三目标混合骨干粒子群算法的电力系统无功优化  

Three-objective Hybrid Bare-bones Particle Swarm Optimization for Reactive Power Optimization

作　　者：马立新[1] 王继银[1] 项庆[1] 黄阳龙 

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093

出　　处：《电力科学与工程》2015年第11期18-23,共6页Electric Power Science and Engineering

基　　金：沪江基金(C14002);上海市张江国家自主创新重点项目(201310-PI-B2-008)

摘　　要：电力系统无功优化通常以降低有功网损和减小电压偏移为目标,建立了综合考虑有功网损和电压偏移最小及电压稳定裕度最大的三目标无功优化模型。首次引入混合骨干粒子群算法用于解决电力系统无功优化问题。该算法利用关于粒子个体极值和全局极值的高斯分布对粒子位置进行更新,再通过K-均值聚类的方式,引入单纯形法对有代表性的粒子进行单纯形搜索,使算法既能够具备较强的全局搜索能力,又能够提高收敛速度和精度。将该算法和其他算法应用于IEEE-14节点系统中进行无功优化,通过数据的计算和比较,结果验证了该模型和算法用于解决多目标电力系统无功优化问题的优越性和实用性。The reactive power optimization in power system usually aimes at reducing active network losses and voltage deviation,while the currently established three-objective reactive power optimization model considers the minimization of active network losses,voltage deviation and the maximum of voltage stability margin. To solve the reactive power optimization problem,hybrid bare-bones particle swarm optimization is introduced in this essay for the first time. The algorithm employs Gaussian distribution of particle individual extremum and global extremum to update particle location. Then it introduces simple method to search for the representative particles by the way of Kmean clustering. This algorithm not only tends to have strong global search ability,but also can improve the convergence speed and accuracy. The results of calculation and data comparison justify the superiority and practicality of this model and algorithm in solving multi-objective power system reactive power optimization problem. It is superior to algorithms when they are implemented on IEEE-14 node system.

关 键 词：骨干粒子群 K-均值 单纯形 三目标优化 电压稳定裕度 

分 类 号：TM714[电气工程—电力系统及自动化]