Safety Analysis For The New Train Control System Based On Statistical Model Checking

With the continuous development of train control systems,the new “onboard-equipment-centric” train control system integrates the functions of the traditional ground route control and movement authority calculation into the on-board equipment,and adopts all-IP and train-to-train communication technologies,which has become a research hotspot.As a typical safety critical system,the new train control system can cause significant or even catastrophic losses to human lives,properties and the environment,once it fails in the operation process.Therefore,the safety verification analysis is of great significance for the new train control system.However,in the new train control system,on the one hand,the on-board equipment undertakes more functions of the ground equipment,and the coupling between the fault modes withing the system becomes much higher;On the other hand,it is difficult to identify external communication faults of the system after the introduction of all-IP and train-to-train communication technologies.Aiming at the safety characteristics of the new train control system,this dissertation proposes a safety analysis method based on statistical model checking(SMC),which abstracts the behavior of the new train control system into a network of priced timed automata(NPTA)model.The statistical sample data of system safety analysis is generated by "model simulation",and the qualitative and quantitative analysis of safety is carried out by the statistical model checking algorithm.Based on this,typical operating scenarios are selected to examine the effectiveness of this method.The main work of this dissertation is as follows:(1)The safety of the new train control system is discussed.This dissertation analyzes the structure and function of the new train control system,while focusing on the general operation scenarios of it.By comparing different safety analysis methods for train control systems,statistical model checking is proposed to analyze the safety of the new train control system.(2)A safety analysis framework based on statistical model checking is proposed for the new train control system.Firstly,the operation scenario of the new train control system is analyzed,and the hazard and operability study(HAZOP)is carried out for this system at both the interface level and function level to obtain the list of hazard sources.Secondly,combined with the list of hazard sources,NPTA model of new train control system operation scenario behavior is established.Sample data of system safety analysis is generated through "model simulation".Safety analysis of new train control system is conducted from qualitative and quantitative aspects by using statistical model checking algorithm(hypothesis testing and probability estimation).Finally,based on the NPTA model of the operation scenario,the simulation data of the system NPTA model is derived for deeply discussing.(3)A case study of the system safety analysis for the typical operation scenarios of the new train control system is carried out.Based on the safety analysis of the new train control system's entering into the station and tracking in the section,the hazard identification of the top-level requirements of the new train control system is carried out based on HAZOP.The subsystem fault identification model and communication fault identification model are introduced according to the hazards identified and the corresponding probability distribution.The operation scenario model is constructed based on the network of priced timed automata(NPTA),and the safety of the new train control system is verified qualitatively and quantitatively based on the model.(4)An automatic analysis tool is developed for the statistical model checking of this system.The data characteristics and probability distribution of entering into the station and tracking in the section are analyzed quantitatively by this tool.And on this basis,based on the random forest regression data analysis model,the prediction of normal and fault output probabilities at a specified time point is realized,which provides a reference for the further study of the new train control system.The research results show that the safety analysis framework of the train control system based on statistical model checking can better simulate the information interactions and state transitions between the modules in the system operation scenarios.And this dissertation achieves the qualitative and quantitative safety analysis of the new train control system through the probabilities of the failure and normal outputs of the model simulation system.This dissertation has 79 figures,30 tables and 85 references.