Wavelet-Based Time-Reassigned Synchroextracting Transform With Application to Fault Diagnosis of Flexible Thin-Wall Bearing

A healthy flexible thin-wall bearing (FTB) generates a series of impulses during operation, which will overwhelm its failure characteristics. To accurately capture the fault feature frequency of FTB, a postprocessing time–frequency analysis (TFA) method named wavelet-based time-reassigned synchroextracting transform (WTSET) is proposed, which can significantly enhance the energy concentration of the time–frequency representation (TFR) while maintaining invertibility. First, the diffused time–frequency energy problem of the wavelet transform (WT) is explored. Second, an operator based on group delay and instantaneous frequency is presented to characterize the transient components in the signals. Third, the blurry time–frequency energy of WT is squeezed into the group delay position, and only those related to the transient components are preserved. Finally, an extraction method based on the correlated Gini index (CGI) is proposed to extract the fault characteristic frequency of FTB, which has proved its worth in diagnosing bearing failures. The numerical simulations and experimental results confirm the efficacy of the proposed WTSET algorithm and the extraction method.