ROLLING BEARING FAULT DIAGNOSIS USING EMD ENERGY ENTROPY METHOD AND RBF NUERAL NETWORK

Abstract

This paper presents a rolling bearing fault diagnosis method with the combination of EMD (Empirical Mode Decomposition) with radial basis function neural network (RBFN). Firstly, original vibration signals are decomposed into some components IMFs (Intrinsic Mode Functions) by EMD method, then the concept of EMD energy entropy is applied to extract feature vectors from a number of IMFs that contained the most dominant fault information. Therefore, to identify rolling bearing fault patterns, the feature vectors could serve as input vectors of RBF network. The analysis results from rolling bearing vibration signals (Normal, Inner-race fault, Ball fault, and Out-race fault) by EMD and RBFN show that the EMD-RBFN can identify rolling bearing fault patterns accurately and effectively and is superior to the combination of wavelet packet with RBFN.