Fully Automatic Operation System Leading Indicator Identification Based On STAMP

In recent years,fully automatic operation(FAO)has gradually been popularized in urban rail transit in major cities,and has become the main development trend of modern urban rail transit.However,with the improvement of the automation of urban rail system,the operation safety also faces new challenges.Therefore,in order to prevent the occurrence of safety accidents and ensure the safe operation of the FAO system,higher requirements are put forward for its safety risk management.The thesis introduces the concept of leading indicator as the safety monitoring parameter of system operation,which is used to identify the increase of accident risk,so as to take measures to prevent the accident before it occurs.At present,the research on the leading indicator of rail transit system,especially that of operation control system,is still immature.Therefore,based on the systems-theoretical accident modeling and processes(STAMP)model,this thesis proposes a leading indicator identification method,which can effectively identify the leading indicator in various scenarios of the FAO system,thus improving the dynamic detection capability of the FAO system's operational safety risks.The main innovations of the thesis are as follows:(1)Aiming at the unclear control relation in the case of multiple controllers of STAMP model,the thesis combines the security defense characteristics of FAO system,and proposes to apply Bow-tie thought to guide the construction of STAMP hierarchical control model,so as to endow STAMP hierarchy with richer connotation and form hierarchical security defense model.Compared with the traditional STAMP hierarchical control structure,the new model can not only clarify the cooperation relationship between controllers,but also show the dynamic change process of safety risks under the action of internal and external deviations.(2)Aiming at the problem that STAMP model cannot accurately describe the complex interaction of system behaviors and lacks formal syntax and data structure support.The thesis transforms the security defense model into a colored petri net(CPN)model and formalizes the security defense model,which not only avoids the ambiguity of natural language,but also describes the dynamic behavior and state changes of the system.(3)According to the current situation that the definition of leading indicator is not uniform,this thesis defines and classifies leading indicator based on the characteristics of FAO system.On the basis of this definition,a leading indicator identification method is proposed by tracing analysis of the reachable path of security defense layer failure.In order to solve the problem of low efficiency and error-prone problem of manual identification of reachable paths,tools are developed to realize automatic search of failure paths in the security defense layer,thus assisting the leading indicator identification of the FAO system.Finally,the leading indicator identification method proposed in this thesis is applied to the leading indicator identification of the platform departure scenario of the FAO system,and the automatically searched reachable path of safety defense layer failure is analyzed to realize the leading indicator identification,which proves the feasibility of the method proposed in this thesis.