Maximum Fault Information Envelope Spectrum Based on the Spectral Coherence for Bearing Diagnosis

As the core component of rotating machinery, bearings inevitably experience failures due to the complexity of internal mechanical systems and the harshness of their operating environments. Efficiently identifying incipient faults in bearings can enhance equipment operational efficiency and reliability while reducing production costs and risks. Therefore, there has been a surge of research in both academia and industry. Among the available techniques, envelope analysis is one of the most popular, found in nearly all commercial software. However, due to its reliance on the assumption of signal stationarity and the inherent limitations in frequency band optimization, envelope analysis often struggles to achieve satisfactory results in industrial environments with strong interference and heavy background noise. Cyclostationarity-based analysis breached the usual assumption of stationary, positing that bearing fault signals exhibit cyclostationary nature, thus providing another tool. Therefore, a new maximum fault information envelope spectrum (MFIES) based on the spectral coherence for bearing diagnosis is proposed to overcome the above limitations. In this work, firstly, the bi-spectral map is obtained. Then, a fault symptom index is introduced to assess the amount of fault information contained in each frequency slice. Finally, the frequency slice corresponding to the maximum index value is selected to generate the MFIES. This way, it enhances the fault features while suppressing other components. Moreover, Simulation analysis and experimental data validation have validated the effectiveness of this method. The results indicate that this method possesses strong capabilities for extracting bearing fault characteristics even in the presence of severe interference.