Abstract: Refill conflicts among the cascade reservoirs in the lower Jinsha River and Three Gorges have become increasingly prominent in dry years. To raise water use efficiency, this study constructs a multi-objective impoundment scheduling model and solves it using the Non-dominated Sorting Genetic Algorithm II (NSGA-II). Based on the daily reservoir inflow data restored for Aug. 1 to Oct. 31 in 1940-2020, we simulate and compare three impoundment scheduling schemes for these reservoirs, i.e., original design scheme, joint operation scheme, and optimal operation scheme. Results indicate the optimal scheme is the highest in average annual power generation and storage rate. Its power generation is increased by 12.258 billion kWh (or 10.35% improvement) and its refill rate increased from 95.06% to 98.81% compared with the original scheme. In normal and dry years, its flood control pressure is relatively lower, and its comprehensive benefits are more significant. Generally, a mutual constraint exists between power generation and refill rate. In the dry years, we can achieve more benefits from a better joint impoundment scheduling of cascade reservoirs by starting reservoir impoundment earlier and raising the flood control water level.

Key words: optimal operation, joint impoundment, comparative study, cascade reservoirs, comprehensive benefits