Analysis and prediction of double-carriage train wheel wear based on SIMPACK and neural networks

Wheel and rail wear seriously affects the safety reliability of train operations. In this study single-carriage double-carriage models considering connecting unit a high-speed are developed to investigate normal forces, lateral displacements wheelsets. Based on results from these models, Archard model is employed predict wheel wear. addition, based daily measured data, nonlinear autoregulatory (NAR) wavelet neural network (WNN) over longer time period. The simulation show that, compared with model, wheelsets close in increase certain extent dependent speed. predictions that near slightly larger than far unit. mean square error, NAR has somewhat better performance prediction WNN model. research represent an important contribution maintenance safe operation trains.