Reliability Estimation for the Braking Systems of High-Speed Electric Multiple Units Based on Bayes Inference and the GO Method

摘要

To solve the reliability estimation problem of the braking systems of electric multiple units (EMUs), component and system models are evaluated by the use of two types of Bayes methods and the GO method in this paper. Beginning with cell-level components, the two Bayes methods, including the multi-Bayes method and a Bayes method based on the Markov chain Monte Carlo (MCMC) algorithm, are applied to estimate the parameters and the failure rates of components whose lifetimes obey a two-parameter exponential distribution model. On the basis of component estimation, application of the GO method achieves a quantitative analysis of the reliability of EMU braking systems based on the reliability characteristics of the braking system components and the logical relations between system operators. Finally, the capability of the two Bayes methods and the GO method to solve the reliability evaluation problem of the braking system is verified by experimental and test data, which provides a theoretical basis for the safe operation and maintenance of EMU braking systems.