T-S Fuzzy Data-Driven ToMFIR With Application to Incipient Fault Detection and Isolation for High-Speed Rail Vehicle Suspension Systems
Yunkai Wu, Yu Su, Yu‐Long Wang, Peng Shi
Abstract
This paper addresses the incipient fault detection and isolation (FDI) problem for high-speed rail vehicle suspension systems and explores further results of data-driven total measurable fault information residual (ToMFIR) with Takagi-Sugeno (T-S) fuzzy dynamical mode. Firstly, the T-S fuzzy model is constructed to represent the global nonlinear dynamics of a China Railway High-speed (CRH) trailer car. Secondly, the data-driven ToMFIR based on T-S fuzzy theory and system identification is designed with the help of system input/output(I/O) data models. Moreover, Jensen-Shannon (JS) divergence based evaluation function is proposed for monitoring the slight changes of data-driven ToMFIR. Furthermore, generalized likelihood ratio (LR) reconstruction method combined with data-driven ToMFIR is designed for incipient isolation of suspension actuator and sensor faults. Finally, simulation studies conducted on SIMPACK-MATLAB/Simulink Co-simulation environment are given to demonstrate the effectiveness of the developed FDI scheme for both slowly developing faults and incipient faults with intermittent characteristics.