考虑回水顶托影响的梯级库群日前调峰MILP模型

A MILP model for day-ahead peak operation of cascade hydropower stations considering backwater

水电日前调峰优化调度需要同时考虑电网、水库及机组多重约束,本质上属于复杂的多维、动态、非凸、非线性规划问题,将回水顶托影响考虑到模型当中进一步加剧了模型的求解难度。本文建立考虑回水顶托的梯级水库混合整数线性规划(MILP)96点调峰模型,通过将回水顶托水库中多个非线性约束聚合成单个四维超曲面约束的方式减少模型整体非线性约束个数,并采用SOS2约束对该模型进行线性化处理,最终实现模型高精度高效求解。以西南地区某梯级库群为工程背景,选取汛枯期4种典型日方案验证所提模型及算法的有效性,并分析考虑回水顶托的必要性。结果表明,所提模型较传统模型在提高精度的同时也显著提升了求解效率,并在各场景下均能达到较好的调峰效果,对水库精细化调度具有重要意义。

In order to realize day-ahead peak regulation of cascade hydropower stations,multiple constraints of power grids,reservoirs and the hydro units should be considered simultaneously.Essentially,the day-ahead peak regulation problem is a complex multi-dimensional,dynamic,non-convex and non-linear programming problem.After considering the impact of backwater,the difficulty of solving the problem increases significantly.This paper established a mixed integer linear programming(MILP)model for 96-period peak-shaving problem with considering the impact of backwater.In order to reduce the number of non-linear constraints,all the non-linear constraints are aggregated into a single four-dimensional hypersurface constraint,which is linearized by SOS2 constraints method.Taking a hydropower cascade system in Southwest China as an engineering background,four typical days in dry and wet seasons are selected to verify the effectiveness of the proposed model and analyze the necessity of considering backwater.The results show that the proposed model can consider both the accuracy and efficiency of the solution,and obtain better results in all scenarios,which is of great significance to the refined reservoir operation.