Research On Comprehensive Evaluation Model Of Urban Rail Transit Operation Safety

Because traffic demand is infinite while urban road resources are limited, the development of urban rail transit has attracted much attention. Especially, since cities such as Beijing, Shanghai and Guangzhou have entered network operation period, static evaluation and dynamic micro reorganization have become important projects for domestic and foreign scholars. Analysis and reorganization of safety issues of urban rail transit is massive, multi-index, complex system engineering. It is a task combing macro and micro, static and dynamic, qualitative and quantitative, the depth and breadth of which will influence the status of urban rail transit in urban transport systems,. Therefore the research subject is significant.In this context, the article clearly raises problems of screening indices, weight determination in static evaluation, and problems of diagram evolution and quantitative analysis as focal point in this research. Secondly, according statistic analysis of 78 foreign and 859 domestic operational incidents of rail transit, reasons of incident types, time, space, levels and delay and network passenger flow are systematic described and explained. Then, based on accident statistic research, main five factors of staff, facilities, environment, management, network operation are analyzed. Concept index system and selected index system including 5 first class indicators,21 second class indicators and 15 third class indicators are proposed. Combing entropy weight models which could represent more differentiation and matter-element models which could assign according to index difference, static evaluation models of urban rail transit safety based on entropy weight-extension and matter-element are developed. As an example, Line 4 of Beijing is calculated and the result shows rationality and validity of the model. Finally, considering the static models having shortcomings such as much macro and less specific, cause-consequence analysis which could analyze top event probability utilizing accident tree and reflect dynamic deduction in time aspect is introduced. Dynamic perceptual model of urban rail transit accidents based on CCA models is developed. As entry points of motor overheat causing fire, causal analysis charts and reliable computing method are used to achieve quantified analysis in accident period. The result indicates automatic fire extinguishing period and whether operation staff successfully extinguishes fire, as two important factors, are efficient endpoints, thus demonstrating model effectiveness.