Fault Diagnosis of Motor Bearings Based on Ensemble Empirical Mode Decomposition and Minimum Support Vector Machine

This paper proposes a bearing fault diagnosis method based on integrated empirical mode decomposition and Bayesian criterion using least squares support vector machine to address the issues of poor accuracy and reliability in the diagnosis of wind turbine motor bearing faults. The collected vibration signals are processed using wavelet denoising method, and the integrated empirical mode decomposition method is used to extract features from the signals. The intrinsic mode functions of each state are selected based on correlation as the evaluation index. Then, the energy entropy of the corresponding mode is calculated as the characteristic phasor of the vibration signal. Finally, a Bayesian based least squares support vector machine classifier is established to complete fault diagnosis. The experimental results show that the EEMD based and Bayesian based least squares support vector machine methods can effectively improve the accuracy of bearing fault diagnosis.