基于多目标优化模型的省级电网月度发电计划  

Monthly Generation Scheduling Method Based on Multi-Objective Optimization Model for Provincial Power Grid

作　　者：梁志飞[1] 夏清[1] 许洪强[2] 朱明祥[2] 张健[2] 杨明辉[2] 

机构地区：[1]清华大学电机系电力系统及发电设备控制和仿真国家重点实验室,北京市 海淀区 100084 [2]山东电力调度中心,山东省 济南市 250001

出　　处：《电网技术》2009年第13期90-95,共6页Power System Technology

基　　金：国家自然科学基金资助项目(50877041);新世纪优秀人才支持计划资助(NCET-07-0484)~~

摘　　要：针对电力体制新形势下多种调度模式的发电计划科学化管理要求,基于月度电力生产的时空均匀性原则提出了月度发电计划的多目标优化模型。针对以往研究中所忽视的电量供需平衡和潮流安全的同时性问题,把电量分解与安全校核细化到各日的峰、平、谷时段。针对多量纲、多数量级的多目标综合决策的要求,基于理想点贴近度的思路将多目标问题转化为单目标优化问题。根据模型特点,采用邻域搜索算法求解。最后采用山东电网实际算例验证了模型的综合决策性能和邻域搜索算法的优越性。To satisfy the requirement of scientific power generation scheduling management in conditions of multi-dispatching modes under the new situation of electric power sector, based on the principle of spatial-temporal uniformity a multi-objective optimization model for monthly generation scheduling is proposed. Considering the simultaneity of equilibrium of power supply and demand as well as power flow security which was neglected in previous researches, the contact volume distribution and security check are refined on peak, flat and valley period in each day. According to the requirements of comprehensive multi-objective decision possessing multi-dimensions and multi-magnitude orders, based on the thinking of closeness degree of ideal point the multi-objective problem is transformed into single-objective problem. Based on the feature of the proposed model, the proposed model is solved by neighbor search algorithm. The comprehensive decision-making performance of the proposed model and the superiority of neighbor search algorithm are validated by an actual computation example of Shandong provincial power grid.

关 键 词：多目标优化 月度发电计划 理想点贴近度 邻域搜索算法 

分 类 号：TM734[电气工程—电力系统及自动化]