Fault Analysis and Detection Method of the EMU Traction Motor

Abstract

A study proposed in this paper is based on EMU traction motor failure. The incipient periodic fault characteristic signals of rotor bars breaking are very weak, and the useful signals submerged in the noise are difficult to be detected. The fault detection method was combination of Duffing oscillator and wavelet packet energy analysis. The system characters were analyzed and testified by the simulation and the experiment.