Adaptive tunable Q-factor wavelet transform for fault feature extraction of gearbox based on vibration signals

Fault diagnosis of gearbox is of great importance to avoid catastrophic accidents. Feature extraction has always been a key problem for fault diagnosis. In this paper, a novel fault feature extraction method called adaptive tunable Q-factor wavelet transform for gearbox fault diagnosis is proposed. The proposed adaptive method is implemented using the tunable Q-factor wavelet transform (TQWT). Kurtosis as an effective index of impulses is adopted to choose the optimal TQWT basis. The new method can obtain the optimal Q-factor according to the maximum of kurtosis. Thus, the Q-factor of the TQWT can match the oscillatory behavior of signals optimally without artificially specified. The interested fault feature is extracted by the single branch reconstruction of optimal subband. The proposed method is applied to vibration signals analysis of a bevel gear with a scratch defect from an antenna transmission chain and a gearbox from an electric locomotive. The processed results demonstrate that the proposed method can extract weak fault features of gearbox efficiently.