Abstract: In order to systematically promote and guarantee the safety of high-speed trains, based on the safety accidents of high-speed trains, the safety risk index system of high-speed trains is obtained by using fault tree analysis and transformed into Bayesian network. Combining with the fuzzy processing method of cloud model, the qualitative natural evaluation language of experts on each station evaluation index was transformed into corresponding fuzzy numbers and the traditional TOPSIS model was improved. Considering subjective and objective weights comprehensively, the TOPSIS model with optimal combination weights was obtained, and the safety of high-speed railway trains was evaluated accordingly. Based on trapezoidal fuzzy number and Buckley method, the relative prior probability of network nodes was calibrated, and the relative posterior probability of each node was calculated by Netica software when the accident occurred. By comparing the data before and after the analysis, the results show that the main safety risk of Beijing-Shanghai high-speed railway train in the section from Beijing south to Jinan West was fire and explosion. By analyzing the main potential risk sources and proposing the control measures, the safety control of high-speed railway train can come true.

Key words: high-speed railway train, safety evaluation, management and control, technique for order preference by similarity to ideal solution(TOPSIS), Bayesian network