Outer Race Bearing Health Prognosis Using Feature Extraction and Continuous Wavelet Transform

Bearings are the most common components prone to failure in induction motors. To minimize the effects of failure, it is important to monitor the health of bearings, so in this paper we propose a new and advanced technique for predicting bearing failure using extraction and selection features and continuous wavelet transform (CWT) technique. In the proposed approach, vibration signals are represented by spectrograms, and various parameters such as time or frequency domain, (RMS, Power, Standard deviation (Std), Peak2peak, Mean, CrestFactor) are applied in order to process this data. Then CWT is used to directly extract TSP (prediction start time), RUL (remaining life) and threshold. In this study, run test data were used to validate the work.