Resonance-based sparse signal decomposition based on the quality factors optimization and its application of composite fault diagnosis to planetary gearbox

The quality factors determine the resonance of resonance-based sparse signal decomposition (RSSD), and directly affect the performance of RSSD. In the existing RSSD, the selection of approximate values of the quality factors with large subjective randomness, reduces the advantages of this method in mechanical fault diagnosis. To solve this deficiency, a new method, the RSSD based on optimizing the quality factors, is proposed. Compared with the existing RSSD, the proposed method optimizes the values of the quality factors with the global optimization ability of genetic algorithm, and adaptively obtains the quality factors of the high- and low-resonance components to realize the optimal matching between RSSD and fault information according to the input signal. Finally, the proposed method is applied to diagnose the composite faults with the planetary gear and the bearing in a planetary gearbox, and effectively extracts the composite fault characteristics from the vibration signal. Accurate diagnosis validates the validity and practicability of the proposed method.