A Review on the Signal Processing Methods of Rotating Machinery Fault Diagnosis

As the core component of rotating machinery, the complex loads and sustained rotations in the harsh conditions are prone to multiple faults. It is the primary researching that extracting the fault information and identifying the fault pattern form vibration signals. The signal processing methods of rotating machinery fault diagnosis in recent years are summarized in this paper. Firstly, the researching status of time-frequency analysis method in vibration signal processing of rotating machinery are summarized, and the drawbacks of these methods are compared deeply. Then, the fault feature learning methods based on sparse representation are summarized, and the advantages and disadvantages of different methods are compared. And then, the research status of fault feature extraction and fault pattern recognition based on artificial intelligence and deep learning methods are analyzed, and summarized the advantages and disadvantages of each method in vibration signal feature extraction. Finally, the existing problems in these researches of fault diagnosis and the future directions are expounded.