Safety Status Evaluation And Prediction Based On Safety Region For Rail Transit Network

Rail transit systems have become the backbone of urban public transport system, and effectively ease the pressure of urban traffic. Strengthen the research on the safety status recognition and prediction of rail transit network, is to improve the risk control capabilities and safety under the condition of network operation, achieve significant changes from passive safety to active safety, has important theoretical and practical significance to ensure safe and reliable operation of rail transit system. The research on safety status recognition and prediction of rail transit network mainly focus on the following aspects.Firstly, based on the analysis of the development of rail transit, the basic characteristics of Beijing rail transit network show that it has the characteristics of poor overall connectivity and the network is fragile. This study makes an in-depth analysis on the existing safety region theory, safety status recognition and prediction, and puts forward a safety region model based on safety state variables.Secondly, the method of extracting the safety status feature of the rail transit network based on principal component analysis is proposed. Taking Beijing rail transit network as an example, the safety status feature collected and extracted from safety and reliability of the rail transit network. Through analyzing the components, safety state variables have strong correlation and the variables about passengers and key equipments have strong influences on safe operation of rail transit. Two main components are extracted, which keep most of safety status information of rail transit network, and provide data support for the safety status recognition.Thirdly, taking the fuzzy boundary among different safety status into account, the IT2FCM and TOPSIS based safety status evaluation method is put forward and has more accuracy in evaluating the safety status of rail transit network than the typical FCM method. TOPSIS method objectively quantify the safety status of rail transit network and safety region estimation based on safety state variables is better expressed by TOPSIS based safety grade classification of rail transit network.Finally, based on the research results of the safety status recognition, the ARMA and GA-SVR based rail transit network safety status prediction are discussed. The results show that GA-SVR model has more accuracy in predicting the safety status of rail transit network, which achieve significant changes from passive safety to active safety.