基于特征相关性的牵引系统主回路接地故障诊断  

Feature Correlation-based Ground Fault Diagnosis Method for Main Circuit of Traction System

作　　者：陈志文[1,2,3] 李学明 徐绍龙 彭涛 阳春华[1] 桂卫华[1] CHEN Zhi-Wen;LI Xue-Ming;XU Shao-Long;PENG Tao;YANG Chun-Hua;GUI Wei-Hua(School of Automation,Central South University,Changsha 410083;the Peng Cheng Laboratory,Shenzhen 518000;State Key Laboratory of High Performance Complex Manufacturing,Central South University,Changsha 410083;College of Mechanical and Vehicle Engineering,Hunan University,Changsha 410082;Zhuzhou CRRC Times Electric Co.,Ltd.,Zhuzhou 412001)

机构地区：[1]中南大学自动化学院,长沙410083 [2]鹏城实验室,深圳518000 [3]中南大学高性能复杂制造国家重点实验室,长沙410083 [4]湖南大学机械与运载工程学院,长沙410082 [5]株洲中车时代电气股份有限公司,株洲412001

出　　处：《自动化学报》2021年第7期1516-1529,共14页Acta Automatica Sinica

基　　金：国家自然科学基金(61803390,61790571,61773407);轨道交通节能控制与安全监测湖南省重点实验室(2017TP1002);高性能复杂制造国家重点实验室自主研究课题(ZZYJKT2020-14);博士后基金(2019T120713)资助。

摘　　要：本文针对目前机车、动车牵引系统中主回路接地故障的精确定位问题,提出了一种基于特征相关性的故障诊断方法.该方法通过在线计算与故障关联的特征变量,提取相关故障特征指标,并考虑各故障特征指标间的相关性,利用典型相关分析得到残差,以实现快速故障检测.进一步,构建基于残差方向的故障隔离方法,实现准确地故障定位.现场实验表明,与传统基于相关性的故障诊断方法以及实际工程应用方法相比,在存在较大测量噪声与暂态工况变化时,本文所提方法能实现更好的故障检测与隔离性能,具有良好的应用价值.A fault diagnosis method based on feature correlation is proposed in this paper to accurately locate the main circuit ground fault in the traction system of electrical locomotive and electric multiple unit(EMU).The characteristic variables and fault features associated with faults are calculated online,and canonical correlation analysis(CCA)is carried out to generate residual signal based on the correlation among the fault features to achieve fast fault detection.Accurate fault location is achieved based on the residual signal direction method.Field tests show that,compared with traditional CCA-based and on-board fault detection method,the proposed method has better fault detection and isolation performance in the presence of large measurement noise and transient condition changes and is also applicable to practice.

关 键 词：接地故障 特征相关性 典型相关分析 故障隔离 

分 类 号：U269[机械工程—车辆工程]