The Rolling Bearing Fault Feature Extraction Method Under Variable Conditions Based on Hilbert-Huang Transform and Singular Value Decomposition

The fault diagnosis precision for rolling bearings under variable conditions has always been unsatisfactory. For solving this problem, a feature extraction method combing the Hilbert-Huang transform with singular value decomposition was proposed in this paper. The method includes three steps. Firstly, instantaneous amplitude matrices were obtained by Hilbert-Huang transform from rolling bearing signals. Secondly, as the fault feature vector, the singular value vector was acquired by applying singular value decomposition to the instantaneous amplitude matrices. Thirdly, the identification and classification of rolling bearing were achieved by Elman neural network classifier. The experiment shows that this method can effectively classify the rolling bearing fault modes with high precision under different operating conditions.