水电站群联合调峰调度均匀逐步优化方法

Uniform Progressive Optimality Algorithm for Hydropower Station Group Peak Operation

由于受到系统规模逐年扩大、调度要求日趋精细和约束集合日渐庞杂等诸多因素影响,如何快速制定合理的发电调度计划是水电站群联合调度面临的关键科学难题,集中体现在初始解的快速生成机制、决策空间的科学辨识模式以及行之有效的优化调度方法等3个方面。为此,从工程实用化角度出发,提出求解水电站群联合调峰调度问题的均匀逐步优化方法。该方法首先依据水电系统相邻日运行工况估算各水电站可能发电能力,并由上游到下游依次采用切负荷方法来快速生成初始解;然后将多阶段决策问题分解为若干两阶段子问题进行求解,并将均匀试验设计耦入各阶段决策变量集合的构造工作,指导算法优选少数富有代表性的决策变量,大幅减少运算量与存储量,进而降低方法的计算复杂度;同时,在寻优过程中将多重复杂运行约束统一集成至指定约束,以动态辨识可行搜索空间,进而降低算法的计算消耗。工程应用结果表明,所提方法可快速获得满意调度结果,能够切实服务于水电调度运行工作的开展。

Due to the growing system size, the increasing scheduling requirements and complex constraints sets, how to develop scheduling plans is the key problem for the operation of hydropower station group. It was reflected in three primary difficulties： the initial solution generation mechanism, the decision space identification method and the efficient optimization algorithm to improve the scheduling results. Thus, this paper proposed the uniform progressive optimality algorithm for the peak operation of hydropower plants. Firstly, the potential generating capacity of each hydropower plant was estimated by its actual working conditions, then the cutting load method was used for the initial solution. Secondly, the multi-stage problem was divided into several two-stage sub-problems, then the uniform experimental design was used to select some representative decision vectors to cut down the computation complexity, and all the constraints were transformed into the flow restricted condition to reduce the searching space. The method would continue the process until the terminal condition was met. The optimization of hydropower system in Guizhou power grid was used to test the validity of the method. The results show that the method is capable of obtaining solutions with higher quality than conventional optimization method.