改进聚类分析在电网故障定位中的应用  

Application of Improved Cluster Analysis in Grid Fault Positioning

作　　者：孙玎 陈少华[1] 陈琳[1] 

机构地区：[1]广东工业大学自动化学院电气工程系,广东广州510090

出　　处：《广东电力》2014年第8期37-41,共5页Guangdong Electric Power

基　　金：广东省教育厅电力节能与新能源技术重点实验室资助项目(IDSYS200701)

摘　　要：传统故障定位方法利用开关变位和继电保护动作等遥信量,很难满足时间的要求,而使用故障发生时第一时间发生变化的电网各节点电压、电流等遥测量进行定位的传统聚类分析,虽具有一定的时间优势,但仍需加以改进使故障定位更加快速、准确。对此,首先介绍了广域测量系统(wide area measurement system,WAMS)及聚类分析理论,利用WAMS采集的遥测量,改变了聚类分析的输入样本,通过加入了故障类型判别过程的聚类分析理论,对电网中各种短路故障进行定位研究。算例分析表明,改进的聚类分析算法,可以更加快速、准确地实现故障元件定位。By using remote communication such as switch displacement and relay protection action,traditional fault positio-ning method is hard to satisfy demand for time while though raditional cluster analysis on positioning by means of remote measure such as variational various node voltage and current of the grid in first time of fault is provided with time advanta-ges,it is still to be improved in order to make fault positioning more rapid and correct. Therefore,this paper firstly intro-duces wide area measurement system （WAMS）and cluster analysis theory. By using remote measurements collected by WAMS,input samples for cluster analysis was changed. By introducing cluster analysis theory on fault type identification, positioning research on various short-circuit fault of the power grid was conducted. Example analysis indicated that the im-proved cluster analysis algorithm was able to more rapidly and correctly realize fault elements positioning.

关 键 词：广域测量系统 聚类分析 故障类型判别 故障定位 

分 类 号：TM761[电气工程—电力系统及自动化]