基于Fretchet距离与TWSVM的多机牵引道岔故障诊断研究  

Research on fault diagnosis of multi-machine traction turnout based on Fretchet distance and TWSVM

作　　者：吴永成[1] 阳长琼[1] 何涛 WU Yongcheng;YANG Changqiong;HE Tao(Gansu Industrial Traffic Automation Engineering Technology Research Center,Lanzhou Jiaotong Uninversity,Lanzhou 730070,China;Key Laboratory of Opto-Technology and Intelligent Control Ministory of Education,Lanzhou Jiaotong University,Lanzhou 730070,China)

机构地区：[1]兰州交通大学甘肃省工业交通自动化工程技术研究中心,甘肃兰州730070 [2]兰州交通大学光电技术与智能控制教育部重点实验室,甘肃兰州730070

出　　处：《铁道科学与工程学报》2019年第11期2866-2872,共7页Journal of Railway Science and Engineering

基　　金：兰州交通大学青年科学基金资助项目(2017039)

摘　　要：针对多机牵引道岔故障频发、人工诊断困难的问题,提出一种基于Fretchet距离和双向支持向量机的故障诊断方法。首先,设计基于Fretchet距离的相似度指标,通过计算实时曲线与标准曲线的相似度值来定位多机牵引道岔的故障源;其次,利用双向支持向量机一对一型分类算法对已定位的故障多机牵引道岔进行故障类型诊断;通过使用兰州西站采集的数据进行仿真,结果表明该算法能及时、准确诊断多机牵引道岔故障。Aiming at the problem of high fault frequency and difficult manual diagnosis of multi-machine traction turnout,a fault diagnosis method based on Fretchet distance and bidirectional support vector machine was proposed.Firstly,the fault source of multi-machine traction turnout was determined by calculating the similarity between the real-time curve and the standard curve by Fretchet distance.Then,the fault type diagnosis of multi-machine traction turnout was carried out by using two-dimensional support vector machine one-to-one multi-level classification algorithm.By using the data collected from Lanzhou West Railway Station,the simulation results show that the algorithm can diagnose multi-machine traction turnout faults timely and accurately.

关 键 词：多机牵引道岔 Fretchet距离 TWSVM 故障诊断 

分 类 号：U284[交通运输工程—交通信息工程及控制]