基于Bayesian Network列控系统超速故障的分析

Bayesian Network Based Fault Analysis in Train Control System of Train Overspeed

摘要:

列控系统非常复杂,在对其进行故障分析时,自身的特质决定了用传统的故障树在对其进行故障分析时会产生局限性;为此文中引入了贝叶斯网络技术,充分利用其推理算法成熟、理论基础非常完备、学习能力非常强的优势,将事件树中各环节的故障树用贝叶斯网络进行描述;利用贝叶斯网络工具箱(BNT)对列车超速故障进行因果和诊断推理分析,通过数值计算结果得出所在不同场景下故障发生的主要原因,并提出减少故障发生概率相应的措施。

Abstract:

The train control system is a complex system, when on the failure analysis, the train control system itself characteristics determine the using traditional fault tree analysis of the fault can be induced in limitations. Introduced the Bayesian networks for this technology, make full use of its mature reasoning algorithm, learning ability is very strong, very complete theoretical basis of advantages, the fault tree of each section of the event tree using Bayesian network is described. Using Bayesian network toolbox(BNT) to train speeding the causation and fault diagnosis reasoning analysis, through the results of numerical calculation of the main causes of failure under different scenarios, and put forward the corresponding measures to reduce the probability of failure.