Experimental fault diagnosis for cylindrical roller bearings in high-speed railway motors via stator current analysis and demodulation techniques

Abstract

Traction motor roller bearings are one of the critical structures for health monitoring of high-speed train vehicles. The technique of motor current signature analysis (MCSA) is a novel way to monitor the mechanical system with induction machines in recent years. In this paper, a fault diagnosis method for motor bearings in high-speed railway vehicles based on MCSA is proposed. It is a sensor-less way for the health monitoring of motor bearings because the induction machine itself can be considered as a sensor. The weakness that vibration data is sensitive to measuring points is avoided. To verify the feasibility of the proposed method, a set of experiments using an induction machine with faulty bearings from a real high-speed train are processed. Compared with the examples under healthy conditions in the time domain, the RMS (Root Mean Square) values of motor stator currents under faulty conditions show an around 20% increase. The modulation phenomena usually happen on motor currents according to both theoretical analysis and practical measurements. A demodulation preprocessing is applied to capture the fault-related harmonic components in the frequency domain. To eliminate the influence of preprocessing algorithms on the result, three different demodulation techniques (HT, WT, and TEO) are compared. The signal processing results prove that it is feasible to monitor the health status of motor bearings in high-speed train based on MCSA.