Data-Driven control design by prediction error identification for a refrigeration system based on vapor compression

Abstract

This paper deals with data-driven control design in a Model Reference (MR) framework for multivariable systems. Based on a batch of input-output data collected on the process, a fixed structure controller is estimated without using a process model, by embedding the control design problem in the Prediction Error (PE) identification of an optimal controller. A multivariable extension of the OCI (Optimal Controller Identification) method is applied in the design of PID controllers for a refrigeration system based on vapor compression, which is the subject of the benchmark process challenge of the IFAC PID 2018 conference. Simulation results show the obtained controllers perform significantly better than the ones provided by the benchmark challenge.