Application of DWT and PDD for Bearing Fault Diagnosis Using Vibration Signal

Abstract

Condition monitoring and fault detection of electrical machinery are one of the important issue for commercial enterprises. Incipient fault detection in hardware can spare a large number of rupee in emergency maintenance cost. In present article an efficient real time vibration measurement of an induction motor bearing on full load has been presented. The method used for the analysis and diagnosis of the fault in induction motor bearing are probability density distribution (PDD), Discrete Wavelet Transform (DWT) as a qualitative and, statistical parameters, Fast Fourier Transform (FFT) as a quantitative has been used. The discrete wavelet transform is used to process the accelerometer signals and discrete wavelet coefficient is processed to determine the spectral energy for different frequency bands containing the harmonics due to fault. The higher spectral energy signal from DWT used for the Fourier analysis to get the location of fault. The statistical parameters of detail coefficients are calculated for different levels of wavelet for faulty and healthy bearings. The outcomes got have demonstrated that this methodology is successful for bearing fault detection and analysis.