基于线路暂态重心频率的配电网故障区段定位  

Fault Section Location of the Distribution Network Based on Transient Center Frequency

作　　者：张姝[1] 杨健维[1] 何正友[1] 李小鹏[1] 

机构地区：[1]西南交通大学电气工程学院,四川省成都市610031

出　　处：《中国电机工程学报》2015年第10期2463-2470,共8页Proceedings of the CSEE

基　　金：国家自然科学基金青年基金(51207130);中央高校基本科研业务费专项资金(2682013CX015)~~

摘　　要：针对小电流接地系统接地故障区段定位问题,提出一种基于线路暂态重心频率的配电网故障区段定位算法。在推导故障线路上游和下游导纳随频率变化关系的基础上,利用故障点两侧线路零序电流重心频率各异的特征,实现故障区段的判定。为减小快速傅里叶变换频谱分析存在的频率提取误差,采用K-means聚类算法,对各区段线路暂态重心频率进行聚类分析。当在特殊故障点时,通过补充重心频率的幅值信息确定故障区段。PSCAD/EMTDC仿真表明,无论是中性点不接地系统还是谐振接地系统,所提方法均能准确判定出故障位置。该方法只需要采集零序电流的信息,不需要各检测点时间同步,具有上传信息量少和一定容错能力的优点。A new fault section location algorithm was proposed in order to solve the problem of the fault section location in neutral non-effective grounded system of the distribution network. The relationship between oscillation frequency of the admittance and the location of fault section was derived in a single-phase ground fault. The fault section was determined based on the difference of the center frequency of the zero-sequence transient current in the both sides of the fault point. This paper used K-means cluster to analyze the center frequency to reduce the error of frequency extraction by fast Fourier transformation （FFT） spectrum analysis. When the fault occurred in the special point, the amplitude information of the center frequency was obtained to determine the fault section. Simulations by PSCAD/EMTDC show that the proposed algorithm can determine the fault location accurately in the neutral ungrounded system and the resonant grounded system. The method only requires the information of zero-sequence current without the time synchronization for each measurement points. Its advantage is less upload information required and fault tolerance.

关 键 词：配电网 故障区段 重心频率 小电流接地系统 零序电流 K-MEANS聚类 

分 类 号：TM77[电气工程—电力系统及自动化]