Research on rolling bearings fault diagnosis method based on EEMD morphological spectrum and kernel fuzzy C-means clustering

Aiming at the fault diagnosis of rolling bearings, a fusion method based on ensemble empirical mode decomposition (EEMD), morphological spectrum and kernel fuzzy C-means clustering (KFCMC) clustering is proposed. Firstly, a vibration signal is decomposed by EEMD to get several intrinsic mode functions (IMFs) which have physical meanings. Secondly, with a fusion evaluation method based on kurtosis, power and standard deviation, the three IMFs which are rich in fault features are selected as data source, the mean values of morphological spectrums in some scales of the three IMFs are extracted, and then the three values constitute a sample, thus sample set can be got. Lastly, all the samples of different working conditions are clustered by the KFCMC to diagnose the rolling bearing faults. In addition, the signals are also decomposed by empirical mode decomposition (EMD), and the samples are also clustered by fuzzy C-means clustering (FCMC), and the results show that the proposed method performs better than EMD and FCMC. The signals of the rolling bearings are tested and verified, and the conclusion is that the fusion method of EEMD and KFCMC is superior to that of EMD and FCMC. The proposed method can diagnosis the faults of rolling bearings efficiently. ©, 2015, Nanjing University of Aeronautics an Astronautics. All right reserved.