Semi-Supervised Transfer Learning Method for Bearing Fault Diagnosis with Imbalanced Data

Fault diagnosis is essential for assuring the safety and dependability of rotating machinery systems. Several emerging techniques, especially artificial intelligence-based technologies, are used to overcome difficulties in this field. In most engineering scenarios, machines perform normal conditions, which implies that fault data may be hard acquire limited. Therefore, imbalance deficiency labels practical challenges bearings. Among mainstream methods, transfer learning-based highly effective, as it transfers results previous studies integrates existing resources. The knowledge from source domain transferred via Domain Adversarial Training Neural Networks (DANN) while dataset target partially labeled. A semi-supervised framework based on uncertainty-aware pseudo-label selection (UPS) adopted parallel improve model performance by utilizing abundant unlabeled data. Through experiments two bearing datasets, accuracy classification surpassed independent approaches.