Fault Diagnosis of Rolling Bearing Based on CNN with Attention Mechanism and Dynamic Learning Rate

It is of great importance to diagnose early failures of rolling bearings accurately and timely in industrial production. However, early faults have the characteristics of weak feature information and low signal-to-noise ratio, which increases the difficulty of fault identification. Therefore, based on deep learning theory, this paper proposes a convolutional neural network model based on attention mechanism and dynamic learning rate for intelligent diagnosis and recognition of early faults of rolling bearings. The model introduces the attention mechanism to improve the learning performance of the convolutional network for effective information, so that the model can learn the discriminative and important features. A dynamic adjustment strategy of learning rate is designed to improve the convergence speed and performance of the model. Finally, the proposed method is analyzed using XJTU-SY rolling bearing dataset. Experimental results indicate that compared with other fault intelligent diagnosis methods, the proposed method can effectively identify the early fault of rolling bearings under different operating conditions, and has higher diagnostic accuracy.