Remaining life prediction of locomotive wheel based on two-stage stochastic Wiener process
To study the problems related to two-stage degradation characteristics of locomotive wheels,a two-stage Wiener process-based method for predicting the remaining service life was proposed.The wheel rim degradation model was established by using a two-stage Wiener process model,and the individual differences in the wheel degradation process were characterized by the random drift coefficient.The expectation maximum(EM)algorithm and Bayesian theory were used to achieve offline parameter estimation and online updating of the degradation model parameters.The change point in the wheel degradation process was determined and found via the Schwarz information criterion(SIC).Finally,an example of wheel rim degradation data from a certain locomotive was used for validation.The results show that,compared with the single-stage degradation model,the two-stage degradation model considering the change point is more in line with the field reality,and the prediction accuracy at 80%of the life quantile of the wheel is improved by 9.42%.The prediction of the remaining service life can provide a certain theoretical basis for the optimization of the wheel turning cycle.
remaining wheel lifetwo-stage stochastic Wiener processEM algorithmBayesian theorySIC guidelines
NETL
NSTL
万方数据
