Element analysis and its application in rotating machinery fault diagnosis

Conventional signal processing methods make it difficult to extract the fault impulse features from the target signal, and the time-frequency representation has energy ambiguity. Thus, it is critical to develop new approaches for mechanical fault diagnostics. In this paper, the element analysis method, which was originally utilized in the marine field, is applied to the field of mechanical fault diagnosis for the first time. A de-noising technique of rotating machinery signals based on the element analysis method is proposed. The proposed method first determines the corresponding wavelet parameters according to the mechanical fault signals and constructs the element model. Then the method performs the Morse wavelet transform on the element model, and calculates the signal impulse point from the wavelet transform to obtain the signal's fault characteristic frequency. Furthermore, the method can also reconstruct the signal by utilizing a small number of solitary points in the time or scale plane. The performance of the method is verified by analyzing simulated signals and mechanical vibration signals collected from different experimental platforms. The results demonstrate that the method has excellent signal characteristic extraction capability and successfully diagnoses different kinds of rotating machinery faults.