Multi-objective Predictive Control for cascade canal system considering constraints and objectives related to gate regulation

This paper addresses the complex constraints and multi-objective requirements faced in the operation process of the Cascade Canal System (CCS), and proposes the Multi-Objective Model Predictive Control (CCS-MOMPC) method that can simultaneously consider the gate control constraints and gate regulation frequency. The proposed method modifies the original predictive control objective function by incorporating a gate regulation penalty term, and directly constrains both the gate deadband and the gate regulation frequency. Furthermore, the multi-objective function is converted into a single-objective function using the weighted method, which is then solved employing the Particle Swarm Optimization (PSO) algorithm. The proposed method is applied in the last eight canal pools of the Middle Route Project of South-to-North Water Diversion (MRP-SNWD). The results show that under the experimental conditions, compared with the traditional method, the proposed method can reduce the maximum water level deviation at control point from 0.15 m to 0.10 m, and decrease the total gate control frequency by 37.1%. In the case of unknown secondary disturbance, the proposed method can reduce the final action time of the gate by 77.5%. The results of this paper show that the improved control method can significantly improve the water level control accuracy and reduce the frequency of gate regulation.