Roller Bearing Fault diagnosis Based on EMD Sample Entropy

摘要

A roller bearing fault diagnosis method has been proposed based on Empirical Mode Decomposition (EMD) sample entropy (SampEn), in order to deal with the nonlinearity existing in bearing vibration signals. Firstly, original vibration signals are decomposed into a number of intrinsic mode functions (IMFs). Then the SampEn values of first numbers of IMFs that contained the most dominant fault information are calculated and serve as the feature vector for bearing fault diagnosis. The analysis results from EMD SampEn of different vibration signals show that the SampEn is an effective feature. Finally, SVM is used to identify the work condition of the roller bearing. Experimental results with CWRU data show that the diagnosis approach based on SVM by using EMD SampEn as features can identify roller bearing fault patterns accurately.