Faults in district heating customer installations and ways to approach them: Experiences from Swedish utilities

Energy 180 (2019) 163e174

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Faults in district heating customer installations and ways to approach them: Experiences from Swedish utilities Sara Månsson a, b, *, Per-Olof Johansson Kallioniemi a, Marcus Thern a, Tijs Van Oevelen b, c, Kerstin Sernhed a a b c

Lund University, Faculty of Engineering, Energy Sciences, Box 118, SE-221 00, Lund, Sweden VITO, Boeretang 200, BE-2400, Mol, Belgium EnergyVille, Thor Park 8310, BE-3600, Genk, Belgium

a r t i c l e i n f o

a b s t r a c t

Article history: Received 31 December 2018 Received in revised form 30 March 2019 Accepted 30 April 2019 Available online 9 May 2019

The district heating (DH) customer installations in current DH systems contain a variety of different faults that cause the return temperatures of the systems to increase. This is a major problem, since the focus in the DH sector is to decrease the system temperatures in order to utilize more low-temperature heat. Therefore, this study has focused on how utilities are currently working to keep their temperatures low, how they involve their customers in this work, and what faults are most common today. This was done by conducting a combined interview and survey study, where Swedish DH utilities participated. The results showed that the two most important elements to obtain low return temperatures were to have physical access to and mandate of the customer installations, and to maintain a good and close customer relationship. The results also showed that many faults occur in the customers’ internal heating systems, or were due to leakages somewhere in the installation. Overall, the results showed that it is indeed possible to work close to and affect the customers to obtain lower return temperatures from the customer installations. It was also clear that the most common faults were rather easy to eliminate as long as the utilities gained physical access to the entire customer installation. © 2019 Elsevier Ltd. All rights reserved.

Keywords: District heating substations Experience from industry Fault diagnosis Faults in substations Poor substation performance

1. Introduction District heating (DH) has been identified as an important part of the future smart energy systems, since it enables an increase of the energy efficiency of the entire energy system [1]. This is due to the ability of DH systems to utilize local fuel or heat resources that would otherwise go to waste [2]. However, the current DH systems are not working as well as they could. One way to improve the system efficiency is to reduce the distribution temperatures of the system [3]. The main advantages of lower system temperatures are obtained when decreasing the return temperature of the system first. This enables a lower supply temperature without a significant decrease of the temperature difference between supply and return temperature [2]. A large temperature difference is important to avoid large flows in the system. Low distribution temperatures

* Corresponding author. Lund University, Faculty of Engineering, Energy Sciences, Box 118, SE-221 00, Lund, Sweden. E-mail address: [email protected] (S. Månsson). https://doi.org/10.1016/j.energy.2019.04.220 0360-5442/© 2019 Elsevier Ltd. All rights reserved.

decrease heat losses from the DH systems to the surroundings [4], and the COP of large-scale heat pumps increase [5]. Low return temperatures increase the efficiency in production sites where flue gas condensation is installed [6]. Low supply temperatures increase the power output in combined heat and power plants (CHP), and increase the heat recovery in industrial excess heat and geothermal heat [2,7e9]. Decreased system temperatures are also one of the main characteristics of the 4th generation of DH (4GDH), which was defined in a paper written by Lund et al. [4]. According to the paper, the temperature levels of 4GDH should be approximately 50/20  C supply/return temperature [4]. This would among else make it possible to integrate heat sources such as industrial waste heat, solar thermal energy, and geothermal energy [1,10,11]. Since these heat sources deliver heat at low temperature levels, it will become extremely important to maintain the low system temperatures in the 4GDH systems. To reach and maintain lower distribution temperatures, it is important to identify and eliminate faults resulting in high return temperatures soon after they have occurred. Faults in the

164

S. Månsson et al. / Energy 180 (2019) 163e174

customers' installations constitute a large share of the problems causing excessive return temperature levels [2,3,12]. These faults may occur in the customers' internal heating systems or in the customers’ substations. By eliminating these faults, the return temperatures from the customer installations could be reduced. This enables a decrease of the system return temperature [2]. Therefore, this study investigates what the most common faults are in current DH customer installations, and what is currently being done by the DH utilities to eliminate them. Previous research show that many DH utilities have approached the problem by shifting the responsibility of fault elimination to the customers [12]. The main reason to this is that the customers usually own both the internal heating system and the substation, and should therefore fix their own equipment [2]. Therefore, the utilities have created monetary incentives for their customers to get them interested in improving their installations’ performances [2]. A common way is to have a flow component in the price model, since an increased flow through the substation results in a high return temperature. This means that the customer is charged for the increased return temperature from the installation [13,14]. Another approach is to implement a bonus-malus system, which is revenue neutral to the DH utilities. In such a system, customers with high return temperatures pay a fee to the DH utility that is used to pay a bonus to customers with low return temperatures [15]. However, a study conducted by Sernhed et al. showed that many customers did not understand the purpose of the flow components [16]. This indicates that the flow component itself may not be a sufficient incentive for the customers to fix faults in the installations. Recent studies have shown that focus is now shifting towards that the DH utilities should work more service orientated towards their customers [17]. This includes offering the customers help to make sure that the customer installations are performing well [18]. Many DH utilities offer some sort of service agreement or review of the installation's performance, and invite their customers to yearly meetings where the DH systems and issues concerning it are discussed. Another common way to work with customer installation performance is to analyse customer data in some way to detect installations that are not performing well. The historically most common approach is to analyse each customer's overflow [2], but current research presents a number of novel fault detection methods utilizing customer data [3,12,19e24]. By implementing these methods, the utilities may improve their fault handling schemes even further. A conclusion when reading previous studies is that they have mainly focused on individual aspects of improving the return temperatures from the customer installations. The aim of this study is instead to provide a holistic overview of measures, and combinations of measures, that are implemented by utilities today to reach lower temperatures. The approach taken has been to investigate what strategies utilities with low return temperatures are working with to maintain their low system temperatures. What are the key successful measures the utilities are currently taking to reach lower return temperatures, and how are their customers involved in the process? The study has also investigated how the utilities are actively working to eliminate faults in the customer installations, and what faults are most frequently identified when visiting the installations. The method of the study has been to conduct a survey study, as well as a series of interviews and literature reviews. In this way, it is possible to obtain both qualitative and quantitative results regarding, for example, different faults, fault handling processes, and fault elimination methods. The utilities participating in the study were located in Sweden and hence the results should be interpreted with a Swedish context in mind. However, large parts of the DH technology are very similar

in other countries. There are also similarities in how the DH systems are managed. Therefore, the results may be transferred to the contexts of other countries as well. 2. The customer installation 2.1. Basics of the customer installation The design of the customer installations varies, depending on what country and DH system the installation is located in. In some countries, it is common practice not to have any hydraulic separation at all, so called direct connection. However, most countries have implemented an indirect connection where one or more heat exchanger(s) separates the water in the internal heating system from the water in the DH system [2]. This study is conducted in a Swedish context, where indirect connection is most common. The substation consists of a number of different components, and the design and connection principle of the substation also varies. The two most common connection principles are the parallel-connected and the 2-stage connected substation [25]. Fig. 1 displays the general outline of a parallel-connected substation and an internal heating system with hot water circulation in the domestic hot water (DHW) system. The Figure includes the most common components of the customer installation. The components are in general described to include the heat exchanger(s), circulation pumps, control valves, temperature sensors, controller, room heaters (e.g. radiators), and a heat meter which consists of supply and return temperature sensors and a flow meter. 2.2. Faults in customer installations A number of different faults may appear in a customer installation. However, several faults do not affect the comfort of the customer. This means that a fault might be present in the installation without the customer experiencing any issues with the space heating or the DHW preparation [26]. However, the fault still affects the return temperature from the installation and therefore the utilities should work actively to eliminate these faults as well. The different faults may be divided into different categories. This may be done in a number of different ways. In this study, the faults are divided into five different categories depending on what component or part of the customer installation that the fault occurs in. The five different categories are: (i) heat exchangers, (ii) control system and controller, (iii) actuators, (iv) control valves, and (v) internal heating system of the customer. Below follows a review of faults that may occur in the five different categories. 2.2.1. Heat exchangers The heat exchanger category includes faults that are related to the heat exchanger(s) of the substation. Depending on the design of the substation, there may be one, two, or more heat exchangers installed. The literature shows that faults related to the heat exchangers themselves most commonly include fouling of heat exchangers [2,26e29]. This can be described as deposits on heat transferring surfaces which cause a resistance to heat transfer and flow in the heat exchanger [30]. Another fault reducing the heat transferring capability is that some heat exchangers are installed incorrectly. This may occur if the heat exchanger is installed so that the DH water and the water in the customer's internal heating system are flowing co-current instead of counter current [27,28]. Other faults in this category include leakages from the heat exchanger itself, and uneven flows in heat exchangers connected in parallel [21,28]. One other problem that may arise is shell and tube heat exchangers, which are commonly present in older DH systems. Shell and tube heat exchangers used to be the state of the art in DH

S. Månsson et al. / Energy 180 (2019) 163e174

165

Fig. 1. General outline of parallel-connected substation with hot water circulation.

systems. Today, the most common solution is to install plate heat exchangers (PHE) instead. However, there are still shell and tube heat exchangers installed in older systems since it is not economically feasible to exchange all heat exchangers at once. One common issue in shell and tube heat exchangers is that they often start leaking after some years [2]. 2.2.2. Control system and controller The category of the control system and the controller includes the controller itself, temperature sensors, connections between sensors and controller, and connections between actuators and controller. This means that there are a number of different faults that may occur in this category. The controller itself may break down causing a completely uncontrolled system, which means that a change of heat demand will not be satisfied by the heating system [27]. The controller might also be installed incorrectly in the system if, e.g., the wrong input signal is connected to the wrong port on the controller. The controller receives an incorrect signal and the control sequence of the installation will be disturbed [28]. There are a number of different temperature sensors in the control system:  Outdoor temperature sensor  Indoor temperature sensor  Temperature sensor measuring supply temperatures in the space heating system  Temperature sensor measuring supply temperatures in the DHW system Faults in these sensors may cause large deviations in how heat is transferred to the customer installation. One fault is that the measurements from the temperature sensors are distorted by noise or drift (bias change) [31]. Temperature sensors may be broken, not sending any signal at all to the controller [27]. If, e.g., the outdoor temperature sensor is broken, the outdoor compensation curve will not be fulfilled and the supply temperature in the space heating system will be constant [32]. The temperature sensors measuring the supply temperatures in the space heating and DHW systems may be placed on the wrong pipe, leading to a disturbed control sequence since the sensor is measuring the incorrect temperature [2,21]. They may also be

assembled incorrectly, so that they are mounted loosely to the pipe [27,28]. In smaller customer installations it is common not to have hot water circulation in the DHW system. When this is the case, the water in the DHW system will be stationary in the supply pipes and will cool down over time. The temperature sensor in the DHW system will sense this and send a signal to the controller to increase the temperature. Since there will be a time delay before the newly heated water reaches the sensor, the control signal to increase the water temperature will continue for longer than necessary. This causes the DHW supply temperature to overshoot the desired value [33]. To avoid this problem, the temperature sensor measuring the supply temperature in the DHW system should be placed as close to the heat exchanger as possible [25]. 2.2.3. Actuators The actuator is the component that control the position of the control valves in the customer installation. The actuators are connected to the valve via a valve stem. Due to wear and tear, the connection between the stem and the valve may become poor, or the actuator may break down. If this happens, the control valve will be stuck in the position it was when the actuator broke down. This cause uncontrolled flows in the installation, and no response to the heat demand in the building [34]. The actuators have a number of different parameters that need to be taken into consideration when sizing the actuator for the installation. Among else, it is important that the actuator is sized for the correct pressure so that it has enough driving force to change the position of the valve [25]. The actuators also have different stroke times. In the DHW system, the actuators should have a short stroke time since the valves need to be able to open or close instantly when a change of the demand for tap water occurs. In the space heating system, the actuators need to have a long stroke time since the supply temperature should change slowly to obtain the desired room temperature. These actuators may be interchanged by mistake at installation, causing poor control of both parts of the internal heating system [27]. 2.2.4. Control valves The faults related to the control valves in the customer installation may arise if the valve is oversized. This fault is especially

166

S. Månsson et al. / Energy 180 (2019) 163e174

common for the DHW system in older buildings, due to a historical practice to oversize the valves. An oversized valve causes an undesirably large change in the flow when the valve position changes, and makes it difficult to control small flows through the valve [21,26]. The valve may also be too small for the flow that is needed in the installation. When this occurs, it will not be possible to obtain larger system flows than the valve is able to transmit [21]. Since the space heating system is not used during warmer months, the valve in the space heating system will be in the same position during this time. This may cause the valve to get stuck in a closed position [35]. It may also happen that the valves seize in an open position or that they are completely stuck [27]. Valves are, like actuators, susceptible to wear and tear. Over time, the disc of the valve may erode due to cavitation if there is a large differential pressure over the valve [36], causing the valve to leak in a closed position [27]. 2.2.5. Internal heating system of the customer The fault category related to the internal heating system of the customer include a large variety of faults, since it contains many different components. In some DHW systems an additional heat exchanger, a pre-heater, is installed in series with the heat exchanger seen in Fig. 1. If this is the case, the circulation pump should be installed between the pre-heater and the final heat exchanger to avoid increased return temperatures [27]. If the DHW circulation is missing or broken, the water in the DHW system will be standing still in the pipes. As described in section 2.2.2, this may cause a poor control of the DHW temperature. It is also important that the flow in the DHW circulation is not too large, since the cooling in the heat exchanger will decrease. The temperature demands in the customer installation is determined by the temperature levels for which the space heating system is sized, and by the set point value that the DHW system requires to prepare hot water [25]. If the temperature demands are high, maybe even higher than the temperature of the DH water, the flow on the DH side of the heat exchanger will always be large, causing high return temperatures [2]. When a building is heated using a hydronic heating system, this needs to be hydronically balanced to make sure that it is able to deliver the correct amount of heat to all heat delivery points in the system. This is done by adjusting the balancing valves in the system so that the opening position of the valves closest to the heat exchangers is smaller than the valves located further away in the building. This cause the flow to travel with the same flow rate to all radiators and that a sufficient amount of heat is delivered to all radiators, irrespective of the distance from the substation. If this is not done correctly, the return temperature from the installation will increase [37]. The radiators in the space heating system are most commonly controlled by thermostatic radiator valves. If these are not working correctly, or are missing, the flow in the radiator system will be uncontrolled [28]. 3. Methodology This study investigated both technical and organisational aspects of faults in customer installations. As previously mentioned, the study was conducted in a Swedish context. The technical aspect was covered by conducting a literature review of the most common faults, after which a survey study was made where Swedish DH utilities were asked to report their most frequently occurring faults. The organisational aspects were mainly covered by a series of qualitative interviews where representatives from six Swedish DH utilities participated.

3.1. Survey study To investigate the faults that occur in the current DH system, a survey study was conducted. The first part of the study was to establish what faults to include in the survey. This was done by conducting a literature review, as well as an interview with a leading district heating expert. The main purpose of the interview was to create a fundamental understanding of the faults, as well as identifying the faults that were most likely to occur in modern DH systems. The literature review gave a comprehensive overview of the historical fault distribution, and what faults should be investigated further in the study. Information about the different faults was primarily collected among scientific articles and research reports conducted by various research institutes. The publications referred to in the study are predominantly conducted and written by Swedish researchers. These references are a result of the large number of DH research projects that have been conducted in Sweden since the mid-1970s, which has generated a number of high-quality research reports and papers that are of interest for the subject of this study. However, to mainly include Swedish references was not an active choice, but when conducting the literature review these were the references that were found on the subject. These results were then used to create a survey which was sent to DH utilities in Sweden. A number of questions regarding the five fault categories identified in the literature review and the specific faults were created. The survey included questions about how common certain faults were, how the utilities identified the different faults, and what the fault distribution in the DH system looked like. The survey also included questions regarding information about the utilities themselves, and their price models and service agreements. There was also plenty of space for the utilities to add additional faults that were not included in the survey to begin with, and the possibility to write down comments and explanations about the different faults. The survey was sent to 139 utilities that were all members of the Swedish district heating association, Swedenergy. 56 of the utilities answered the survey. In addition to these utilities, three utilities answered that they do not work with the maintenance of the installations in their systems themselves. Instead, they utilized subcontractors to perform this work. The data obtained from the survey were then analysed. Particular attention was paid to the different fault categories and what faults were reported as most frequently occurring. 3.2. Interview study The interviews were conducted during the fall of 2018, and representatives from six DH utilities were interviewed during the study. The participating utilities were all located in Sweden. The utilities were of varying sizes, ranging from approximately 300 customers to several thousands, and delivered varying amounts of heat to their customers. Five of the utilities were selected due their low annual average return temperatures. The reasoning behind this selection was that utilities with low return temperatures should already be working actively with their customers’ return temperatures. The last utility that participated had higher return temperatures than the rest. This utility was included to investigate whether there were any clear differences in the way of working of a utility that had low return temperatures, compared to a utility with high return temperatures. The interviewees were people involved with, or responsible for, the operation and maintenance of DH production, distribution or end user. The interview study was conducted using a qualitative, semi structured approach in order to obtain in-depth knowledge of how the utilities were working with their customers in the context of decreasing the return temperature levels. Before the interviews

S. Månsson et al. / Energy 180 (2019) 163e174

were conducted, a number of themes and questions were prepared in order to facilitate the interview process. However, the semi structured approach allowed for the interviewer to follow up the responses of the interviewee during the interview and ask new questions if needed. All six interviews were conducted by the same interviewer. Four out of the six interviews were conducted face-to-face with the interviewer and the interviewee(s). The interviews were recorded and took approximately 1 h to complete. The two remaining interviews were conducted as phone interviews and recorded via the phone. The interviews were transcribed word for word, and the material obtained was then analysed by dividing the answers thematically into four different categories: (i) incentives for district heating utilities to reduce return temperatures, (ii) incentives to make district heating customers reduce their return temperatures, (iii) how utilities gained access to and mandate of the maintenance of the customer installations, and (iv) analysis methods for finding poorly performing customer installations. 4. Results and analysis 4.1. Faults reported as most frequently occurring in survey study One of the most interesting aspects of the different faults is what type of faults the utilities experience that they most frequently encounter in their customer installations. This was investigated by analyzing the results from the survey study. Fig. 2 shows how the utilities answered when asked what fault category the most frequently occurring fault belonged to. As can be seen in the Figure, there were two additional fault categories that emerged in the answers, that were not included in the survey from the beginning: leakages and inferior gaskets. The faults that were included in the leakages category included leakages in all parts of the customer installation and substation, and inferior gaskets included all gaskets that are used in a substation. Fig. 2 shows that the largest fault category according to the survey was leakages (33%), closely followed by faults in the customers’ internal heating systems (31%). This shows that it might be hard for the utilities to gain access to all faults in the customer installation since the utilities in most cases only have access to the substation and not the internal heating system. Below follows a detailed description of the fault distribution for each of the five fault categories, based on the results from the survey study. 4.1.1. Reported heat exchanger faults Fig. 3 displays the distribution of the reported faults in the heat exchanger category. As can be seen in the Figure, it is clear that

Fig. 2. Distribution of the most frequently occurring fault categories.

167

Fig. 3. Distribution of the heat exchanger faults that were reported as most frequently occurring.

leakages were experienced to be the most common reason to faults in heat exchangers. 77% of the utilities stated that this was the most common fault in the category. The utilities stated that the leakages occurred due to incorrect installation of the heat exchanger, and leakages in the pipings connected to the heat exchangers. When comparing these results to the overall fault distribution in Fig. 2, it is clear that the utilities may have answered “leakages” instead of “heat exchanger faults” when asked what category constitutes the largest fault category in their utilities. A reason to this might be due to the way the questions in the survey were formulated. In the survey, the heat exchanger category is meant to include the heat exchanger as an individual component. However, when reading the word “heat exchanger”, many people think of the entire DH substation, and not only the singular component. Therefore, the respondents may have interpreted the questions regarding the heat exchanger category as if they concerned the entire substation. This interpretation may have caused an overrepresentation of leakages in the heat exchanger category since it may be that some of the leakages that the utilities refer to actually appear elsewhere in the system. The second most common fault was cracking of the heat exchanger plates (7%), which the utilities have experienced when water hammer occurs in the system. They also experienced that this type of cracking lead to leakages. Other issues that the utilities experienced were fouling of heat exchangers, shell and tube heat exchangers that leaked, and broken heat exchanger plates. These faults corresponded to 4% each of the reported faults. Fouling of heat exchangers was by most utilities considered as an uncommon fault. However, a small share of the utilities stated that this was their overall most common fault. The utilities all agreed that the amount of fouling depended on the water quality in the systems. Broken shell and tube heat exchangers were reported by utilities in charge of DH systems that were built when shell and tube heat exchangers were still state of the art. The utilities stated that these exchangers break due to their age and have to be replaced. There were also utilities reporting issues with broken heat transfer plates in the PHE. The problem occurred when the connection between the DH system and the heat exchanger was built incorrectly so that the incoming DH water blasted straight onto the heat transferring plates. If unnoticed, the water blasts could cause holes in the plates. Finally, 4% of the utilities stated that there was almost never an issue with the heat exchangers as a component. 4.1.2. Reported control system and controller faults Faults in the control system and controller corresponded to approximately 5% of the overall fault distribution. This indicated that faults in the control system and controller were not the most

168

S. Månsson et al. / Energy 180 (2019) 163e174

frequently occurring faults, but almost all utilities stated that faults in this category did occur on a regular basis. Despite this, the utilities stated that they did not consider them to constitute the most frequently occurring faults when comparing to the rest of the fault categories. Fig. 4 displays the fault distribution of the control system and controller category. As can be seen in the Figure, 49% of the utilities answered that the most common fault was that the controller was broken. The second most common fault was broken temperature sensors (16%), and the third most common fault was temperature sensors giving the wrong signal (11%). This indicated that a large share of the faults in the control system and controller was an effect of broken or poorly performing components. The utilities also experienced that incorrectly placed temperature sensors were a frequently occurring problem. In total, incorrectly placed outdoor temperature sensor, temperature sensors placed on the wrong pipe, and temperature sensors placed too far away from the heat exchanger corresponded to a total of 13% of the faults. All of these faults lead to poor control of the internal heating system of the customer, causing abnormal temperature levels. According to the utilities, they should all be recognized by the customer since they affect the customer comfort; either in terms of an unwanted change of the indoor climate, or in terms of unwanted DHW temperature levels. However, the utilities agreed that it was beneficial for the utilities to be able to detect these faults at an early stage, even before the customers. One interesting fault that emerged in the survey was that many utilities experienced that the customer itself may occur as a fault in the system. This problem occurred if the customer changed the set point values or outdoor compensation curve of the controller manually. The utilities experienced that this happened when the customers felt that their indoor temperature was too low, and changed the settings in order to try and increase the temperature. In most cases, this did not give any effect since the low indoor temperature was due to faults located elsewhere in the installation and not due to how the system was controlled. The only effect of the change in the controller settings was that the flow through the substations increased, with increased return temperature levels as a consequence. Incorrect control sequences were reported by 5% of the utilities as the most frequently occurring fault in the fault category. These faults mainly occurred if the settings in the controller were not adjusted to the building the controller was located in. 4.1.3. Reported actuator faults The fault distribution of the faults related to the actuators can be seen in Fig. 5. As can be seen in the Figure, 70% of the actuator faults

Fig. 4. Distribution of the control system and controller faults that were reported as most frequently occurring.

Fig. 5. Distribution of the actuator faults that were reported as the most frequently occurring.

were a result of broken actuators. Some of the utilities experienced that this fault may occur in the DHW system if the control valve it was connected to was too large for the system in question. When this was the case, an oversized valve caused the actuator to operate constantly to attempt to control the flow, which caused the actuator to break down prematurely. This was mainly an issue in the older DH systems, where historically oversized control valves was an issue. The second most common fault in the category was actuators that could not change the position of the valve and corresponded to 22% of the utilities’ answers. This fault was commonly due to oversized valves or that actuators were not dimensioned for the current pressure. The utilities experienced that this was a fault that the customer may not notice if the faulty actuator was located in the space heating system. This was due to the fact that the room heaters, especially radiators, in many space heating systems were controlled by thermostatic radiator valves. If so, the final heat supply to a room was controlled by the thermostats. This means that the heat deliveries to the room heaters were controlled even though the flow through the substation was too large. Therefore, the utilities wanted to detect this type of fault at an early stage. 4% of the utilities reported that actuators with incorrect stroke times were the most frequently occurring fault. The utilities experienced that this fault was most urgent to detect in the DHW system, since it caused large temperature variations which increased over time. Actuators that did not close the valve completely corresponded to 2%, and poor connection between actuator and control valve was reported by 2% of the utilities as the most common actuator fault. The utilities stated that poor connection caused poor control of the control valve, leading to either a very high or a very low flow through the valve that did not change according to the control signals. 4.1.4. Reported control valve faults Fig. 6 displays the fault distribution for the faults related to the control valves. In this category, oversized control valves were reported as the most frequently occurring fault (35%). However, the occurrence of this fault varied greatly between the utilities. Some of the utilities had never experienced that this was a problem while some of the utilities experienced that this was their most common fault. This depended on the age of the DH systems, and how much effort the utilities had put into changing the historically oversized valves. The two faults that were reported as second most common in the category were control valves leaking in a closed position (27%) and control valves seizing in a closed position (27%). The leaking control valves were most commonly due to wear and tear of the

S. Månsson et al. / Energy 180 (2019) 163e174

Fig. 6. Distribution of the control valve faults that were reported as the most frequently occurring.

control valve itself, or in the connection between the valve stem and the actuator. The seizing control valves most frequently occurred in the space heating system at the beginning of the heating season. The two remaining faults that were reported as most frequently occurring for control valves were leaking gaskets (7%) and control valves seizing in an open position (4%). 4.1.5. Reported faults in the customers' internal heating systems Fig. 7 shows the fault distribution for the faults related to the customers' internal heating systems. Most of the utilities experienced that the most common issue was poor balancing of the radiator system (32%). The utilities experienced that the main reason to poor balancing was that the customers were lacking knowledge of how to perform the system balancing, and that they in many cases did not ask for professional help. The utilities also stated that the balancing of the radiator system was the customers’ responsibility but that they would be happy to advise the customers about the issue. The second most common fault was missing or broken thermostatic radiator valves (23%). The utilities responded that many of the faulty thermostatic valves were located in older buildings where the valves had broken down due to wear and tear, causing poor space heating. Systems without DHW circulation was also mentioned as a frequently occurring issue (16%), primarily in smaller houses. The consequences of this issue were increased return temperatures and poor DHW preparation. Although this is not necessarily a fault rather than a design choice, it causes high return temperatures from the customers’ installations. Another frequently occurring fault was that the set point values

Fig. 7. Distribution of the faults in the customers' internal heating systems that were reported as the most frequently occurring.

169

in the customers’ systems were close to, or higher than, the DH supply temperatures (12%). Some of the utilities mentioned that this issue may occur during the summer months due to low DH supply temperatures in the outer parts of the DH system, where it was harder to maintain higher supply temperatures. If the customers had the same set point values during the entire year, the set point value would be higher than the DH system temperature during these periods. The utilities also stated that this was an issue which caused high over-consumption of flow in the installation. Therefore, the utilities wanted to identify the installations with high set point values as soon as possible. The utilities reported that they had issues with unnecessarily large flows in the DHW circulation (5% of the utilities). This caused the return temperature in the customers' internal heating systems to increase, causing increased return temperatures from the installation. This was also the case when there was an uncontrolled flow in the secondary circuits, which 5% of the utilities stated was the most common fault in their customers’ internal heating systems. 4.1.6. Interpretation of results from survey The results from the survey gave an insight into what faults the utilities experienced as most frequently occurring in their systems. Some of the faults in the individual categories were reported as “very rare” which might mean that the utilities almost never encounter these faults. However, it might also be that the utilities simply do not find these faults when investigating a customer installation. This may have been the case for some faults in the internal heating system category, since the utilities normally don't have access to their customers' internal heating systems. The utilities that answered the survey were all members of Swedenergy, which include almost all DH utilities in Sweden. This means that the responding utilities were located all over Sweden, from north to south. The size of the systems and temperature levels varied depending on their location and age of the system, and their building stock varied depending on the DH system. Therefore, the answers from the survey included a wide variety of faults. For example, older systems experienced more issues with shell and tube heat exchangers and oversized control valves than the more modern systems did. When analyzing the results, it is also clear that the utilities may have associated some faults with other fault categories than the authors had in mind when creating the survey. An example of this is that faults in the actuator and control valve categories were very closely related to each other. This means that some of the actuator faults may have been reported in the control valve category instead, and vice versa. A clear example of this was the fault “actuators cannot change position of valve” in the actuator category. Some of the utilities stated that this fault appeared when the control valve was stuck, which might as well could have been interpreted as a fault in the control valve category. This kind of interpretations may have resulted in skewed fault distributions in the different categories. Another issue was that many of the utilities reported the effect of a fault, rather than the actual fault, as a fault. This was especially clear for faults considered as leakages, which constituted the largest share of the overall fault distribution. However, when analyzing the results for the heat exchanger category, the utilities stated that, e.g., “these leakages occurred due to inferior gaskets” and “poor welding caused leakages from the heat exchanger”. This indicates that the utilities have in deed stated the effect of a fault rather than the cause of a fault. It is also important to keep in mind that the participating utilities were located in Sweden. This may have affected the results, since it is possible that some faults do not occur in the DH systems

170

S. Månsson et al. / Energy 180 (2019) 163e174

due to national conditions. One example of this is that fouling of heat exchangers seemed to be a relatively unusual problem according to the utilities, while it is mentioned as one of the most common faults in the literature. One reason to that fouling is not as frequently occurring in the survey as expected may be that the overall water quality in the Swedish DH systems is high due to the measures that have been taken to improve the quality. Another reason may be that the study has been conducted in Sweden, where it is common practice to use compact plate heat exchangers which generally are less exposed to fouling. 4.2. Incentives for district heating utilities to reduce return temperatures The DH utilities that participated in the interview study were all working with their customers' return temperatures in different ways. Some of the utilities put a lot of emphasis and work into reducing the costumers’ return temperatures, while others considered their low return temperatures as an added bonus when working actively with other issues in their systems. In general, the utilities agreed that it was most urgent to identify customers that had high return temperatures combined with a large heat demand. These customers were identified to have the largest impact on the system performance. The interviews showed that there were different incentives for DH utilities to work with their customers' return temperatures. One common incentive was that the systems had flue gas condensation in their production facilities. There were also some incentives related to how the heat was used in the DH systems. In cases when a utility served as both heat distributor and heat transmitter, i.e. when a large amount of heat was transferred through the DH system each year to another system, it was important to decrease the customers’ return temperatures mainly to prevent the need of increased pumping in the system. 4.3. Incentives to make district heating customers reduce their return temperatures From the interviews, it was clear that there are many different incentives that could be used to get the customers to address problems and faults in their installations. Some of the utilities made use of their customer relationships to create incentives, while other created incentives in terms of additional costs if the installations were not working correctly. The latter could also be seen in the survey study, where many of the utilities with low return temperatures had some sort of flow component in their price models. The overall most important factor to increase the customers’ willingness to contribute to and work towards lower return temperatures was identified to be the relationship between the DH utility and their customers. The experience was that the customers were willing to pay the costs of correcting faults and problems in their installations, as long as the information about why and how

they should do this was sufficient and easy to understand. This information was distributed to the customers in different ways, depending on what other measures the utility took to create incentives for their customers to work with their return temperatures. In the smaller DH systems, it was common practice to visit all customer installations once per year to examine the installations and substations on site. This provided the utilities with an excellent opportunity to talk to the customers and improve their customer relationships. On the contrary, this was one of the factors that the utility with high return temperatures experienced was hardest to obtain. This was mainly due to the size of the DH system, which meant that the utility unfortunately did not have time to establish close customer relationships. One common monetary incentive was to have a flow component in the basic price models for one or several types of customers. This meant an additional cost for the customers if a large amount of water passed through the substation during one month. The flow component did not constitute a large share of the final invoice, but the experience was that it was enough for the customers to become interested in different ways of reducing the flow through their substations. However, the experience was that the flow component itself was not enough to create an incentive for the customers; it was also of great importance to inform the customers of the reasons behind the flow fee and what they could do to decrease the amount of flow passing through their substation. Without this information, the flow component was seen as an additional cost only that the customers did not understand and therefore did not want to pay for. The results from the survey showed that there is a large variety in how the flow component is designed. As can be seen in Table 1, there were utilities that charged their customers according to the amount of DH water that passed through the substation (SEK/m3). Some of the utilities had this flow component for all of their customers, while some introduced it to certain customer groups. The groups were based on different criteria, which can be seen in the second and third column of the table. For example, some utilities had a flow component for customers with return temperatures above a certain target value. Another solution was that some utilities only had a flow component during parts of the year, normally during the heating season. There were also a number of utilities that used a bonus-malus system. These were normally based on the customer's return temperature level - if the customer installation had a return temperature above a certain reference value, the customer had to pay a flow fee, and if the return temperature was below the reference value the customer received a bonus. In some utilities, the flow fee was used as a last and final action, if no other actions from the utilities encouraged the customer to improve the installation performance. The utilities emphasized that it was important to initially inform the customers about the high return temperature and what this meant to the customer and DH system. The utility should also provide assistance to find a solution to the problem, before introducing this type of flow component. If

Table 1 Flow component designs according to the survey. Price component

Description

Types of customers

Flow cost, SEK/m3

Flow component in price model

All customers All customers except single family houses All types of customers Only non household customers All types of customers All types of customers All customers except single family houses All types of customers All customers except single family houses

Flow Flow Flow Flow

cost, cost, cost, cost,

SEK/m3 SEK/m3 SEK/m3 SEK/m3

Bonus-malus, SEK/m3

Flow Flow Flow Flow

component for all customers with contracted loads above a certain amount component for customers with Tr above certain value component for all customers with DT below certain value component. Only included during parts of the year, normally during heating season

Customers with Tr below reference value gets bonus, customers with Tr above reference value pay fee

S. Månsson et al. / Energy 180 (2019) 163e174

the customer would not be susceptible to the information and proposed solution, a flow component could be introduced. The customer would then have to pay for the flow until the problem or fault in the installation had been solved. The utilities experienced that the need for this type of measures mostly occurred if the installation was not used by the owner of the installation but rented out to someone else. If so, the additional cost did not affect the owner him or herself since an increase of the rent for the tenant would cover the additional heating costs for the owner. In general, the interviews showed that the customers were willing to work with the utilities to decrease their return temperatures, even though no clear incentive for each individual existed. However, there was a difference between the smaller and the bigger utilities where the smaller utilities experienced that it was easier to convince all of their customers to work with their installations. This was due to the fact that they were more likely to have time to visit all of their customers, and that they had time to give their customers in-depth knowledge about the customer installations. The smaller utilities also experienced that their customers were willing to work with their installations due to the collective responsibility to keep the DH prices down. The larger utilities experienced that it was relatively easy to target their largest customers, but stated that it was harder to work with and convince their smaller customers to improve their installation performance. This was both due to the large amount of man-hours that would be needed to target all customer installations, as well as the fact that smaller installations have a smaller system impact when performing poorly. Hence, it was more cost efficient for the larger utilities to target the larger customers first. 4.4. Importance of getting access to and mandate of the maintenance of the customer installations One of the most important issues regarding the work with the customers' installations was to gain insight into how the installations performed, and get physical access to the customer substation. The general conclusion was that the utilities gained a lot by being allowed into the customers’ properties to physically examine the substation. They also experienced that their customer relationships improved significantly since they not only gained physical access to the installation and substation - they also gained access to the customer. This meant that they were able to show and explain to the customer in person what happened in an installation if a fault occurred, and why it is beneficial to have an installation that performs as well as possible. One common way of gaining access to the customers' installation was for the utilities to sign service agreements with the customers. This created a natural reason for technicians from the DH utilities to visit the customers’ installations and make sure that the installations were performing well. The experience was that the outcome of the service visits was most beneficial when doing the visits at a regular basis, with at least one visit every other year. During the visits, the technicians investigated the performance of the installation by performing an inspection where the different parts of the substation were tested and made sure to work as they should. If something in the installation was broken, it was common practice to offer the customer some sort of solution. This could be that the utility replaced the broken component free of charge, or that they gave the customer an offer about the cost to replace the component. This gave the customer the possibility to decide if the utility should perform the work, or if someone from a plumbing company should get involved. The utilities stated that most customers chose to accept the offer from the DH utility rather than involving a third party. Other ways of gaining access to the customers’ installations at a

171

regular basis were to offer the customers a free of charge survey of the installation when needed, and to allow the customers to call the utility at any time if something in their installations was not working correctly. This gave the customers a sense of security and trust towards the DH utility, which the utilities considered to be a big advantage in their customer relationships. 4.5. Analysis methods to detect deviating installation behaviour The results from the interviews and survey showed that it was common practice to use different varieties of analysis methods to find the customers with undesirably high return temperatures. The methods varied in complexity, but common traits were that the methods were based on customer data from heat meters that were normally used for billing purpose. The utilities had structured their analysis work in different ways. Some of the utilities performed analysis a few times per year, while others received instant alarms from their measurement systems if a large deviation from the normal installation behaviour occurred. Others performed monthly controls of their customers’ data to investigate whether the behaviour of the installation deviated significantly from its normal behaviour. In general, the larger utilities were most interested in developing and utilizing different data analysis methods. The reason to this was that they benefited a lot from obtaining automated methods that sped up their fault detection process. The different methods that were identified in the interviews are summarized in the bullet list below and further explanation follows.     

Monthly checks of customer billing data Quality index based on the installation performance Analysis of return temperature levels Analysis of flow Over-consumption method

Monthly checks of customer billing data were used to identify whether a customer deviated significantly from its normal behaviour. In one utility, customer data was utilized to create a quality index which was based on the installation performance, as well as the location of the installation in the system. The quality index was then analysed manually on a daily basis and if the index changed rapidly or the index pattern started to deviate from the normal behaviour, a service technician contacted the customer immediately to investigate the installation further. The analysis methods were in most cases based on the temperature levels at the customer substation. The return temperatures or cooling performance of each individual customer installation were investigated, as well as the flow, to decide what customers to prioritize. The over-consumption method was also used by some utilities. Over-consumption can be described as the additional amount of DH water that has to pass through the substation in order to deliver enough heat when the installation contains a fault [2]. The over-consumption method takes into account the delta T of the substation, the flow through the substation, and the amount of heat being delivered to the customer. The advantage of the over-consumption method is that it gives a possibility to rank the installations according to how large their impacts on the system are. This is because an installation with a larger heat demand will have a larger over-consumption if the installation is poorly performing in some way [3]. In the survey, 50 of the 56 participating utilities stated that they performed analysis of customer data in some way to detect deviating installation behaviours. This data was primarily data normally used for billing purposes. Fig. 8 displays what variables the

172

S. Månsson et al. / Energy 180 (2019) 163e174

Fig. 8. Different measurements/values the utilities in the survey study used to identify poorly performing customer installations.

different utilities primarily analysed. As can be seen in the Figure, 19 of the utilities primarily performed analysis of the return temperature levels to identify the poorly performing installations. The main method was to check the temperature levels, identify the installations with the highest return temperatures, and contact the customers in question. Some of the utilities administered the customer contact themselves, while others used subcontractors in the plumbing industry. Whether the utility performed the work themselves, or used a subcontractor, depended on the size of the DH system and if the number of employees at the utility allowed for doing this kind of maintenance work. Nine of the utilities performed analysis of the cooling of the substation, while seven used the over-consumption method to identify installations that were not performing well. The utilities that investigated their customers’ heat consumption (six utilities) often controlled how much heat their customers consumed during one month, compared to an expected heat consumption in the installation. If the consumption deviated clearly from the expected behaviour, the utilities got an indication that something may be wrong. However, further investigation of the installation was needed since a change in the heat demand may occur due to natural reasons, e.g., new residents or an unusually cold month. However, the utilities also stated that in some cases it was not a change in the customer behaviour that had occurred; rather an incorrect heat demand estimation done by the utility when determining “the expected behaviour”. 17 of the utilities stated that they performed analysis on a combination of measurement values from the customers’ meters. This meant that they utilized, e.g., the return temperatures and the consumption, or the return temperatures and the overconsumption, to identify poorly performing installations. Often the analysis was made as a series of different analysis steps. An example of this is that some utilities initially investigated the overconsumption, and then investigated the return temperature levels of the customers that were most poorly performing according to their over-consumption level. As can be seen in Fig. 8, three of the utilities utilized so called QW values to investigate the performance of their customers’ installations. The QW values were calculated by dividing the flow rate with the heat used in the building during, e.g., one month. This provided the utilities with a figure that could be used to perform comparative analysis of installation data from installations of different sizes and heat demands. 5. Discussion The work conducted in this study clearly shows that high return

temperatures and faults in the customer installations were viewed as important issues that need to be addressed in order to enhance system efficiency in DH systems. It will also be of great importance to work with these issues in the 4GDH systems, since faults in customer installations giving higher return temperatures will have a larger system impact in the future systems. The interviews showed that the current DH utilities have a number of different work procedures for fault handling in customer installations that could be implemented in the future systems as well. From the results, it was clear that different utilities worked with faults in different ways depending on size of the DH system, number of employees at the utility, and the amount of resources that could be allocated to this work. However, one of the most important aspects to eliminate faults resulting in high return temperatures was to create a close and good relationship with the customers. In new and future DH systems, this could be done already from the start in order to create a good foundation to have well performing customer installations continuously during the operation of the DH system. The second aspect of the fault handling process that the utilities stressed was the importance to gain physical access to the customers’ installations. This was primarily achieved in two ways: by signing service agreements with each individual customer, or to include a yearly inspection of the installation in the DH price. In some cases, the utilities also had mandate to fix minor faults free of charge for the customer, such as replacing a broken temperature sensor, which gave enhanced customer satisfaction while improving the overall system efficiency. Therefore, it might be a good idea to include service agreements and/or free-of-charge inspections in the price model in future DH systems. It was also interesting to see that there were no clear differences in the way utilities with low return temperatures were working, compared to utilities with high return temperatures. The largest difference was identified in the way they were working with their customer relationships. The utilities with low return temperatures involved their customers in many different ways to maintain the low temperatures. The utility with high return temperatures experienced that this was harder to obtain, mainly due to the number of customers in the DH system. This is an indication that it might be valuable for future DH utilities to have more personnel dedicated to working with customer relationships. The survey showed that leakages were experienced to be the most common category of faults, followed by faults in the customers’ internal heating systems. Since these faults may be hard to detect during service visits, it would be beneficial to be able to detect these faults by investigating customer data. This requires automated and accurate fault detection methods using data analysis that are easy to understand and interpret. Hence, such methods should be utilized in the future DH systems. However, some of the faults may be hard to detect in the data that is available today, which in most cases is the data used for billing purposes. This indicates that it might be a good idea to implement more measurement points in the customer installations in the future systems. Regarding the results from the interview study, they should be seen as examples of how the Swedish DH utilities are currently working with fault handling in their customer installations. They may also be seen as good examples of how a utility may work with their customers and customer installations to keep the DH system temperatures down. It is possible that more interviews would have resulted in new responses. However, the last interviews did not produce any results that had not already occurred during the previous interviews. The same was true for the interview with the utility with high return temperatures. This indicated that the study was saturated, and that further interviews might not have contributed to a bigger picture of the issues investigated in this

S. Månsson et al. / Energy 180 (2019) 163e174

study. One of the major points of improvement would have been to include more utilities with high return temperatures in the study. It is also possible that the results would have been different and more diverse if the utilities were chosen randomly. However, since the study focused on finding key successful measures for decreasing customer return temperatures, the selection criterion was considered sufficient. The results from the survey may have been affected by the response rate. 56 out of 139 utilities answered the survey, corresponding to a response rate of approximately 40%. It might be that only utilities that experienced a frequent occurrence of faults answered the survey. However, when analyzing the survey responses, it was clear that the participating utilities had different fault occurrence levels and that the responses also included utilities that did not experience a large amount of faults in their systems. The results may also have been affected by the fact that the study was conducted in a Swedish context. Some faults, e.g., fouling in heat exchangers, are frequently occurring in other countries but did not emerge as common faults in the survey. However, many of the components in both substations and internal heating systems are the same no matter what country the customer installation is located in. This means that the results are applicable for international contexts as well.

6. Conclusions This study has investigated how DH utilities are currently working with fault handling in customer installations in order to reach lower return temperature levels. The study has also investigated what the most frequently reported faults are and how the utilities are working to identify these faults. The results of the study may be seen as good examples of how this work can be conducted, and what key aspects of fault handling utilities in charge of future DH systems may consider to implement. The main conclusions of the study are described in the bullet list below.  To obtain low return temperature levels from the customer installations, it is of significant importance to gain physical access to the installations. This gives the opportunity to fix minor faults, while improving customer relationships.  Utilities with low return temperature levels work actively to obtain good customer relationships. Common ways to obtain this is to have service agreements with the customers or to include free of charge inspections in the DH agreement.  It is important to create clear incentives to why the customer should work with faults in the installation. The incentives should be easy to understand and combined with substantial information.  DH utilities benefit from helping their customer to identify and correct minor faults in the installation. This encourages the customers to contact the utilities the next time there is an issue in the installation, further improving the fault detection rate.  Most of the utilities utilized customer billing data in some way to identify poorly performing substations. This will be a crucial aspect of fault handling in future DH systems.  The most frequently reported fault categories in the study are leakages and faults in the customers' internal heating system.

Acknowledgments This work is part of the TEMPO project which is funded by the European Unions Horizon 2020 Programme under Grant Agreement no. 768936.

173

References € ller B, [1] Mathiesen BV, Lund H, Connolly D, Wenzel H, Østergaard PA, Mo Nielsen S, Ridjan I, Karnøe P, Sperling K, Hvelplund F. Smart energy systems for coherent 100% renewable energy and transport solutions. Appl Energy 2018;145:139e45. https://doi.org/10.1016/j.apenergy.2015.01.075. [2] Frederiksen S, Werner S. District heating and cooling. Lund, Sweden: Studentlitteratur; 2013. [3] Gadd H, Werner S. Achieving low return temperatures from district heating substations. Appl Energy 2014;136:59e67. https://doi.org/10.1016/ j.apenergy.2014.09.022. [4] Lund H, Werner S, Wiltshire R, Svendsen S, Thorsen JE, Hvelplund F, Mathiesen BV. 4th generation district heating (4GDH): integrating smart thermal grids into future sustainable energy systems. Energy 2014;68:1e11. https://doi.org/10.1016/j.energy.2014.02.089. [5] Lund H, Østergaard DS, Yang X, Mathiesen BV. Comparison of lowtemperature district heating concepts in a long-term energy system perspective. Int J Sustain Energy Plann Manag 2017;12:5e18. https://doi.org/ 10.5278/ijsepm.2017.12.2. [6] Bühler F, Petrovi c S, Karlsson K, Elmegaard B. Industrial excess heat for district heating in Denmark. Appl Energy 2017;205:991e1001. https://doi.org/ 10.1016/j.apenergy.2017.08.032. [7] Lauenburg P. Improved supply of district heat to hydronic space heating systems. Doctoral dissertation, Department of Energy Sciences, Faculty of Engineering, Lund University; 2009. [8] van Ruitenbeek J, van der Kamp J. Efficient use of residual heat in combined heat & power systems. Eur Energy Innovat Autumn 2013;2013:10e2. [9] Langendries R. Low return temperature (LRT) in district heating. Energy Build 1988;12:191e200. https://doi.org/10.1016/0378-7788(88)90064-3. [10] Lund H, Andersen AN, Østergaard PA, Mathiesen BV, Connolly D. From electricity smart grids to smart energy systems - a market operation based approach and understanding. Energy 2012;42:96e102. https://doi.org/ 10.1016/j.energy.2012.04.003. [11] Lund H, Hvelplund F, Mathiesen BV, Østergaard PA, Christensen P, Connolly D, et al. Coherent energy and environmental system analysis (CEESA). Research report. Denmark: Department of development and planning, Aalborg University; 2011. ISSN/ISBN: 978-87-91404-15-3. [12] Gadd H, Werner S. Fault detection in district heating substations. Appl Energy 2015;157:51e9. https://doi.org/10.1016/j.apenergy.2015.07.061. [13] Li H, Sun Q, Zhang Q, Wallin F. A review of the pricing mechanisms for district heating systems. Renew Sustain Energy Rev 2015;42:56e65. https://doi.org/ 10.1016/j.rser.2014.10.003. [14] Song J, Wallin F, Li H, Karlsson B. Price models of district heating in Sweden. Energy Procedia 2016;88:100e5. https://doi.org/10.1016/ j.egypro.2016.06.031. €despremie och returtemperatur [15] Petersson S, Larsson CD. Samband mellan flo (Relation between flow price component and return temperature), Research €rrva €rme AB; 2013. Swedish. report 2013:25. Stockholm, Sweden: Svensk fja [16] Sernhed K, Gåverud H, Sandgren A. Customer perspectives on district heating price models. Int J Sustain Plann Manag 2017;13:47e60. https://doi.org/ 10.5278/ijsepm.2017.13.4. [17] Lygnerud K. Challenges for business change in district heating. Sustain Soc 2018;8. https://doi.org/10.1186/s13705-018-0161-4. [18] Euroheat & Power, District heating in buildings, Tech. rep. Brussels, Belgium: Euroheat & Power; 2011. https://www.buildup.eu/sites/default/files/content/ District/Heating/in/buildings_final.pdf. [19] Johansson C, Wernstedt F. N-dimensional fault detection and operational analysis with performance metrics. In: The 13th international symposium on district heating and cooling, copenhagen, Denmark; 2012. [20] Xue P, Zhou Z, Fang X, Chen X, Liu L, Liu Y, Liu J. Fault detection and operation optimization in district heating substations based on data mining techniques. Appl Energy 2017;205:926e40. [21] Yliniemi K. Fault detection in district heating substations. Licentiate thesis. Luleå University of Technology - Department of Computer Science and Electrical Engineering; 2005. [22] Sandin F, Gustafsson J, Delsing J. Fault detection with hourly district energy €rrv€ data, Research report 2013, vol. 27. Stockholm, Sweden: Svensk Fja arme AB; 2013. [23] Månsson S, Johansson Kallioniemi P-O, Sernhed K, Thern M. A machine learning approach to fault detection in district heating substations. Energy Procedia 2018;149:226e35. https://doi.org/10.1016/j.egypro.2018.08.187. [24] Månsson S, Davidsson K, Lauenburg P, Thern M. Automated statistical methods for fault detection in district heating customer installations. Energies 2018;12. https://doi.org/10.3390/en12010113. €rrva €rme. Fa €rrva €rmecentralen - utfo €rande och installation (The dis[25] Svensk fja trict heating substation - implementation and installation), Technical regu€rrva €rme AB; 2014. Swedish. lations F:101. Stockholm, Sweden: Svensk fja €rrva €rmefo €reningen, Avkylningen i ett fja €rrva €rmesystem. The [26] Svenska Fja cooling in a district heating system), Project report FVF 2000:03. Stockholm, €rrva €rmefo €reningens Service AB; 2000. Swedish. Sweden: Svenska Fja €rrva €rmecentraler strukturerade (Types of faults in [27] Werner S. Feltyper i fja €gskola; 2013. Swedish. district heating substations structured). Halmstad Ho [28] Averfalk H. Morgondagens effektiva fj€ arrv€ arme: en beskrivande litteratur€gskola; 2014. studie (The efficient district heating of tomorrow). Halmstad Ho

174

S. Månsson et al. / Energy 180 (2019) 163e174

Swedish. [29] Geni c SB, Ja cimovi c BM, Mandi c D, Petrovi c D. Experimental determination of fouling factor on plate heat exchangers in district heating system. Energy Build 2012;50:204e11. https://doi.org/10.1016/j.enbuild.2012.03.039. [30] Bott TR. Fouling, 10.1615/AtoZ.f.fouling, Online. 2011. . [Accessed 25 October 2018]. [31] Yliniemi K, van Deventer J, Delsing J. Sensor fault detection in a district heating substation. In: Proceedings of the 10th IMEKO TC10 international conference on technical diagnostics; 2006. [32] Djuric N, Novakovic V, Frydenlund F. Heating system performance estimation using optimization tool and BEMS data. Energy Build 2008;40:1367e76. https://doi.org/10.1016/j.enbuild.2007.12.006. € m KJ. Reglerteknik e en elementa €r introduktion. Kapitel 8: [33] Åstro €dtidskompensering (Automatic control engineering e an elementary do

[34]

[35]

[36] [37]

introduction. Chapter 8: dead time compensation), report TFRT-3173. Lund, Sweden: Department of Automatic Control, Lund Institute of Technology (LTH); 1983. Swedish. € teborg energi AB 1995-2004 (Optimizing Fransson A. Avkylningsarbete på Go of the district heating system in Gothenburg 1995-2004 - practical application €rrva €rme AB; 2005. of R&D knowledge), Report FOU 2005:132. Svensk Fja Alsbjer M, Wahlgren P. Fj€ arrv€ armecentraler 10 år - håller de måttet? (District heating substations 10 years - do they measure up?), Report 2011:9. Stockholm, Sweden: Svensk Fj€ arrv€ arme AB; 2011. Swedish. online Pumpportalen.se, K€ agelventil (Globe valve). 2015. Swedish, http:// www.pumpportalen.se/kagelventil/. [Accessed 30 November 2018]. Energimyndigheten. Energieffektivisering i stora v€ armesystem (Energy efficiency in large heating systems), Publication ET2013:20. 2013. Eskilstuna, Sweden, in Swedish.