| With the continuous development of China's high-speed railway technology,the performance of various components of EMU train have been greatly improved.Axle box bearing is the most critical and basic components of EMU train's transmission system,due to its complicated operating environment and severe load impact,it cannot work smoothly anytime,which may affect the safety of EMU Train.At present,the axle bearings of EMU train are all imported bearings,which means the cost and maintenance cost are very expensive.In order to meet the needs of driving safety and economy at the same time,it is urgent to find a method that can effectively predict the remaining useful life of the bearings,and use this method to make reasonable maintenance decisions.The research and application of PHM technology is the basis and guarantee for the development of modern mechanical equipment in the direction of high performance,high precision and high efficiency.Choosing the appropriate state monitoring method,establishing a complete residual life prediction model and conducting accurate trend prediction are the core of PHM technology.In the process of performance degradation throughout bearing's whole life,if its operation state can be effectively monitored and evaluated,and the state information can be used to predict its remaining useful Life and evalute its reliability,then the cost of maintenance also can be reduced greatly under the premise of ensuring the safe operation of the train.This paper has carried out the following work on how to establish the RUL prediction model and how to predict RUL based on the bearing acceleration performance degradation test data:(1)The axle box bearing structure and several typical failure modes which often appear in the process of using are analyzed.The extraction method of bearing's performance degradation characteristic parameters based on vibration signal is selected,by analyzing and comparing the four commonly used state monitoring methods.Six performance degradation parameters which are sensitive to bearing's performance degradation are extracted and screened in the vibration signal's time domain,frequency domain and time-frequency domain,which lays a theoretical foundation for the subsequent construction of bearing performance degradation indicators.(2)In order to solve the problem that the information redundancy of the six training bearing's performance degradation parameters is too high,a method based on PCA for bearing performance degradation index is proposed.Firstly,the performance degradation feature set is constructed by using the selected six performance degradation feature parameters,and then processed by PCA.Finally,the index which can fully characterize the bearing performance degradation process is obtained,which lays a data foundation for the establishment of the RUL prediction model.(3)In order to solve the problem that the prediction accuracy of traditional life prediction model is not high and the accuracy is poor,a residual life prediction method based on Weibull distribution proportional failure rate model is proposed.The future prediction is predicted by GM(1,1)gray prediction model.Bearing performance degradation index,and the performance degradation index in the future as a covariate input model,the predicted RUL is compared with the real life to verify the validity of the model,and finally based on the RUL model and its reliability function.A corresponding maintenance decision is made and supplemented by the current planned maintenance decision. |
PDF Full Text Request