Bearing Fault Diagnosis for Imbalanced Data Based on Multi-Scale Data Generation and Ensemble Modeling

Le résumé fourni par la source

Accurate fault diagnosis of bearings is of utmost importance, as bearings are crucial components in rotating machinery. Deep Learning (DL) has developed rapidly in recent years and is widely used in the field of fault diagnosis. The bearing fault diagnosis method based on DL needs a lot of fault data, but it is difficult to obtain in the actual engineering scene. This problem of data imbalance seriously affects the accuracy of fault diagnosis. To solve this problem, a bearing fault diagnosis model based on Multi-Scale Generative Adversarial Network (MSGAN) and Ensemble Convolutional Neural Network (ECNN) is proposed in this paper. Firstly, MSGAN is used to generate fault samples with different scale features and balance the corresponding scale feature samples. Then an ECNN fault diagnosis model is established to improve the diagnosis accuracy. During training, the model is able to learn features of different scales from samples. During training, the model is capable of learning features at different scales from the samples. During testing, it outputs classification results based on the average probability. The experimental results demonstrate that this method achieves high fault diagnosis accuracy under imbalanced conditions.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.