Early fault feature extraction of rolling bearing based on ICD and tunable Q-factor wavelet transform

When a fault occurs on bearings, the measured bearing fault signals contain both high Q-factor oscillation component and low Q-factor periodic impact component. TQWT is the improvement of the traditional single Q-factor wavelet transform, which is very suitable for separating the low Q-factor component from the high Q-factor component. However, the accuracy of its decomposition heavily depended on the selection of Q-factors. There is no reported simple but effective method to select the Q-factors with enough accuracy. This study aims to develop a strategy to diagnostic the early fault of rolling bearings. In this paper, a characteristic frequency ratio (CFR) is used to optimize Q-factors of TQWT (OTQWT). However, directly application of OTQWT is difficult to extract fault signatures at early stage due to the weak fault symptoms and strong noise. A strategy of combination of intrinsic characteristic-scale decomposition (ICD) and TQWT is proposed. ICD owns significant advantages on computation efficiency and alleviation of mode mixing. The effectiveness of the proposed strategy is tested with both simulated and experimental vibration signals. Meanwhile, comparisons are conducted between the proposed method and other methods like: envelope demodulation and EEMD-TQWT. Results show that the proposed method has superior performance in extracting fault features of defective bearings at an early stage.