Zaimin Ren ¹ Bin He ^2* Chenchen Yao ¹ Xiaoshuai Lv ¹ Xihong Wang ²

• ¹ PowerChina Huadong Engineering Corporation Limited, Hangzhou, Jiangsu Province, China
• ² Dalian University of Technology, Dalian, China

The comprehensive cascade reservoir system is extensively used for flood control and power generation. Traditional reservoir scheduling often treats these tasks as competing objectives, prioritizing flood control during the flood season and neglecting power generation benefits. This approach, primarily applied to large floods, leads to hydraulic resource wastage and power generation loss during minor and moderate floods. The lower Jinsha River cascade reservoirs have significant flood control storage capacity, allowing part of the capacity to be used for maintaining higher water levels without compromising safety under minor and moderate floods within a 20-year return period. This paper proposes an optimal scheduling model that considers both flood control and power generation tasks for the lower Jinsha River cascade reservoirs. The Multi-objective Particle Swarm Optimization (MOPSO) algorithm was used to find non-inferior solution sets. An evaluation index system was developed to select optimal solutions under different flood frequencies, using the Minimum Discriminant Information Principle and VIKOR model. The results indicate that the optimal scheduling scheme under 20-year, 10-year, and 5-year return period floods can enhance power generation by 1.64, 1.71, and 1.35 billion kWh, respectively, compared to conventional scheduling. This approach supports coordinated flood control and power generation scheduling, contributing to the high-quality development of the Yangtze River Economic Belt.