Bearing damage detection via wavelet packet decomposition of stator current

The induction motors are used heavily in industrial manufacturing plants. Bearing faults, insulation faults, and rotor faults are the major causes of induction motor failures. Even though the replacement of bearings, greatest single cause of motor faults, is the cheapest fix among the three conditions, it is the most difficult one to detect. This dissertation deals with the detection of incipient bearing defects.; Motor current signature analysis, one of currently available predictive maintenance methods, provides continuous monitoring in a non-intrusive way. In this method, the motor current signal is analyzed to detect any defects. The wavelet packet decomposition of the stator current is the proposed method of analysis in this dissertation. The bearing condition monitoring can be easily implemented in a power quality monitor if computationally efficient filter banks are used in the wavelet packet decomposition of the stator current.; The use of wavelet packet decomposition enables the analysis of frequency bands that can accommodate the rotational speed dependence of the bearing defect frequencies. The wavelet packet decomposition also offers a better treatment of non-stationary stator current than currently used Fourier techniques.