Rolling bearing fault diagnosis method based on dispersion entropy and SVM

Abstract

According to the different characteristics of different fault vibration signals of rolling bearing, a method of bearing fault diagnosis based on dispersion entropy (DE) and support vector machine (SVM) is proposed. The intrinsic mode function (IMF) component of the bearing vibration signal is obtained by improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) decomposition, and the dispersion entropy of the first few IMF components containing the main fault information is calculated. The eigenvectors are constructed by calculating the DE values of the first few IMF components which contain the main fault information and trained as the input of the SVM, the classification of bearing faults is realized. Compared with permutation entropy (PE), approximate entropy (AE) and sample entropy (SE), DE has a higher accuracy.