor and controls)The calculated investment costs of a pump includingmotor, control, and pump are shown Fig. 1.28394 %Fig. 2. An example of lifetime costs of three different sizedpumps.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAs we can see the pump‘s efficiency plays a huge rolebecause the lifetime costs mainly consist of theoperational energy (84–94 per cent). For that reason alot of attention should be paid to the efficiency whenmaking the investment. Pumps with a low efficiencymay eat into the savings of the investment many fold.1.3 Booster pump station and costsA booster pump station should be considered when theprimary pumps of an energy station do not haveenough capacity to ensure the pressure difference atthe last customer. Typical reasons for the building of abooster pump station can be: long transmission lines,expansion of network, optimization of pumping energyand controlling of pressure level.The investments of a booster pump station includingpump, motor, frequency converter, building, automationsystems, etc. are shown in the following figures 3and 4.500 000 €480 000 €460 000 €440 000 €420 000 €400 000 €380 000 €360 000 €340 000 €320 000 €The investment costs of a booster pump station300 000 €0 50 100 150 200 250 300 350 400 450 500pow er [kW]cent or 2*70 per cent in parallel connection withindividual rotating speed controls. In that way thepumping of maximum heat load can be managed andthere is a room for possible expansion of the districtheating network. The other pump will act as a ―summerpump‖ so that the efficiency of pumping remains highalso when the heat load is low.By dividing the pumping capacity between manypumps it is possible to save pumping energy even ifpumping is handled from one point or from the heatproduction plant and the booster pump station. Thepossibilities to divide the pumping must be examinedcase by case by taking into account every single thingthat might have an effect on the costs.3. MAXIMUM WATER FLOWThe actual cooling of the district heating system inoperational conditions of pumping should be taken intoaccount when determining the calculated maximumwater flow. It is worthwhile to specify the water flowaccording to slightly worse cooling than the actualconditions require so that there is some design marginfor unusual conditions.4. OPERATION POINTTo change the rotating speed of a pump with afrequency converter is a good way regarding energyefficiency because the pump‘s efficiency often remainson high level within the whole adjusting area but theneed for power reduces strongly when the rotatingspeed goes down.Fig. 3. The investment costs of a booster pump station(only 1 pump).The investment cost of a booster pump station (flow + return)760 000 €710 000 €660 000 €610 000 €560 000 €510 000 €460 000 €410 000 €360 000 €0 50 100 150 200 250 300 350 400 450 500pow er [kW]Fig. 5. An example functional diagram of a pump.Fig. 4. The investment costs of a booster pump station(with 2 pumps).2. PUMPING ARRANGEMENTSAt the primary station it is usually sensible to divide thepumping between a few pumps, for example 2*60 per284An example functional diagram is shown in Fig. 5.When the rotating speed changes, the efficiencyremains good regardless of the changing rotatingspeed. The pumping of a district heating networkfollows this theoretical situation very well. Howeverwhen choosing a district heating pump it is important to

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniapay attention to its rotating speed which should be atthe minimum from 50 to 60 per cent of the nominalrotating speed.PART 2.INTRODUCTIONThe goal of the second part was to motivate the districtheating companies to analyse their pumping methodsand, hopefully, to lower their pumping costs.Total savings potential in district heating pumping inFinland was also estimated.This part is based on the Finnish district heatingstatistics of the year 2007 [1]. The statistics cover thedata of nearly 200 district heating companies but onlyabout 60 of them have reported the electric power usedin district heating pumping.The biggest companies have reported the pumpingenergy, thus, the pumping figure is available tocompanies which supply almost 70 per cent of alldistrict heat in Finland.STATISTICAL FINDINGSThe used pumping energy in different companies wasanalyzed by comparing the pumping energy to thefollowing parameters:Heat supply (sold heat + losses)Length of the district heating networkHeat density (supplied heat energy divided by thelength of the DH network)The following parameters were also examined but noclear correlation was to be seen, and the results aretherefore not reported in this paper:Heating output density (daily maximum heatingoutput divided by the length of the DH network)CHP productionShare of small (< 30 kW) consumersPeak load utilization timeLosses of DH network1. Heat supplyHeat supply is the same as sold heat + losses.The assumption used was that the bigger the companythe smaller the proportional pumping energy.The situation is presented in figures 6a and 6b, whereonly the companies which supply less than2500 GWh/a (Helsinki not included) are shown.Specific pumping energy (electrical power /heat supply)Fig. 6a. Example Electricity used for pumping in relation tothe size of a district heating company, heat supply0–2 500 GWh/a.Specific pumping energy (electrical power /heat supply)Fig. 6b. Electricity used for pumping in relation to the sizeof a district heating company, heat supply 0–300 GWh/a.The average pumping energy is 0.6 per cent of theheat supply. The bigger the company, the smaller theproportional pumping energy.Companies with exceptionally high pumping energy aremarked with a circle.2. Length of the district heating networkSpecific pumping energy (electrical power/ heat supply)1.6 %1.4 %1.2 %1.0 %0.8 %0.6 %0.4 %0.2 %1.6 %1.4 %1.2 %1.0 %0.8 %0.6 %0.4 %0.2 %1.7 %1.5 %1.3 %1.1 %0.9 %0.7 %0.5 %0.3 %Specific pumping energy vs. heat supplySupply 0 - 2 500 GWh/a0.0 %0 500 1000 1500 2000 2500Heat supply, GWh/aSpecific pumping energy vs. heat supplySupply 0 - 300 GWh/a0.0 %0 50 100 150 200 250 300Heat supply, GWh/aSpecific pumping energy vs. length of the DH netLength 0 - 500 km0.1 %0 100 200 300 400 500Length of the DH net, kmFig. 7a. Electricity used for pumping in relation to thelength of the district heating network, 0–1 300 km285

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSpecific pumping energy (electrical power/ heat supply)Fig. 7b. Electricity used for pumping in relation to thelength of the district heating network, 0–500 km.Specific pumping energy (electrical power /heat supply)1.7 %1.5 %1.3 %1.1 %0.9 %0.7 %0.5 %0.3 %1.6 %1.4 %1.2 %1.0 %0.8 %0.6 %0.4 %0.2 %Fig. 7c. Electricity used for pumping in relation to thelength of the district heating network, 0–70 km.For big companies the proportional pumping energy isalmost constant 0.5 per cent of heat supply.The longer the DH network, the smaller the proportionalpumping energy. The result is partly the same as in theprevious chapter: the bigger companies have smallerproportional pumping energies.If a company seems to have a high proportionalpumping energy in figures 7a–7c it may be due to poorheating density (lots of pipes in areas with not so muchconsumers).3. Heat densitySpecific pumping energy vs. length of the DH netLength 0 - 70 km0.1 %0 10 20 30 40 50 60 70Length of the DH net, kmSpecific pumping energy vs. length of the DH net0.0 %0 200 400 600 800 1000 1200 1400Length of the DH net, kmHeat density is the heat supply divided by the length ofthe district heating net.Fig. 8. Electricity used for pumping in relation to the heatdensity.It is natural that in a DH network with not too manypipes in proportion to sold heat the need for pumping ofDH water is lower.FURTHER INFORMATION:Pöyry Finland OyPL 93 (Tekniikantie 4 A)FI-02151 EspooFinlandantti.hakulinen@poyry.comCONCLUSIONDistrict heating networks enlarge and changecontinuously and therefore the conditions of pumpingwill also change. For that reason, it is important tocheck every now and then if the actual operating pointof the pump is as designed and what the efficiency ofthe present operating point is. The pumping could stillwork technically well but the pumps could be operatingwith low efficiency.The most important issues in designing and operatingof district heating pumping are:Specific pumping energy (electrical power /heat supply)1.7 %1.5 %1.3 %1.1 %0.9 %0.7 %0.5 %0.3 %Specific pumping energy vs. heat density0.1 %0 1 2 3 4 5 6 7Heat density, GWh/kmA sufficient but not too big pressure differencemust be guaranteed for customers.There must be enough pressure in all parts of thenetwork at all circumstances and at the same timethe maximum pressure level must not beexceeded.When designing pumping it is important to studyall possible pumping cases.Good operating point should be verified whendesigning and operating pumps.Pumping energy is dependent on certain parameters.The best parameter is considered to be heat density onwhich pumping energy is clearly dependent. And this isa quantity every district heating company measures.286

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe figure below is the same as the Fig. 8 added with ared line to help the reader estimate the pumping energyof his own plant. If the pumping energy is above the redline some measures ought to be taken.Specific pumping energy (electrical power / heatsupply)1.7 %1.5 %1.3 %1.1 %0.9 %0.7 %0.5 %0.3 %Specific pumping energy vs. heat density0.1 %0.1 %0 1 2 3 4 5 6 7 8Heat density, GWh/kmFig. 9. Electricity used for pumping in relation to the heatdensity of the district heating network + trend lineFigure 9 shows that on average the electricity neededfor district heating pumping should not be over 0.5 percent of the total energy supply (=sold+losses). If thedensity (supply/length of the network) of the districtheating network is less than 3 GWh/km, the energyneeded for pumping may rise. In any case theproportional pumping energy should be lower than1 percent.1.7 %1.5 %1.3 %1.1 %0.9 %0.7 %0.5 %0.3 %The following figure illustrates an example case inwhich the heat density is over 2.5 GWh/km. The figurecan be utilized when estimating the losses in realmoney if the proportional pumping energy is over theaverage of 0.5 percent.Value of excess pumping energy, 1 000 EUR/a400350300250200150100500Value of "excess" pumping energyHeat density > 2.5 GWh/km, Value of power 60 EUR/MWhProportional share of pumping energy 0.8 %Proportional share of pumping energy 0.7 %Proportional share of pumping energy 0.6 %100 600 1100 1600 2100Heat supply, GWh/aFor example, if the heat supply of the company is 1.1TWh/a and the proportional pumping energy is0.7 percent, the losses of unnecessarily high pumpingenergy is € 130 000 per year.Some of the pumping energy is converted to heat. Thisdecreases the value of the losses.In total, the potential savings in all Finnish districtheating companies are approximately 20 percent of thecurrent pumping energy, i.e. 30 GWh/a. This isequivalent to a yearly saving of approximately€ 2 million.REFERENCES[1] DH statistics 2007, Energiateollisuus ry, 2008287

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaMODELLING DISTRICT HEATING COOPERATIONS IN STOCKHOLM – ANINTERDISCIPLINARY STUDY OF A REGIONAL ENERGY SYSTEMD. Magnusson 1 , D. Djuric Ilic 21 Department of Thematic Studies – Technology and Social Change, Linköping University,SE-581 83 Linköping, Sweden2 Department of Mechanical Engineering, Division of Energy Systems, Linköping University,SE-581 83 Linköping, SwedenABSTRACTIn this paper, a combination of methods from socialscience (interviews) and technical science (modelling)have been used to analyse the potential forcooperation in the present and future district heatingsystem in Stockholm. The aim of the paper is to explorebarriers and driving forces for energy cooperation in theStockholm district heating system and to analyse thepotential for combined heat and power generation inthe system. In the study it was found that with betterconnectivity in existing systems, the annual systemcost would decrease by approximately 10 million €, andwith new CHP plants a similar potential exists. There isalso a large potential for decreasing the local andglobal emissions of CO2 with CHP plants. The resultsfrom the interviews showed that the existingcooperation has a long history and is working welltoday. The advantages are higher supply security andeconomic benefits, while disadvantages are a need formore administration and control because of a morecomplex system. That the barriers to cooperation areseldom technical is another conclusion. With thecombination of methods, we have gained a betterunderstanding of the actual potential for thedevelopment of the system.NOMENCLATURECO 2 – carbon dioxide;LECO 2 – local emissions of CO 2 ;GECO 2 – global emissions of CO 2 ;CHP – combined heat and power;BCHP – CHP plants fuelled by solid biomass;NGCHP – CHP plants fuelled by natural gas;TGC – tradable green certificates;GHG – greenhousegas.1. INTRODUCTIONSwedish district heating has a long history and is todayone of the dominant heating forms with approximately55% of market share, and an annual energy productionof approximately 55 TWh[1]. The first system was builtin Karlstad in 1948 and during the following decadesthe largest cities built their own systems, as was thecase in Stockholm [2]. Because of the large amount ofenergy in the systems, the fuel used in the plants has amajor impact on greenhouse gas (GHG) emissions,and there is also a large potential for using combinedheat and power (CHP) technology in the systems. CHPtechnology is becoming more important as a part ofcreating sustainable energy systems, which forexample can be seen in the EU directive for promotionof cogeneration [3]. In Sweden, as well as inStockholm, large investments are made in building newCHP plants, in large part thanks to the electricitycertificate system [1]. Another important potential withCHP generation is through the Electricity Directive of1996, in which the EU prescribed common rules forcreation of an open and competitive electricity market[4]. With a fully integrated electricity market, theSwedish prices of electricity can be expected toincrease. However, as long as they are lower thanEurope‘s there is a large potential for exportingelectricity. From a marginal power productionperspective, which will be discussed further in thepaper, there is a potential for decreasing globalemissions of CO 2 , if the exported electricity comes fromnon-fossil fuels.A large enough system is an important prerequisite forinvestment in CHP plants, in order to take advantage ofthe economy of scale of district heating and CHPgeneration. In Stockholm, the largest urban region inSweden, there are already well-developed districtheating systems. The systems started as smaller unitsthat gradually have been interconnected and todayconsist of three large networks. However, since thereare eight different energy companies in the city region,a working cooperation between the energy companiesis important. With this in mind we will analyze how theactors perceive existing and future cooperation. Thestudy is conducted with an interdisciplinary approachwhere interviews have been combined with modellingthe systems' performance with present and possiblefuture interconnections, present plants and future CHPplants, and finally with a hypothetical introduction ofnatural gas. The aim of the paper is to explore barriersand driving forces for energy cooperation in the288

Stockholm district heating system and to analyse thepotential for CHP generation in the system.2. CASE STUDYThere are three large district heating networks inStockholm that deliver more than 12 TWh of heatannually, produced in some 70 heating plants [5].Table I shows the heat production, types of baseproduction and installed heat and electricity capacity inthose networks. Six of the plants in the system areCHP plants with total installed electricity capacity ofabout 600 MW, which gives a possibility for productionof over 2 TWh of electricity annually [5].Table I. – Major district heating networks in Stockholm. [5]SouthcentralNorthwestSoutheastHeat production inthe year 2005 [TWh] 9.4 2.2 0.53Installed heatcapacity [MW] 4000 700 300Installed electricitycapacity [MW] 493 105 20Base production3. METHODSCHPwaste,CHPcoalBCHPNGCHPA combination of methods from social science andtechnical science has been used; modelling withMODEST and semi-structured interviews withrepresentatives from the largest energy companies.MODEST is an energy-system optimisation model withtime-dependent components that was developed atLinköping University in Sweden. MODEST uses linearprogramming to calculate the most profitablecombination of existing and potential new facilities andshows which investment options are financiallyviable [5].3.1 The structure of the interviewsThe interviews were conducted during the spring of2009, with representatives from the five largest energycompanies producing and/or distributing district heatingin the Stockholm region. These are Fortum, Norrenergi,Söderenergi, E.ON and Vattenfall. The representativeswere chosen by the companies themselves, since theycould better decide who would be most appropriate toanswer questions regarding interconnections,cooperation and future strategies. We decided to let therespondents remain anonymous, since one of theinterviewees wanted this. The interviews were semistructured,as we had similar questions for most ofThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia289them, although with some differences depending on thecompany. Because a semi-structured interview is aqualitative method, the possibility of using openquestions is an advantage, and since we are interestedin a specific situation, the interviewees have thechance to give their opinion. It also gives theopportunity to analyze the answers in different ways, tounderstand the opinions expressed [7].3.2 Modelling Stockholm’s district heating systemBased on the data from ―Open district heating networkin greater Stockholm‖ [5] a model of Stockholm‘sdistrict heating system has been constructed.Purchases and sales prices of electricity, taxes andtradable green certificates (TGC) are included in themodel (Table II) as well as the operating andmaintenance costs for all plants and fuel prices.However, due to agreements with the contact personsfrom the district heating companies, the prices for fuelare not presented in the paper.Table II – The average annual purchases and sales pricesof electricity, including all taxes and TGC. [8], [9], [1]Current price of electricity [€/MWh]Purchase Sale Sale with TGCincluded70.10 35.46 67.56European price of electricity [€/MWh]Purchase Sale Sale with TGCincluded83.30 48.65 80.77Carbon dioxide emissions used in this paper are shownin Table III [10]. However, since the greenhouse effectis a global problem, carbon dioxide (CO 2 ) emissionsare not simply analysed from a local perspective butalso in regard to a global perspective. The globalemissions of CO 2 (GECO 2 ) of the system are calculatedwith the assumption that electricity produced in theplants is going to replace marginal power production inthe integrated European electricity market. Since coalfiredcondensing power plants have the highestvariable cost compared with other sources of electricityin the EU, they work as the marginal powerproduction [11]. When assuming that the coal-firedcondensing power plants have an electricity efficiencyof 33%, each megawatt-hour of electricity generated insuch a plant releases approximately one tonne of CO 2 .According to that, any increase in electricity productionin Stockholm‘s district heating system can lead toreduced production in the marginal coal condensingpower plants, and consequently to a reduction ofone tonne of CO 2 emissions. However, it is necessaryto mention that considering the EU Emissions TradingScheme (EU ETS), the decrease of CO 2 emissions in

electricity production sector does not necessarily tolead to reduction of GECO2 [12]. But the marginalelectricity concept still has significance for futuremeasurement of and planning for future limitations ofCO 2 emissions and the future trading system.Table III. – Net emissions of CO 2 [10].FuelEmissions kg/MWhfuelOil 280Coal 330Waste 100Biomass 0Electricity 950Natural gas 2303.3 Description of chosen scenariosNine different scenarios have been analysedconsidering the possible future cases (Table IV), withspecial attention to economic and environmentalaspects.The existing district heating system (scenario 1) andthe system with three new CHP plants that are plannedto bee built according to the interviews and documents(scenario 4) have been analysed. Since the baseproductions in the networks differ, the differencesbetween the production‘s costs in different parts of thesystem are notable. Because of that, in both cases(scenarios 1 and 4) the influences of a betterconnectivity between networks have been studied(scenarios 2 and 5).Table IV. – List of the chosen scenarios.Sc.Plantsin thedistrictheatingsystemConnectivityElectricitypriceTGC1 existing existing Nordpool exist2 existing one Nordpool existnetwork 13 existing existing EU exist4 + new existing Nordpool existCHPP5 + new one network Nordpool existCHPP6 BCHP one network Nordpool exist7 BCHP one network EU exist8 BCHP one network EU do notexist9 NGCHP one network EU -1) Interconnections between the south-central and thenorth-west networks have been introduced as well asinterconnections between south-central and south-eastnetworks. Capacities for existing pipes have beenincreased.The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia290Since electricity generation will probably be the primaryproduction in all district heating companies in thefuture, when the Swedish electricity price becomes ashigh as the typical European price, in scenarios 6-9 ourresearch focuses on the cogeneration potential inStockholm's district heating system. Scenario 1 hasbeen used as a reference scenario for scenario 6.Scenario 3, where the influences of a higher electricityprice on the system with the existing plants have beenanalysed, has been used as a reference scenario forscenarios 7–9. Scenarios 6-9 are analysed as possiblefuture cases that may exist more than 10 years fromtoday. Because of that, all plants in the scenarios arenew so the investment costs for all plants areconsidered. While in the scenarios 6, 7 and 8 thesystem consists of 31 CHP plants fuelled by solidbiomass (BCHP), there are a total of 46 CHP plantsfuelled by natural gas (NGCHP) in scenario 9. Inscenario 9 it is assumed that the natural gas networkexists along the Swedish east cost.The characteristics of the CHP plants that have beenintegrated in the model of the district heating system inscenarios 4–9 are presented in Table V [13].Table V. – The characteristics of the new integrated CHPplants in scenarios 4–9 [13].Sc.4–5Technical characteristicsFuelElectricaloutputMWe %FuelefficiencyΑ*biomass 30 110 0.45waste 20 91 0.32biomass 80 110 0.466–8 biomass 80 113 0.519 natural gas 150 89 1.414–5Economic characteristicsProcessplant costOperating and maintenance€/KWe % of PPC €/MWh fuel2 745 1.5 2.455 440 3 9.312 110 1.5 2.456–8 2 110 1.5 2.459 715 2.5 0.9* electrical/thermal output3.4 Previous studiesTwo studies regarding Stockholm´s district heatingsystem were done in the years 2005 [10] and 2006[14], and the results showed that benefits for betterconnectivity between some parts of the system existed.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaIt was also shown that if all plants in the system arereplaced with BCHP plants, with an electricity-to-heatoutput ratio 0.46, up to 10TWh electricity can beproduced and the potential for decrease of GECO 2 ofthe system would be 3 tons CO 2 annually. If all plantsin the system are replaced with NGCHP plants, with anelectricity-to-heat output ratio 1.2, the electricitygeneration in the system can increase to 11TWh andthe potential for decrease of the GECO 2 of the systemwould be about 5 tons CO 2 annually. However, sincethese two studies were done, a new connectivitybetween networks has been built and the total installedelectricity capacity in the system has increased by 20%[5]. Furthermore, new CHP technologies are constantlybeing developed, which enable greater electricityefficiency and consequently greater benefits fromeconomic, energy and environmental viewpoints.Regarding interconnection and cooperation of DHsystems, some studies have been conducted in aSwedish context. However, none of them have focusedon cooperation between energy companies. They haveinstead focused on cooperation between energycompanies and industry.Thollander et al. [15] found that technical aspects areseldom barriers to cooperation. The barriers are ratherrisk, different aspects of information duringnegotiations, and other social factors such as inertiaamong personnel. Driving forces have been economicfactors such as an aim for lower costs and means ofcontrol, as well as environmental values. In a study witha similar aim, Fors [16] found the same results, thattechnical aspects are seldom barriers. Informationduring negotiations, stable contracts and theimportance of involving the personnel at the plants inthe process are important factors. It is also importantthat the cooperation benefits both parties. Grönkvistet al.[17] reached a similar conclusion in a study thatemphasises the importance of the willingness of peopleon both sides to cooperate. The main advantages ofthe cooperation are lower costs and benefits for theenvironment, while the main disadvantages are lessflexibility as both parties work under contracts.Historically, interconnection of technical systems hasbeen seen in the theory of Large Technical Systems asone way for systems to grow. Systems start in a localcontext, but when the technology is transferred to othergeographic areas, the systems grow and can then beinterconnected as they often have grown into eachother. Interconnection of systems can also beexplained through the fact that larger systems have ahigher load factor and better economic mix [18], [19],[20].4. RESULTS OF THE SCENARIOSThe results from the scenarios are presented inTable VI and Table VII.According to the optimisation results, if betterconnectivity is introduced, some economic benefitsexist. In both cases the case with only existing plants inthe system and the case where the new plants areintroduced in the model (scenarios 2 and 5) thedecrease in system costs would be about 10 million €annually. The potential for decrease of theenvironmental impact of the system is more notable. Ifbetter connectivity were introduced in the systemtoday, the biomass share in total fuel use would be 8%higher and consequently both the local emissions ofCO2 (LECO2) and GECO2 of the system would beabout 0.25 million tons lower annually. The potential fordecrease of GECO2 of the system if better connectivityis introduced after the building of new CHP plants(scenarios 4 and 5) is 0.4 million tons annually.Table VI. – Results for the scenarios – economic aspects.Sc.AnnualsystemcostsCHP heatproductionshareElectricityAnnualproductionTheincomefromelectricitymillion € % TWh Million €1 258 47 2.30 1222 245 47 2.31 1253 243 48 2.35 1504 204 58 2.96 1645 192 62 3.15 1766 403 4247 344 100 6.39 4828 546 2819 504 100 17.66 777As the electricity price increases, the system wouldearn extra income from the electricity sold, and thus theheat production cost would decrease (scenarios 1, 3).This gives an even bigger advantage to CHPgeneration compared with pure heat production.291

Table VII. – Results for the scenarios – environmentalaspects.Sc.Biomassshare inthesystemLECO 2% [milliontons/year]1 48 2.50 0.322 52 2.25 0.063 49 2.46 0.23GECO 2 ofthe system[milliontons/year]4 52 2.12 – 0.695 55 1.91 – 1.0867 100 0 – 6.0789 0 7.80 – 8.98The income from the electricity sold in scenario 3 isabout 30 million € higher then the income in scenario 1,and because of that the system cost is 6% lower. Thedifference between the electricity production inscenarios 1 and 3 is not significant, but in spite of that,the decrease of GECO2 of the system in scenario 3 isalmost 100%. The reason is higher biomass share inthe total fuel used in the system in scenario 3, andconsequently lower LECO2 in the system.The introduction of three new plants in the system(scenario 4) would lead to a significant reduction of theheat 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 wouldbe 20% lower. This confirms that heat production inCHP plants has a major influence on the economicefficiency of the district heating system. With theassumption that the electricity produced would replacethe marginal electricity in the European electricitymarket, reduction of GECO2 of the system would bealmost 1 million tons annually.If all plants in the system are BCHP (scenarios 6–8) orNGCHP (scenario 9) plants, the annual electricityproduction would be as high as 4.5% and 12% of thetotal electricity production in Sweden, which was about145 TWh in the year 2008 [1]. The annual income fromthe electricity sold in those scenarios is much higherthen the income from the electricity sold in the otherscenarios. In the scenarios with typical Europeanelectricity price, (scenarios 7–9), the income from theelectricity sold is 220%, 90% and even 420% higherthen in scenario 3, where the system with the existingplants is analysed with the higher electricity price. It isThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia292also notable that in scenarios 6, 7 and 9 the annualincome from electricity is higher than the annualsystem costs. However, since all plants in thosescenarios are new, the total investments are high.Because of that, if the analysed time period is just10 years, the annual system costs are much higherthen today.The lowest GECO2 of the system are in the scenarioswhere all plants in the system are BCHP (scenarios 6-8) and NGCHP (scenario 9) plants. In those two casesGECO2 in Sweden, which is about 60 million tonsannually [21] would be reduced by approximately 9%and 15% respectively, with the assumption that theelectricity produced would replace the marginalelectricity. LECO2 in the system is highest in thescenario where all plants are NGCHP but at the sametime GECO2 of the system is lower because of the highelectricity production.5. RESULTS FROM THE INTERVIEWSIn the following section the results from the interviewswill be presented. The interconnections between thesystems make it possible to cooperate regarding heatproduction and distribution.5.1 The system todayThe interviews show that the interconnections have ahistorical background. Most of them were made duringa period when a regional energy company calledSTOSEB (Greater Stockholm Energy Company)existed, where the municipalities, which to a largeextent owned the systems then, were represented. Themain reason for the interconnections then was supplysecurity. When the systems were interconnected, thecompanies could help each other during stops, and thisis still the case. All representatives say this, and therepresentative from Söderenergi expresses it this way:…At the same time it is a common good. It is good thatthe systems are interconnected. It is an extra security ifone plant should stop for some reason [22].The advantages historically and foremost today arealso economic. The emissions trading makes itadvantageous, since the companies can use theproduction better by making ―capacity trades‖ and evenout the production cost between the companies:We see that we can use existing production moreeffectively. Most of the trades are a trade to mid-priceso to speak. You can say that we split the profit.Capacity trading (effektköp) is also common. Like wehave here with Söderenergi, we have partlya production cooperation and partly we buy capacity.They have more capacity than they need today [23].

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe main advantage with capacity trading is to avoidpeak load, which often is oil-based, which is costly bothfor the fuel price but also because of the emissions.Another point is that, as the Fortum representativesaid, it is possible to even out effect between systems.One system may have cheaper base load than theother, and for tax reasons it may be cheaper to buyfrom the other than to use peak load.One factor that is pointed out for a successfulcooperation is that both parties can benefit from it. Asin all business, it is important that the cooperation becorrect from a business standpoint and that both partsare satisfied [24].The extent of cooperation varies between thecompanies. Some have more extensive cooperationwith daily trades, like Fortum and Söderenergi, whileothers, for example Fortum and E.ON, do not tradeevery day. In the latter case, they normally do not tradeas much during winter, although sometimes when peakload is needed it is decided quickly [25]. Anotheradvantage with the interconnections is that thecompanies can cooperate regarding revisions of theplants. While one company has revision duringsummer, the other can produce for the other company.The factors that are seen as barriers are seldomtechnical. The companies think that the technicalproblems often can be solved while making theinterconnection and at that point there is a need tonegotiate certain aspects. For example, who providesthe electrical energy for the pumps and takesresponsibility for the regulation of the water pressure inthe culverts and repairing the system in a joint part ofthe system? However, this is often solved:Yes, the other things we can handle while building thetechnical parts. At that point we hopefully haveidentified all technical barriers so that they can betaken into account. They should not appear duringproduction. Settlement of account and such things,they are not a big problem although complicated.However, it is nothing that makes you pass on aprofitable cooperation [23].In the above quote, we see one of the disadvantageswith today's cooperation, on which all the companiesagree, and that is the settlement of accounts. It iscomplicated to control the systems and the trades, andit requires staff to do so.5.2 Barriers towards more co-operationsIn the interviews the companies expressed satisfactionwith the present cooperation. Few actual barriers assuch were expressed, except the ones that today‘ssituation creates. For example, it is almostgeographically impossible to expand the systems tosmaller systems nearby. As could be seen in themodelling, the systems are also already wellinterconnected:Yes, the principal structure is already established. (…)It is this connection, between the central and thenorthwest system, it is the only one. That is not solvedyet [23].This particular connection would interconnect the twomain systems, and has been discussed in someinvestigations [26], [27]. However, it is yet to be done.This connection is most important for Fortum, as forexample E.ON thought that it made little difference tothem.The other main connection still missing is connectionbetween the south system and Vattenfall's system inthe southeast. Vattenfall thinks that the question hasbeen raised on occasion, although never realized. Theygive no specific reason for this; they state that allcooperation is important and that differentinvestigations have shown the advantages, although itis difficult to quantify what it means practically [28].Stockholms Energi (now Fortum) previously owned oneof the plants, and there were plans to interconnect thesystems then. Fortum gives no explanation for why theinterconnection has not been done earlier or now.Although no direct comments regarding the lack ofinterconnection were made, one of the intervieweeswho previously worked at Vattenfall said that there wasan opinion at Vattenfall that they prefer to keep tothemselves, without interconnections, and should notwork towards cooperation. Comments without aspecific direction were also expressed in interviews thatthere was a lack of will to cooperate from somecompanies. There is also a history of rivalry betweenVattenfall and the former Stockholms Energi [29]. It ispossible that this rivalry stills exists. Fortum alsoexpressed opinions about the fact that other companiesare building their own CHP plants instead of trying tofind regional solutions.5.3 Building CHP in the systemAs seen in the scenarios, in the near future in Swedenmany CHP plants are planned and will start to be built.In Stockholm most of the companies have plans forCHP, and two of them have already built in the lastyears, for example Igelsta (Söderenergi) and Jordbro(Vattenfall). Other companies are making plans, suchas Norrenergi, EON and Fortum. The reasons forbuilding CHP are varied, but the most clear is that theysee economic advantages in selling electricity, and ourstagnating heat load ahead. By selling electricity thereis a possibility to keep profits high, even with astagnating heat load. The system is also relatively oldand well established; the potential for furtherconnections are getting smaller as saturation in the293

heating market for district heating makes it moredifficult to expand:A rough rule of thumb has been that the expansion withnew customers that have been, (...), has been eaten upby the efficiency we could achieve together with thecustomers in their buildings. So basically, the heat loadhas 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 wecan not expand. We have our two customers anddistrict heating is not a new thing in the municipalitiesso the chance of getting new customers is limited [22].The other representatives are of a similar opinion, thata stagnating load can be expected, and CHP is a wayto keep profits high. The Swedish certificate systemalso makes it advantageous to build new bio-fuelledCHP-plants. Another reason, arguably of a morerhetorical character, is that building CHP is moreeconomically and environmentally correct since the fuelefficiency is higher with CHPs. As scenarios 4–8 show,there is major potential for reducing local and globalCO 2 emissions.In the interviews we also asked questions about thepossibility of an introduction of natural gas in theregion. Investigations have been made earlier by theabove mentioned STOSEB; however, the plans nevercame to reality. Generally the representatives did notthink that an introduction would happen. Since most ofthem also have strategies to be climate neutral, naturalgas probably is not an option. The large investments ininfrastructure are another barrier:These are such large infrastructure investments andnatural gas is not especially cheap either. It is difficultto come in with natural gas in this energy system. It israther stable [23].What the representative here points at is also theinertia in the system. In LTS terms it is calledmomentum: as the system is stable, it is difficult tochange the structure [18], [19].6. CONCLUDING DISCUSSIONThe study has shown the advantages of aninterdisciplinary approach. Advantages withinterconnections and CHP have been shown in themodelling; however, as there are many different actorsinvolved, there is a need for a will to cooperate. Theinterconnections have a historical background, with anaim for higher supply security, and today most of themcontinue to cooperate, despite the fact that thestructure and ownership of the companies in somecases have changed since the deregulation of theelectricity market in 1996. As previous studies haveshown, the main advantages with cooperation havebeen economic, as is also the case in this system. TheThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia294actors say that they can optimise the system'sperformance, and our scenarios have shown that morecooperation could benefit them even moreeconomically. Even though the gain is not extremelyhigh, since the lower system cost would beapproximately 5%, there is potential. However, sincethere seems to be reluctance to cooperate betweensome actors, it is difficult to fulfil the potential.Advantages with the cooperation are said to be apossibility to even out the production in the system andthus avoid peak load. The disadvantages with thecooperation are the need for more administrative workto control the system and the trades; the control of thesystem becomes more complex. This study alsoconfirms previous studies that have pointed out thattechnical aspects are seldom barriers to cooperation.Most things are solved while the systems are beinginterconnected, and the will of the persons involved tocooperate is important.There is a large potential in building new CHP plants,both from an economic and an environmentalperspective. If all the plants in the system werereplaced by BCHP or NGCHP, the electricity producedcould make up to 4.5 or 12% of total Swedish electricityproduction, based on the fact that total production inSweden in 2008 was 146 TWh [1]. The reason that thedifference between the electricity productions in thosetwo cases is so large is a big difference between theelectrical/thermal outputs (see Table V). The introductionof NGCHP is a less likely future since it can beconsidered only with the assumption that the naturalgas network already exists along the Swedish eastcost. On the other hand, introducing more BCHP in thedistrict heating system would increase the system‘sdependence on biomass availability and the heatproduction cost would become highly sensitive to thesolid biomass cost. The actors are highly aware of thepotential for CHPs. Since they are expecting astagnating heat load, the sale of electricity is a way tokeep profits high. However, none of them think thatnatural gas will become a reality in the near future, andeven if it did, the introduction is expected to besomewhat problematic, since the fuel can beconsidered fossil fuel and substantial infrastructure isneeded.The study has shown a potential for decreased LECO 2and GECO 2 . The largest potential from a localperspective is from BCHP; so, since the LECO 2 wouldbe low and with high electricity production, the potentialfor lower GECO 2 would exist. The high electricity-toheatoutput ratio in NGCHP has a high potential fordecreasing GECO 2 of the system. If all plants in thesystem would be replaced with NGCHP the GECO2 ofthe system would be -9 million tons annually. However,in that case LECO 2 would be much higher than today.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaModelling of a system gives one side of the truth, asdoes interviewing the actors involved. When combiningthe methods there is a possibility of getting a better anddeeper understanding of the actual potential forcooperation. The historical and social aspects cannotbe neglected; they can in many cases explain whypotentially beneficial cooperation is or is not done,while modelling can show the actual potential.7. ACKNOWLEDGEMENTSThis article has been carried out in two PhD projects inthe Energy System Program, financed by the SwedishEnergy Agency. The authors would also like to thankJenny Palm (Linköping University) and Louise Trygg(Linköping University) for valuable comments on thepaper.8. REFERENCES[1] Statens energimyndighet (Swedish EnergyAgency), Energy in Sweden 2009. SwedishNational Energy Administration, ET 2009:30,Eskilstuna, Sweden (2009)[2] S. Werner, Fjärrvärmens utbredning och utveckling(The development and expansion of districtheating), Värmeföreningen, Stockholm (1989)[3] European Union, Directive 2004/8/EC, ‗Directive onthe promotion of cogeneration based on a usefulheat demand in the internal energy market‘,(2010), homepage:http://www.managenergy.net/products/R81.htm,2010-04-28[4] European Union, , Directive 96/92/EC ‗Secondreport to the Council and the European Parliamenton harmonisation requirements concerningcommon rules for the internal market in electricity‘,(2010), homepage:http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:32003L0054:EN:NOT , 2010-04-28[5] B. Dahlroth, Öppnade fjärrvärmenät iStorstockholm (Open district heating network ingreater Stockholm), Stockholm, Sweden:Fastighetsägarna Stockholm, (2009)[6] D. Henning, Optimisation of Local and NationalEnergy Systems. Development and use of theMODEST model. Dissertation No. 559. Linköpingsuniversitet, Linköping (1999)[7] S. Kvale and S. Brinkmann, Den kvalitativaforskningsintervjun (The qualitative researchinterview), Studentlitteratur, Lund (2009)[8] Nordpool. Nordpool electricity spot market, (2009),homepage: http://www. nordpool.com. 2009-12-18.[9] M. Melkerson and S-O. Söderberg, Dynamiskaelpriser – elprissättning på en integrerad europeiskelmarknad (Dynamic electricity prices – pricing inan integrated European electricity market – inSwedish), Sweden: Institute of Technology, Dept ofMech Eng, Linköping University, Linköping (2004)[10] N. Levinson and R. Freiman, Optimal kraftvärmeoch nätinvestering I Stockholms fjärrvärmesystem(Optimal CHP and district heating networkinvestments in Stockholm), Linköping University,Linköping (2005)[11] Statens energimyndighet (Swedish EnergyAgency). ‗Marginal elproduktion och CO 2 -utsläpp iSverige‘ (Marginal electricity production and CO 2 -emissions in Sweden, in Swedish), SwedishNational Energy Administration, ER 14:2002,Eskilstuna, Sweden. (2002)[12] E. Dotzauer, Greenhouse gas emissions frompower generation and consumption in a nordicperspective, Energy Policy 2010, Vol. 38(2), pp.701-704.[13] H. Hansson, S-E. Larsson, O. Nyström, F. Olssonand B. Ridell, El från nya anläggningar (ElectricPower from New Plants), Elforsk, Stockholm(2007)[14] M. Danestig, A. Gebremehdin and B. Karlsson,Stockholms CHP potential – An opportunity forCO 2 reductions? Energy Policy 2007, Vol. 35(9),pp. 4650-4660.[15] P. Thollander and I-L. Svensson, ‖Vägen tillframgångsrika värmesamarbeten – en fallstudie‖(Road to succesful heating co-operations – a casestudy), In: L. Trygg, L. et al (2009) optimalafjärrvärmesystemi symbios med industri ochsamhälle, Rapport 2009:13, Svensk fjärrvärme(Swedish district heating association) (2009)[16] J. Fors, ‖Spillvärme från industri till fjärrvärmenät –sammanfattning av intervjuer på 5 orter‖, (Excessheat from industry to district heating systems)Rapport 2004:5, Svensk fjärrvärme (Swedishdistrict heating association) (2004)[17] S. Grönkvist and S. Sandberg, ―Driving forces andobstacles with regard to co-operationbetweenmunicipal energy companies and processindustries in Sweden‖, Energy Policy 2006, Vol. 34,pp. 1508–1519.[18] T.P. Hughes, Networks of Power: Electrification inWestern Society 1880 – 1930, John HopkinsUniversity Press, Baltimore (1983)[19] B. Joerges, Large Technical Systems: Conceptsand issues, In: R. Mayntz and T.P. Hughes, The295

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaDevelopment of Large Technical Systems,Campus Verlag, Frankfurt (1988)[20] J. Summerton, Changing Large TechnicalSystems, Westview Press, Boulder, CO (1994)[21] SCB (Statistics Sweden). Utsläpp av växhusgaser,(2010), homepage:http://www. scb.se. 2010-02-17.[22] Söderenergi, Production manager, 090318[23] Fortum, Site manager and Senior advisor, 090325[24] Norrenergi, Production manager, 030304[25] E.ON., Group manager production, 090323[26] STOSEB, STOSEB 92 – Energiframtider förStockholms län (STOSEB 92 – Energy futures forStockholms County), STOSEB, Stockholm (1992)[27] Fortum & Stadsbyggnadskontoret, Möjlighetsstudie:Nätintegration Storstockholm, (Possibilitystudy: Netintegration in greater Stockholm) Fortum& Stadsbyggnadskontoret, Stockholm (2005)[28] Vattenfall, Head of business development andSenior advisor, 090320[29] STOSEB, 25 Energiska år – Om Stor-StockholmsEnergi AB (25 Energic years – about greaterStockholms Energy AB), STOSEB: Stockholm,(2003)296

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaCUTTING COSTS OF DISTRICT HEATING SYSTEMSBY USING OPTIMIZED LAYING TECHNIQUESAlexander Goebel 1 , Dr. Stefan Holler 11MVV Energie AG, Mannheim, GermanyABSTRACTThe soil covered plastic jacket pipe is the commonstate of the art laying technique in the district heatingsector: A preferable shallow trench is dug out andbackfilled with cable sand after the installation of thetwo pipes. Alternative possibilities concerning thedigging of the trench, the backfill and the piping itselfare evaluated in this paper. Results show, that anoptimized laying technique can save construction orrunning costs under the right boundary conditions:Backfill materials with insulation properties can reducethe heat losses by about 25 %. Using glass-reinforcedplastic pipes (GRP) instead of steel pipes leads topump energy savings of about 40 %.INTRODUCTIONIn the first place, excavation costs could be cut bydigging smaller and shallower trenches. However, thisis only possible if the location of the construction site isappropriate. In an urban area the situation is completelydifferent from a rural area concerning space andregulations. The paper describes the boundary conditionsand compares different methods from the technicalas well as the economical perspective using theexample of the district heating system in Mannheim,Germany.The second approach which will be presented in thepaper is the potential to reuse the excavated materialand to use self-compacting material when refilling thetrench. Furthermore, it is also possible to use newmaterials with better insulation properties in order to cutdown heat losses. In the paper the different propertiesof the new materials will be compared and evaluated.A third possibility to reduce costs is the use ofspecialized piping systems wherever possible.Nowadays a wide range of products is available on themarket. In many cases a specialized system fits someapplications better than a standard system does. Notonly insulation properties but also compensation,ductility and friction losses are important characteristicsof modern piping systems. In the paper it will be shown,how costs could be reduced by using less or nocompensation measures (cold laying, flexible pipes,fibre pipes), by avoiding welding measures (flexiblepipes for house connections, fibre pipes) or by reducingfriction losses (fibre pipes).MATERIALS AND METHODSThe cost saving potentials of the alternatives, concerningthe digging of the trench and the backfill, aremainly evaluated by outlining the results of researchreports. Calculations are used in order to estimate theinsulation properties of special backfill material. Alsothe cost saving potentials of pipes with low frictionlosses are evaluated with simple equations.RESULTSReuse of the excavated material [1], [4]Earlier research activities have proven, that plasticjacket pipes could be used with backfill materialshowing a greater grit size than cable sand. Specialprotection material is available not only for the mufflesbut also for the pipes. Field tests have shown, thatthere are promising money saving potentials becauseof the significant reduction of transport and disposalcosts. A consideration of reusing the excavated soil isalso reasonable from an environmental point of view.Following points are important, when it comes to anevaluation of this possibility at an individualconstruction site:the grading of the excavated materialsandy or cohesive groundcompacting propertiesthe friction between the ground and the jacket pipeprotection measures for muffles and the pipeunderground construction regulationsa place for the storage of the excavated material(beside the trench, container or any place near theconstruction site)formation of dust (especially in the summer)contaminationpH-valuean improvement of the excavated material withlime (especially with cohesive ground)a removal of the coarse materiala separation of the material, if the ground is split upin different layersThe use of self-compacting material [1], [4], [5]It is important to distinguish between the following twotypes of self-compacting material:stabilised sand mix297

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia excavated material mixed with water and specialadditives in order to get a self-compactingbehaviourThe use of self-compacting materials offers a widerange of advantages and applications:it is possible to dig out a narrower trench, becauseno machines are needed for the critical compactionaround the pipesthe backfill process is significant faster – after thetrench is filled up, it takes normally only one dayuntil the material is hard enough to walk onself-compaction is more reliable within difficultconditions (many crossing pipes etc)without the use of compaction machines, buildingsnearby the construction site are stressed less (novibrations)there is less inconvenience for residents livingnearby the construction site, because of the noisereductionin combination with the pipeline laying technique,the sheeting can be omitted, because nobodyneeds to work in the trenchA common problem is the local availability of thetechnology. The price is also an issue, if the reason ofthe application is the approach to save money.Another problem concerning the dimensioning of thecompensation measures is the bad predictability of thefriction between the jacket pipe and the selfcompactingmaterial. Depending on whether the pipesare taken into service during or after the hardeningtime, which is about a month long, a more or lesscrucial tunnel effect is observed [4].The reuse of the excavated soil as base material ismore elegant, than the stabilised sand mix, because ofthe recycling aspect. Research projects have evenshown that sharp particles are less problematic,because they are enclosed in the self-compactingmass. An advantage of the stabilised sand mix is theeasier application.If the district heating line does not run under a street,compaction measures around the pipes can be avoidedsimply by watering the cable sand, which is filled inlayers into the trench.Cost saving potentials of backfill material withinsulation propertiesIf a reduction of the heat losses comes into consideration,the change to a higher insulation series isevaluated. Calculations show that in most cases aneconomical justification is not given for this measure.The idea of filling the trench around the pipes withmaterial that provides an additional insulation seems tobe promising. Like in the case of the self-compactingmaterial, the local availability is the greatest problem.An economical justification is only achievable, if thetransport costs are low and the heat price is high. Froma technical point of view, the compaction behaviour hasto meet the requirements and regulations. The jackettemperature must not exceed the maximum of 50 °Cand the friction between pipe and the material shouldbe in the common range.A calculation method for heat losses of plastic jacketpipes is described in EN 13941 ANNEX D [2]. Figure 1shows the influence of the thermal conductivity of soilλ s on the heat losses. Normally the value of λ s lies inbetween 1,0 and 2,0 W/m*K [2]. The curve becomesvery non-linear below a value of 1,0 W/m*K. Thisindicates that it is necessary to customise thecalculation method in order to get realistic results. Theheat losses are cut down by 30%, if the λ s is reducedfrom 1,5 to 0,35 W/m*K.heat losses of the flow and return pipeΦf + Φr [W/m]75706560555045400.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00Thermal conductivity of the soil λ S [W/m*K]Fig. 1 Heat losses of a district heating line as a function ofλ s (DN 250, 120/50 °C, Z = 0,6 m, C = 0,55 m,λ i = 0,03 W/m*K)The insulation material should solely be integrated inthe calculation as an additional thermal resistivity(R λ,embedment ), since the soil around the pipes is notmade completely out of it. The heat dependency of theinsulation foam‘s thermal resistivity should also betaken into account. Figure 2 illustrates, what is meantwith ―addi tional insulation layer‖.Fig. 2 The different layers of the heat conductivity problem298

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe modified calculation procedure is made up of thefollowing equations:RRR1 ,steeliD i12 steel1 D ln2 i D Doln Diinsulationo(1)(2)(3)RUUhflow14 returnR RsR Rsembedmenti,flowi,return 2(Z R sln1 DC sembedment R Rh R Rh1 R,steel1 R,steel0) R R2 ,jacket,jacket R R(7),embedment(8),embedment(9)R,jacket12 C D ln DCinsulation(4) i i , 50C0 ,0001T i , average 50K(10)RR,embedments12 embedment DlnC 2sD1 4 ( Z R0 s) ln2 s DC 2 sembedmentembedmentC(5)(6)TflowTreturnf r Uflow UreturnTsoil (11) 2 Ti, average Tfluid QRi R R,steel R,jacket(12)heat losses of the district heating line[W/m]18016014012010080604020without insulation materialwith insulation materialreduction in %30%25%20%15%heat loss reduction010%0 200 400 600 800 1000nominal diameter [DN]Fig. 3 Reduction of the heat losses for DN 15 to DN 1000The average temperature of the insulation wascalculated with following equation and put back into(10). A VBA script was used to iterate five times.Fig. 3 shows the results of the calculation. Theinsulation material was taken into account with a valueof 0,33 W/m*K (λ embedment ). Around and in between theflow and the return pipe a space of 0,2 m for each pipesize was chosen (s embedment ). The depth of cover had avalue of 1 m (Z). Like in the previous example, the flowtemperature was at 120 °C and the return temperaturewas at 50 °C.Fig. 3 shows, that savings are significant lower withsmall diameters. Also the specific thickness of the PURinsulation, which differs because of standardised jacketpipe diameters, has an impact.The heat losses of a DN 250 pipe are reduced from 67to 51 W/m (24%). This means, that the heat lossreduction is 6% less compared to Fig.1.299

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaSince the use of an insulating backfill is more efficientwith huge diameters, a DN 700 pipe was chosen for anexample scenario. An annual average for the flow andreturn temperature was taken into account. For thecalculation of the required backfill volume in theembedment, 0.2 m space in every direction of the pipeswas estimated. It is important for the calculation to takeonly the additional costs into account. That means theprice difference between cable sand and the insulationmaterial including the transportation costs.A common value of 6% was chosen for the requiredrate of return (i).The net present value C 0 was calculated with thefollowing equation:CTt0 I (R t) (1 i)t 1(13)The internal rate of return shown in Fig. 4 wascalculated with the IRR- function in Excel.Table 1 Scenario for insulation materialParameter Value UnitLength of the districtheating line:5000 mNominal diameter: DN 700Average flowtemperature:Average returntemperature:Annual hours ofoperation:Heat price (at the time ofthe invest):Required volume ofinsulation material:95 °C50 °C8760 h15 €/MWh th1.85 m 3Additional specific costs: 16 - 17 €/m 3Required volume for thewhole line:Heat loss with use of thematerial:Heat loss without use ofthe material:9250 m 369.6 W/m90.7 W/mEnergy savings: 23.3 %Annual savings of thewhole line:924 MWh thAdditional investment (I): 148,000 – 157,000 €Required rate of return: 6.0 %Time of cash flow: 20 anet present value (20 years) [€]60,00050,00040,00030,00020,00016 €/m 3 17 €/m 316 €/m 3NPV10,000internal17 €/m 3rate ofreturn00.0% 1.0% 2.0% 3.0%growth rate of the heat price10.0%8.0%6.0%4.0%2.0%0.0%Fig. 4 NPV and IRR of the scenario defined in Table 1depending on the additional specific costs of the insulationmaterial.The results in Fig. 4 show, that the additional specificcosts should be below 17 €/m 3 in order to get a positivevalue spread, assumed that the required rate of returnis 6 %. A reduction of specific costs of 5% (16 €/m 3 )results in an increase of the value spread by 1%. Thenet present value after 20 years rises about 10,000 €.The growth rate of the heat price is difficult to predict,but has an important influence within the given periodof 20 years. Presumably the heat price is mainlyinfluenced by emission trading, governmental subsidiesand the development of the fossil fuel price.Other scenarios may estimate higher growth rates, butin order to get realistic results, the rate was varied from0% to 3,5%.Table 2 gives an example of materials with low heatconductivity that could be interesting to use as backfill.It is obvious to look for natural products, because of theprice and environmental regulations.internal rate of returnAnnual growth rate ofthe heat price:0 – 3.5 %300

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTable 2 Heat conductivity of different materialsFlexible pipesFlexible pipe systems, which are defined in EN 15632[3], are mainly distinguished by the material of theservice pipe:hard plaster [9]plastic (e.g. PE-Xa, Polybuten)coppermild steelcorrugated stainless steelW/m*K600 kg/m 3 0.18900 kg/m 3 0.301200 kg/m 3 0.431500 kg/m 3 0.56light sediment natural stone [9] 0.85porous rock, e.g. lava [9] 0.55natural pumice [9] 0.12bitumen [10]2100 kg/m 3 0.70as matter, 1050 kg/m 3 0.17membrane, 1100 kg/m 3 0.23expanded volcanic rock (perlite) [11]loose perlite, 50 - 130 kg/m 3 0.07perlite compressed with filaments,170 - 200 kg/m 3 0.06Thermosand © [8] 0.33Flexible pipes have a significant higher operatingpressure (16 or 25 bar) [3], if the service pipe is madeof metal. Also the maximum and continuous operatingtemperatures differ much. Because of this fact,systems with plastic service pipes could normally notbe used within huge district heating networks. In asmaller network with lower flow temperatures, whiche.g. was built to distribute the heat of a small blockheating station, a system with a plastic service pipemight have an application.Flexible systems with a corrugated service pipe havesignificant higher friction losses, which has to be takeninto consideration (dimensioning).When it comes to money saving potentials, the mostimportant properties of flexible pipes are the following:less welding measuresself-compensatingless insulation workless work concerning the monitoring systemless head access holes, because of the reducedwelding measuresless risk of leaks, because of less weld joinsfaster laying of the pipesPipes with low friction lossesService pipes made of glass-reinforced plastic (GRP)have significant less friction losses than steel servicepipes. Because of their chemical resistance, GRPpipes are used mainly in the chemical industry. It isimportant do distinguish between filament-wound pipesand centrifugally cast pipes. Because of the Poisson‘seffect almost no compensations measures are needed,if filament-wound pipes are used. Centrifugally castpipes need to be compensated, but have an evensmoother inner surface, which means the lowestpossible friction losses. Also the temperatureresistance is a little bit higher. The greatest problem ofGRP pipes is the fact that the service life is cut downby high temperatures in combination with highpressures (derating factor).Fig. 5 shows the possible savings, if a GRP pipe with asurface roughness of k = 0,01 mm is compared to asteel pipe with a roughness of k = 0,2 mm.The following equations were used:The calculations of the Reynolds number:w dRe (14)The value of the kinematic viscosity () was taken with2.941*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 thefollowing equation [6]: k 68 0,11 d Re 0,25(15)The pressure loss was calculated with the followingequation: L wp d 22(16)301

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe equation for the pump power:P pumpVp(17)Equation (17) shows, that the correlation betweenpump power and pressure loss is linear. The savingsare expressed as a percentage. They do not dependon the diameter or length of the pipe, because only thefriction factor differs.reduction of pressure loss47.5%45.0%42.5%40.0%37.5%35.0%32.5%0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6flow speed [m/s]Fig. 5 Pump energy savings of a GRP pipeAnother important aspect of GRP pipes are the joints:the pipe ends are glued together with a two componentadhesive, which is heated up for the curing process.This can be an advantage, because welding measureson a construction site are often problematic (lack ofspace, wind). Statistics show that in most cases leaksare caused by bad weld seams [12].CONCLUSIONA consideration of an alternative laying technique isusually worth the work, because the money savingpotential is often higher than expected. It dependsstrongly on the single project and the local boundaryconditions (heat prices, rural or urban area, availabilityof technologies/materials etc), whether a differenttechnique makes sense from an economical point ofview.Containing a bunch of alternatives in district layingtechniques, Table 3 gives a rough overview with asimple rating. Techniques appearing in the table, whichare 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 wordalternative is used, it means every technical aspect,which differs from the standard laying techniquedefined in the abstract.Table 3 Overview of alternative laying techniques++ highly recommended to take intoconsideration from an economical point ofview+ a closer look seems promising0 an economical benefit can be achieved, ifspecial boundary conditions are given- because of technological or economicalreasons, the use of the technology is notrecommended-- the technology is not available or can notbe applied reasonable under the givenboundary conditions302

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaGRP pipesflexible pipepipeline laying techniqueself-compacting materialreuse of excavated soilcombined laying withother supply pipesstacked layinginsulation as a castingcompoundinsulating backfilltrench-less layingtechniquesnominal diameter< DN 150nominal diameter> DN 150+ ++ 0 ++ + 0 + ++ - +++ -- ++ + ++ 0 ++ 0 ++ ++new district heating line ++ + + ++ ++ ++ ++ - + ++renovation measure 0 - 0 ++ ++ 0 -- ++ + --construction site in anurban areaconstruction site in arural areayet to be built housingestate+ ++ -- ++ 0 + + + 0 +++ + ++ 0 ++ ++ ++ + + 0+ + ++ + + ++ ++ - + 0existing housing estate + ++ -- ++ + + + - 0 ++NOMENCLATURER λ,jacketinsulance of the jacket pipeλ sthe coefficient of thermal conductivity forthe soilR αinsulance of the convective heat transferinside the pipeλ iλ i,50 °Cλ Cλ steelthe coefficient of thermal conductivity forthe PUR insulationthe coefficient of thermal conductivity forthe PUR insulation at 50 °Cthe coefficient of thermal conductivity forthe jacket pipethe coefficient of thermal conductivity forthe service pipeR 0R λ,embedmentD iD 0D insulationD Csurface transition insulanceinsulance of the insulating material usedin the embedmentinner diameter of the service pipeouter diameter of the service pipeouter diameter of the PUR insulationouter diameter of the jacket pipeλ embedmenαR hR sR iR λ,steelt the coefficient of thermal conductivity forthe insulating material used in theembedmentheat transfer coefficientinsulance of the heat exchange betweenflow and return pipeinsulance of the soilinsulance of the insulation materialinsulance of the service pipeT flowT returnT soilT i,averageT fluidU flowU returnflow temperaturereturn temperaturetemperature of the soilaverage temperature of the PURinsulationflow or return temperatureheat loss coefficient for the flow pipeheat loss coefficient for the return pipe303

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaZCs embedmentΦ f + Φ rC 0IR tiTtkRewdλΔpLρP pumpVηdistance from surface to the middle of thepipedistance between the centre lines of thetwo pipesspace between the pipe and the trenchwallheat loss per pipe pairnet present valueinvestmentnet cash flow (annual savings)required rate of returngiven periodthe time of the cash flowsurface roughness of the service pipeReynolds numberflow speedinner diameter of the service pipethe kinematic viscositythe pipe friction factorpressure losslength of the pipethe density of the heating waterpump powerflow ratepump efficiencyREFERENCES[1] Alexander Goebel, Alternative Fernwärme-Verlegesysteme, Mannheim (2010)[2] European Committee for Standardization,EN 13941 ANNEX D, Brussels (2009)[3] European Committee for Standardization,EN 15632, Brussels (2009)[4] Dipl.-Ing. Heinz-Werner Hoffmann, Dipl.-Ing.Torsten Göhler and Dr.-Ing. Manfred Klöpsch,Fernwärmeleitungsbau mit Recyclingmaterial, MVVForschungsbericht, Mannheim (2006)[5] Dipl.-Ing. Heinz-Werner Hoffmann and ZoltanDioszeghy-Günter, Neuartige Verlegetechniken fürdas Kunststoff-Verbundmantelrohr-System Band 1,MVV Forschungsbericht, Mannheim (1995)[6] Fratzscher et. al., Energiewirtschaft fürVerfahrenstechniker, VEB Deutscher Verlag fürGrundstoffindustrie, 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 derWärmetechnik, C. F. Müller Verlag, Heidelberg(2008), p. 26[12] Dipl.-Ing. (FH) Frank Espig, Schadensstatistik KMR2007 des AGFW, article published in the„EuroHeat&Power―(2008), issue 10304

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaANALYSIS OF HEAT TRANSFER IN HEAT EXCHANGERS BY USINGTHE NTU METHOD AND EMPIRICAL RELATIONSO. Gudmundsson, O. P. Palsson and H. Palsson 1Faculty of Industrial Engineering, Mechanical Engineering and ComputerScience Hjardarhagi 2-6, IS-107 Reykjavik, IcelandABSTRACTHeat exchangers are widely used in domestic andindustrial applications involving transfer of energy fromone fluid to another, for example in district heatingsystems. The wide usage underlines the importance tohave a good technique to detect if the effectiveness ofan heat exchanger is diminishing. There are number ofthings that can cause diminishing effectiveness of anheat exchanger, for example fouling, changes in fluidproperties as well as corrosion. In many cases thefouling is a particular problem, for example whengeothermal water is used. Geothermal water is verymineral rich which can cause serious fouling problems.The method presented in this paper is simple and easyto use and can be used to detect a diminishing heattransfer coefficient in many types of heat exchangers,in this paper the method is used on cross flow heatexchanger. The method uses measurements of theinlet and outlet temperatures as well as the mass flows,these measurements are usually easy to gather undernormal operation. The method uses the well knownNumber of Transfer Units (NTU) method as well asempirical relations to estimate the overall heat transfercoefficient, which is then statistically analyzed. Thedata used in this study was gathered from a simulatedcross-flow heat exchanger where the overall heattransfer coefficient was gradually decreased tosimulate diminishing effectiveness of the heatexchanger. The conclusion of this study shows that thederived detection method can detect fouling based onthe data from a simulated cross-flow heat exchanger,with a good accuracy and consistency. Further analysison real data is scheduled.INTRODUCTIONHeat exchangers are widely used in domestic andindustrial applications involving transfer of energy fromone fluid to another. General classification of heatexchangers are parallel flow, counter flow and crossflow. Their size and complexity can also vary greatly.Their operating conditions can be classified into twomain classes, steady operation where mass flow andtemperatures are relatively constant and dynamicoperation where mass flow and temperatures can varygreatly with time.During operation it is important to have someknowledge of the condition of the heat exchanger. Fora steady state condition it has proven to be relativelysimple, since analytical and empirical relations can bederived for different heat exchanger types and used forall necessary calculation regarding time invariantconditions, see e.g. [1]. If a dynamic operation exists itbecomes more complex to monitor the condition of theheat exchanger and more complex models are used,see e.g. [2] and [3].In this study, a mathematical model is used that hasbeen developed to simulate accurately the temperatureand flow transients in a cross flow heat exchanger. Themodel is based on the finite volume method (FVM)where a mathematical representation of a generalcross flow heat exchanger is solved numerically. Onepossible application of such a model is to generatedata that can be used to compare and tune moresimple dynamic models based on either black boxmethods or state space modelling. An importantapplication in this context involves methods to detectfouling in heat exchangers under dynamic operation.Description of the model can be seen in [4].Fouling in heat exchanger can be categorized in thefollowing categories, precipitation fouling, chemicalreaction fouling, corrosion fouling, particulate fouling,biological fouling and freezing fouling. Usually fouling isa combination of the categories. The fouling process inheat exchanger can be described as a process wherethe separating metal inside the heat exchangeraccumulates deposits from the fluids. This is verycommon and poses problems and results in reducedefficiency of the heat exchangers. There are numerousmethods available to address the effect of fouling, see[5–8]. Finally, decrease in the thermal efficiency of aheat exchanger due to property changes in a workingfluid will have similar effect on the heat exchanger asfouling.There are number of ways to detect fouling butaccording to [9], classical methods involvea) examination of the heat transfer coefficient,b) simultaneous observations of pressure drops andmass flow rates, c) temperature measurements,d) ultrasonic or electrical measurements ande) weighing of the heat exchanger plates. Methodsa–c) require the heat exchanger to be operating insteady state condition, d) can only monitor local foulingand e) requires the process to be stopped. Theserestrictions can be too strict or costly. Another305

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaapproach is to model the heat exchanger and look fordiscrepancy between model predictions and what isactually measured, see [10] and [4]. The method usedin this study falls into category a). To make the methodvalid for dynamic operating conditions, empiricalrelations for the mass flow rates are furthermore used.Although district heating systems usually operate inrelatively steady state it can be argued that methodsthat work well to detect diminishing efficiency underdynamic operation should work very well under steadystate condition.DATA USEDThe data used in this study was the same data as wasused in [4]. The data was generated by a simulatorrepresenting an unmixed cross flow heat exchanger.The advantage of using simulated data is that it ispossible to control when and how much fouling willoccur in addition to controlling the inlet temperaturesand the mass flows. The data used had temperaturesfor the hot side in the interval [53, 67] °C and the coldside [12, 27] °C, the mass flow rates for the hot andcold side were in the interval [0.30, 1.45] kg/s.Description of the simulator can be found in [4].FoulingDuring design a heat exchanger is commonly designedto operate under mild fouling by assuming a foulingfactor in the interval 0.0001 to 0.0007. According to [11]and [12] there is usually an induction time before anoticeable amount of fouling has accumulated. In [13] itis shown that the fouling will grow with increased rateduring the fouling period. Figure 1 shows the evolutionof the fouling factor from the time the heat exchangerstarts to accumulate fouling until the simulation isstopped. A dimensionless time is used to make easycomparison between different lengths of data series.allowed to progress to a maximum of R f =0.00033,which corresponds to 25% decrease in the overall heattransfer coefficient.THE DETECTION METHODThe fouling detection is done by estimating the overallheat transfer coefficient, U, by using NTU relations andmonitor the means of U for shift that can be related todiminishing efficiency either because of accumulationof fouling or property changes of the working fluid.NTU method is commonly known and a description of itcan be seen in [1].It is known that effectiveness of a heat exchanger canbe calculated byThe minimum fluid is the fluid that has the minimumvalue of the production of mass flow and specific heat,. Effectiveness for a unmixed cross flow heatexchanger can also be calculated by the followingrelations of the effectiveness to NTU.In normal use, the overall heat transfer is usuallyunknown and it is therefore not possible to calculateNTU directly. It is therefore necessary to estimate NTUfrom the relation between NTU and the effectiveness.The estimation is done by minimizing a score functionwith respect to NTU. The minimizationwas done by using the minimization routine fmincon inMatlab, see [14].The parameter NTU is defined by(1)(2)(3)From Eq. (3) it is easy to derive the formula for U(4)EMPIRICAL RELATIONSFigure 1. Evolution of the fouling factor from the timeThe simulated data sets used in this study include 200sets without fouling and 200 sets with fouling, the datasets are further divided equally between slow and fastfouling. In the fouled cases the data set was withoutfouling for the first 25% and then the fouling factor wasIn the case of heat exchanger under dynamic operationwhere big variations can occur during operation, it ishard to see shift in the overall heat transfer coefficientthat can be related to diminishing efficiency in the heatexchanger. In [15] it is proposed to use empiricalrelations of U to make a heat exchanger model validover a wide range of operating conditions. The heattransfer coefficient can be written as306

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaWhere i is constant, ii is temperature dependent and iiiis mass flow dependent. In this study only mass flowdependency was considered, since it has been shownby previously by [15] that the temperature dependencycan be neglected. The relation for the heat transfercoefficient can therefore be written as(5)The effect of the empirical relations can be seen inFigure 2. It can clearly be seen that including theempirical relations really helps to reduce the variationsin the overall heat transfer coefficient.By assuming that Eq. (6) applies to both the hot andthe cold side and neglecting the thermal resistance inthe separating metal, the overall heat transfercoefficient, U, can be written as(6)where y is the exponent of the Reynolds number. In [1]it is recommended to use y=0.8 for turbulent flow,which is expected in a heat exchanger.It can be practical to normalize U with a referencemass flow.The overall heat transfer coefficient according to thereference mass flow and is similarlyAfter inserting Eq. (7) and (8) into Eq. (4) to make itmass flow dependent and normalizing, the estimatedoverall heat transfer coefficient will become(8)(7)Figure 2. The figure shows the evolution of the number oftransfer units and the overall heat transfer coefficient withand without the empirical relations.To detect fouling a CuSum chart is used, see [16]. TheCuSum chart was chosen since it is known to beeffective to detect shift in mean values. When usingCuSum charts it is necessary to define two CuSumparameters, a decision limit to prevent false detectionand a reference value for deviations. Detection is madewhen the cumulative sum of deviations goes over thedecision limit.It can be seen in Figure 3 that the method is veryconsistent in detecting diminishing efficiency. Figure 4shows the detection if no empirical relations are used.The overall heat transfer coefficient in Eq. (9) is thevariable that is used to detect the fouling in the heatexchanger.(9)RESULTSAs mentioned above the method was applied to thesame data set as was used in [4].Measurement errors were added to the inlet and outlettemperatures as well as the mass flows to make themeasurements more realistic. Measurement errors of0.2 °C were assumed on the temperatures and 1–2%measurement errors to the mass flows.Figure 3. The CuSum chart quickly detects the shift in theoverall heat transfer coefficient.307

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniamethod seems also to be very stable in detecting thefouling.DISCUSSION AND CONCLUSIONThe results indicate that the method proposed can beused to detect fouling in cross flow heat exchangersoperating under dynamic condition by usingmeasurements that can be obtained under normaloperation. The detection method is based on the wellknown method of Number of Transfer Units, withaddition of empirical relations to make the method validover wide range of mass flow rates.Figure 4. The CuSum chart quickly detects the shift in theoverall heat transfer coefficient.Comparison of Figures 3 and 4 shows that it is possibleto detect fouling in heat exchangers operating indynamic condition with quite good accuracy by usingthe NTU method and empirical relations.In Table 1 a comparison between the method in [4] andthe method presented in this paper is shown. From thetable it is apparent that the method presented in thispaper gives better results. The fouling detection intervalfor the drift corresponds to fouling factors on theintervals [0.00002, 0.00004] and [0.00001, 0.00003]respectively for the fast and slow fouling. The methodis therefore giving considerable better results than themethod described in [4].Table 1: Comparison of detection time between the twomethods, where method 1 is from [4]PercentilesMethod 1 Method 2Fast2.5% 0.59 0.2650% 0.83 0.3597.5% 0.98 0.40PercentilesSlow2.5% 0.63 0.2350% 0.81 0.3097.5% 0.93 0.35Typical fouling factors are, as stated above, on theinterval [0.0001, 0.0007]. The results therefore indicatethat the method can be used to detect fouling in crossflow heat exchangers that are operating in non-steadystate condition prior to the time a typical fouling factorheat exchangers are designed for is reached. TheBy monitoring the calculated overall heat transfercoefficient, it is possible to detect changes that are dueto fouling or changes in the working fluid. Unlikeconventional methods, this method can detect foulingin heat exchangers that are not operated in steadystate conditions. The fouling detection is performedwithin the designed fouling factor interval.Further work will include application of the method ondata from a real heat exchanger.ACKNOWLEDGEMENTThis work has been supported by the Environmentaland Energy Research Fund of Orkuveita Reykjavíkur,National Energy Fund and Energy Research Fund ofLandsvirkjun.REFERENCES[1] J. P. Holman Heat Transfer. Ninth edition, McGrawHill, 2002.[2] M. Mishra, P. K. Das and S. Sarangi. "Effect oftemperature and flow non-uniformity on transientbehaviour of crossflow heat exchanger".<strong>International</strong> Journal of Heat and Mass Transfer,2008, p. 2583-2592.[3] H. Kou and P. Yuan. "Thermal performance ofcrossflow heat exchanger with nonuniform inlettemperatures". <strong>International</strong> Communications inHeat and Mass Transfer, 1997; 51(9-10):357-370.[4] O. Gudmundsson, H. Palsson and O. P. Palsson."Simulation of fouling in cross-flow heat exchangerand a fouling detection based on physicalmodeling". In: Proceeding of The 50th Conferenceon Simulation and Modelling, Fredericia, Denmark,7-8th of October, 2009.[5] W. L. Pope, H. S. Pines, R. L. Fulton and P. A.Doyle. "Heat exchanger design "why guess afouling factor when it can be optimized?". EnergyTechnology Conference and Exhibition. Huston,Texas, 1978.308

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia[6] A. Nejim, C. Jeynes, Q. Zhao and H. Müller-Steinhagen. "Ion implantation of stainless steelheater alloys for anti-fouling applications". In:Proceedings of the <strong>International</strong> Conference on IonImplantation Technology, 1999;2:869-872.[7] P. K. Nema and A. K. Datta. "A computer basedsolution to check the drop in milk outlettemperature due to fouling in a tubular heatexchanger". Journal of Food Engineering.2005;71:133-142.[8] S. Sanaye and B. Niroomand. "Simulation of heatexchanger network (HEN) and planning theoptimum cleaning schedule". Energy Conversionand Management. 2007; 48:1450-1461.[9] G. R. Jonsson, S. Lalot, O. P. Palsson and B.Desmet. "Use of extended Kalman filtering indetecting fouling in heat exchangers". <strong>International</strong>Journal of Heat and Mass Transfer, July,2007;50(13-14):2643-2655.[10] O. Gudmundsson, O. P. Palsson, H. Palsson andS. Lalot. "Fouling detection in a cross flow heatexchanger based on physical modeling". In:Proceeding of Heat Exchanger Fouling andCleaning, Schladming, Austria, 14-19th of June,2009.[11] B. Bansal and X. D. Chen. "Fouling of heatexchangers by dairy fluids – a review". In:Proceeding of Heat Exchanger Fouling andCleaning – Challenges and Opportunities, KlosterIrsee, Germany, June 5-10, 2005.[12] F. Fahiminia, A. P. Watkinson and N. Epstein."Calcium sulfate scaling delay times under sensibleheating conditions". In: Proceeding of HeatExchanger Fouling and Cleaning – Challenges andOpportunities, Kloster Irsee, Germany, June 5-10,2005.[13] M. W. Bohnet. "Crystallization fouling on heattransfer surfaces – 25 Years research inBraunschweig". In: Proceeding of Heat ExchangerFouling and Cleaning - Challenges andOpportunities, Kloster Irsee, Germany, 5-10th ofJune, 2005.[14] MathWorks http://www.mathworks.com/. 20th ofApril 2010.[15] G. R. Jonsson and O. P. Palsson. "Use ofempirical relations in the parameters of heatexchangermodels". Industrial and EngineeringChemistry Research, June, 1991;30(6):1193-1199.[16] NIST/SEMATECH e-Handbook of StatisticalMethods, April 30, 2009,http://www.itl.nist.gov/div898/handbook/.309

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaHEAT LOSS ANALYSIS AND OPTIMIZATION OF A FLEXIBLE PIPING SYSTEMJ. Korsman 1 , I.M. Smits 1 and E.J.H.M. van der Ven 21Liandon B.V.2 Thermaflex <strong>International</strong> Holding B.V.ABSTRACTThe object of this paper is to evaluate heat losses of aflexible PB-PE-PE piping system in the field, comparedto a conventional rigid Steel-PUR-PE piping system.The flexible system is optimized in both insulationquantity (thickness) and quality.The heat loss for pairs of pipes in the field, with 70 °Csupply and 40 °C return temperature, is based on heatloss measurements in the laboratory and has beenevaluated using the multipole method.Since the hydraulic properties of Polybutylene andsteel medium pipes differ, hydraulic calculations of ademonstration distribution network, fitted with eithersystem, are made.Total system heat losses for this demonstrationnetwork are calculated by summing the product of theheat loss per pair of pipe and the amount of pipe used.INTRODUCTIONFlexible piping systems for district heating and coolinghave several advantages when compared to rigidpiping, mainly during installation, some even in use.Flexible pipe can be utilized much like cable, arrives onlarge reels, requires less engineering and fewer hasjoints. However, flexibility comes at a price. It seemsharder to reach comparable levels of insulation, seeSmits et al. 2010 [1].The reason for this lies in the specific properties of thematerial most commonly used for insulation:Polyurethane foam. PUR foam has a crystallinestructure and tends to be quite rigid. It is not verysuitable for flexible applications. Bending may lead to abreakdown of the crystalline structure and may alsocompromise the bonding between foam and mediumpipe, thus creating a channel. This channel mayaccelerate the exchange of foaming agent and air withthe environment, thereby speeding up the ageingprocess. Flexible variants of PUR are available, but donot seem quite as good. Insulation foam made ofpolyolefins show ample flexibility and quite goodinsulation properties for small diameters. Furthermore,aging typically is a faster process than in rigid systems.As for the medium pipe, metals may be flexible enoughfor the smaller diameters, but are too rigid for thebigger pipes. Again, using polyolefins like Polyethylene(PE), cross linked Polyethylene (PE-X), Polypropylene310(PP) or Polybutylene (PB) improves flexibility. Fromthese, PE does not have adequate strength at highertemperatures and PP is rather stiff. This leaves PE-Xand PB, of which the latter can be welded withoutdifficulty. This is therefore the material of choice for thisstudy. In accordance with the temperature durationprofile mentioned in the BRL5609/EN15632, PB issuitable up to a maximum temperature of 95 °C.As with other plastics, PB is prone to some diffusion ofoxygen and water vapor. These effects have beeninvestigated by Korsman et al. 2008 [7]. To preventoxygen diffusion, an EVOH oxygen diffusion barriermay be used. Unless fully submerged for years, thediffusion of water vapor will not be much of a problem.When kept completely under water, it will take at least30 years for all cells to fill with condensate.It is not all that easy to compare the heat loss of aconventional Steel–PUR–PE piping system to a flexiblePB–PE–PE piping system. Internal diameters differ, asdoes the friction coefficient, because PB is smootherthan steel. A pipe for pipe comparison yields skewedresults. One way around this problem is to comparecomplete distribution systems, as was done inKorsman et al. 2008 [2]. A demonstration (or reference)network is used to design and compare similarnetworks. For reference purposes, the same network isused in this study.The differing properties that complicate comparisonbetween piping systems, can also be used to minimizedistribution system heat loss. The object of this study isto reach comparable heat loss for the flexible system,by exploiting specific properties, whilst transporting thesame amount of heat with comparable pressure losses.1. HEAT LOSS IN THE GROUNDHeat losses have been measured on test rigs asdescribed by van Wijnkoop et al. 2010 [3] and havebeen evaluated by van der Ven et al. 2010 [4].With the results of these tests, the in-ground heatlosses are calculated using the multipole method byJohan Cleasson and Camilla Persson in 2005 [5]. Notethat the mentioned heat losses are calculated for a pairof pipes, run at 70 °C supply and 40 °C returntemperature.For rigid piping, some room between the pipes isrequired for welding, see Fig. 1a.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThe installation of a supply and return closely togetherin the ground has a small, but positive effect on theheat loss of the pair, as is shown in Fig. 3.PB std insulation thicknessPB closeFig. 1a, Supply and return pipeFor flexible systems, there is no space requirementbetween the pipes for installation purposes, and pipesare best installed right next to each other, see Fig. 1b.Heat Loss [W/m]5045403530252015105016 20 25 32 40 50 63 75 90 110Nominal diameter [mm]Fig. 3, Heat loss per pair, standard and close togetherFor the rest of this paper, flexible pipes are supposedto be installed closely together.3. PIPE PER PIPE COMPARISONFig. 1b, Supply and return closely togetherWhen supply and return are installed closely togetherthe temperature profile in the ground is altered inbenefit of the in-ground heat loss.Even though internal diameters and friction coefficientsare different between Steel-PUR-PE and PB-PE-PEand therefore will lead to a different selection ofdiameters in the engineering process, an approximatecomparison can be made, see Fig. 4.Dependant on the refill, it may be difficult to achievedefined compaction when supply and return areinstalled too closely together. However, similartemperature effects can be reached by installinglikewise in vertical orientation.As mentioned before, the in-ground heat losses arecalculated using the multipole method by JohanCleasson and Camilla Persson in 2005 [5].See Fig. 2 for a calculated temperature profile.Heat Loss [W/m]50454035302520151050St std PB close16 20 25 32 40 50 63 75 90 110Nominal diameter [mm]Fig. 2, Temperature profile supply and return closelytogether in the ground311Fig. 4, Heat loss per pair, flexible PB versus SteelOn the left of the graph, two diameters are included forwhich no steel counterpart has been incorporated. Thereason for this is that PB allows for higher fluidvelocities. However, in the current range, using thesmaller diameters does not create a heat lossreduction.

4. SELECTION OF INSULATION THICKNESSThe reason for the relatively high heat losses for thetwo smallest diameters is explained by Fig. 5.Thickness insulation [mm]PB std insulation thickness454035302520151050PB extended insulation thickness16 20 25 32 40 50 63 75 90 110Nominal diameter [mm]Fig. 5, Insulation thickness, standard and increasedIn red, this graph contains the insulation thickness ofthe current PB-PE-PE product range. The somewhaterratic distribution of insulation thickness over therange is caused by the use of customary dimensionsfor the outer casing. It can be seen that for the twosmallest diameters the insulation is rather thin, whichexplains the relatively high heat losses.In purple, the graph in Fig. 5 shows a modified range,with increased insulation thickness for some of thesmaller diameters, as in general it is easier to achieve agood insulation quality for the smaller dimensions. Theheat losses of the modified range were calculated andare presented in Fig. 6.Heat Loss [W/m]50454035302520151050St stdPB std insulation thicknessPB extended insulation thickness16 20 25 32 40 50 63 75 90 110Nominal diameter [mm]Fig. 6, Heat loss per pair, including increased insulationthicknessThe <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia312The graph in Fig. 6 shows that the modified range (inblue) is much closer to the reference (Steel-PUR-PE, inpurple). Heat loss is (almost) proportional to thediameter, which seems about right. However, themodified range of insulation thickness in Fig. 5 (purple)still shows a somewhat erratic distribution, whichsuggests that the current range of customarydimensions for the outer casing does not lead to anoptimal distribution of insulation thickness. It may proveworthwhile to develop a new range of outerdimensions, adapted to the diameter of the mediumpipe.5. IMPROVEMENT OF INSULATION QUALITYDue to the new testing facilities described by vanWijnkoop et al. in [3], the process of productimprovement has been speeded up considerably.During the course of the investigations, resulting in thispaper, it is becoming clear that further improvement ofthe insulation quality is feasible. The measurementprinciple used for the determination of heat losses doesnot allow for direct measurement of the insulationproperties of the foam; however, some sort of―equivalent lambda‖ can be derived from the data bycalculation. As explained by van der Ven et al. in [4],insulation quality differs for different diameters. Forproduction reasons, it is not expected that insulationquality will reach the same level over the entire productrange. Typically, the higher values will be reached inthe smaller dimensions. Still, an educated guess canbe made as to which levels are feasible from atechnical viewpoint, see Table 1.Table 1, Improved insulation quality, ―equivalent‖ or―synthetic‖ lambda at 50 °C mean temperatureType Area Lambda fresh Lambda Degassed50A25 1074 0.0283 0.032663A32 1701 0.0287 0.033075A40 2364 0.0291 0.033590A50 2993 0.0295 0.034090A40 3670 0.0300 0.034590A32 3886 0.0301 0.0346125A63 6204 0.0316 0.0363160A90 10790 0.0345 0.0397160A75 12611 0.0357 0.0411200A110 16879 0.0385 0.0442Please note: The lambda values in Table 1 are not themeasured lambdas of samples of the insulation foam,and may not be interpreted as actual physicalproperties of the insulation material. The values werecalculated on the basis of heat loss measurements ofsections of pipe according to EN15632, and thereforeare some sort of ―synthetic system lambdas‖.

The values presented do of course largely depend onthe actual physical lambdas (in W/m.K) of the insulationmaterial, but the underlying measurement data suggestthat other factors come into play as well, such as thegeometry of foam in combination with the temperaturedependence of the ―physical‖ lambda of the foam.Therefore, the values in Table 1 are valid only forcalculation / prediction purposes, in exactly the samecalculation model from which they were derived (alsoaccording to EN15632). The values in Table 1 aresupported by experimental data on four samples at thetime of writing this paper.When all parameters are known, equation 1 can beused to calculate heat loss: i Where:1ln sd 2d 1 2 T probe T casing1 ln id 3d 21 ln cT probe, T casing represent probe (medium) and casingtemperatured 1 to d 4 represent inner/outer diameters of servicepipe and casingλ s , λ i , λ c = heat coefficient of service pipe,insulation and casingIn this case, λ s and λ c are known: λ s = 0.19 W/m.Kand λ c = 0.40 W/m.K. On a test rig, T probe , T casing andheat loss are measured, so for specific test samples,eq. 1 can be used backwards to calculate ―synthetic‖values for λ i in Table 1.For the Steel-PUR-PE reference, see [1], values canbe determined in a similar fashion. ―Synthetic‖ λ ivalues for PUR foam, determined frommeasurement of samples, were typically in the rangeof 0.030 to 0.032 W/m.K, with λ s for steel 50 W/m.KHeat Loss [W/m]50454035302520151050St stdPB std insulation thicknessPB extended insulation thicknessPB impr. freshd 4d 316 20 25 32 40 50 63 75 90 110Nominal diameter [mm]Fig. 7, Heat loss per pair, including improved insulationquality(1)The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia313The red graph in Fig. 7 represents the predicted heatloss values for the combined effect of both increasedinsulation thickness and insulation qualityimprovement. For most diameters, these are on parwith or slightly better than the reference in Steel-PUR-PE. These data are valid only for the recently producedor ―fresh‖ product. As there is no experimental dataavailable on the rate of degassing and therefore therate of ageing, it is difficult to predict heat loss over thelife time of the product.However, it is possible to speed up the process ofageing artificially, until all the foaming agent has beenreplaced by air. The predicted values for this conditionare also presented in Table 1, as lambda degassed.These are ―synthetic‖ as well, and suitable forcalculation purposes only. Calculated heat loss resultswith these values are presented in Fig. 8.Pump I, power: 16 kWHeat Loss [W/m]454035302520151050St std PB impr fresh PB impr degassed84 %14 %2 %16 20 25 32 40 50 63 75 90 110Nominal diameter [mm]Fig. 8, Heat loss per pair, including improved insulationquality, fully degassedIn 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 fullydegassed PB-PE-PE. The values vary a bit, but aregenerally in the same range. During the lifetime of theproduct, heat loss is expected to increase from the redvalues to the green values.Of course, ageing is also applicable to the referenceproduct, but not included here for two reasons. First,the ageing process for rigid systems is expected to besignificantly slower than for flexible systems, andsecond, the reference samples were not fresh, as couldbe judged by the gas content. Therefore, it is not likelythat the values presented for the reference system willdeteriorate much further during lifetime.Ageing can be slowed down considerably if measuresare taken to prevent the exchange of blowing agentwith the environment. If successful, these measures

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniawould result in the ―red‖ heat loss values duringlifetime. Moreover, a new generation of blowing agentsis under development. These new agents aim at lowerconductivity values for the gas and larger molecules.This may result in lower conductivity values for theproduct as well as a slower ageing process.6. HYDRAULIC CALCULATIONSThe pipe per pipe comparison between Steel-PUR-PEand PB-PE-PE as demonstrated in Fig. 8, gives anindication of field results, but is not conclusive. Internaldiameters differ, as do friction coefficients. Therefore,for the comparison between distribution systems fittedwith either pipe, hydraulic calculations are needed. Tothis end, a reference network is introduced in Korsmanet al. 2008 [2]. The same network is used here. It isinstalled in a housing estate near Arnhem, theNetherlands, and has been designed using Pipelab,developed by Prof. Dr. Pàll Valdimarsson in 1995 [6].See www.pipelab.nl. Standard design criteria wereused. A total of 247 houses are connected by 3.02 kmof DH network (6.05 km of pipe), 12.2 m per house.The graph in Fig. 10 represents the pressure in thesupply network (in m water column), as a function ofthe distance from the source. For standard symmetricalnetworks, the return network is similar, but mirroredover a horizontal axis.Using the flexible and smooth PB pipes allows forsmaller diameters, mainly because PB is less prone tothe transmission of hydraulic noises. This is due to thelow modulus of elasticity of PB when compared tosteel. In contrast, a steel pipe filled with water is quite agood conductor of sound. To prevent noise caused byhigh flow velocities, these are limited in the design forsteel networks to 1 m/s.A network, specifically designed for PB, is shown inFig. 11. Smaller diameters in the periphery of thenetwork as a result of a higher permitted fluid velocitycauses higher pressure drops. This has to becompensated by bigger pipes closer to the source toreach the same overall pressure drop.Fig. 9, Aerial photograph of reference housing estateFig. 11, Design pressure drop PB networkFig. 10 shows an output graph of Pipelab.Fig. 10, Design pressure drop steel networkIn the design of district heating networks, the maximumdesign point is chosen considerably below the sum ofthe installed power in the connected buildings. It is notuncommon to have a design point of 50% of the totalinstalled power for larger numbers of connections,depending on the experience and the courage of thedesigner. A design point of 50% of the total installedpower was used in both designs in this paper. Inpractice, no problems have arisen with this designpoint, partly because not all installed power is used atthe same time. However, this statistical effect does notapply to individual connections. Therefore, a designtrick is used in the periphery of the network, to preventproblems in the service pipes connecting the buildings.The flow in these pipes is raised artificially above thedesign point, up to 100% load. The result of thiscalculation is shown in Fig. 12, which can be comparedto Fig. 10.314

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia1200(1) ST.PUR.PE Ref (2) PB as measured std eng.1000Pipe Length [m]800600400200Fig. 12, Pressure drop steel network with increasedservice pipe flowThere is not a lot of difference between both graphs inFig. 10 and Fig. 12. The reason for this is that thedesign maximum fluid velocity is rather low for steel.This may prove different for the PB network, which isdesigned with smaller diameters in the periphery. SeeFig. 13, which can be compared to Fig. 11.The graph in Fig. 13 indeed shows an increasedpressure drop in the service pipes connecting thehouses, when the flow in those pipes is artificiallyincreased to 100% of the installed power. However, thetotal pressure drop stays within the same limits as doesthe steel network under similar conditions (Fig. 12).016 20 25 32 40 50 63 75 90 110Nominal diameter [mm]Fig. 14, Pipe length histogram Steel and PBAs a result of the use of smaller diameters with PB, thedistribution of pipe lengths generally shifts to the left inthe pipe histogram. As heat loss increases withdiameter (see Fig. 8) this should have a positive effecton the total distribution system heat loss.This shift to the left may be taken one step further,since the wall thickness of the smallest PB mediumpipes currently is chosen a bit larger than the strengthclass (SDR11) requires. This is done for ease ofinstallation. If the thickness of these pipes is chosen nolarger than SDR11, there is a slight additional shift tothe left, see Fig. 15.(1) ST.PUR.PE Ref (2) PB as measured std eng.(3) PB impr. Fresh all SDR1112001000Pipe Length [m]800600400200Fig. 13, Pressure drop PB network with increased servicepipe flowThe result of both design calculations is plotted inFig. 14, steel in red and PB in green.016 20 25 32 40 50 63 75 90 110Nominal diameter [mm]Fig. 15, Pipe length histogram steel, PB and PB SDR11networks315

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia6. TOTAL SYSTEM HEAT LOSSTo calculate the total system heat loss, pipe lengths asshown in the pipe length histograms are to bemultiplied by the respective heat losses per pipe pair,as shown in the heat loss value histograms.Fig. 15 ―multiplied‖ by Fig. 6 leads to the total systemheat losses in Fig. 16.Heat Loss [kW]80.070.060.050.040.030.020.010.00.0ST.PUR.PE RefPB as measured std eng.PB increased insulation thicknessPB increased insulation thickness all SDR11system [-]Fig. 16, Total system heat loss, current insulation qualityThe graphs in Fig. 16 show the reference heat loss forsteel-PUR-PE in red and the currently measured heatloss for PB-PE-PE with non-optimized insulationthickness in orange. In green, total system heat loss isshown for current insulation quality but with optimizedinsulation thickness. The exclusive use of SDR11(blue) has a rather small effect.Improving insulation quality, as described in paragraph5 and shown in Fig. 7, leads to slightly lower totalsystem heat loss for freshly produced PB-PE-PE whencompared to the reference, see Fig. 17.Heat Loss [kW]60.050.040.030.020.010.00.0ST.PUR.PE RefPB impr. insulation thickness/qualitysystem [-]There is currently no experimental data on the rate ofageing of PB-PE-PE as a result of the exchange ofblowing agent with air. However, it is possible tocalculate a worst case situation (see fig. 18), using thepredicted values plotted in fig. 8.Heat Loss [kW]70.060.050.040.030.020.010.00.0ST.PUR.PE RefPB impr. insulation thickness/qualityPB impr. Insulation thickness/quality degassedsystem [-]Fig. 18, Total system heat loss, including worst caseIn practice and over time, the predicted total systemheat loss will slowly shift from the fresh value in purpleto the worst case value in blue. Average heat lossduring lifetime will be somewhere in-between.FUTURE RESEARCHHydraulic calculations in combination with insulationthickness form an interesting optimization problem:what diameter to select and which insulation thicknessto choose?Current design strategies for hydraulic networks,aiming at linear pressure drop with distance, seem tooadventitious to be optimal. In addition, heat losscalculations using standard casing dimensions showrapidly diminishing yields with each step up ininsulation thickness, suggesting the optimum issomewhere in-between.First attempts have been made to use Pipelab [6] in adouble optimization routine, trying to find optimalhydraulic performance in combination with optimalinsulation thickness distribution over the network.Given the specific hydraulic properties of PB (high fluidvelocities permitted) and the specific insulationproperties of PE foam (better at small size), this maylead to rather different design strategies whencompared to conventional rigid piping systems fordistrict heating and cooling.Fig. 17, Total system heat loss, reference and predictionfor improved insulation quality316

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFURTHER INFORMATIONQuestions concerning the paper can be addressed to:Liandon B.V.www.liandon.comwww.pipelab.nlDijkgraaf 46920 AB DuivenThe NetherlandsThermaflex <strong>International</strong> Holding B.V.www.thermaflex.comVeerweg 15145 NS WaalwijkThe NetherlandsCONCLUSIONFlexible Polybutylene piping, insulated withPolyethylene foam is a recent development, leavingample room for product improvement. Experimentaldata shows rapid improvement in heat lossperformance.By optimizing both quantity (thickness) and quality offlexible Polyethylene foam, and by using the specifichydraulic properties of Polybutylene piping, the heatloss performance of conventional rigid steel pipingsystems insulated with polyurethane foam, is withinreach.If the current rate of improvement of the PB-PE-PEflexible piping is maintained, total distribution systemheat losses will be comparable to conventional rigidSteel-PUR-PE piping. The evident benefits of flexibilitywould become available without the current heat losspenalty.ACKNOWLEDGEMENTSpecial thanks are due to Ivo Smits who did all thecalculations for this paper and made all the graphs. Asmall change in approach may result in a lot ofrecalculation. Thanks also to Camilla Persson forsupplying us with a MathCad implementation of themultipole method back in 2008 and Pàll Valdimarssonfor the invention of Pipelab, which set me off on thisroad at the symposium in Helsinki, 1995.REFERENCES[1] I.M. Smits, J.T. van Wijnkoop, E.J.H.M. van derVen, ―Comparison of competitive (semi) flexiblepiping systems by means of heat lossmeasurement‖, in Proc. of the <strong>12th</strong> <strong>International</strong><strong>Symposium</strong> on District Heating and Cooling,Tallinn, Estonia (2010).[2] J. Korsman, I.M. Smits, S. de Boer, ―Systemoptimization of a new plastic piping system‖, inProc. of the 11th <strong>International</strong> <strong>Symposium</strong> onDistrict Heating and Cooling, Reykjavik, Iceland(2008).[3] J.T. van Wijnkoop, E.J.H.M. van der Ven,―Verification of heat loss measurements‖, in Proc.of the <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on DistrictHeating and Cooling, Tallinn, Estonia (2010).[4] E.J.H.M. van der Ven, R.J. van Arendonk, ―Heatloss of flexible plastic pipe systems, analysis andoptimization‖, in Proc. of the <strong>12th</strong> <strong>International</strong><strong>Symposium</strong> on District Heating and Cooling,Tallinn, Estonia (2010).[5] J. Claesson, C. Persson, ―Steady-state thermalproblem of insulated pipes solved with themultipole method‖, Chalmers University ofTechnology, Report 2005:3. (2005)[6] P. Valdimarsson, "Graph-theoretical calculationmodel for simulation of water and energy flow indistrict heating systems", in Proc. of the 5th<strong>International</strong> <strong>Symposium</strong> on Automation of DistrictHeating Systems, Helsinki, Finland. (1995).[7] J. Korsman, S. de Boer, I.M. Smits, ―Cost benefitsand long term behaviour of a new all plastic pipingsystem‖, IEA DHC|CHP Annex VIII research report(2008)317

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaFREE OPTIMIZATION TOOLS FOR DISTRICT HEATING SYSTEMSStefan Gnüchtel 1 , Sebastian Groß 11Institute of Power Engineering, Technische Universität Dresden, 01062 Dresden, GermanyABSTRACTAt the Technische Universität Dresden, Institute ofPower Engineering, Chair of Power SystemsEngineering as part of the project ―LowEx Fernwärme– Multilevel District Heating‖ [1] supporting by theFederal Ministry of economy and technology (FKZ0327400B), two public available and cost free softwaretools have been developed, which enable the user tofindA) the optimal unit commitment of district heatinggenerators: FreeOptmodify the system configuration to check how thesystem reacts under new conditions and how theoperating costs will change. In all cases the tool givesvaluable information.So FreeOpt provides help for any municipality thatneed a first guess on the feasibility and operating costsof a new district heating network or who need toimprove the operation of an existing one. And of courseit does not matter if the network is supposed to beextended or build from scratch or an existing systemjust to be analysed.B) the optimal pipeline route with the optimal pipediameter of district heating networks: STEFaNat a minimum of costs. Both software tools are veryeasy and intuitive to handle. Lead time required tolearn how to operate the programs is short. Both toolsintend to support general design decision pro districtheating systems. In this paper an overview on theiradvantages and fields of application, as well asexample calculations are presented.PROGRAM FREEOPTFreeOpt is an optimization tool to find the optimal unitcommitment of district heating generators for a giventime domain. Optimal decisions are found to minimizetotal costs related to thermal and electric loads.Local district heat networks are becoming morecommon, so it is important to know how to operateeven small systems in terms of minimal costs andhighest efficiency. When should which generator beswitched on or off? How to handle the storage? Whichinfluences have contracts for electric power?There already is a lot of existing software for unitcommitment. As commercial and generalised softwarefor large systems it is mostly very expensive. SoFreeOpt has been developed for any cases of districtheating networks for most efficient operation of all heatand power generators. Mainly operators of smallersupply areas purchasing an expensive softwaresolution would be uneconomical can reach monetarysavings.A stable version of FreeOpt is already finished. Theprogram‘s power is demonstrated with a simpleexample determining cost optimal operation of aspecific district heating network. Furthermore with thehelp of parameters and figure lines it is very easy to318How does the software tool work?First the simulation time is divided into time steps. Forevery time step several variables (i.e. generated heatand electric power, amount of fuel, energyconsumption) exist inside given boundaries as well asindividual costs and proceeds are stated (i.e. for fuel,CHP refund). Different kinds of generators can bechosen like heat plants, combined heat and powerplants, solar thermal plants or heat pumps. Hot waterstorage tanks and electricity transferred in or out thegrid via contracts improve the flexibility of the wholesystem (Fig. 1 ).System boundarySQ fuelQ demBlock heatand powerplantPowerdemandCombinedheat andpowerplantP demHeatplant/boilerSolarthermalsystemHeatpumpHeatdemandHeatstorageFig. 1 System boundary and interaction planGridconnectionThe modular design of FreeOpt allows to form easilyany generator system. All one has to know are therespective figure lines and parameters for all availablegenerators and the network. After fixing the demand for

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniaheat and electric power as well as the parameters theoptimization problem is defined automatically. FreeOptcalculates the minimal costs to satisfy the demand. Fortesting purpose there is no limitation of the values.The mathematical model belongs to the mixed integerproblems. So the objective function and all constraintsare linear, all variables are continuous or discrete. It iswritten in the mathematical modelling language GMPL[2] and solved with the COIN-OR brunch-and-cut solverCBC [3]. Both GMPL and CBC are open-sourcesoftware under GNU GLP license [4]. An intuitive userinterface enables to enter all input data like variableboundaries, cost coefficients, starting values, demandvalues or figure lines in a very easy way. The internaldata flow between user interface, optimization modeland solver is realised with help of txt-files (Fig. 2).The most important ones are the two balanceequations for power (2) and heat (3).stands for the generated electric power,power transferred in or out the grid via contracts,for the electric power demand andconsumption of electric power.stands for the generated thermal power,the heat demand,cooler andheat storage.(2)forfor the own(3)forthe heat disposed in the auxiliaryfor the heat transferred in or out of theExampleAs an example the following heat network of a localenergy supply company is given in Fig. .Fig. 2 FreeOpt user interfaceElectricity networkRunning the generators cause costs. Therefore themain aim is to minimize the total costHeat boiler plantCHP 1CHP 2 CHP 3Heatstoragestorage(1)in whichproceeds):are the following operating costs (andcosts for fuelcosts or proceeds for transferredelectricitycosts for start-up procedurescosts for network accesscosts for maintenancecosts for CO 2 -cerfiticatesCHP-refundEEG-refundproceeds for avoiding network accesscosts for electricity taxpenal costs for balance violation(virtual costs)As already noted several variables exist for every timestep limited by some boundaries and connected byparameters in lots of equations and inequations.319Fig. 3 Flow scheme of a local heat networkThe heat demand of the customer is provided by3 CHPs (Block Heat & Power Plants) – baseloadHeat plant (natural gas) – peak loadHeat storage – used for optimizationand is given for every hour of one year.The electricity demand is not directly consideredbecause of intern clearings inside the energy supplycompany. The whole generated power is transferred inthe grid and refunded as well as used to satisfied theown consumption. It is also possible to transferelectrical power out the grid when all CHPs areswitched off.

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaTab.1 gives an overview on the design parameters of allheat and power generation units as well as on the costassumptions for fuel and the refund of the energy tax.Tab. 1 Design parameters of generation unitsGenerationunitCHP 1 CHP 2 CHP 3 HeatPlant/ kW 911 774 911 -/ kW 911 774 911 -/ kW 1200 1020 1200 200/ kW 1200 1020 1200 2000/ % 0.35OwnconsumptionCosts fuel /€/kWhRefundenergy tax /€/kWhStart upcost /€/start up( )0.35( )0.35( )0.87( )0.76 0.76 0.76 -2.0 %(of )2.0 %(of )2.0 %(of )0.04444 0.04444 0.044440.0055 0.0055 0.00550.5 %One specific characteristic of the energy system is thatall three CHP cannot be operated in part load, so thevalue of the maximum and the minimum power have tobe equal.The base own consumption of the whole energysystem is 50 kW. In the 100 m 3 heat storage it ispossible to store 4000 kWh. Tab.2 gives an overview ofall intern electricity contracts.(of)0.045740.002212 12 12 6Analysis operation modeFirst the real operation mode of whole year 2008 isanalysed retrospectively (case I) and the optimaloperation mode is determined with the help of FreeOpt(case II). Following the annual operation costs arecalculated for both cases (Tab. 3).The calculations in Tab.3 show that the costs for fuelincrease but so the proceeds through electricity saleand CHP-refund increase too.Tab. 3 Comparison annual operating costs and proceeds(case I and case II)Annual operationcosts andproceeds / 10 3 €Real operationmode(case I)Optimaloperationmode(case II)Costs fuel 1326.12 1331.13Proceedselectricitycontracts468.48 473.65CHP-refund 58.17 58.67Start up costs 10.01 6.62Costs networkaccess2.83 2.47Total costs 812.30 807.90Furthermore costs for start ups and network accesscan be reduced. But altogether the total annualoperating costs for case I and case II are nearly equal.In this particular example savings are under 1% of theoperation costs. So the energy system is operatednearly in an optimal way.System configuration changesThe energy supply company considers to replace theCHP 2 through a new one. Of course the new CHP hasnew parameters (Tab. 4).Tab. 4 System configuration changes CHP 2Tab. 2 Electricity contractsCosts /€/kWhElectricity purchase6am – 22pm22pm – 6 amElectricitysale0.08150 0.05700 0.04735The CHP-refund is 0.0056 €/kWh and costs for networkaccess are 0.0386 €/kWh.320Generation unit Old CHP New CHP/ kW 774 404/ kW 774 404/ kW 1020 535/ kW 1020 535/ % 0.35 0.36Start up cost /€/start up12 9.5

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaAll other parameters are the same like the old CHP 2.Main difference is the smaller range of performance.The remaining energy system left unchanged exceptfor the CHP-refund expiring the next year.The optimal operation mode considering the old(case III) and the new (case IV) parameters of theCHP 2 are determined. Case III serves as a referencecase. The calculated operation costs per week areshown in Fig. 4 for both cases.Total cost / 10 3 €35302520151050Operating costs per weekJ F M A M J A S O N DMonthOld CHP 2 (case III) New CHP 2 (case IV)Fig. 4 Operation costs per week (case III and case IV)In Fig. 4 it can be seen that the operation costs forcase IV are below the operation costs for case III inevery week, especial in the summer months becausethe smaller size of the new CHP 2 suits better to theheat demand. Altogether total operating costsamounting to about 27770 € (3.2 % of total costs) canbe saved (Tab.5). Main reasons are the huge fuelsavings which settle easily the decreasing proceedsthrough electricity sales. By given investment cost it isvery simple to check if the renewal of the old CHP 2 iseconomic reasonable.Tab. 5 Comparison annual operating costs and proceeds(case III and case IV)Annual operationcosts andproceeds / 10 3 €Old CHP 2(case III)New CHP 2(case IV)Costs fuel 1323.87 1208.79Proceedselectricitycontracts462.87 367.12Start up costs 7.72 1.20Costs networkaccess1.92 0.00Total costs 870.64 842.87As example the heat balance curve on one summerweek (168 hourly time steps) is shown in Fig. 5. Such acurve helps to determine the recommended operationmode.The red line marks the given heat demand which issatisfied at all time steps. The small-sized CHP 2operates continuous. When the heat demand is higherthan the output of the CHP 2 the heat plant is switchedon or the heat storage is discharged mostly. Chargingthe heat storage takes place in low demand times andby switching on one of the others CHPs (CHP 1 orCHP 3) for up to four hours. Using the CHP is moreeconomic than using the heat plant but start up costsand the size of the heat storage restrict the operation ofa second CHP.heat flow / kW18001600140012001000800600400heatplantHeatbalanceheat demandcharge ofstorageby CHP200CHP productiondischarge of storage00 12 24 36 48 60 72 84 96 108 120 132 144 156 168time / hCHP production storage load heat plant storage unload heat demandFig. 5 heat balance curve on one summer weekFinally it can be summarized that first experiences andcalculations show, how the FreeOpt allows in an easyand quick way to check beforehand if certain systemconfigurations are useful or contra productive.PROGRAM STEFANApplication fieldThe network optimization is a special case of theresearch-main focus „optimization of the technicalstructure of district heating systems―of the 5th energyresearch program of the German Federal Government.Due to the relatively high net costs of district heatingsystems it is necessary (beside the application ofactual piping systems) to optimize the nets concerningtheir design parameters, in particular the pipe diameterand the pipe routing.Therefore the software tool ―STEFaN‖ has beencreated for the combined pipeline routes and diameteroptimization. This Windows program for the support ofthe application of the district heating has interfaces togeographical information systems (GIS) and iscomplementary with these. Its application is possible in3 planning phases:321

1600The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estonia1. In the conceptual design planning phase for thecost estimate and the principle decision for thedistrict heating (yes or no).2. In the detailed planning phase the localization ofthe pipeline route occurs for the approvalplanning.3. In the execution planning phase the finaldetermination of the dimension occurs, but nomechanical calculation (stress-strain analysis) isdone by the program. Still the required proofsaccording to e.g. EN 13941 have to be done.In addition, the program can be used for the hydrauliccalculation of existing district heating networks.ModelThe hydraulic calculations establish the technical basiswhich performs constraints of the optimization model.In district heating systems a distinctive turbulent flowcan be presumed. In this case a good approximationwith the surface roughnesscoefficient of friction is applied (4).of the pipe and theIn the mentioned planning phases the coefficients ofdrag are included blanket into the pressure lossaccording to (5):(4)(5)If the edge is not used for the site development ( ),then holds.As variables are required beside the diameter furthervariable than auxiliary variables to the formulation ofthe constraints: vector of the mass flows of the edges vector of the pressures of the verticesbinary variable to the capture of the jump atThus the constraints can be formulated. These are theequation (6), local and technical limitations as well asequations. (8) and (9).First Kirchhoff‘s law: Point rule. The sum of all massflows in a vertex is equal zero. ( - vertex matrix)Second Kirchhoff‘s law: Mesh rule. The sum of thepressure losses along a mesh is equal zero ( - meshmatrix):This mathematical model is simple to describe, butdifficult to solve (already for medium-sized graphs). Ifthe diameters are eliminated by the equation (6) as avariable, the variables and whose impact on theobjective function is discussed in detail in [5] remain.The principal dependency of the objective function onthe vector of the mass flows is displayed in Fig 6schematically.(8)(9)whereas a extra charge of length.For a pipe of the length and the diameterfor the pressure loss of a plain pipe.(6) arises(6)Thus the following mathematical optimization modelarises:The investment costs of every new route come into theobjective function (investment costs, annual costs ornet present value). They are included in the form of (7)in the model.The bracket of the first summand contains theinvestment costs of the route per meter as a total lumpsumprice and must be multiplied according to by thelength of pipeline. In the second summand"obstacles" can be included as direct costs dependentfrom the diameter. The exponent is set =1 in thepresent program version for linear dependence. Theparameters and are input data.(7)K konkavK BaumFig. 6 Schematic dependence of the objective function onthe mass flowOn the abscissa the circulatory mass flowof a meshis displayed. The ordinate shows the non-convexobjective function which shows jumps by the binaryvariables with the rhombuses (the filled rhombus is thefunction value).322

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, EstoniaThis shape leads to the fact that as a solution under therequirements mentioned above a graph free ofmeshes, thus a tree, arises.This complex course of the objective function requiresespecially suitable methods.Development of mathematical proceduresDifferent mathematical methods are used:a) The classical non-linear optimization which isapplied for the diameter optimization for fixeddevelopment ways.b) Topological optimizations to the determination ofthe shortest ways (shortest path problem) and theshortest networks (spanning tree problem) whichare combined under use of the procedure from a)to a special iteration process.c) Stochastic methods for the improvement of theoptimization results of the algorithm of b): A specialimplementing of the Monte Carlo Method and aspecial implementing of the Evolutionary Algorithm.Project processingThe help file and the user's manual of this programcontain detailed instructions to its operation and for theproject processing which occurs typically in six steps inthe change of STEFaN (steps 2, 3, 5 and 6 a) and of aGIS (steps 1, 4 and 6 b).These necessary steps 1 to 6 are demonstrated at afictive example. As a GIS system the recommendedand provided program ShapeUp (www.nilione.com) isused. It is freeware too.ExampleStep 1: Gathering of Geographical informationWith the GIS layers (themes) with geo-referencedinformation (vertices for the source and for thecustomer as well as edges for consisting and possibleroutes) invested and in a special standardized format(MIF – MapInfo Interchange format) exports:a) The geo-referenced background image is imported(© OpenStreetMap, pale colors in Fig. 7). Thefigures of the buildings (darkly) and the courses ofthe streets (white) allow a good orientation.Fig. 7 step 1323

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniab) Directly input (or import) of the heat source (redpentagon on the top left in the Fig. 7, through violetcircle marked) and the sinks (customer – yellowflags in the Fig. 7, as house service connectionstations in the house lines arranged) with the givenattribute for the heat demand.c) The input of possible routes (thick green lines inFig. 7) by using the mouse and the assignment ofthe attributes (table in the right section of Fig. 7),for quite available pipes with attribute(branch pipe from the source inDN 150 – input value:ground,channel).for the laying procedure () andfor street andfor cellar corridor and availableStep 2: Providing the non-geographical dataThe files with the non-geographical data (generalentries to the network as for example mediatemperatures as well as economic data) are entered onforms.Step 3: Generating the network topographyAfter the import of the files invested by the GIS thenetwork topography of the possible routes is created bythe program. Gaps between the inputted routes (step1c) are complemented by the program to a graph withentire and varied development.Step 4: Verification of the generated networktopographyWith this step the files generated by the program canbe imported in the GIS: The generated edges (thin bluelines in Fig. 8) and vertices (blue dots in fig. 8)complete the entered network topography and can bechecked.Step 5: Determination of the optimal developmentThe route optimization is carried out by the program.Step 6: Evaluationa) Output of a result report and export of theoptimization results to the GIS.Fig. 8 step 4324

The <strong>12th</strong> <strong>International</strong> <strong>Symposium</strong> on District Heating and Cooling,September 5 th to September 7 th , 2010, Tallinn, Estoniab) The local representation and if necessarytreatment of the results by the GIS: The result canbe visualized in the GIS (Fig. 9). The differentcolored lines in the left section show theascertained route planning, and in the right sectionof Fig. 9 a part of the data base with the siterelateddata is displayed: Length of pipeline in m,the diameter in mm, the mass flow in kg/sand the pressure differencein bar.Fig 9 step 6CONCLUSIONStable versions of FreeOpt and of STEFaN are alreadyfinished. The program‘s power was demonstrated withsimple examples. They can be downloaded from [6].FreeOpt calculates the optimal operating solution ofdistrict heating networks at a minimum of costs to estimatesaving potentials. With the help of parameters and figurelines it is very easy to modify the system configuration tocheck how the system reacts under new conditions andhow the operating costs change. In all cases the tool givesvaluable information.Unfortunately, the user guide, the help file and themanual are only available in German at the moment forboth programs.REFERENCES[1] www.bmwi.de: LowEx Fernwärme – MultilevelDistrict Heating, Fördergeber: Bundesministeriumfür Wirtschaft und Technologie, FKZ 0327400B.[2] www.gnu.org/software/glpk:GNU Linear .Pro-gramming Kit[3] www.coin-or.org: Computational Infrastructure forOperations Research.[4] www.gnu.org/licenses/gpl.html: GNU GeneralPublic License[5] S. Gnüchtel, Ein Beitrag zur Strukturoptimierungvon Fernheiznetzen, PhD thesis TU Dresden(1981)[6] http://tu-dresden.de/die_tu_dresden/fakultaeten/fakultaet_maschinenwesen/iet/ew/forschung_und_projekte/mldh/download_ml325