Design of Multiple Input-Multiple Output Model Predictive Controller for Precise Temperature Control of Variable Speed Refrigeration System

Variable speed refrigeration systems are composed of multiple input multiple output systems with several state variables. These input/output variables and state variables typically have various constraints, making it challenging to design optimal control considering these constraints. However, model predictive control (MPC) has been gaining attention in process control recently due to its ability to optimally control a multi-input multi-output system with a number of constraints and early constraints reflection. It can predict future outputs while taking into account current constraints through quadratic programming optimization with a mathematical algorithm. Subsequently, it can determine the optimal control input. To verify the validity of MPC, the control performance of MPC was compared with conventional PI control satisfying constraints through Matlab simulation and experiment on the oil cooler system.