Inferable Deep Distilled Attention Network for Diagnosing Multiple Motor Bearing Faults

Bearing, as a vital component in electric powertrains, is increasingly used globally, such as in electric vehicle (EV). Their damages and faults may bring huge cost loss to the industry and even threaten personal safety. This article proposes an inferable deep distilled attention network (IDDAN) method, which is a self-attention (SA) mechanism and a transfer learning-based method to diagnose and classify multiple bearing faults in various motor drive systems efficiently and accurately. Compared with the convolutional networks, the SA-based network can better extract the global feature information and easier to benefit from large amounts of pretraining data. Its significance is to accurately classify various faults of the target machine when the labeled data of the target machine are not enough to directly train the diagnosis model. First, this article attempts to apply the SA-based network to build an advanced fault diagnosis model. Second, this article optimizes the structure of networks through a knowledge distillation (KD) technique to require a lighter and fast model. Third, this article proposes a new data augmentation (DA) strategy for 1-D vibration signals to provide large-scale pretraining samples for IDDAN. Experiments show that the SA mechanism-based model is more likely to benefit from large-scale data. After testing, compared with many methods and other exist similar methods, the proposed method achieves higher classification accuracy and better performance.