Data-driven based ToMFIR Design with Application to Incipient Fault Detection in High-speed Rail Vehicle Suspension System

Abstract

Suspension system, the base of the traction system, whose reliability is of critical importance to the safety of the CRH (China Railway High-speed) train. Incipient FDD (Fault Detection and Diagnosis) of suspension faults can improve the reliability and avoid serious failures in the entire traction systems. Such that, the incipient FDD of spring and damper, the core component of suspension system is urgently demanded. In this paper, a discrete mathematical model for suspension system of a three-car multiple unit is established firstly, based on which, the track irregularities (deterministic disturbances) and then the incipient reduction of spring and damper coefficient (incipient component faults) are further modeled. After that, a data-driven total measurable fault information residual (ToMFIR) based incipient fault detection method is proposed for the possible incipient spring and damper faults. Finally, the proposed theoretical results are applied to a CRH supension simulation system. Simulation results show that the incipient fault detection scheme proposed in this paper provides high sensitivity to incipient suspension faults (spring and damper coefficients reduction of 15% slowly), which has certain reference value in practical engineering applications.