基于DBN的区域计算机联锁系统可靠性分析

区域计算机联锁设备是实现区域内行车安全、保证运输效率的核心设备,对其可靠性研究具有重要意义。结合目前存在的2种区域联锁制式,采用一种新的联锁方案,即在主控站和从控站(选择其中1站或多站)均设置联锁设备。综合考虑联锁系统的共因故障和可维修等因素,利用动态贝叶斯网络对其进行可靠性分析。首先,从系统故障■安全和危险输出的角度出发,建立区域两联锁单元和三联锁单元的动态故障树,并将其转换为相应的动态贝叶斯网络模型;然后利用动态贝叶斯网络的推理特性,对区域联锁系统进行可靠性分析;最后比较了该方法与基于静态贝叶斯网络和动态故障树分析法的结果。计算结果表明:主控站和其中之一的从控站均设置联锁设备是实现区域联锁的较佳方式;且基于动态贝叶斯网络的系统可靠性分析较上述两种方法在计算准确度和时间复杂度方面均有明显优势;并通过动态贝叶斯网络的诊断推理可知,共因故障是系统故障的主要原因,因此应重点防范以降低事故发生的概率。

Reliability analysis of regional computer interlocking system based on dynamic Bayesian network

Regional computer interlocking equipment is the core equipment to ensure the safety of regional traffic and transport efficiency, and the reliability research of it is of great significance. Combining the two existing regional interlocking schemes, a new interlocking scheme is proposed, that is, both the main control station and the slave control station (choose one or more stations) are equipped with interlocking equipment. Taking the common cause fault and maintainability of interlocking system into account, the dynamic Bayesian network is used to analyze the reliability. Firstly, from the perspective of system fault-safety and dangerous output, the dynamic fault tree of two interlocking units and triple interlocking units are established and then it is transformed into the corresponding dynamic Bayesian network models. By using the reasoning characteristic of dynamic Bayesian network, the reliability of regional interlocking equipment is analyzed. Finally, the results of this method are compared with those of static Bayesian network and dynamic fault tree analysis. The results show that it''s the best way to set up interlocking equipment in the main control station and one of the slave control stations and the reliability analysis based on dynamic Bayesian network has obvious advantages over the above two methods in terms of calculation accuracy and time complexity. Through the diagnosis and reasoning of dynamic Bayesian network, it is known that common cause fault is the main cause of system fault, so we should focus on prevention of it to reduce the probability of accidents.