PID control loop tuning for three-port mixing valves within building energy distribution systems : application and assessment of different tuning methods under real operation conditions

摘要

We assess different control loop tuning methods to tackle suboptimal building performance due to suboptimal PID control within building automation systems. Advanced control research for building energy systems is usually conducted either simulation-based or with the help of experiments via test benches. Therefore, it is limited in terms of applicability to real systems. Consequently, we conduct real tuning experiments with different PID tuning methods under real operation conditions within a state-of-the-art building energy distribution system. We use experimental step response tests to identify transfer functions for different three-port valves. We theoretically specify controlled systems’ transmission behaviour by functional diagrams, as well as mathematically describe them via characteristic variables according to Ziegler and Nichols and via fitted transfer functions. In order to ameliorate the actual control quality, we apply different tuning rules and adjustment standards: namely lambda-tuning, absolute value optimum method, symmetrical optimum method, Ziegler and Nichols, optimization on different control quality indicators and cognitive parameterization. For each method-system-combination, we experimentally conduct step experiments to evaluate with reference to control quality. Our results show an improvement of control quality up to 88%. Controller parameterization based on cognitive parameterization and lambda-tuning are the most effective methods. By conduction of field experiments under operating conditions, we show how to monitor, evaluate and ameliorate control quality of three-port valves. Therewith, we derive the most effective tuning methods. Further, we conclude that control quality amelioration crucially depends on the applied control tuning method and on knowledge about the behaviour and characteristics of controlled systems.