Modular hydronic subsystem models for testing and improving control algorithms of air-handling units

Abstract

To reduce the energy consumption of building energy systems, efficient and well-tuned control strategies are needed. However, developing and benchmarking control strategies is a challenging task. Here, simulation-based controller tuning and testing is beneficial and more cost-efficient compared to experiments. However, generating a model for this purpose is time-consuming and profound expert knowledge is needed. An approach to reduce the modeling effort is to divide the overall system into frequently used subsystems and develop one model for each type of subsystem. In this paper, we present modular models for frequently used hydronic subsystems in air-handling units (AHU). The models are implemented in the modeling language Modelica with a high level of detail and provide a standardized connection interface to different controller models for testing and tuning. The developed subsystem models are validated using measured data of a real-world air-handling unit. The results show that the models have nearly the same static and dynamic behavior as the real system. The mean absolute percentage error (MAPE) between the simulation and measured data is in the range of 0.14% to 11.3% for all compared sensor values. In two case studies, the usage of the models for controller testing and tuning is demonstrated: the first case study deals with the tuning of a PI controller, and in the second case study we compare two control strategies regarding the control quality and energy consumption.