华东交通大学学报2017，Vol.34Issue(3)：109-117,9.
基于K-means聚类法的牵引供电隔离开关故障状态监测
Fault Monitoring of OCS Isolation Switch Based on K-means Clustering
摘要
Abstract
The operation reliability of isolation switches in the overhead contact system(OCS) is of great signifi-cance for the stable operation of the traction power supply system. According to the mechanical characteristics of the isolated switch,the function relation between the motor current and the operating torque was deduced in this paper. It adopted the motor current acquisition system to obtain the current signal of the isolation switch and by using the Wolf algorithm it found out the maximum Lyapunov index of the current signal is positive,which shows that the current signal has chaotic characteristics. The current signal was reconstructed through the phase space reconstruction theory,which was then analyzed by K-means clustering to obtain the cluster center, and the fault condition of the switch was judged by the change of the cluster center positions. On the above basis,the fault condition of the isolated switch was classified,which may provide reference for fault monitoring and state evalua-tion of isolation switches.关键词
隔离开关/K-means聚类/相空间重构/电流信号Key words
isolation switch/K-means clustering/phase space reconstruction/current signal分类
交通工程引用本文复制引用
刘仕兵,葛俊祥..基于K-means聚类法的牵引供电隔离开关故障状态监测[J].华东交通大学学报,2017,34(3):109-117,9.基金项目
国家自然基金项目(11162006) （11162006）
江西省教育厅科技项目(GJJ150530) （GJJ150530）
江西省教育厅科技项目(GJJ160488) （GJJ160488）
