An Intelligent Fault Diagnosis Method Based on Empirical Mode Decomposition and Support Vector Machine

A novel fault diagnosis method based on empirical mode decomposition (EMD) and multi-features fusion support vector machine (SVM) is proposed in this paper. Firstly, the given signal is decomposed into a number of intrinsic mode functions (IMFs) by EMD. Choose the first several energy-dominating IMFs, and extract their wavelet packet features, respectively. So, a series of feature sub-spaces are obtained base on each of IMFs, respectively. Then, BT-SVM based weak node classifiers are trained in each of feature sub-spaces corresponding to each of IMFs. Finally, based on the energy proportion of each of IMFs, strong node classifier is obtained by weighted fusion of these weak node classifiers. The experiment part, the application in fault diagnosis of diesel engine shows the feasible and efficient ability of the proposed method.