Research on Life Prediction Method of Motor Bearings

In order to better achieve the life prediction and reliability analysis of motor bearings, a bearing life prediction method based on deep learning is proposed in this paper. The method firstly extracts features from the original vibration signal of the bearing and uses Weibull distribution to fit the extracted features. Then the fitted features are applied to the training phase of the SFAM(Simplified Fuzzy ARTMAP) neural network, and the extracted original features are applied to the testing phase, after SFAM neural network classification, a category representing the bearing degradation rate is given for each input vector. Finally, the classification results are made more continuous by a smoothing algorithm. The results show that this method can realize the life prediction and reliability analysis of motor bearings, and it is universal.