区域计算机联锁系统安全性分析的动态故障树模型与方法研究

摘要

At present,the main method to analyze safety of dynamic redundancy system structures at home and abroad is based on the Markov model.However,when using the Markov model to analyze the regional comput-er interlocking system,the scope of state space is quite large such that the modeling process and solving proce-dure become very complex.To solve this issue,this paper established the Dynamic Fault Tree model of region-al computer interlocking system from the perspective of system failures,and applied the probabilistic approxi-mation method to solve the probability of falling safety and probability of falling danger,and eventually per-formed a comparison with ones of the Markov method.The researches show as follows:the DFT probabilistic approximation method possesses roughly the same outcome with the Markov method,but its calculation process is quite simple and storage capacity is reletively small;this indicates that the DFT probability approxi-mation method is feasible to calculate the relevant safety indexes for railway signal safety critical systems fea-tured by low failure rates and short maintenance time,and it avoids the heavy and complicated process of the Markov method and provides a new solution to complex dynamic redundancy system safety analysis.%目前，国内外关于动态冗余系统安全性分析的主要方法是 Markov模型。对于区域计算机联锁系统，使用Markov模型进行分析时，其状态空间规模较大，致使其建模及求解过程十分繁琐。为了解决这个问题，本文从系统失效的角度建立区域计算机联锁系统动态故障树模型，采用概率近似法求出系统的安全失效概率和危险失效概率，并将所得结果与 Markov方法所求指标进行比较。研究结果表明，故障树概率近似法和 Markov 方法计算结果十分接近，但计算过程简单且对存储量要求小。说明对于低失效率和短维护时间的铁路信号安全苛求系统，采用故障树概率近似法计算有关安全性能指标切实可行，避免了利用 Markov 方法求解的繁琐过程，为复杂冗余动态系统的安全性分析与计算提供了新的解决方案。