Modeling And Assessment Research On Reliability Of High-speed EMUs Braking System

The safety and reliability of the high-speed train which is an important means of modern transportation, especially the function of braking system, have been increasingly emphasized,and some scholars have made some research on the related question. At present, for high-speed EMUs braking system and its components, the research on failure modeling and assessment is not common, and most of the research is mainly applied to reliability evaluation on small amounts of railway vehicle components valve.In order to solve the reliability evaluation problem of the braking system of EMUs,beginning with cell-level components, GO-Ba method is put forward in this paper to establish the system reliability model by combining with the failure analysis of braking system and that of components. At the same time, the braking system components are product with high reliability, complicated structure and high cost, so this paper makes use of small sample theory to estimate lifetime and failure parameters for key components based on Bayes estimation and truncation test. And through quantitative analysis for braking system reliability,the lifetime model of system can be obtained to realize the reliability assessment of EMUs braking system.The paper mainly introduces the related reliability research as follows:1. On the basis of different reliability evaluation methods summarized, this paper makes use of multi-Bayes theory under small sample condition to estimate the failure rate of exponential distribution for reliability life assessment of key component brake controller of braking system. And then using type-I life test data of the brake controller, the model can be verified so that reliability mathematical model of the parameter failure rate is established under small sample condition.2. For a posteriori distribution of Bayes analysis with the complicated high-dimensional integral operation, this paper employs Markov Chain Monte Carlo integration algorithm by combining with the modified likelihood function to establish reliability mathematical model of component EP valve whose lifetime obeys two-parameter exponential distribution. And then, it compares the reliability mathematical model of EP valve obtained through Bayes estimation method based on MCMC algorithm with that of modified maximum likelihood and multi-Bayes estimation methods. The compared result shows that the former can preferably solve the life assessment problem of the brake system components, and proves that the Bayes method can take advantage of MCMC algorithm for parameter assessment of two-parameterexponential distribution.3. Based on the truncation test and multi-resources information of prior information, the paper puts forward a method combining Bayes theory with GO method, which is referred to as the GO-Ba method for short. On the basis of reliability estimation making use of Bayes theory for key component, GO method is used to establish the system reliability model, so that the reliability estimation of EMUs brake control system is conducted, which can provide a certain theoretical basis for the stable and safe operation of high-speed EMUs.