基于二层规划的电力系统无功优化混合算法研究  

A Hybrid Algorithm for Reactive Power Optimization Based on Bi-Level Programming

作　　者：唐永红[1,2] 李旻[3] 姜振超[1,2] 路轶[3] 范宏[4] 季昕雨 陈斯[4] 

机构地区：[1]国网四川省电力公司电力科学研究院,四川成都610072 [2]智能电网四川省重点实验室,四川成都610072 [3]国网四川省电力公司调度控制中心,四川成都610041 [4]上海电力学院电气工程学院,上海200090

出　　处：《电网与清洁能源》2017年第2期37-43,共7页Power System and Clean Energy

基　　金：四川省电力公司2015科技支撑项目《AVC应用评价系统开发及全程极限控制策略研究》(52199713506C)~~

摘　　要：结合电力系统电压无功综合调控特点,建立无功优化的二层规划模型。上、下层分别选取系统网损最小和各节点电压偏移最小为目标函数,根据上、下层解空间的不同,上层采用原-对偶内点法,下层采用森林算法。搭建基于电力系统全数字实时仿真装置(ADPSS)的无功优化闭环检测平台,综合对比分析文中提出的二层规划程序和OPEN3000中的嵌入式AVC系统的优化策略,结果表明:该算法可以显著将低电力系统的网损,提高电压合格率。In this paper, a bi-level programming model for reactive power optimization is established based on the characteristics of the grid voltage-reactive power control. The mini-mum grid loss and the minimum voltage offset are selected as t he objective functions for the upper-level and lower-level respectively. According to the differences of the solution space of the two levels, the primal-dual interior point algorithm is re-commended for the upper-level while random forests algorithm （RFA） for the lower-level. A reactive power optimization closed-loop detection platform is established based on advanced digital power system simulator （ADPSS） for a comprehensive comparative analysis of the optimization strategy of the bi--level programming proposed in the paper and the embedded AVC system in the OPEN30000.

关 键 词：二层规划 无功优化 内点法 森林算法 检测平台 

分 类 号：TM761.1[电气工程—电力系统及自动化]