Research on node-improved energy dissipation wear model for fretting fatigue prediction in railway press-fit shaft

Abstract

As a key component of the traveling system of high-speed trains, the axle is crucial for its safe operation. The current research on the fretting wear and fatigue development of shafts suffers from the problems of low local simulation accuracy of the wear model and the lack of detection and validation methods for the dynamic expansion of wear and fatigue. To this end, this study firstly proposes a new node-improved form of energy dissipation wear model, which is more sensitive to the contact behavior of the asperity body in the overfilled region and the energy transfer process; it exhibits wear prediction results that are closer to the actual situation than the traditional model and improves the prediction accuracy of the corresponding relationship between the cycle time and the wear fatigue crack extension. Secondly, the acoustic emission deep learning detection and validation method of shaft wear fatigue under dynamic rotation is developed; 75,480 wear fatigue signal time-frequency map datasets are established with shaft dynamic rotation experiments, and the convolutional self-encoder and multi-head attention mechanism recognition and detection network are trained. The recognition accuracy of the wear fatigue damage stage reaches 87.4 %. This study shows that the node-improved energy dissipation model combined with acoustic emission dynamic rotation wear fatigue detection can accurately assess the effect of shaft fretting wear on fatigue development.