Risk Analysis And Control Of Train Control System Based On Dynamic Fault Tree

The CTCS-3train control system plays a very important role in ensuring traffic safety, which uses a wireless communication system to achieve continuous, two-way transmission of information between the train and ground, so the risk analysis of the communication system between train and ground has important significance. The traditional fault tree analysis is not an effective method to analyze the reliability of system with dynamic random and dependency, therefore this paper research the risk analysis and control of the communication sub-system between train and ground based on dynamic fault tree analysis method.(1) Based on the above basic theory of dynamic fault tree, this paper studies the modular method of dynamic fault tree. The dynamic fault tree is divided into static sub-tree and dynamic sub-tree, then Markov chain and static method is used to analyze. At last, this paper create and analyze dynamic fault tree model of the communication system between train and ground, which prove this method improves the accuracy and completeness of the model.(2) With the explosion problem may be generated by Markov model, this paper introduces a dynamic fault tree analysis method based on Bayesian network. Based on the system probabilistic safety assessment of Bayesian network, this paper assesses and analyzes the simulation results of the communication system between train and ground.(3) With more and more events, and more complex of the system, the Markov chain nodes and state sets present exponential growth, and the event is required to accord with exponential distribution; At the same time, the analysis algorithm based on Bayesian network of is huge workload and low viability. Therefore, this paper introduces the dynamic fault tree analysis method based on the trapezoidal formula. Since this method has not limit to the probability distribution of the events, so it has better adaptability; while this method use approximate method to solve the model explosion problem, reducing the cost of solving the model. This paper also introduces the Monte Carlo simulation method, which can analyze the dynamic fault tree with event which does not obey the exponential distribution.(4) The main work is to control risk after the risk analysis. This paper put forward the fuzzy comprehensive evaluation method of maintenance decision based on the reliability-based maintenance (RCM) theory; Finally, this paper put forward a two-level fuzzy comprehensive evaluation model, and a case is presented to show that this method is feasible.