Research on the Fault Diagnosis of Rotating Machinery Based on Wavelet Analysis and BP Network

The four common faults of rotating machinery, that is imbalance, misalignment, rubbing and oil whirl, were simulated on Bently, then the time-vibrational displacements of the four faults have been got, and the corresponding figures of time-displacement were drawn by using Matlab 7.0. On the basis of wavelet analysis of vibrational displacement signal, a feature extraction method based on scale-energy modulus was introduced and the fault type of extracted characteristic vector was identified by BP network. The results show that this method is effective for fault recognition of rotating machinery, and also have a certain reference value for maintenance of rotating machinery. This method can also be extended to other mechanical fault diagnosis.