| Title |
Kalman Filter and Multi-Input Multi-Output Based Model Predictive Control for Precise Temperature and High Efficiency Control of Variable-Speed Refrigeration System |
| Authors |
Hyeon Seok Jeong ; Seok Kwon Jeong |
| DOI |
https://doi.org/10.6110/KJACR.2024.36.11.521 |
| Keywords |
베이지안 최적화; 칼만 필터; 모델 예측 제어; 다입출력; 가변속 냉동시스템 Bayesian optimization; Kalman filter; Model predictive control; Multi-input multi-output; Variable speed refrigeration system |
| Abstract |
A variable-speed refrigeration system is widely used in various industries because of its high energy savings during partial load. However, because there is an interference between the target temperature and superheat, it is difficult to control these two variables independently and precisely without a large transient phenomenon that causes energy loss in the event of a sudden change in the control input. In this paper, we applied multi-input and multi-output (MIMO) based model predictive control (MPC) to offset the interference. In addition, a Kalman filter was designed to reduce effects of thermal noise mixed into state variables and output variables required for control logic calculations. Especially, Bayesian optimization was applied to obtain the optimal value without trial and error for design parameters of MPC. The validity of the proposed MPC was verified through simulations and experiments with an oil cooler system. Moreover, the effectiveness of the proposed MPC was demonstrated through comparative experiments of control performance with conventional PI controllers. |