Abstract: Aiming at the problem of low recognition accuracy of rolling bearings with different fault types and damage degrees, a fault diagnosis method combining wavelet packet energy entropy, gref wolf optimizer (GWO) and support vector machine (SVM) was proposed. Firstly, the rolling bearing vibration signal was decomposed by three-layer wavelet packet, following reconstructing the wavelet packet decomposition coefficients of each frequency band in the third layer, and the energy entropy of components in each frequency band was extracted to form fault feature vectors. Secondly, GWO was used to optimize SVM parameters. Finally, based on the optimized SVM classification model, the feature vectors of different fault types and damage degrees of rolling bearings in the test set were recognized and diagnosed. The experimental results show that the proposed method has more outstanding fault identification ability for both different fault types and different damage degrees of rolling bearings than other methods in experiments and literatures.

Key words: bearing failure, wavelet packet, energy entropy, GWO, SVM