Intelligent Fault Diagnosis of Engine Bolt Tightening Based on One-dimensional Convolutional Neural Network

Threaded connection is one of the most widely used connection methods in automobile manufacturing industry. The quality of thread connection is directly related to the life safety. With the development of tightening technology and the improvement of tightening equipment manufacturing level. As a method to record the dynamic changes of tightening process, tightening curve technology has been relatively mature. However, it is extremely difficult to identify the failure of tightening process in real time with human experience. In this paper, an intelligent fault diagnosis method based on one-dimensional convolutional neural network (ID-CNN) is proposed, which can effectively reduce the false positive and false negative caused by manual recognition. Experimental results show that the fault diagnosis accuracy is over 98%.