摘要
为了克服配电混合网络线路结构的复杂性及单相接地故障区段定位通信量大的问题,文中提出利用多点测量暂态相电流信息进行故障区段定位的方法。首先将各馈线线路测量点采集的暂态相电流信息上传到对应线路的馈线终端装置(FTU),利用快速傅里叶变换(FFT)对暂态相电流信息进行谐波变换,得到相应的幅值谱和相位谱;然后计算相邻测量点幅值谱的欧氏距离和相位谱的余弦相似度,作为故障区段定位的数据特征,并将处理后的数据特征上传到主站,利用局部异常因子(LOF)算法进行分析,从而实现故障区段定位;最后通过PSCAD仿真实验表明,能够较好地判别不同工况下的故障区段,且能适应于分布式电源并网、调整中性点运行方式和部分测量点被噪声干扰的情况。
For overcoming the complexity of the structure of the distribution hybrid network line and the large traffic location problem of the single-phase earthing fault section,a method of fault section location based on multi-point measurement of transient phase current information is proposed in this paper.First,the information of transient phase current collected from the measurement points of each feeder line is uploaded to the feeder terminal unit(FTU)of the corresponding line. Then,fast Fourier transform(FFT)is used to perform harmonic transformation of the information of transient phase current to obtain the corresponding amplitude spectrum and phase spectrum. After that,the Euclide distance of amplitude spectrum and cosine similarity of the amplitude spectrum and phase spectrum of the adjacent measurement points are calculated as the data characteristics of the fault section location,and the processed data characteristics are uploaded to the master station and the local outlier factor(LOF)algorithm is used for analysis,thus achieving location of fault section. Finally,it is shown by PSCAD simulation experiment that the location technology can identify fault sections under different conditions,can be appliable to interconnection of distributed power supply and adjust neutral point operation mode and interference of partial measurement point by noise.
作者
汪士杰
孙晓茜
杨延钧
张建华
WANG Shijie;SUN Xiaoqian;YANG Yanjun;ZHANG Jianhua(School ofElectricalEngineering,Zhengzhou University,Zhengzhou 450001,China)
出处
《电力电容器与无功补偿》
2022年第1期111-119,共9页
Power Capacitor & Reactive Power Compensation
基金
2018年度河南省高等学校重点科研项目(18A470017)。
关键词
配电混合线路
单相接地故障
快速傅里叶变换
欧氏距离
余弦相似度
局部异常因子
distribution hybrid line
single-phase earth fault
fast Fourier transform
Euclidean distance
cosine similarity
local outlier factor