Research And Engineering Realization On The Fault Diagnosis Of Urban Rail Vehicle Suspension System

Since the suspension system is one of the key components of the urban rail vehicle, it has the vital significance to study on its real-time condition monitor and fault diagnosis. It is not only can improve the safety and reliability of urban rail vehicle’s operation, but also can significantly reduce the cost of the preventive maintenance.This paper bases on Mission6"Integrated distributed fault diagnosis system of urban rail vehicle", which is member of Title3"Online fault diagnosis technology of train’s key equipment based on sensor network", a subtitle of the National863Project "Key technology on online monitoring and security warning of urban rail vehicle"(No:2011AA110503). Here we research on the issue of fault detection and fault isolation of urban rail vehicle’s suspension system using some intelligent algorithms, and apply them to specific project realization. The research mainly includes the following sections.(1) Construction of the simulation platform for urban rail vehicle suspension system and abstraction of the fault feature.In this part, we make a research in the construction and function of urban rail vehicle suspension system. And we also introduce the fault features which will be used for fault isolation.(2) Suspension system fault isolation based on Support Vector Machine (SVM).In this part, we explore fault isolation based on SVM aim at the problem for urban rail vehicle suspension system. And apply SVM to isolate fault that happened in the urban rail vehicle suspension system. By using a Gaussian RBF as Kernel function, SVM can obtain the best performance in the fault isolation.(3) Suspension system fault isolation based on Fuzzy Min-Max Neural Network (FMMNN).In this part, we research in the fault isolation using FMMNN, and apply FMMNN to urban rail vehicle suspension system for fault isolation. FMMNN can obtain the better performance in the isolation for two types of faults.(4) Engineering realization of the fault detection algorithm and separation algorithm.In this part, we apply the research to fifteen vehicles of Guangzhou subway, by manipulating the related function of fault diagnosis, extracting the fault diagnosis feature, and analyzing the characteristics of data. We also apply the SVM and the FMMNN to fault separation to urban rail vehicle suspension system.