Hardware-in-the-Loop for Building Energy Systems
Tomorrow's building energy systems will be more complex and diverse. We bring an entire building energy system with our hardware-in-the-loop approach into our labs, combining the benefits from the field with the advantages of the lab.
Within our hardware-in-the-loop approach for building energy systems, we combine real hardware with real-time dynamic simulation models. The boundary between hardware and simulation is freely selectable. However, complex components of the energy conversion system, for example, a heat pump or a PV inverter, are usually installed as real hardware in the laboratory. Components that are not part of the technical supply system, such as the building envelope, can be represented by a dynamic simulation. After coupling the existing hardware with the modeled software, the states of the components are interdependent. For example, if the heat pump is in a faulty state, the rooms in the building cool down over time. Thus, highly dynamic tests are feasible under broad boundary conditions and high repeatability.
In addition to economic and ecological key figures relating to the energy conversion system, our infrastructure also allows us to evaluate variables from the simulation, such as user comfort. Frequently, annual key figures are of particular importance and thus pose great challenges to laboratories, which we address scientifically. Our developed methodology uses a mathematical clustering procedure to determine representative days to which a building energy system is subject over the year. Our hardware-in-the-loop method allows us to determine annual key figures within a few days.
In a nutshell, we can connect a component or software under test to any building at any location to form a building energy system and evaluate it by holistic key figures related to one year.
Test benches for emulation of the boundary conditions
The boundary between real hardware and digital components can be flexible in our process. At each boundary intersection, the current digital data must be transferred to a real emulation. For the heat distribution system, for example, the calculated heat flows must be provided through a water volume flow measurement and a temperature reading from a test bench. The ambient conditions must be emulated, which can significantly influence the heat generation system. A uniform network protocol is used for digital interfaces.
These challenges are solved in our laboratories by three hardware-in-the-loop test benches. The test benches consist of a hydraulic test bench, a climate chamber, and a cloud data infrastructure. The sensor measurements and actuator values are read out and controlled at a high frequency using state-of-the-art PLC technology. In addition, innovative control algorithms in higher programming languages are used to meet the high requirements of the dynamic measurement method.
The hydraulic test benches consist of up to eight circuits that can be operated in parallel and are characterized by different performance classes. This allows different performance classes to be covered, which are used to emulate the heating and domestic hot water requirements of single-family and multi-family buildings. The coupling of the test benches to the district heating and cooling networks of the RWTH campus enables a stable energy supply.
The climatic chambers of the test benches allow precise control of the air temperature and humidity. For this purpose, a room air conditioning system is used, which can generate an ambient temperature between -20 °C and +40 °C and relative humidity between 20 % and 95 %. For this purpose, electric heating coils, steam humidifiers, and several heat exchangers are used. A low-temperature refrigeration machine supplies for this the air conditioning system.
In addition, we use a test bench for electrical components such as PV systems and battery systems. This test bench handles the distribution of electrical energy and records all electrical energy flows with high precision. Both electrical generation and electricity consumption can be emulated.
Simulation
High-resolution, dynamic simulation models represent the simulation side in the hardware-in-the-loop process. For this purpose, models in the Modelica modeling language are predominantly used. Programming languages such as Python or Matlab/Simulink can also be used for digitally integrated controllers.
Buildings can be displayed in room resolution at the lowest level of detail. External boundary conditions such as temperature and solar radiation are precisely considered. Models are available for the heat transfer system for radiator systems, underfloor heating systems, and concrete core activation. Each heating circuit can be controlled by a controller integrated into the model or by other software. Other components that can be simulatively integrated are thermal storage units, photovoltaic and solar thermal systems, battery systems, and geothermal probes.
Communication and data infrastructure
On the one hand, the communication and data infrastructure connects all components and forms a monitoring and surveillance platform. The data streams are transmitted via the open-source network protocol MQTT. All data is sent to a central broker and distributed from there. The storage of all data is handled by time series and SQL databases. For visualization, modern web applications are implemented, which are accessible from the World-Wide-Web. This enables real-time monitoring and visualization, also for our project partners.