Approaching simulation-based controller design: heat exchanger case study

Abstract

Supported by an identification experiment using random-phase multisines, uncertain parameters in a grey-box model for a multiple-input multiple-output laboratory-scale heat exchanger are fitted to experimental data. By defining desired trajectories for the controlled system concerning setpoint changes, simulations and a cost function taking control signal activity into account, we determine both a linear and a nonlinear PI-controller. The resulting control systems are evaluated through practical experiments and analysis with encouraging results. The approach to modelling and controller design raises questions about what is needed from an educational point of view, e.g., what skills are needed for simulation-based control design and analysis?