Fault isolation for rail vehicle suspension systems based on PSD distance feature

Abstract

The Suspension system is crucial for safety and steady operation of high-speed vehicles, for which the research on fault diagnosis and isolation is necessary. In this paper, the PSD distance feature extraction method is presented. The vehicle dynamic model is built in SIMPACK and the external force models are built in SIMULINK. Different fault levels of the vehicle suspension components are simulated based on the co-simulation between SIMPACK and SIMULINK. The Fuzzy Possibilistic C-Means Clustering (FPCM) and BP neural network are combined for isolating different component faults. The simulation results show that the method proposed in this paper achieves sound performance.