Research On Reliability Modeling Method Of High Speed Railway EMUs

Speeding up passenger transport is one of the directions of China railway.High-speed railway has become a high-density and efficient mode of transportation between cities.As the main vehicle of high-speed passenger transport,the reliability and safety of Electric Multiple Unit(EMU)in operation have become the primary concern in China with the increasing EMUs and their operation mileage.The EMUs have complex working conditions and functional hierarchy structures,so the forms and mechanisms of their faults are complicated.At the same time,due to factors like the lack of statistics such as field and test data and the non-standardization of such data,it is difficult to convey the probability of component failure through accurate numerical values.Furthermore,the fault relationship of EMUs' key components has priority correlation,functional relevance,order correlation and the logical relationship of spare parts(cold and hot spare parts,etc.)caused by system redundancy.Therefore,features,such as the ambiguity of fault status and failure rate of the EMUs' components,the uncertainty of the logic between components and the dynamic sequential,are found in the formation of faults.The traditional reliability analysis and modeling for EMUs can not be used to truly characterize the multi-fault status and dynamic sequential of EMUs.It is of great significance to explore the reliability modeling methods of high-speed railway EMUs for improving their maintenance strategy,enhancing the reliability level of the active EMUs and ensuring their safe operation.Based on the projects of “Reliability Analysis Method of EMUs on Lanzhou-Xinjiang Passenger Dedicated Line” and “Research on Key Technology of High Speed EMU Operation under High Wind Environment”,which are directed or participated by the author,this thesis looks into key issues like the failure probability acquisition and fusion method of key components of EMUs,fuzzy polymorphism reliability modeling of EMUs and fuzzy dynamic reliability modeling of EMUs.In the maintenance of the EMUs on the Lanzhou-Xinjiang dedicated line,a certain amount of operational data has been gathered with the features like small sample,multi-source heterogeneity and uncertainty.It is difficult to accurately express the failure probability of key components of EMUs,which limits the improvement of reliability analysis level of EMUs.In view of the problems above,the author integrates objective operational data and information provided by experts and put forth a data fusion method based on D-S evidence theory.Based on the analysis of current EMUs' detection and fault data,parameters like the major faults of EMU components and their rate are obtained,according to the attributes such as fault level and fault frequency.The expert investigation based on confidence index is used to gain the data about the major faults of EMUs and their rate,and the reliability or objectivity of the data are evaluated by experts themselves.The usability of the data is thus improved,and such data and operational data also effectively supplement each other.Finally,according to the different credibility of operational data and the data from the experts,the fusion weight is determined.The two types of data is combined according to the D-S Evidence Theory,which integrates subjective expert information and the objective data,ensures the availability of the failure rate data of the key components of the EMUs under the conditions of small sample and uncertainty,and lays a solid foundation for the reliability analysis of the EMUs' components.The EMUs have a complex functional hierarchical-structure.Besides,influenced by various conditions such as natural environment,maintenance and operation,the components of working EMUs often have many fault status,and their failure mechanisms are uncertain.A three-state system,including an intermediate fault state,an ideal working state and a completely failed state,is generally used for representating the multi-state systems.Because it has only one intermediate state,it is hard to represent the influence of multiple fault status on the reliability of the EMUs.Therefore,a fuzzy multi-state reliability modeling method for EMU components is proposed in this thesis.In the modeling,the uncertainty of the failure mechanism of complex components of EMUs is considered.A T-S fuzzy polymorphism fault tree model is constructed by T-S fuzzy fault tree and multiple states of faults are represented by fuzzy numbers(0,0.25,0.5,0.75,1).Considering the large computational complexity of the T-S fault tree,the T-S fuzzy multi-state fault tree model is mapped to a fuzzy bayesian network.The reliability analysis and calculation of T-S fault tree are carried out with it,which combines the dvantages of the fault tree and the bayesian network.The reliability parameters of traditional two-state,multi-state fault tree and fuzzy Bayesian network are compared and analyzed through a simple model,and the validity of the model is verified.In view of the dynamic sequential features such as the priority correlation,function correlation,order correlation and the logical relation of spare parts(cold and hot spare parts,etc.)caused by the system redundancy of the EMUs,a fuzzy dynamic reliability modeling for EMUs' components is proposed in this thesis.Considering the uncertainty of the fault mechanism of the EMUs' components,the analytical model of T-S fuzzy dynamic fault tree is constructed by using dynamic fault tree and T-S fuzzy theory,and the conditional probability table of the dynamic gate in the T-S fuzzy dynamic fault tree is determined.Thus the difficulty of calculation using the model is reduced.The reliability parameters of the dynamic fault tree and T-S fuzzy dynamic fault tree are compared and analyzed by a simple dynamic model,and the validity of T-S fuzzy dynamic reliability analysis model is verified.By using the T-S fuzzy multi-state reliability analysis model and fuzzy dynamic reliability analysis model,the reliability analysis of the pantograph system and the air braking system were performed respectively.The conditional probabilities of fuzzy polymorphism and dynamic gates are determined,and the reliability parameters of the EMUs' components are calculated.Thus the high-risk events and weak links of the pantograph system and the air braking system are found,which provides reference for improving the operational reliability of EMUs.