基于K-Means聚类算法的井下电缆双端在线局放定位方法
K-means Clustering-based Method for Dual-end Location of Online Partial Discharge of Underground Cables
程维新1
作者信息
- 1. 窑街煤电集团有限公司金河煤矿,甘肃 兰州 730000
- 折叠
摘要
常规的电缆局放定位方法以反射信号识别与定位为主,时间同步偏差相对较大,影响最终的局放定位精准度,因此设计了基于K-Means聚类算法的井下电缆双端在线局放定位方法.该方法通过提取井下电缆双端行波模量特征,将井下电缆局放信号进行相模变换,分析相应电荷气隙平衡条件,获取更加准确的双端局放位置.基于 K-Means算法构造电缆在线局放定位聚类中心,将空间距离相似的电缆进行局放判断,排除异常定位数据对聚类结果的影响,从而减小定位误差.采用对比实验验证了该方法的定位精准度高,能应用于实际生活中.
Abstract
Conventional methods of locating cable partial discharge mainly rely on reflection signal recognition and ranging,with relatively large time synchronization deviation,which affects overall location accuracy.Therefore an under-ground cable dual-end online partial discharge location method based on K-Means clustering algorithm was designed.By extracting characteristics of traveling wave modulus at both ends of the cable,the method performed phase mode transfor-mation of partial discharge signal and analyzes the corresponding charge air gap balance conditions,thereby obtaining a more accurate dual-end partial discharge location.Based on the K-Means algorithm,a clustering center for online partial discharge was constructed,and cables with similar spatial distances were used for partial discharge judgment to eliminate influences of ranging data anomalies and reduce deviations of results.The proposed method was proved by comparative ex-periment highly accurate and practically applicable.
关键词
K-Means聚类算法/井下电缆/双端/在线局放/定位方法Key words
K-Means clustering algorithm/underground cable/dual-end/online partial discharge/positioning method引用本文复制引用
出版年
2024
NETL
NSTL
维普
万方数据