基于FT-MILP算法的大规模水电中长期优化调度研究

Study on Medium and Long-term Optimal Dispatching of Large-scale Hydropower Based on FT-MILP Algorithm

随着国内水利事业的快速发展,水电系统规模不断扩大,已有的模型设计通用性不强,优化调度维数灾问题日益凸显。为此,采用森林结构(FT)描述水电站群拓扑结构,利用混合整数线性规划算法(MILP)对模型中非线性函数进行分段线性插值处理,构建了基于FT-MILP算法的大规模水电系统中长期优化调度模型。以我国四川地区大型水电基地277座水电站为例,将所提FT-MILP算法与传统POA算法进行优化调度计算,并将优化调度结果与实际运行结果相比。结果表明,在丰、平、枯水年,大规模水电场景下,FT-MILP算法的水库群总发电量分别为传统POA算法的99.27%、99.14%、99.55%,计算效率约为传统POA算法的20倍,丰枯电量比分别由实际运行的2.26∶1、2.13∶1、1.98∶1减少为1.79∶1、1.70∶1、1.62∶1。因此FT-MILP算法在牺牲少量理论发电效益的同时,大幅缩短了计算耗时,且优化了丰枯出力结构,对于研究我国巨大规模的水电优化调度问题有一定的借鉴意义。

With the rapid development of domestic water conservancy,the scale of hydropower system is expanding.The existing model design is limited,and the problem of dimensional disaster in optimal scheduling is becoming more and more prominent.The forest topology(FT)was used to describe the topology of hydropower station group,and the mixed integer linear programming(MILP)algorithm was used to construct a medium and long-term optimal scheduling model for large-scale hydropower system by segmented linear interpolation of nonlinear functions in the model.Taking 277 hydropower stations in large hydropower bases in Sichuan area of China as an example,the proposed FT-MILP algo-rithm and the traditional POA algorithm were used for optimal scheduling calculations,and the results of optimal schedu-ling were compared with the actual operation values.The results show that in the wet,normal and dry years,under the large-scale hydropower scenarios,the total power generation by the FT-MILP algorithm is 99.27%,99.14%and 99.55%of the traditional POA algorithm,respectively,and the computational efficiency is about 20 times of that of the traditional POA algorithm.The dry and wet electric quantity ratio is reduced from 2.26∶1,2.13∶1 and 1.98∶1 in the actual operation to 1.79∶1,1.70∶1 and 1.62∶1.Therefore,the FT-MILP algorithm significantly shortens the compu-tation time while sacrificing a small amount of theoretical power generation benefits,and optimizes the power in wet and dry season,which is of reference significance for the study of China's huge scale hydropower optimal scheduling problems.