A New Algorithm Based on Adaptive SSD-TET and Its Application

A signal processing method combining adaptive Singular Spectrum Decomposition (SSD) with Transient Extracting Transform (TET) is proposed to address the challenge of weak energy in early rolling bearing fault signals, which complicates fault classification based on time-frequency information. Initially, the optimal decomposition level for the SSD algorithm is determined using the instantaneous frequency mean method, and the original signal is decomposed into the corresponding singular spectral components. These components are then reconstructed based on the maximum kurtosis criterion, and transient feature extraction is performed to produce time-frequency maps that highlight concentrated fault information. Finally, fault classification is carried out using the C4.5 decision tree. Results demonstrate that the adaptive SSD-TET algorithm effectively extracts transient features of rolling bearing faults, shows greater noise robustness, and provides clearer time-frequency maps, improving fault diagnosis. In comparative tests, the adaptive SSD-TET algorithm achieved a diagnostic rate of 98.3% for various fault types, showing significant advantages over traditional methods.