A frequency-based approach to detect bearing faults in induction motors using discrete wavelet transform

Detection of faults in induction motors is nowadays is a hot trend in the field of electrical machinery. There are several methods to detect electrical and mechanical faults in an asynchronous motor; fast Fourier transform, short-time Fourier transform, and wavelet transform are the most popular ones. A major deficiency that most of these solutions face is not being able to detect low energy faults, such as mechanical bearing faults. The new solution proposed in this paper focuses on detection and prediction of low energy faults applying discrete wavelet transform (DWT); the output signal is passed through high pass and low pass filters and coefficients are derived consequently. The method offered by the authors of this paper includes deriving frequency spectrum of each level of discretization. Especially in high decomposition levels, inner race bearing faults can be detected earlier by monitoring frequency spectrum of high levels in DWT.