Acoustic Based Fault Analysis of Induction Motor Using FFT and Wavelets- A Comparative Study

Abstract

Detection of fault in induction motors is indispensable for avoiding unforeseen failures. This work intends to diagnose bearing faultin an induction motor. Fault diagnosis is performed on the basis of audio signals emitted by the rotor during rotation of an induction motor. The audio signals which are obtained via a microphone is analysed in Fast Fourier Transform (FFT) and Discrete Wavelet Transform (DWT) methods. It is seen that even though FFT technique was able to differentiate between the frequency spectrum of the faulty motor and the healthy motor, discrete wavelet transform showed a better level of differentiation with respect to the faulty condition and the healthy condition, which were evaluated based on certain statistical parameters.