Fault Diagnosis Based on Vibration and Torque Signals

In recent years, intelligent fault diagnosis technology has become a research hotspot in the field of fault diagnosis by virtue of its powerful performance and broad prospects. However, most of the current research is based on the fault diagnosis of a single vibration signal, and the model trained by single-modal data often has the problems of poor fault tolerance rate, susceptibility to noise and poor generalization. Therefore, in order to explore the connection between different signal fault features and extract and integrate complementary fault features, this paper proposes a new method to fuse vibration and torque signals, called VTCNN-LSTM, which is experimentally verified on the bearing dataset Paderborn of the University of Paderborn in Germany. The results show that VTCNN-LSTM has nearly 100% diagnostic accuracy in a noise-free environment. It also has good performance in high noise environment, which shows the effectiveness of the method.