Diagnosis Method of Combing Feature Extraction Based on Time-Frequency Analysis and Intelligent Classifier

摘要

In the process of using neural network to carry out intelligent fault type identification, how to extract sensitive fault features from the original data is quite important for an accurate diagnosis result. An intelligent fault diagnosis method was proposed, which combined time domain analysis and wavelet analysis method to extract features from vibration data of a motor bearing. The resulting vector obtained from the feature extraction was used as samples to train the BP neural network intelligent classifier to enable the classifier to identify fault type. The comparison of experiment results showed that the proposed diagnosis method was effective.