基于EMD聚类算法的配电网单相接地故障选线方法  

Faulty Line Selection Method Based on EMD and Cluster-algorithm in Distribution Network

作　　者：廖桂源 李彩林[1] 施伟[1] 孙新望 

机构地区：[1]桂林电子科技大学机电工程学院,广西桂林541004

出　　处：《电力科学与工程》2014年第2期57-61,共5页Electric Power Science and Engineering

基　　金：广西省自然科学基金资助项目(2012GXNSFAA053195);广西区基金(2013GXNSFAA019301)

摘　　要：谐振接地系统发生单相接地故障时故障信号不易处理。为准确选出故障线路,提出一种基于EMD算法和以欧氏距离为基础的聚类算法相结合的故障选线方法。根据谐振接地系统单相接地故障后零序暂态电流的特性——故障线路的零序暂态电流幅值远大于各健全线路的零序暂态电流幅值并且极性相反,首先对各馈线暂态零序电流进行EMD分解并提取主振荡频带电流信号,然后求取各馈线主振荡频带电流信号的综合欧氏距离,以此为基础对各馈线进行排序和聚类,由聚类结果确定故障线路。仿真结果表明,该方法准确、可靠。The single-phase-to-ground fault signal is not processed easily in resonant earthed system. In order to choose the fault feeder accurately,a new method of detection based on the empirical mode decomposition （EMD） and the cluster-algorithm that based on Euclidean distance is proposed. According to the characteristic of zero se-quence transient current of resonant earthed system with single-phase-to-ground fault：the zero sequence transient current amplitude of the fault feeder is much larger than the perfect feeders＆#39;,and the current polarity is opposite between the fault feeder and the perfect feeders;first,the zero sequence transient current of the feeder is decom-posed with EMD to get its main oscillation frequency current,and then the comprehensive Euclidean distance of the main oscillation frequency current is calculated to sort and cluster the feeders ,so the fault feeder can be detec-ted by cluster-algorithm. The simulation results show that the method is reliable and accurate.

关 键 词：谐振接地系统 故障选线 EMD算法 欧氏距离 聚类算法 

分 类 号：TM773[电气工程—电力系统及自动化]