Fault Diagnosis Of Urban Rail Train Suspension System Based On An Experimental Platform

The suspension system is a key part of urban rail vehicle. Its state is related to the train ride comfort, safety and reliability, directly. Meanwhile, the suspension system of urban rail trains’ conventional detection is mainly operated by regular inspection with artificial observation, which is a waste of manpower, material and financial resources. This paper is supported by the Chinese863program "On-line condition monitoring and detecting techniques for urban rail vehicles key components based on sensor network". Researches on the fault diagnosis of suspension system are carried out in this paper. Different fault conditions of the train suspension system are simulated based on experimental platform simulation, then, the fault detection and fault isolation method which have guiding significance to the vehicle suspension system fault diagnosis are proposed and verified with track irregularity and track step excitation, respectively. Therefore, this paper has carried out the following research contents, mainly includes:1. A suspension system simulation platform of rail transit train is built according to the suspension system structure and other characteristics of urban rail train, combined with the train dynamics theory, suspension system fault simulation experiment platform of the actual rail transit train is set up. Different failure modes of the suspension system key components are simulated on this experiment platform. At the same time, the relevant experimental data are collected by means of reasonable layout sensors, which provide data support for the follow-up fault detection and isolation algorithm.2. In the excitation generated by track irregularity, fault detection tests of the rail transit train suspension system are carried out using the DPCA algorithm and PLS algorithm. First of all, principles, steps of DPCA and PLS algorithms are introduced. And then, the suspension system fault detection program based on this method is produced with MATLAB. Finally, fault detection experiments of the urban rail train suspension system are done based on the collected data from the experimental platform.3. In the excitation generated by track irregularity, fault isolation tests of the rail transit train suspension system are carried out using D-S evidence theory. First of all, the principles, steps of D-S evidence theory are introduced. Then, the suspension system fault isolation program is produced by MATLAB. Finally, fault isolation experiments of the rail transit train suspension system are done based on the collected data with the experimental platform. 4. A new fault diagnosis model of track step incentive is proposed. The feasibility of track step method is proved through the mechanical analysis of the vehicle in the track step. Fault isolation of the rail transit train suspension system in the excitation generated by track step is implemented based on the collected data from the experimental platform.5.Comparations of fault diagnosis with two kinds of excitation show that both methods mentioned above can complete detection and isolation of different kinds of faults, effectively. However, applicability of the two methods is quite different. So, which method to choose for fault diagnosis should be considered with the actual situation.