Safety Research On Virtual Coupling Scenario Based On Dynamic Risk Analysis

As a breakthrough innovation in the rail transit industry,virtual coupling mainly uses low-latency vehicle-to-vehicle communication technology and precise train control algorithms to reduce the tracking distance of trains,and improve the operating efficiency of the line.However,with the shortening of safety margin between trains and the increase of system complexity,the operation safety of the system is required to be higher.At the same time,virtual coupling needs to take into account the line,vehicle,signal,communication and other aspects of the conditions,in order to break through the original constraints to complete the function of virtual coupling and line capacity requirements.Once a reasonable decision cannot be made within a limited time during the hazard processing of virtual coupling,it will likely cause huge economic losses.At the same time,according to the train running state and equipment condition,the possible hazard factors that may exist in the running process can be dynamically detected,providing decision support for dispatchers,which is of great significance to improve the safety of train virtual coupling operation.Based on the equipment status,interactive information and train running status,the paper constructs the fault propagation model in the causal scenario.Build dynamic Bayesian networks capable of probabilistic inference,the dynamic risk assessment of virtual coupling scenarios is realized.The main contributions of this dissertation are as follows:(1)Based on the comparative analysis between the vehicle-to-vehicle communication and the traditional train operation mode,according to the functional requirements of train virtual coupling,design the train control system framework applied to virtual coupling.The train running state under virtual coupling and the mutual conversion relationship between states are determined.The functional relationship between the train running interval of the virtual coupling and the driving speed of the front and rear vehicles,the communication delay and the train running model is clarified.Determine the virtual coupling control methods in different scenarios,and establish a security protection strategy based on elastic adjustment.(2)After safety analysis based on STPA and leading indicator identification,early warning is carried out by identifying the equipment status in the actual operation of the train and the safety leading indicators in the information exchange process.Establish a safety monitoring mode based on leading indicators to visually display the dynamic changes of leading indicators of train operation.Analyze equipment status and information interaction process,predict possible dangerous events and provide awareness of risk situations.(3)According to the monitoring process of the leading indicators,determine the fault propagation process in the causal scenario,and based on the establishment of a dynamic Bayesian network model,logical inferences are made on the occurrence probability of accidents.Using Ge NIe2.0,a scientific computing tool based on Bayesian network,to make posterior probability inference for dynamic Bayesian network.(4)Taking the western section of Beijing Metro Line 11 as a case simulation background,set up a causal scenario under the actual line state.The process of the method is demonstrated through the application of the risk assessment in the established virtual coupling scenario and result validity.By constructing causal scenarios and risk assessment models,the dynamic risks of trains in specific coupling scenarios are obtained.The results show that the method proposed in this paper can identify the hazard signs before the hazard actually occurs,which can be used as a theoretical basis for the relevant personnel making decision.This thesis includes 55 figures,22 tables,72 references.