A Contribution to the Efficient Control of Heating and Cooling Supply Systems over the Annual Operating Cycle – Development of a System Controller

Key words: System controller, Heat pump systems, Building retrofit, Multivalent operation, Energy optimization

This paper introduces an approach for implementing such a control concept and presents the development of a system controller based on a holistic methodology, covering all stages from conceptual design and simulation studies to field testing under real operating conditions

Building services systems were traditionally designed for a single function, resulting in separate concepts and control strategies for heating and cooling. With the increasing deployment of heat pump technology as the main energy supply component in both new and existing buildings, these two functions can now be integrated within a single system. However, the resulting increase in system flexibility is accompanied by higher dynamics and stronger interactions between subsystems. Therefore, advanced control approaches are required to coordinate generation, storage and distribution across all operating modes while optimising efficiency, operating costs, emissions and thermal comfort simultaneously.

Background and Motivation

The scientific foundation of the following analyses is based on the research findings presented in [1]. These studies demonstrated that existing heating systems, in which heat is typically distributed via conventional radiators or free heating surfaces, can in principle also be used for space cooling. This provides a cost-effective and resource-efficient approach to extending existing residential buildings with a cooling function. The results were published, among others, in [2] and serve as the basis for the subsequent research project KUEHASystem [3], which focuses on the practical testing of the system solution developed in [1].

The accompanying long-term monitoring revealed that the year-round operation of existing heating systems with combined heating, domestic hot water and cooling functions leads to high system complexity. This results from the simultaneous coordination of multiple heat and cold generation units, the integration of weather and load forecasts, dynamic energy and CO₂ tariffs, the assurance of domestic hot water hygiene, dew point-controlled cooling via free heating surfaces, and the utilisation of waste heat generated during cooling operation. These findings highlight the need for a superordinate control concept that integrates all subsystems of a building, considers their interactions and enables a holistically optimised mode of operation.

This paper introduces a manufacturer-independent system controller that addresses these requirements. It combines heating, domestic hot water and cooling functions within a transparent and adaptive operational framework that equally considers efficiency, flexibility and grid compatibility. The objective is to ensure thermal comfort throughout the entire year while achieving reductions in both operating costs and emissions, thereby contributing to the transformation of the building stock towards a climate-neutral heat supply.

Conceptual Design

The developed system controller is based on a hierarchical and modular architecture that is divided into three levels: the decision level (Level 1), the process level (Level 2) and the field level (Level 3). Each level performs specific functions within the overall control structure and is interconnected through defined communication interfaces.

The following section focuses on the conceptual design of the decision level, as illustrated in Figure 1. This level forms the central logic of the controller and is fully implemented in Python. The software architecture follows a modular structure, in which each functional module fulfils a clearly defined task within the control process. It comprises all modules for data acquisition, evaluation, forecasting and decision-making.

At the beginning, parameter.py initializes the operating parameters and imports user-defined settings and database content. Subsequently, data_fetch.py reads current measurements such as temperatures and meter readings, which are complemented in forecast.py by weather and load forecasts. These data sets are then processed in set_mode.py for operating mode selection and in availability.py for availability checks, where energy sources, generator states, dynamic tariffs and possible faults are assessed. Based on this information, decision.py selects the optimal heat generation system, while budget.py ensures weather-adjusted budget tracking for heating and domestic hot water operation.

Figure 1. Flowchart of the system controller (macro level).

Communication between the decision level and the process level is established via the modbus.py module, which serves as the interface to the system peripherals and transmits control commands both digitally and analogously. In addition, all operating data are continuously analysed, visualised and processed in observer.py to evaluate the system’s operational status.

The process level (Level 2) is responsible for the signal-based execution of control commands and ensures a defined fallback strategy in the event of faults or communication failures. In such cases, a local controller maintains heating and domestic hot water operation. The field level (Level 3) represents the physical interface to the plant components, where heat generators, pumps and valves are actuated directly and all relevant measurement data are recorded.

A specific module, cooling.py, is integrated for cooling operation. It implements a dew point-controlled regulation of the supply temperature, allowing existing heating surfaces to be used for cooling without the risk of condensation. If geothermal sources are available, the heat extracted during cooling operation is directed back into the borehole field, enabling seasonal regeneration and improving the efficiency of the heat pump system during the subsequent heating period.

The presented controller architecture thus integrates all key functions required for cross-sectoral system optimisation within an open and extensible system structure. It forms the foundation for the methodology described in the following chapter, which details the development and validation process of the controller.

Methodology

The development and validation of the controller were carried out in a multi-stage, iterative process based on the combination of Hardware-in-the-Loop (HIL), Software-in-the-Loop (SIL) and field testing. The objective of this methodological approach is to first validate the interfaces, test the functions and then gradually evaluate the controller’s functionality under realistic conditions while continuously optimising the control logic.

In the HIL phase, a demonstrator was set up as shown in Figure 2, replicating the real communication and signal structure of the system controller. The objective was to verify the interfaces, analyse the data flow between the control levels and perform initial validation of the controller’s functional response to changing operating conditions. In this process, typical operating states and fault scenarios were primarily simulated.

Figure 2. Hardware-in-the-loop demonstrator for interface validation.

The setup of the demonstrator follows the three-level architecture illustrated in Figure 3. At the supervisory level (Level 1), a Raspberry Pi 5 serves as the central processing unit running the Python-based system controller. Local parameterisation and visualisation are carried out via a connected touch display. Communication with the automation level (Level 2) is established via Modbus TCP[1] through an Ethernet switch.

The automation level is implemented using a C.M.I.-module[2] and a UVR16x2 controller from Technische Alternative. It functions as a signal gateway between the supervisory level and the field level (Level 3). Data exchange between the UVR16x2 and the field devices takes place via digital and analogue signal lines, as well as sensor cables connecting temperature, humidity and flow sensors. An M-Bus[3] extension integrates the heat, gas and electricity meters. Power supply and switching logic are provided via 230 V/24 V and 0–10 V signal and supply lines, respectively.

The field level includes the connected control units of the heat generators (heat pump and boiler control), pumps, valves and various sensors for temperature, humidity, metering and flow measurement. This setup reproduces the complete signal flow of the real system within the demonstrator.

Figure 3. Schematic representation of the demonstrator for interface testing (HIL).

The data flow diagram shown in Figure 4 illustrates the digital coupling of the HIL environment with external services. The system controller at Level 1 reads measurement data such as temperatures and meter readings every 60 seconds via Modbus-TCP and stores them locally in an InfluxDB^[4]. The measurement and control data stored there are visualised using Grafana^[5]. If required, this information can also be transmitted to other services via MQTT^[6].

The demonstrator thus enables a comprehensive examination of the hardware and software interfaces, protocol integration, and temporal data behaviour between all control levels. By combining real hardware and virtual simulation, control algorithms and communication routines can be validated, tested, and optimised prior to field implementation.

Figure 4. Data flow diagram (HIL).

Building on the HIL phase, the controller was subsequently tested in a simulation-based environment using Polysun [4]. The simulation software was extended by a specially developed plug-in that integrates the complete Python implementation of the system controller. This allowed the actual controller code to be executed directly within the simulation environment, establishing a seamless link between system simulation and control algorithm. With this coupling, the entire system configuration of the field test installation could be modelled in detail (Figure 5) while simultaneously testing the developed control algorithm under reproducible boundary conditions. In this way, both the functional logic and the interface communication, as well as the control sequences, could be verified under virtual operating conditions.

The dynamic behaviour of the controller could be observed in real time, and the interaction with the simulated system components could be analysed with high precision. In the future, the simulation environment is to be extended by coupling Polysun with TRNSYS-TUD [5]. While Polysun provides a realistic representation of the thermo-hydraulic system components, TRNSYS-TUD enables a detailed simulation of the building’s thermal behaviour, including dynamic heat and moisture storage processes. This will make it possible to analyse the interaction between building and system operation within a unified, controller-driven environment, further optimising the control strategy and preparing it for field implementation. For this coupling, the existing controller code of the system controller will be extended to enable bidirectional data exchange between both simulation platforms. The controller thereby acts as a central interface, ensuring consistent simulation of the overall behaviour

Figure 5. Simulation model in Polysun (SIL).

Since the end of October 2025, the system controller has been operating in parallel with the SIL phase under real system conditions. The final validation is being carried out in a multi-family residential building with 36 dwelling units located in Leipzig. The building is representative of a large proportion of the German housing stock and has been equipped with system technology that allows comprehensive testing of the developed controller under real operating conditions. The recorded measurement data are used both to assess the energy performance and to evaluate the control accuracy and user comfort parameters. All operating data are collected and analysed via the existing InfluxDB/Grafana environment, enabling continuous monitoring and real-time analysis of system performance.

The system configuration of the field test installation is shown in Figure 6. It consists of two heat pump units connected on the source side to a borehole field with 16 probes, each 100 m deep. In addition, a gas condensing boiler is installed, which operates as a backup system for both space heating and domestic hot water (DHW) preparation. Targeted hydraulic adjustments ensure that each of the three energy supply components can operate either as a stand-alone system or as part of a redundant configuration. Furthermore, an electric heating element is available as an additional backup for DHW generation.

The building is cooled using the existing free heating surfaces. The borehole field serves as the heat sink, enabling cooling operation independently of the heat pump compressors. During the summer months, the extracted heat is utilised either for regenerating the borehole field or as a heat source for DHW preparation via one of the heat pumps. Figure 7 shows the installed heat pumps of the field test installation, with the new system controller on the right and the previous standard control unit on the left. The field test installation enables a comprehensive evaluation of the controller under real operating conditions and provides the basis for comparing its control performance with the existing control logic, thereby quantifying the potential of the developed approach.

Figure 6. 3D model of the field test installation.

Figure 7. System controller in the field test installation

Summary and Perspectives

The aim of this work was to develop a universally applicable controller concept that enables year-round operation of heating, cooling and domestic hot water systems based on a hierarchical architecture. The focus of the development work was on existing buildings, particularly multi-family dwellings, in which conventional energy supply systems are increasingly being complemented by installations utilising renewable energy sources. The concept allows for the superordinate control of such bi- or multivalent energy supply systems and ensures target-oriented operation in accordance with operator-defined priorities.

For validation, a multi-stage development process was implemented, systematically combining hardware-in-the-loop experiments, simulation studies and field testing. The resulting knowledge transfer between model-based analysis and experimental investigation enables a comprehensive evaluation of the controller concept under both reproducible and real operating conditions. The simulation studies conducted so far indicate that the modular, hierarchically structured control logic supports stable and adaptive system operation while improving efficiency, cost-effectiveness and user comfort. Moreover, the combination of decision, process and field levels creates a transparent and traceable operational structure that offers new possibilities for both scientific analysis and practical application. The ongoing field tests serve to verify these assumptions and to assess control accuracy, robustness and practical applicability.

The findings of these investigations form the basis for the final evaluation of the system controller and contribute to the further development of the concept. In doing so, an approach has been established that supports the transformation of the existing building stock towards a climate-neutral heat supply without requiring extensive system modifications. Future work will focus on extending the controller with adaptive and self-learning components that enable enhanced responsiveness to changing boundary conditions.

References

[1]     EnOB: KUEHA – Testing and demonstration of an innovative system solution for summer space cooling with special consideration of energy efficiency and practical applicability (Project duration: 06/2017 – 05/2020). Funding indicator 03ET1461A.

[2]     M. Arendt, A. Kremonke, R. Gritzki, A. Perschk, L. Haupt, and C. Felsmann, “The KUEHA project – A new solution for space cooling during summer,” REHVA Journal, vol. 55, no. 2, pp. 39–44, Apr. 2018.

[3]     Joint Research Project: KUEHASystem – Year-round overall system optimization for reducing CO₂ emissions of existing heating systems – Demonstration of a system solution for heating and cooling; Subproject: System analysis. Funding indicator 03EN6010A.

[4]     Vela Solaris AG, Polysun Simulation Software.

[5]     TU Dresden, Chair of Building Energy Systems and Heating Supply, TRNSYS-TUD – Extension of the TRNSYS simulation program for building services applications.