Research on digital twin-assisted dual-channel parallel convolutional neural network-transformer rolling bearing fault diagnosis method

The existing data-driven fault diagnosis methods face some significant problems in practical applications. Many traditional methods rely on a large number of high-quality labeled data for training, but in the industrial environment, the actual fault data obtained is often limited and unbalanced. This data scarcity seriously limits the diagnostic ability of the model and is prone to insufficient diagnostic accuracy. In addition, the data-driven method has a strong dependence on data, and it is prone to misjudgment in the face of complex environments such as noise interference and equipment state changes. These problems jointly restrict the application effect of fault diagnosis methods in industrial actual scenarios. Based on this, this paper proposes a new method of rolling bearing fault diagnosis based on digital twin technology and improved convolutional neural network (CNN)-Transformer deep learning model. Firstly, the geometric characteristics and motion mechanism of rolling bearings are analyzed in depth, and a high-fidelity virtual twin model is established. A balanced simulation data set is generated by numerical simulation. Secondly, we improve the traditional CNN, combined with the Transformer deep learning framework, to enhance the ability of the network to extract features. By performing wavelet transform on the test data obtained from the rolling bearing acceleration test bench and the simulation data generated by the twin model, a dual-channel signal of parallel convolution is formed, and a fault diagnosis model based on dual-channel parallel CNN-Transformer is constructed. Finally, the effectiveness of the proposed method is verified by ablation experiments. The results show that the proposed method can accurately and efficiently identify different rolling bearing fault modes and has superior diagnostic performance. At the same time, the model can also be further extended to related fields to provide new ideas and technical references for fault diagnosis of other mechanical equipment.