Detection of bearing faults in mechanical systems using motor current signature and acoustic signatures

Abstract

Bearings are one of the critical components in rotating machinery. The need of an easy and effective fault diagnosis technique has led to the increasing use of motor current signature analysis (MCSA). Bearing faults in the mechanical system run by an induction motor causes change in its stator current spectrum. The faults in the bearings cause variations of load irregularities in the magnetic field which in turn change the mutual and self inductance causing side bands across the line frequency. The objective of this paper is to detect bearing faults (outer race fault) in a mechanical system using motor current signature. Fast Fourier Transform (FFT) is initially employed for a first comparison between a healthy and a defective bearing. Six wavelets are considered out of which three are real valued and remaining three are complex valued. Base wavelet has been selected on the basis of wavelet selection criteria - Maximum Relative wavelet energy. Then, 2D wavelet scalogram has been used for the detection and occurrence of outer race faults of various sizes in ball bearings of mechanical systems using motor current signatures of induction motor.