Abstract:The cascade reservoirs in the upstream Yellow River consist of a long-period regulating reservoir, an annually regulating reservoir, and five daily regulating reservoirs. In order to optimize the operating policies of the multipurpose cascade reservoirs given a deterministic inflow forecast, a long-term optimal operation model is developed in this paper. The model's objective is to maximize the total power generation of the cascade reservoirs. The constraints set include some strict equality constraints on the water level, the release and the power generation of the reservoirs, as well as other normal constraints. By employing the penalty functions to relax these strict equality constraints, an improved optimization method combining discrete dynamic programming (DDP) and discrete differential DP (DDDP) is applied to the cascade reservoirs operation. The case study shows that the improved optimization method can effectively handle the deterministic equality constraints and can be used to optimize the long-term operation for cascade multipurpose reservoirs.

An Improved Optimization Method for Long-Term Operation of a Cascade of Multipurpose Reservoirs on the Yellow River, China