Data-driven based predictive controller design for vapor compression refrigeration cycle systems

Data-driven control approaches have been widely applied in the modern industrial process control. A multivariable data-driven based controller using model predictive control strategy for the vapor compression refrigeration cycle system is proposed in this paper. For the purpose of further simplification of the controller design, a 3rd-order model has been derived by model identification method, which is based on the input/output data, and its accuracy has been confirmed by the comparisons of dynamic response characteristics among the nonlinear model, full order linearized model and the reduced-order model. The effectiveness of the proposed controller is verified on an experimental system.