Fault detection method based on SVD and WPD

Abstract

A fault detection method based on singular value decomposition(SVD) and wavelet packet decomposition(WPD) is designed. The basis principle of this method is the difference of energy distribution of the vibration signals in the frequency domain when the machine works in the normal mode and the failure mode. Firstly, the signal is de-noised by SVD. Secondly, the signal samples are decomposed by using the wavelet packet, and then the wavelet packet coefficients are reconstructed and the energy of different layers is extracted. The energy extracted from different frequency domain is taken as a feature vector, and it will be the input sample of the support vector machine(SVM). Finally, the correctness and feasibility of SVM are proved through practical examples. The results of experiment show that the method based on SVD and WPD can identify the faults of the vacuum pump well.

Keywords

• fault of vacuum pump
• singular value decomposition(SVD)
• wavelet packet decomposition(WPD)
• support vector machine(SVM)