Fault diagnosis of rolling bearing considering different damage degrees based on wavelet entropy and SVM

The feature extraction in time domain and frequency domain analysis of nonlinear vibration signals is insufficient for the early fault signal of rolling bearing. A rolling bearing diagnosis method is proposed considering different damage degrees based on wavelet entropy feature extraction and support vector machine (SVM). The feature extraction of wavelet entropy mainly includes the extraction of wavelet energy entropy (WEE), wavelet correlation feature scale entropy (WCFSE), and wavelet singular entropy (WSE). The bearing data in this paper adopts the open bearing database of Case Western Reserve University in the United States, considering a variety of damage degrees, which increases the complexity of the data to be diagnosed. Firstly, the bearing vibration signal is decomposed by three layers of wavelet packet, and the wavelet entropy analysis is carried out. Then, taking the wavelet entropy features as the input, the learning model of the support vector machine is established. Finally, the fault location of the rolling bearing is determined. The results show that the joint model can effectively identify the rolling bearing fault.