Incipient multi-fault diagnosis of rolling bearing using improved TQWT and sparse representation approach

Abstract

Aiming at the issue of extracting the incipient multi-fault of rolling bearing from the nonlinear and non-stationary vibration signals with a strong background noise, a novel fault diagnosis method based on improved tunable Q-factor wavelet transform (TQWT) combined with sparse representation is proposed. Firstly, the optimal Q-factor and decomposition scale can be determined by the improved TQWT according to kurtosis maximum principle. Then the double transient impulse dictionary atom model is established based upon the physical mechanism of periodic transient impulses. Finally, the matching pursuit (MP) algorithm is adopted to reconstruct transient impulses from the maximum kurtosis peak point generated by improved TQWT. Experimental results of accelerated lifetime test of rolling bearing indicate that the proposed method is more effective and applicable for incipient multi-fault diagnosis.