Research On Safety Early Warning And Linkage Subject Selection Based On Subway Operation Accidents

The increasingly prominent urban traffic problems,with the continuous acceleration of urbanization,have been observed,which will not only negatively impact residents’ travel but also have significant impacts on economy,environment,society,and other aspects.The subway,given significant importance by many cities as a means to solve urban traffic problems,for its large volume of traffic,high punctuality rate,fast running speed,and low energy consumption.However,the unbalanced development of subway construction and operation,resulting from faster construction speed and a lack of experience in relevant operation management departments,has led to frequent operation accidents.Therefore,to ensure the safe operation of the subway,it is of great significance to deeply explore the causal factors of subway operation accidents,accurately provide early warning,and timely implement joint rescue measures.First of all,historical operation accident data of domestic subway were collected in this study,and operation accidents were classified and analyzed from four aspects of personnel,equipment,environment,management,and the severity of accident consequences.The impact of different accident causes on operation accidents was also investigated.Secondly,node attribute selection and data imbalance processing were carried out on the sample data.A subway operation accident severity prediction model was built using Bayesian network.The accuracy rate,accuracy rate,recall rate,and other indicators were selected for model evaluation.The data after unbalanced processing showed that all indexes could reach more than 85%.In particular,the prediction results before data imbalance processing were greatly improved.The accuracy of the training set and test set can reach 90.2% and 90.8%,respectively,with the prediction effect of major accidents verified.The comprehensive index F3 of the training set and test set can reach 91.5% and 90.1%,respectively.Based on the results predicted by the model,a corresponding warning strategy was formulated.The accident warning level was obtained by inputting the operation accident information,which could provide intelligent auxiliary decision-making for the subway management department to make accurate accident warnings at the initial stage of the accident.Finally,the emergency linkage handling of the subway system was taken as the research object in this study.An accident scenario module was built according to the accident warning information,and an accident handling module was built based on the situation module information.The accident handling process was analyzed based on the accident handling module.The main points of information were extracted from the accident situation information to determine the needs of accident rescue.The linkage subjects needed in different accident scenarios were determined based on the matching of responsibilities and needs of corresponding linkage subjects.The time and form of the linkage subjects participating in the linkage under different response levels were selected to provide auxiliary decisions for the subway emergency management department to deal with emergencies and improve the efficiency of accident disposal,thereby reducing the waste of rescue resources.