Bearing failure diagnosis in three-phase induction motor by chirp-Z transform and zoom-MUSIC

Abstract

This paper introduces a new approach of the stator current spectral analysis, based on chirp-Z transform (CZT) and Zoom Multiple signal classification (ZMUSIC), for diagnosing the occurrence of bearing faults in three-phase induction motor. In contrast with traditional FFT, CZT and ZMUSIC are suitable when it is necessary to analyze a not stationary signal such as the current signal of the induction motor. In this paper, firstly, the supply frequency is evaluated by CZT, and the fundamental component is turned to be DC component by shifting it to 0Hz. Then the narrow frequency bands around the characteristic bearing fault frequencies are analyzed by using ZMUSIC. The fault characteristic frequency components in the spectrum are clear visible by the elimination of the dominant fundamental component and noise. The proposed technique is verified by applying to diagnose the bearing failures in a three-phase induction motor.