基于模拟退火的粗糙集K均值电力负荷聚类分析  

Clustering Analysis of Power Load by Rough Set and K-means Based on Simulated Annealing Method

作　　者：刘建华[1] 王进[1] 孟颖[2] 王文生[1] 

机构地区：[1]长沙理工大学电气与信息工程学院,湖南长沙410114 [2]长沙理工大学计算机与通信工程学院,湖南长沙410114

出　　处：《现代电力》2012年第1期10-14,共5页Modern Electric Power

基　　金：国家自然科学基金(71071025)

摘　　要：为了建立合适的变电站负荷模型,将聚类方法引入到负荷特性分析,提出了一种基于模拟退火的粗糙集K均值电力负荷综合聚类算法。该算法将粗糙集理论、模拟退火思想与K均值相结合,通过模拟退火思想优化K均值聚类算法,采用最大最小原则动态生成K均值聚类数和初始聚类中心,同时结合粗糙集理论的上逼近和下逼近处理边界对象。最后,对变电站综合负荷静态特性进行聚类分析,比较类间距离和类内距离,结果验证了本文算法的可行性和有效性。To establish a proper load model of substation,in this paper,a comprehensive clustering algorithm of power load by rough set and K-means based on simulated annealing method is proposed by introducing clustering method to analyze load characteristics.By combining rough set theory,the ideal of simulated annealing and K-means,optimizing K-means clustering algorithm by simulated annealing method,K-means cluster number and initial cluster centers can be obtained dynamically with the maximum and minimum principle,and boundary objectives are dealt with by upper and lower approximation of rough set theory.In the end,through clustering analysis on static characteristics of comprehensive load for substation,the inter-class distance and the intra-class distance are compared,which verifies the feasibility and effectiveness of the proposed algorithm.

关 键 词：粗糙集 K均值算法 模拟退火 电力负荷 聚类分析 

分 类 号：TM714[电气工程—电力系统及自动化]