大规模混联水库群长期优化调度

Optimal Operation Method of Large-scale Multi-reservoir Group

混联水库群长期优化调度问题是一个典型的高度非线性、多维、多时段的优化问题。针对问题特点,应用水力关联矩阵描述混联水库群之间的水力联系,进而应用逐次优化算法(POA)和动态规划逐次渐近(DPSA)算法对该问题在时间和空间上进行降维。以金沙江、雅砻江和大渡河干流10座具有季调节能力以上调节性能的混联水库群作为研究对象,分别采用DPSA和POA-DPSA两种算法对问题进行求解,并初步分析了联合调度对径流的影响。优化结果相对于设计值都有一定程度的提高,且寻优速度较快。

Over the past decades,operation optimization of multi-reservoir group has been a major area of water resource studies. It is a typical problem of highly nonlinear,multi-dimensional and multi-periods optimizations. In order to solve the problems mentioned above,a hydropower correlation matching matrix,which can better describe the spatial location relationship and hydraulic connection of reservoirs is proposed. The idea is applied to 10 reservoirs built on Three River （Jinsha River,Yalong River,Dadu River） Basin. After establishing the hydropower correlation matching matrix of the 10 reservoirs,the DPSA and POA-DPSA are used to reduce the time-space dimensions. The optimization results demonstrate a better solution and faster calculation speed in comparion with the old method.