Reliability Modeling And Re-profiling Time Prediction For Track Wheel Based On Proportional Hazard Model

Wheels of railway wagon have to bear the weight of the body and cargo during train operation,and a large number of impact loads as well.Therefore,its reliability would directly affect the running quality and safety of railway wagons.In some extreme cases,it may even cause major safety accidents and economic losses.At present,methods used for predicting the remaining life of railway wheels around the world are mainly based on the dynamic simulation models and data-driven methods.It is difficult for researchers to build an accurate physical model for simulation due to the uncertainty of the actual environment where trains operate,the various failure forms of parts and the ambiguity of the failure mechanism.Regarding data-driven method on railway wagon wheels,the accuracy of the research results is also not guaranteed because incomplete data of wheel life and the differences in the wheel material and model have been seldom considered.Based on what mentioned above,this thesis explores the relationship between the running state and reliability of railway wagon wheels,and then builds a railway truck wheel proportional hazard model based on the existing maintenance and repair information and status information,in order to achieve realize reliability modeling and re-profiling time prediction for track wheel.The proportional hazard model enables operators to link the covariates of the model and material of the railway wagon wheel with data related to wheel reliability,so as to establish more accurate reliability models of different wheels,determine the state of the wheels at each running time,and maintain the railway wagon wheels.This model also provides a theoretical basis for maintaining the wheels of railway wagons.Firstly,this thesis established a wheel proportional hazard model based on the general idea of data collection and cleaning,confirming the distribution of life-time forms,selection of covariates and hypothesis testing.Secondly,considering the complexity of the model,the maximum likelihood estimation method and the Newton's method are used to estimate the parameters of the model,and the point estimates of the instantaneous failure rate and reliability of the railway truck wheels are obtained.Thirdly,in order to obtain the confidence interval of the re-profiling time of railway wagons,the parameters of the proportional hazard model are estimated using the bootstrap method.And the influence of each material and wheel model on the re-profiling time of railway wagon wheels is analyzed,finally obtain the reliability and interval estimation of life.