During the rainy season, open-channel irrigation systems (OCISs) in the hilly regions of southern China simultaneously undertake flood discharge and storage tasks, which are critical for flood mitigation, rainwater resource utilization, and long-term water security in climate-vulnerable monsoon regions. However, existing methods typically adopt a decoupled framework that separates optimization calculations from rule corrections, often leading to repeated “optimize–correct–reoptimize” iterations and struggling to coordinate the coupling between channel water level evolution and gate operation rules, resulting in frequent gate movements, intensified water level fluctuations, and elevated operational risks. To address these challenges, this study proposes a hybrid model predictive control method (HyMPC) for flood regulation in irrigation canal systems. The method jointly optimizes discrete gate opening and closing states with continuous water level dynamics within a receding prediction horizon. It employs discrete variables to represent gate states and water level zoning, continuous variables to describe channel water level processes, and an integrator-delay model to establish bidirectional coupling between them, enabling coordinated gate group control under combined flood discharge and storage conditions. Taking the flood event from 17 to 20 July 2020, in the Shi River Irrigation District, Anhui Province, China, as a case study, the proposed method was validated through comparative experiments. Results show that, compared with conventional MPC-based canal control models, the method improves gate regulation smoothness (13.33% reduction in the dimensionless integrated absolute flow change), water level stability (26.08% reduction in the high-frequency component of water level fluctuations), and rainwater resource utilization efficiency (6.98% improvement). Scenario analysis further demonstrates that the method can effectively enhance regulation stability and rainwater resource utilization while ensuring flood safety, providing a robust technical pathway and quantifiable tool for adaptive, integrated flood–drought management in irrigation canal systems.
Hou et al. (Wed,) studied this question.