Optimized Spectral Amplitude Modulation based on Generalized Envelope for Bearing Compound Fault Diagnosis

As a critical component, rolling bearings are widely adopted in rotating machinery. Therefore, bearing fault diagnosis has been a topic of intensive research in recent decades. However, current methods have focused on demodulating an optimal narrow band to detect potential bearing fault characteristic frequency (FCF). In the case of fault information being scattered over multiple frequency bands or compound damages, these methods may underdiagnose or even fail to diagnose. To tackle this issue, a optimize spectral amplitude modulation (SAM) is proposed to detect bearing compound faults. Inspired by envelope analysis, a generalized log envelope analysis is first constructed to enhance the bearing damage components masked in multiple interferences. Subsequently, a novel envelope sparsity index is developed to quantify fault information. On this basis, a weight optimization strategy is introduced to improve SAM detection accuracy, after which more obvious FCF can be identified. In this way, all fault-related information is enhanced while other components are suppressed. Finally, the advantages of the proposed method are validated by experimental data acquired from a locomotive bearing test bench. The analysis results show that the proposed method may provide an effective tool for bearing compound fault diagnosis.