Energy-efficient heating and cooling systems as well as intelligent systems for energy distribution are urgently required in order to be able to meet the ambitious goals of the European Union to reduce greenhouse gas emissions. The present article is intended to show that intelligent system extensions for the area of heating, cooling and electricity production for the industrial sector can lead to significant increase in efficiency. For this purpose, a simulation study for the expansion of a combined heat and power (CHP) plant with 2 MW thermal output using a 1.4 MW absorption chiller has been carried out. This shows that a heat-controlled CHP unit can significantly increase its running time. A system model was created for the initial situation and validated with existing measurement data. In the second step, this model was expanded to include the ACM module. The simulation was able to prove that in the event of a system expansion, the run time of the CHP unit can be increased by 35%. In addition to then increase of energy efficiency in the supply system, the analysis also focuses on the efficiency of the energy distribution via thermal networks in an industrial environment. The presented paper therefore also highlights the optimization potentials in the operation of thermal supply networks for industrial applications. For this purpose, a mathematical model has been developed which in addition to the components of the thermal network itself also comprises the producers and consumers. The specific construction of thermal networks for the supply of industrial properties requires adapted solutions for the simulation of such systems. Therefore, amongst other things, in the paper, solutions are shown for the modelling of direct flow local heating networks as well as for the operation of a cascade-controlled pump group.
9.1 Introduction
In order to achieve the ambitious goals of the European Union (EU) of reducing greenhouse gas emissions by 80% below the 1990 level by 2050 (European Commission, 2011), the efficiency of energy generation and its distribution must be significantly improved. Around half of the EU’s total final energy consumption is used for heating and cooling buildings and the use of thermal energy in industrial environments (European Commission, 2016). The cooling of office buildings and industrial processes is responsible for a significant part of the total electricity consumption. This proportion is expected to increase further in the coming years (RESCUE, 2015). Special weather situations such as persistent heat waves are putting considerable strain on the power grid already (ENTSOE, 2017).
In addition to efficient cooling, intelligent heat generation and in particular heat distribution are equally important. This also includes a sensible combination of different forms of energy, for example, through combined heat and power. The operation of thermal networks is changing. The endeavour to lower the flow temperature in heating networks as much as possible and in return to increase it in cold water networks poses new challenges for planners and operators. The flexibilization of thermal networks is therefore an important factor in achieving the goals set by 2050.
Advertisement
Against this background, the iCity project deals within the sub-project 3.1, “Increasing efficiency through intelligent networking of producers and consumers in commercial areas using the example of Robert Bosch GmbH at the Schwieberdingen location”, with various questions about increasing the efficiency of an industrial site. Robert Bosch GmbH’s Schwieberdingen industrial site serves as a case study for the development and testing of general approaches.
Two main topics are dealt with in the present work. On the one hand, this is the development of a simulation tool for calculating thermal networks. The specific requirements of the heating network of the present case study must be considered. On the other hand, the increase in efficiency of the operation of an industrial combined heat and power plant (CHP) through the establishment of a combined cooling, heat and power system (CCHP) is considered. It has been shown that CHP plants in Europe get a very useful extension to combined cooling, heating and power systems by combining them with absorption chillers (ACM) (Haider, M., et al., 7/2005a). With modern absorption chillers and an intelligent control concept, the running times of a CHP unit can be significantly extended (Albers, J. et al., 2017). To analyse the combined operation and to evaluate the optimal sizing, a detailed dynamic simulation model has been developed in TRNSYS. Monitoring data collected for the heating and cooling supply of the whole site has been used as a basis for the simulations.
9.2 Case Study Description
The Schwieberdingen location has been operated by Robert Bosch GmbH as a development centre since 1968 and currently employs around 6800 people. The site covers more than 45 hectares, with currently around 60 buildings. Figure 9.1 shows an aerial photo of the industrial site from 2018.
Fig. 9.1
Industrial site Schwieberdingen of Robert Bosch GmbH
×
The structure of the industrial site is in constant change. Ongoing construction projects continuously change both the building landscape and the building use. The location has an independent heating and cooling supply. The required thermal heating and cooling energy is generated in two energy centres and distributed to the consumers via the on-site networks. Figure 9.2 shows the two energy centres 509 (central 1) and 307 (central 2) beside the buildings of the industrial site.
Fig. 9.2
Location scheme of the Bosch Schwieberdingen industrial site with the two energy centres 509 (central 1) and 307 (central 2), as well as the buildings on the supply “Trasse West” (Robert Bosch GmbH)
×
Advertisement
Through the iCity project, a comprehensive measurement data recording of all energy flows is being set up for the Bosch Schwieberdingen industrial area. The energetic data of all thermal generators and the distribution of the generated energy via the heating or cooling network at the location are recorded using measurement technology. The cross-system measurement data are recorded centrally via the building management system at the site and historized in databases. With the “EmTool”, a software developed at Stuttgart Technology University of Applied Sciences, these databases are summarized and continuously updated. The EmTool is able to generate synchronized time series across all systems, which enables a detailed analysis of the measurement data. For the present work, time series as input data for simulations and their validation are generated with the EmTool.
The heating energy consumption of approx. 20 GWh per year and the cooling energy consumption of approx. 40 GWh per year of the entire industrial site are covered by two heating and cooling centrals and distributed through heating and cooling networks. In case of the cooling, some additional decentralized chillers are used in buildings which are not connected to the cooling grid. To cover the high energy demand, the total installed heating capacity in the two energy centres is approx. 23 MW and is provided by natural gas boilers and a heat pump for the low-temperature heating distribution. Part of the heating energy is generated by a combined heat and power plant, which varies between 8 and 10 GWh per year. The total installed cooling capacity at the location is approx. 21 MW and is mainly covered by compression chillers with dry or hybrid air coolers and wet cooling towers. The property’s annual electricity consumption is around 65 GWh (including electricity for cooling). At the site, seven photovoltaic systems and the CHP unit are currently operated to generate their proper electricity. In-house power generation for 2019 (CHP and five PV systems) was around 8 GWh. Figure 9.3 shows the measured data of the site for the electrical energy demand, heating and cooling energy demand as well as the property’s own energy production for heat and electricity by the CHP and photovoltaics.
Fig. 9.3
Energy consumption data as well as own production at the site by CHP and photovoltaics
×
Figure 9.4 shows the thermal networks for supplying the industrial site with the required heating and cooling. A 7°/13° cooling network and two heating networks with different temperature levels are operated via the two energy centres. In addition to the 90°/70° high-temperature network from central 2, a 60°/40° low-temperature network is operated for the eastern part of the site via a heat pump in central 1.
Fig. 9.4
Heating and cooling networks at the industrial site (Robert Bosch GmbH)
×
9.3 Study to Increase the Run Time of a CHP by an Absorption Chiller
9.3.1 Initial Status Analysis
The following sections deal with analyses on the operation of the above-mentioned CHP plant in the industrial environment. The gas-fired combined heat and power plant has a thermal capacity of 2 MW. The electrical capacity is thereby approximately 1.6 MW. The co-generation unit is located in the central 2 heating and cooling centres, where the majority of the heating energy for the site is generated. In the central 1, however, most of the cooling energy is provided. The location of the two power stations is shown in Fig. 9.4 (in the context of heating and cooling network) and Fig. 9.2 (in the context of the site buildings).
The CHP is heat-controlled, i.e. operation depends on the seasonally fluctuating heat load at the site. Bosch’s decision to operate the CHP unit heat-controlled is based on ecological considerations. The CHP unit couples the heat generated into the heating network all year round, thus covering the base load of the heating requirement. However, in summer or during off-peak periods, the load in the heating network regularly falls below the thermal part load operation of the CHP unit, which leads to power modulations and a clocked operation. Figure 9.5 shows the typical course of CHP operation on two exemplary days in winter and in summer 2017. The diagram below clearly shows the fluctuating operation of the CHP in summer. This results from the high heat output of the CHP unit with simultaneous low network consumption. In winter, on the other hand, the CHP unit is operated at full load for long periods, which is usually only interrupted briefly, as can be seen in the upper of the two diagrams.
Fig. 9.5
Typical behavior of the CHP operation for two exemplary days in winter and in summer
×
Due to the clocked operation of the CHP in summer, the theoretically possible running times of the plant are not achieved. The diagram in Fig. 9.6 shows that electricity and heat generation in the summer months is significantly lower than in the winter months. The present study aims to investigate the possibilities of increasing the running time of the CHP.
Fig. 9.6
Monthly net electricity and hot water generation of the CHP for the years 2016 and 2017
×
All considerations regarding the optimization of running time assume that the CHP unit continues to be operated in a heat-controlled manner. This requirement means that a solution must be worked out, from which the additionally generated heat can be used under energetically, ecologically and economically sensible aspects. Initial considerations in this regard included a concept to supply heat in excess of the plant’s own requirements to a residential area close to the site via a district heating connection. After initial studies, however, this concept had to be rejected. Further scenarios aimed at storing the excess heat in long-term storage facilities. Within the framework of the project, extensive studies were carried out on the so-called seasonal storage systems. Projects that have already been implemented but are also still in the planning stage were examined for their applicability to the case study. As an example for the concept implementation of seasonal storage, a project in Friedrichshafen Wiggenhausen (Morgenbrodt, L. 2008) can be mentioned, which is one of the largest in Germany. The parameters of the storage facility and its characteristics provide an overview of the potential applications of this technology. The seasonal storage facility has a volume of 12,000 m3 with a storage capacity of 675 MWh. The planned construction costs of the project were estimated at about 1.35 million euros. Irrespective of the above-mentioned costs, implementation cannot be considered reasonable from a technical point of view. With a storage capacity of 675 MWh, the co-generation plant will be extended by just under a month in summer. Therefore, such a solution is not reasonable from an economic point of view.
Another promising idea was to extend the operating time of the CHP by means of a combined cooling, heat and power system. For this purpose, it should be investigated whether the integration of an absorption chiller, which is operated by means of the heat generated by the CHP unit into the energy production process, can be realized. This concept was generally considered to be promising and should therefore be investigated in more detail in this study.
Figure 9.7 shows a very simplified scheme for heat generation in the central 2 by the CHP. A 38 m3 hot water tank is located between the CHP unit and the hot water network. This is dynamically loaded when the heat load on the hot water network falls below the minimum output of the CHP unit. The CHP is operated in two stages. Normal operation corresponds to the nominal thermal output of 2 MW, and in partial load operation, the CHP unit achieves an output of 1.5 MW. In the summer months, the CHP unit starts about six times within 24 h with a respective running time of about 75 min per cycle. This means that the CHP unit starts up almost 1000 times during the summer months because it is heat-controlled. During the summer months, the entire heating load of the site is carried by the CHP. In the winter months, the gas boilers 1–4 are dynamically switched on according to the existing heating load. The economic efficiency of CCHP plants depends significantly on the utilization period of the units (annual full load utilization hours) (Haider, M., et al., 8/2005b). Due to the different operating conditions, a precise analysis of the load curves for electricity demand, heating demand and cooling demand at the industrial site is essential.
Fig. 9.7
Simplified scheme of the heating center central 2
×
The diagram in Fig. 9.8 shows the total monthly heat energy provided by the CHP and the gas boilers in central 2 for the year 2017. During the summer months, the heat demand of the industrial site drops under 500 MWh, which is about 1000 MWh less than the monthly heat production of the CHP. Accordingly, during the summer months, nearly 1000 MWh more could be produced via the CHP. From this, a theoretical (maximum) heat potential for the operation of the CHP unit can be determined from a static point of view. This is the difference between the amount of heat required, i.e. the existing load profile, and the heat that can theoretically be generated by the CHP unit. In the diagram, this potential is represented by the line with the circle symbols. For the year 2017, this value is approx. 8750 MWh.
Fig. 9.8
Theoretical heat potential for the operation of an ACM through extended running times of the CHP
×
With this purely static approach, however, it is not possible to examine the extent to which the operation of an ACM can be integrated into the energy supply of an industrial site in terms of systems technology.
In the present study, therefore, the operating mode of a planned system is examined in detail with the aid of dynamic simulation. For this simulation study, the existing load curves and system-related boundary conditions of the case study are to be considered. Thus, an exemplary operation of the CHP can be investigated based on realistic data. The results obtained in this way can be used for further considerations such as a CO2 balance, an economic analysis or an operation optimization. Furthermore, the scenario developed can be used for planning and implementation at other industrial sites.
9.3.2 Simulation Study
For the creation of the simulation models mentioned here, the software TRNSYS in version 18 was used. First, the basic methodology is explained. In the following, the created simulation model is explained in detail. Finally, the results are presented and discussed.
Methodology
A numerical model of the existing CHP plant was created as a basis for the simulation study on combined heat and power generation. In a first step, the basic model was validated with the measurement data available on site. The aim of this fundamental work was to obtain a basic model of the CHP system validated with load curve data available at the site, which could be extended in the second part of the study by the module of the absorption chiller. For the CHP base model, the CHP and the associated 38 m3 storage tank were considered, as shown in Fig. 9.7 in simplified form. The gas boilers 1–4 as further heat generators were not yet part of this model.
For the simulation study, an extended model was created in the next step. For this, the basic model had to be extended by the module of the absorption chiller. This system includes the ACM itself, three cooling towers for recooling the chiller and a significant buffer storage expansion in the form of a second tank with a volume of 160 m3. For the study, the technically possible maximum storage size for the site was used. This is a result of the maximum size of the storage components that can still be transported on normal traffic routes. With these components, a simple system was created with a focus on functionality. First of all, this model was to be used to investigate the basic feasibility as well as an expected extension of the operating time of the CHP. For the study, a lithium bromide absorption chiller with a nominal cooling capacity of 1.4 MW and a hot water-side power consumption of 2 MW at a temperature spread of 90/55 °C was selected. In order to be able to take the dynamic heating and cooling load of the case study into account in the model, it had to be extended with the components of heat and cold generation. Therefore, in the last step of the modelling, the gas boilers as heat generators and the compression chillers with the associated recooling as cooling generators were implemented into the overall model. In addition to the thermal simulation, the parameters of the electrical generators and consumers were also included in the model. Figure 9.9 shows a simplified schematic of all components that are part of the simulation model, with the exception of the gas boilers. In the illustration, the components of the absorption chiller module are outlined with a dashed line.
Fig. 9.9
Simplified scheme for the integration of an ACM into the existing CHP plant system
×
Model Description
As mentioned at the beginning, the simulation model was developed in TRNSYS 18. In the following, some basic information on the most important modules of the model is explained.
The central CHP model was created with the Type 907 “IC Engine Generator”. Water is used as heat transfer medium. Its density and specific heat capacity are considered in the model depending on the temperature. The heat exchanger is calculated with Type 5b, and the heat transfer of exhaust gas and heat pump (cabin cooling) is calculated with TEES Type 682 “Heating and Cooling Loads Imposed on a Flow Stream” (Tess Library, TESS 2017). The pipe connections for fluid transport are made with Type 31. Type 2d (On/Off differential controller) was used to model the pumps. The two heat accumulators were modelled with Type 158. The absorption chiller was modelled with the TRNSYS Type 107, and for the three cooling towers of the recooling system, the Type 162b was used. The three compression chillers used in the central 2 energy centre differ in their design and performance. The smaller plant with a cooling capacity of 1 MW was modelled with TESS Type 666 (Tess Library, TESS 2017), and the two larger 2.5 MW plants were modelled with TESS Type 668.
Control Strategies
A control concept for dynamic simulation was developed for the basic model of CHP operation. For this purpose, the control concept of the existing plant had to be divided into defined operating states with the respective parameters. For the final model, with ACM, further operating states were defined according to this procedure. The relevant parameters for plant control are anchored in the respective states. A total of six states were thus defined with which the entire plant operation can be represented.
Table 9.1 shows in a simple matrix the respective states of the plant operation considered for the study. States 1 to 4.1 represent the transition months and the summer. State 5 is pure winter operation. Although there is also a cooling requirement here due to the test stands operated all year round at the site, the heat produced by the CHP unit is completely fed into the heating network and not used for the operation of the ACM. However, during the winter months, the refrigeration required for this purpose is provided entirely by central 1.
Table 9.1
Operating states from which the control strategy is defined
Status
CHP
ACM
Storage
Hot water network
Cold water network
1
ON (full power)
OFF
Partial loading
Demand
Demand
2
ON (partial power)
OFF
Partial loading
Demand
Demand
3
OFF
OFF
Unloading
Demand
NO demand
4
ON (full power)
ON
Partial loading
Demand
Demand
4.1
ON (full power)
ON
Partial unloading
Demand
Demand
5
ON (full power)
OFF
OFF
Demand
NO demand
For the operation of the gas boilers and the compression refrigeration machines, operating states were also defined and implemented in a control matrix. These states are downstream of the respective superordinate state from the overall model.
Simulation Results
The two scenarios—(1) initial state with a pure CHP operation and (2) plant extension with an absorption chiller—were simulated with input data from the case study for the period April 2018 to March 2019. For a comparability of the two cases, they must be based on the same data of the thermal energy supply. Therefore, the measured load curve data for heat and cold production in the period under consideration were used as input data for the simulations. A time step of 5 min was chosen for the simulation. The system limit for the following considerations was set at the central 2 energy centre.
The diagram in Fig. 9.10 shows a summary of the central results from both scenarios. The “cooling generation of central 2” is the amount of cooling energy distributed from energy central 2 via the cooling network to the connected consumers. This amounts to 6214 MWh in the period considered and is generated conventionally via the existing compression chillers. Not listed here is the part of the cooling generation that comes from energy central 1. Central 1 also supplies the site with the required cooling energy during the winter months.
Fig. 9.10
Electricity and cold production to be expected through the combined heat and power generation
×
For the simulation of the second scenario, the already explained control scheme was applied. The parameters used were selected in such a way that operation of the complex plant network is ensured. The focus here is on extending the operating time of the CHP. Accordingly, the scenario does not yet represent the optimal ACM operation. From this point of view, the cooling quantity produced by ACM is 2658 MWh. This means that approx. 43% of the total cooling demand, which must be covered by the central 2 energy centre, can be produced via the ACM.
The diagram also shows the electricity production achieved by the CHP for both scenarios. As expected, the ACM operation significantly increases the operating time of the CHP unit in summer and thus also the electricity production. In this scenario, an additional production of electrical energy for the site of 3570 MWh can be achieved for the period under review. This corresponds to an increase in CHP electricity production of around 49%.
Ideally, the increase in the site’s own electricity production takes place during the period when demand in the energy centre is at its highest. In the summer months, the electricity consumption for energy production is highest over the year as the compression chillers are increasingly in operation during this period. The diagrams in Fig. 9.11 compare the simulation results for electricity consumption and electricity production for the two scenarios. The diagram on top shows the initial situation. Here it becomes clear that the low level of CHP electricity production in the summer months cannot cover the needs of the energy centre. This is divided into three groups: electricity consumption by the CHP (in the diagram below electricity consumption_CHP-ACM), electricity consumption by the heat generators and electricity consumption for conventional cooling, which accounts for the largest share. The consumption of the generating plants themselves and the auxiliary energy consumption that can be allocated, such as that caused by pumps, are taken into account here. In comparison to the initial scenario, this scenario with ACM shows that the total electricity demand for refrigeration is decreasing. Although the electricity demand for the CHP-ACM system increases, the demand for conventional refrigeration decreases at the same time. Due to the longer running time of the CHP, the own power production increases, as already shown, so that the demand can be covered by the central 2 energy centre.
Fig. 9.11
Power consumption and power production for the cases with (bottom) and without (top) absorption chiller
×
Figure 9.11 shows the impact of the plant expansion on electricity production and consumption. The diagrams in Fig. 9.12 show the comparison for thermal energy. The red line on the diagram on top shows the heat requirement of the energy centre during the period under consideration and how this is covered by the CHP and gas boilers. In the period under consideration, from April 2018 to March 2019, the heat requirement is 15,465 MWh, of which 7672 MWh is covered by the heat production of the CHP unit. Due to the heat supply of the CHP unit, this diagram again shows the low running time of the CHP unit in the summer months. The potential heat production for the CHP unit due to the maximum expected running time is 9848 MWh for this period. The diagram below in Fig. 9.12 shows the results for the simulation of the CCHP system. For the simulation, the real load curves of central 2 for heating and cooling are used. The heat production by the CHP unit is divided into the share for covering the heating demand (Qhw - CHPsim) and the share of heat production for the ACM operation (Qhw - ACMsim). In total, the heat production by the CHP unit in this scenario amounts to 13,885 MWh, which is shown in the diagram with the line Qhw - CHP/ACMsim. Thus, an increase of 6213 MWh is achieved for the CHP unit through the additional use of the ACM which results in 3570 MWh additional electricity generation. The first simulations for a combined heat and cooling system thus show that the theoretically possible potential can thus be exploited to a good 60%.
Fig. 9.12
Thermal comparison of CHP operation with (bottom) and without (top) absorption chiller
×
The operation of absorption chillers depends on various factors. One limiting factor is the load curve of the cooling demand of the site. If the mass flow of the cold water network at the evaporator of the ACM is too low, the ACM switches to shutdown mode due to the lack of storage facilities for the cold produced. In addition, the heat requirement influences the running times of the ACM. To show this, it is necessary to consider the mass flows. The diagram in Fig. 9.13 shows the simulated mass flows for a selected period in the summer. The summer operation of the heating network at the site is usually characterized by domestic hot water (DHW) preparation. The heating energy required for this is provided exclusively by the CHP. The DHW preparation is given priority over the ACM operation. In real operation, this prioritization is solved purely hydraulically. If the heating demand increases, the CHP unit decreases the lateral mass flow for charging the storage tank and is, however, directly fed into the hot water network. Additional heat demand is covered by the hot water stored in the tank. This has the consequence that the hot water level in the storage tank drops during the ACM operation. If there is an increased demand from the hot water network over a longer period, the storage tank temperature drops below the critical value for ACM operation, and the absorber goes into shutdown mode. The diagram in Fig. 9.13 shows these relationships between mass flows and storage temperatures. The periods in which the mass flow_ACM (line with circles) goes to zero mark the standstill times of the absorption chiller. These periods are marked in the diagram for clarification.
Fig. 9.13
Distribution of the hot water mass flows that are generated by the CHP plant
×
9.3.3 Conclusions
The present contribution shows the results of a simulation study which analyses the CHP operation and its extension to a combined cooling, heating and power systems by combining them with absorption chillers. The plant concept and measurement data from the case study “Entwicklungszentrum Robert Bosch GmbH Standort Schwieberdingen” were used. The CHP plant considered in the study has a thermal capacity of 2 MW, and the electrical capacity is 1.6 MW. The cooling capacity of the extension plant is 1.4 MW. The measurement data available in detail in the case study show that in the years 2016–2019, the CHP unit was able to produce an average of 9000 MWh of heat and 6800 MWh of electricity per year. This means that the CHP unit achieves a capacity utilization of a good 52% of the maximum possible capacity utilization in continuous operation. The reason for this is that the CHP unit is operated on a purely heat-controlled basis for ecological reasons of the site operator. As a result, the CHP unit has an extremely low capacity utilization in the summer months. This study therefore analyses the extent to which longer plant operating times can be achieved by adding an absorption chiller to the CHP unit. The present work investigates the operation of this overall system based on a detailed plant simulation in TRNSYS. The dynamic simulation provides information on how the CHP-ACM operation fits into the energetic overall operation of the case study.
The simulation results show that a significant increase in the operating times of the CHP unit can be expected with the additional absorption chiller. For the simulation, a defined period from April 2018 to March 2019 was considered. For this period, there was an increase from 44% to 79% of the maximum possible heat production by the CHP unit and thus also its capacity utilization. Downtimes due to maintenance work or revisions of the CHP and ACM are not included in this consideration. A detailed analysis of the complex plant operation shows that for ACM operation, downtimes are to be expected for certain boundary conditions even in the summer months.
The work presented here shows that for complex plant systems, such as those presented here, a static consideration can only provide approximate results. The detailed plant model for the simulation of the overall system provides valuable operating information at this point and represents a basis for plant planning and later system optimization. The chosen approach can be transferred to similar constellations in other real estate operations. The findings on modelling gained here will simplify this for future applications. However, the simulation of complex systems, such as the plant available here, is always comparatively time-consuming. For the validation of the basic model, a database that is as consistent as possible is also required.
9.4 Development of a Simulation Programme for Modelling and Calculation of a Thermal Local Heat Supply
9.4.1 Initial Status Analysis
The previous part of this article deals with increasing the efficiency of power generation and its supply in an industrial context. Just as important as an energetically optimized energy production is the efficient transport of energy to the respective consumer. Terms such as lowering the supply temperature (Leoni, 2020 or Tol, 2021), decentralized feed-in (Monsalvete, 2017 or Paulick, 2018) or generation change (Haoran Li, 2018) in the field of thermal networks characterize the current developments in this sector. In this chapter, the energy distribution, here in particular the heating distribution by means of a district heating network on an industrial scale, is analysed.
One of the main topics of the present iCity sub-project is the analysis of the heating network at the Schwieberdingen site. The production of heat and cold is divided into two energy centres, which are connected to the properties at the site via two separate distribution networks. In the following, the subject of the analysis is a sub-area of the Schwieberdingen heating network (“Trasse West”). This area needs a simulation model representing the hydraulic and thermal conditions as accurately as possible. In Fig. 9.2, the buildings along the supply route “Trasse West” are marked with light-grey hatched pattern. Due to the numerous measuring points for temperature, pressure and volume flows within the heating network, the simulated values can be compared with the measured values. With a calibrated model of the heat distribution of the supply route “Trasse West”, scenarios are simulated to find opportunities to reduce the return temperatures of the heating network. Lower return line temperatures mean a higher spread between supply flow and return flow temperature. Through a higher temperature spread, electricity for the operation of the circulation pumps can be saved even though the supplied heating energy remains the same. There are also reduced thermal losses of the network to the environment (media channels or soil). So the lower return temperature saves thermal and electrical energy during the operation of the network. In comparable studies, a saving potential in the order of 0.15 euro/(K*MWh) was realized (Köfinger, 2017).
Using the example of the “Trasse West”, it will also be analysed to what extent it is possible to reduce the supply temperature without endangering the thermal comfort within the buildings. As a practical implementation measure, the area is examined for possibilities of cascaded connections of consumers (Köfinger 2017). As a criterion for a cascaded connection, the return temperature of the buildings will also be used. However, the focus of this investigation is on finding the highest possible return temperatures that could satisfy a nearby consumer as a supply temperature.
9.4.2 Simulation Study of “Trasse West”
The graphical programming software INSEL, which was co-developed at Stuttgart Technology University of Applied Sciences, was chosen as the simulation tool for the investigation of the “Trasse West”. INSEL is a modular programming language similar to TRNSYS. In INSEL, different plant configurations can be displayed on a graphical user interface by interconnecting blocks. The “blocks” in INSEL correspond in functionality and semantics to the “types” in TRNSYS. INSEL is open for the integration of experimental user blocks via dynamic link libraries (“.dll”), which allows the simulation of more complex plant components. The user block “spHeat” described in the following is a block developed at the University for the simulation of thermal networks in INSEL.
Methodology
Network Types and Their Representability in “spHeat”
“spHeat” is a simulation tool for mapping the hydraulic and thermal properties of first- to third-generation district heating networks (see Fig. 9.14). For first-generation heating networks, usually only one central heat generator is provided, which is connected to consumers by means of a distribution network. Networks of the second generation include additional heat generators in the network structure, and the flow temperatures are usually lower.
Fig. 9.14
Evolution in the system development of district heating networks (1st–3rd generation with central circulation pump at the heating plant, decentralised feed-in only planned from third generation). https://de.wikipedia.org/wiki/Fernw%C3%A4rme#/media/Datei:Generations_of_district_heating_systems_DE.svg. Based on Ashreeta Prasanna et al.: Optimisation of a district energy system with a low temperature network. In: Energy and Henrik Lund et al.: fourth Generation District Heating (4GDH)
×
In third-generation networks, the flow temperature of the networks is further reduced. As a result of this development, it will be possible for the first time to integrate heat from renewable sources in these network types. Possible energy sources include biogenic fuels (biogas, wood chips, etc.) with high combustion temperatures and heat from solar collectors (Heymann, 2018). When selecting the renewable energy sources, it must be ensured that the temperature requirements of the networks of approx. 70 °C are achieved (valid for supply line or supply line feed-in). For return temperature increases in pure heating networks, without combined heat and power generation, energy sources with lower source temperatures can also be considered. Common to both types of feed-in is the central connection of the renewable heat. In these network architectures, decentralized feed-in is not provided for and is therefore not supported by the standard version of “spHeat”.
Image of the Topography of a Network in “spHeat”
Decisive physical parameters for the description of a thermal network are the pressure distribution in the network and the resulting mass flows as well as the flow and return temperatures of the heat transfer medium. The mathematical equivalent of the real network is realized by a subdivision into network nodes and the edges in between (Fig. 9.15).
A network node can represent, e.g. a simple valve, a pipe diameter change, a consumer with a transfer station or a change of direction of the line in the locality. Network edges generally represent pipe sections with constant diameter and constant material and without additional installations. Once the pressure conditions in the network have been calculated and the heat extraction in the case of consumers or the heat input in the case of a decentralized heat generator is known, the temperatures at the network nodes can be calculated by energy balances.
Programme Sequence for the Calculation of the Variable Sizes of a Mesh in “spHeat”
A quasi-dynamic approach is used to determine the different pressures in the network. The flow and pressure are calculated with the help of a static flow model in the subprogramme “spHydro”. Since the relationship between flow resistance and resulting mass flow is not linear, a numerical approximation method must be used for the calculation (Newton iteration). The temperature, however, is calculated dynamically in “spThermo” depending on the flow velocity and various boundary conditions such as soil and ambient temperature. To solve the differential equations of heat transfer along the pipes, the backward differential method is implemented. In “spHeat”, the network is divided into flow and return flow plane. Both subnetworks are calculated separately (Fig. 9.16).
Fig. 9.16
Program flow diagram of “spHeat” with hand over of variables from program sections
×
Model Description
Hydraulically Separated Systems (Normal Case) or Direct Flow District Heating Networks
A basic assumption made in the simulation with “spHeat” is the division of the network into primary and secondary side and the coupling of both by heat exchangers in the transfer stations (see Fig. 9.17).
Thus, the simulation assumes two hydraulically separated circuits, the district heating network on the primary side and the internal heating circuit on the secondary side. The block “spHeat” only calculates the primary circuit.
In order to be able to map the network of the Schwieberdingen site mathematically, an adaptation of the source code of “spHeat” is necessary, because the primary and secondary circuit is not separated in the Schwieberdingen plant (see Fig. 9.18). Rather, there is only one circuit whose heat transfer medium (water) flows directly through the radiators in the properties and the water heaters in the energy centre Si307. This special feature of the system has far-reaching consequences for the calculation of the flow conditions on the primary side.
Fig. 9.18
Local heating connection without transfer station (Situation local heating network Bosch Schwieberdingen). (Originally: Verein Energiepfad Grabs/FirstMedia Solutions GmbH St. Gallen)
×
In classical heating networks with hydraulic separation between primary and secondary circuit, the change of flow conditions on the secondary side, e.g. by opening radiator valves, has no hydraulic effect on the primary side. But here these flow changes would have a considerable effect on the flow resistance of the entire system.
To prevent this flow-related influence of “building services components” on the hydraulics of the distribution network of the “Trasse West”, decentralized circulation pumps are used at the Schwieberdingen site (see Fig. 9.18). These are dimensioned and controlled in such a way that a constant differential pressure between flow and return can be maintained at the heating distributor.
To simulate the heating network in Schwieberdingen as a direct flow network with the existing software, a virtual heat exchanger as a new interface (see section “Creating the Simulation Model in the INSEL-GUI”) had to be created for “spHeat”. This interface allows the mapping of dynamic change of the flow conditions due to the radiator regulation and the calculation of the hydraulics on the primary side.
Circulation Pumps (Single Pump or Pump Phalanx)
During the original development of the “spHeat” programme, certain system components were mapped in a simplified form. The modelling depth of these components had to be adapted to the existing conditions in Schwieberdingen. One example is the internal mapping of circulating pumps for the primary circuit.
“spHeat” calculates the coefficients of the quadratic pump characteristic curve based on three operating points of a pump, defined by the delivery head in metres above the volume flow (Fig. 9.19 light-grey sloping curve).
Fig. 9.19
Pump characteristic curve (light-grey sloping curve) meets pipe network characteristic curve (dark-grey rising curve) at operating point B source: https://heizung.de/ call 20.10.2020 14:30 h. (Viessmann Climate Solutions Berlin GmbH)
×
Single pumps can be modelled with this method. When using two or more pumps in parallel operation, this method is insufficient.
The heat network in Schwieberdingen is operated with a pump phalanx with one, two or three pumps of the same type running in parallel. The starting and stopping of pumps is controlled by the differential pressure and depends on the actual flow rate.
The integration of the control of the pump phalanx in “spHeat” was not considered to be the best solution, but the opposite way to extract the calculation algorithm for the characteristic curve determination from “spHeat” and transfer it into a separate INSEL block. The interfaces between the pump modelling and “spHeat” are then no longer the operating points of the feed pump (delivery head and volume flow) but the coefficients of the quadratic equation.
Advantages of this solution are the increased flexibility in the modelling of pump components and the simplification of the “spHeat” source code. Also, this deconcentration of the programme routine of “spHeat” leads to the possibility to model decentralized supply by the so-called prosumers (consumers that can also supply heat to the network) with their own pumps.
The parameters the programme routine uses to distinguish between a consumer (without pump) and a prosumer (with pump) are the coefficients of the square pump curve (cA, cB, cC), which are all set to zero, in the case of a consumer. The pressure difference at a given operating point of the pump (given volume flow) is calculated as follows:
$$ \Delta p= cA\ast \dot{V^2}+ cB\ast \dot{V}+ cC $$
Δp: pressure difference in metres of pump head (m).
cA: quadratic coefficient of pump pressure equation (m/(m3/s)2).
cB: linear coefficient of pump pressure equation (m/(m3/s)).
cC: offset of pump pressure equation (m).
\( \dot{V} \): volume flow (m3/s).
Image of the Pump Control in Si307
The pump control is shown in detail in Fig. 9.20. On the left side, you can see the input files of the pump control separately for operating modes “one pump”, “two pumps in parallel” or “three pumps in parallel”. For the continuous simulation of the “Trasse West”, a seamless switching of the operating modes during the simulation run is essential. The simulation blocks on the right side of Fig. 9.20 allow exactly this.
Fig. 9.20
Pump control from the INSEL model for the “Trasse West” (left, the READ blocks for reading in the pump control signal differential pressure and volume flow; centre, pump blocks with calculation algorithm of cA, cB and cC; right, output files and plot block)
×
To get an idea of the variable pump control modes, Fig. 9.21 shows the pump characteristics for a single pump (solid line with diamond marking) and the operation of two pumps in parallel (solid line with triangle marking) or three pumps in parallel (solid line with square marking).
Fig. 9.21
Pump curves for a single small type pump, two small pumps in parallel and three small parallel pumps in operating modes 100%, 80% and 40% of pump speed
×
While the maximum head is nearly identical for all curves, the three operating modes differ considerably in the achievable flow rates at a fixed pressure difference. The changes of the pump line coefficients cA, cB and cC during the simulation run are shown in Fig. 9.22. This figure shows how these quantities change as a function of time. During the simulation run, the jump functions of cA, cB and cC represent the switching from operation with one pump 100% (solid line with diamond marking) to two pumps in parallel operation also 100% (solid line with triangle marking (Fig. 9.21)).
Fig. 9.22
The pump parameter coefficients cA, cB and cC as variables depending on the switching on and off of individual pumps (representation of the time course in January for approx. 2 operating hours, switching from 100% single pump to 100% double pump)
×
For clarification, the switching of two operating states was chosen, which would not follow each other in real operation, because in the chosen example, a speed control of the pumps was left out (dotted curves (Fig. 9.21)). The adaption of the pump speed would be the first try to match the requirement of a fixed pressure difference between supply line (SL) and return line (RL) within the district heating. In this case, the change of the pump coefficients cA, cB and cC would be continuous instead of forming a step function as shown in Fig. 9.22.
Creating the Simulation Model in the INSEL-GUI
Topology of the “Trasse West”
The simulation model for the “Trasse West” is to be divided thematically into the heat distribution and consumption via the “Trasse West” and its connected consumers and the heat generation in the energy centre Si307. The special features of the plant in Schwieberdingen require changes at the simulation block “spHeat” for a congruent modelling in INSEL.
Heat Distribution and Consumption
Definition of a “Virtual Heat Exchanger”
Since the “spHeat” routine assumes hydraulically separated circuits (primary and secondary circuit), the variable flow resistance of the “virtual heat exchanger” must be updated from time step to time step. For this purpose, the measured differential pressure between supply and return flow at the heating manifolds is converted into an additional flow resistance, which is added to the values from the pressure loss calculation of the pipe network. In this way, the hydraulic changes can be transferred to the primary circuit by opening and closing the valves, without having to make changes to the architecture of two separate circuits defined for “spHeat”. The thermal boundary conditions for the “virtual heat exchanger” can also be determined from the measured variables at the heating manifolds (measured SL/RL temperature, measured volume flow). These values can be used to balance the thermal power that is delivered to the “virtual secondary side”.
Validation of Image of Pump Control and “Virtual Heat Exchanger”
Figure 9.23 shows the structure of the simulation model for testing the virtual heat exchanger and the pump phalanx on the reliability of the representation of the real conditions. The core of the modelling is the combination of all connected consumers of the local heating system of the Bosch site in Schwieberdingen into one (“2_Big_Consumer”). In order to determine the hydraulic resistance of the entire network, it is necessary to calculate the pressure loss from the differential pressure between the supply and return line at the Si307 energy centre (“1_Producer”). This variable, supplemented by the time step values for the pump characteristics cA, cB and cC, is the input variable for modelling the network hydraulics (“spHeat” block in the centre of the GUI).
Fig. 9.23
Scheme of a simplified model for the DH in Schwieberdingen. The whole network is simulated as one virtual consumer (named here as “Big_Consumer”)
×
Simulation results of this setup are compared to measured data to prove the correct modelling of the hydraulic conditions of the network.
The comparison between the values of the simulated and measured volume flow for the hour 0–5 on September 1, 2020, is shown in Fig. 9.24. The two curves match almost identical during the analysed period. The comparison of the two values over a 5-day period is shown in Fig. 9.25. Also, in this period the results of the simulation match with the measured data at the Schwieberdingen site.
Fig. 9.24
Match of calculated and measured value of total mass flow for the DH in Schwieberdingen (here hour 0–5 of September 1, 2020)
Fig. 9.25
Match of calculated and measured value of total mass flow for the DH in Schwieberdingen (here September 1–5, 2020)
×
×
Figure 9.26 shows the deviation of each simulated value compared to the measured value for the 5-day period (dots). The maximum deviation between the two values is +7 or −8% which only occurs at very few data points.
Fig. 9.26
Deviation of simulated values for total mass flow compared to measured ones (September 1–5, 2020)
×
Most of the data lies within the standard deviation of 0 to +2% which is a confirmation that the simulation performance is very adequate in comparison to the real circumstances. With this deviation, the detail of the big heat consumer can be raised in the next step of the investigation.
Heat Generation
Layout of the Distribution Circuits in the Heating System Si307
Layout of Manifold Heating System Si307
The “Trasse West” is a distribution network of the heating circuits supplied by the Si307 energy centre (see Fig. 9.27). Further connected distribution circuits are the branch for the planned “Trasse Süd”, the heating centre Si307 itself (own consumption “Si307_consumer”) as well as the consumer “Si121” and the “Trasse Ost”. All circuits are connected via the main manifold in Fig. 9.27 with the node numbers 2–3, 5–6, 8–9, 11–12 and 14 marked with the pump phalanx and the water heaters of the energy centre (node 1).
Fig. 9.27
Heating distributor in the Si307 energy centre (to simulate the “Trasse West”, it is necessary to model the other consumers with load data, as the circulation pumps and pressure maintenance include all connected consumers)
×
Since the pump phalanx with up to three pumps is connected to the manifold in parallel operation and directly supplies all connected distribution circuits without hydraulic switches, the flow conditions in all distribution circuits must be considered in the hydraulic and thermal balancing of the “Trasse West”.
The principle of the “virtual heat exchanger” is also used here to represent the other outlets of the heating circuit distributor. Except for the “Trasse West”, the hydraulic resistances of all outlets (“Trasse Süd”, “Si307_consumer”, etc.) are modelled via the differential pressure at the respective flow and return lines at the heating manifold (for the “Trasse Süd” at node 5, for “Si307_consumer” at node 8, etc.). From the measured volume flow and the SL/RL temperatures, the thermal boundary conditions of these balance limits can be determined for the simulation model.
Layout of “Trasse West”
From the metrologically recorded values at the heating distributors of the buildings (the location is shown in Fig. 9.28 in the centre of the buildings), the necessary calculations can be made, which allow the modelling of a “virtual heat exchanger”.
Fig. 9.28
Topology of the “Trasse West” of the heating network in Schwieberdingen (the main pipeline network with DN 300, the sevice pipes to the heating distributors, diameter varying depending on the connected load)
×
At the “virtual heat exchanger” interface, the flow temperature (SLsim) calculated by the “spHeat” block is reduced by the amount of thermal power transferred to a return temperature (RLsim), which is then compared with the measured values of the building’s return temperature. If both return temperatures match within the error tolerance, the simulation model is brought into agreement with the measured data.
Calculation Routine of “spHeat”
Starting from a flow rate for the “Trasse West” (node “4” in Figs. 9.28 and 9.29) at the Si307 heating system, the actual mass flows are determined iteratively by calculating the pressure loss along the network’s pipelines.
Fig. 9.29
Mathematical model of the “Trasse West” in the graphic programming language INSEL (with corresponding designation of the network nodes in the block diagram; see Fig. 9.12)
×
The Newton iteration method is used, because hydraulic resistances (unlike, e.g. electrical resistances) are not linearly related to the driving force but are measured in a quadratic relationship. Consequently, the relevant systems of equations for the description of the pressure distribution or the flow velocities cannot be solved analytically but can be approximated exclusively by numerical means. The iteration loop is aborted if the deviation of the calculated total mass flow to the preset mass flow (measured at the heating system Si307) is within the set error limits. If there is no convergence between the two quantities after several iteration steps specified by the user, the simulation is aborted with an error message.
In the case of convergence, the iteration loop has been passed successfully. The mass flows of all network edges and the differential pressures between forward and return flow at all network nodes (nodes 16–34 in Fig. 9.28 or Fig. 9.29) were determined.
In a further step, all temperatures of the network nodes are defined by balancing the thermal losses of the pipe network in connection with load data for the extraction of heat by the consumers (nodes 18, 19, 22, 25, etc. in Fig. 9.29).
Thus, the continuous simulation is based purely on the input values at the outlet of the “Trasse West” in the heating system (node 4 in Fig. 9.29) and the load data of the consumers (“Si225”, “Si226”, “Si125”, etc.). The distribution network in between (shown in Fig. 9.29 with thick dark-grey line) is clearly defined due to the material properties and input variables determined by the user.
Input Data from Measured Values
The available input data for network simulation is shown in Fig. 9.30. Measured values for the Si225 building are available for supply and return temperatures, flow rate, differential pressures and thermal output at the heating manifold “building-specific” and with high temporal resolution. Not all metres of the properties of the “Trasse West” record as completely as in the example building Si225. In close cooperation with the facility management of the Bosch site in Schwieberdingen, schedules for metre replacement and improvements in data management are being developed. The aim is to record the properties of the “Trasse West” for a sufficiently long period (4–6 weeks minimum) in order to validate the INSEL model with this complete data set.
Fig. 9.30
Course of the input data of “spHeat” in building Si225 for an exemplary day
×
9.4.3 Summary and Outlook
With the changes to the “spHeat” calculation routine, a calculation programme that can map district heating networks of the first to third generation with hydraulic separation between the primary and secondary circuit has been enhanced to emulate networks with direct heating circuit flow.
This required the considerations discussed with the umbrella term “virtual heat exchanger”. The technical changes for the implementation of “virtual heat exchangers” in the INSEL block “spHeat” were successfully integrated, so that the software can simulate considerably more flexible district heating networks with and without hydraulic separation between primary and secondary circuit beyond the concrete application in Schwieberdingen.
The further procedure for the application case Schwieberdingen can be outlined as follows:
After checking and calibrating the two separate simulation sections (heat generation and heat distribution/consumption) with measured values, the two sections are merged into a holistic simulation model of the “Trasse West”. With this holistic simulation model, which is often referred to as “digital twin”, investigations with different boundary conditions are possible.
At the current stage of network expansion, the properties of the “Trasse West” are supplied with a minimum flow temperature of approx. 70 °C for a large part of the existing buildings. The level of the flow temperature, which was chosen when the site was first built in 1968 (the original dimensioning was 120 °C for the SL), is too high for the supply of buildings with current building shells and heating technology.
Since a mix of buildings constructed during the last five decades is supplied with the district heating network at the Schwieberdingen location, especially on the “Trasse West”, the supply temperature of the network cannot be lowered without endangering the supply security of the older buildings or the water heaters in the newer buildings.
With the simulation model, however, scenarios can be calculated without interfering with the ongoing operation, such as how the return temperature of the “Trasse West” could be further reduced by a cascade connection.
A basic requirement for a cascade connection is a high return temperature of one building, combined with a low requirement for the supply temperature level of a second building in the immediate vicinity. A further prerequisite is that the heating capacities of these two buildings allow a cascade connection. If both conditions are fulfilled, the building with a low supply temperature requirement can be connected to the return flow of the building with a high return temperature and thus cool down a return flow which is, e.g. approx. 55 °C to 25 °C.
The advantages of this type of connection are both savings of thermal energy due to lower thermal losses of the pipe network and an increase in the spread between flow and return temperature and the associated savings of electrical drive energy for the pump phalanx.
Further scenarios in which the buildings of the “Trasse West” are also integrated as single zone building models in the calculation model are conceivable. If this modelling depth is reached, statements about the thermal comfort in the buildings would then be possible when the total supply temperature is lowered.
Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.
The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
