| Train control system is a key technical equipment and safe critical system for high-speed railway that automatically controls the safety headway of trains to prevent trains from overspeed.During the operation of the train control system,it covers at least 14 different complex operation scenarios.Under different operating scenarios,the train control system has different interaction behavior and control logic.Therefore,the risk coupling problem caused by complex interactions in potential dangerous causes of train control system is extremely complex and dangerous and has become an urgent problem to be solved in the field of safety analysis of high-speed rail train control system.In this thesis,a security analysis method based on multi-agent simulation is used to set up an RBC handover risk coupling analysis platform for the typical scenario of the CTCS-3 train control system—RBC handover scenario.This platform can provide a comprehensive and systematic analysis for risk coupling problems in RBC handover.The content of this thesis mainly has the following points:1.The thesis introduces the research status of train control system security analysis methods,analyzes the advantages and disadvantages of these methods,and then introduces the basic flow and related theories of the security analysis method based on multi-agent simulation.2.Based on the technical specifications of the train control system,this thesis abstracts the information interaction,function logic and state transition of the participants in RBC handover scenario.UML interactive diagrams are used to describe the information interaction among the participants,UML state diagrams are used to describe the state transition and the functional logic of the participants,and then a UML model of RBC handover scenario is constructed.Based on the model,using Mason-based multi-agent simulation tools and development techniques,a multi-agent simulation model of RBC handover scenario is constructed.This simulation model can describe the hybrid properties,concurrency and nonlinearity of the train control system.And the model can well describe the functional logic,state transitions,and information interaction of each participant in the RBC handover scenario.3.Using HAZOP method to identify the fault events of the RBC handover scenario,and then combined with the FARM fault model to construct the failure mode library of the RBC handover scenario.Then,in order to simulate the situation multiple faults occurring in the multi-agent simulation model dynamically and exploratorily,a fault-injection agent is designed.Using simulation-based fault injection techniques,the fault-injection agent can select faults based on fault scripts,combine them,and inject them into the multi-agent simulation model of RBC handover scenario to explore the impact of fault combinations on the simulation model.Finally,this thesis designs the risk-coupling rule learning agent.This agent can supervise and record the result of the fault injection and get the system simulation data.After all the fault combinations have been injected,the agent can learn the system simulation data and discover the risk coupling rules.4.This thesis establishes a risk coupling analysis platform for RBC handover scenario.Users can control the process of fault injection and observe the response of the simulation model to the fault through the controller.Then,based on the platform,this thesis verifies the correctness and completeness of the multi-agent simulation model.At the same time,this thesis uses the train exceeded the limit speed during the handover process as the monitoring object,and derives the risk coupling rules that lead to the overspeed of the train.Compared with the HAZOP method,it is verified that the security analysis method based on multi-agent simulation can comprehensively and accurately analyze the problem of risk coupling in RBC handover scenario. |
PDF Full Text Request