The Research of the Transient Feature Extraction by Resonance-Based Method Using Double-TQWT

Signal processing aims to extract useful features from signals. However, the useful features are usually so weak, and corrupted by strong background noise, so it is difficult to extract by traditional linear methods. In this paper, a resonance-based method using double tunable Q-factor wavelet transform (TQWT) is applied for transient feature extraction. With the double-TQWT, the non-stationary signal is represented as the mixture of high resonance components and low resonance components based on the different resonance. The transient feature has a low Q-factor and belongs to low resonance components. Results of applications in transient feature extraction for simulation signal and bearing fault signal show the new method outperforms the average filtering method and the wavelet threshold algorithm, which further confirms the validity and superiority of this method for transient feature extraction.