Safety Analysis On Chinese Train Control System3Based On Bayesian Network

Chinese train control system3(CTCS-3) is the core of safety equipment for ensuring the safety and efficiency operation of the high-speed train. The goal of security assessment is to decrease the probability of accidents through identifying and analyzing security risks and to determine safety integrity level of the system, as well as proposing measures that are used to guarantee security of system. Security analysis of CTCS-3is of important theoretical and practical significance for improving the safety of train control system. And the main work of this paper is as follows:(1) The background knowledge for the CTCS-3and safety assessments as well as Bayesian network has been introduced in this paper. The method of converting fault tree model to dynamic Bayesian network related to AND gate and OR gate as well as order-related gate has been introduced briefly. And related formulas of security analysis using dynamic Bayesian network also have been given.(2) In order to analyze safety of CTCS-3comprehensively, we use the fault mode of the speed/distance beyond safety limits as a starting point. Basing on more mature conversion method from static fault tree model to dynamic Bayesian network, and referencing mature fault tree model of the ETCS-2, the safety of CTCS-3has been analyzed from bottom to top through modeling for13fault modes, such as the train positioning error and the positioning speed error. The failure probability of the fault mode and failure probability of corresponding network intermediate node has been obtained. The results show that the objects of study meet the requirements of the SIL4.(3) The problem of dynamic sequential logic failure mode possessed that exists during the train operation has been analyzed and modeled by using dynamic Bayesian networks. For example, hot-standby redundancy architecture and the failure of track detection as well as the emergency braking due to degraded situations have been analyzed and modeled by using dynamic Bayesian networks in detail in this paper. The effectiveness of dynamic Bayesian networks theory that is applied to the safety analysis in CTCS-3has been verified.