A technique based on decomposition using wavelet transform has been proposed for detecting taperred roller bearing outer race defects and measuring its width. Experiments and subsequent analysis reveal that decomposition of vibration signal with Symlet wavelets is suitable for bearing fault analysis because it improves signal sharpness and suprass noise. Sharp and high amplitude impulses are obtained when roller crosses over the entry and exit of the defect. When the roller remains in contact

more »

... th the groove base between entry and exit, the impulses due to roughness of the groove surface are observed. The defect evaluation technique using decomposition extracts roller entry and exit components from the signal with good resolution offering a way for feature extraction. The proposed technique has been successfully implemented for measuring defect width. Results has been also verified by image analysis. The maximum deviation between the calculated and the verified value is low enough to prove the right choice of the feature extraction used in the research.