基于贝叶斯网络考虑共因失效的高速铁路牵引变电所可靠性分析

摘要

Compared with traditional Fault Tree analysis, the method of Bayesian Network based on Fault Tree is more flexible to express some uncertain problems and proceed with some uncertain reasoning. This paper proposes the exponential distribution as the life distribution of railway traction substation and establishes Bayesian Network of main electrical connections for railway traction substation. According to the two state characteristics of the system and the reliability conforming to exponential distribution, βfactor model is selected as the model of common cause failure of railway traction substation and the explicit analysis is taken as the method of analysis. Bayesian Network that addresses common cause failure is programmed based on MATLAB. It also draws the diagram of reliability and instantaneous availability . The results indicate that the reliability becomes lower with consideration of the common cause failure, which means the results are closer to actual value. Besides, the qualitative and quantitative assessment of traction substation is fulfilled, the reliability index of traction substation is calculated and the system weakness is identified.%运用一种基于故障树的贝叶斯网络分析法，相比于传统的故障树分析法，它能够灵活地表示不确定信息，并能进行不确定性推理。并提出牵引变电所的寿命分布为指数分布，建立牵引变电所电气主接线的贝叶斯网络模型，对于系统二态性特征以及系统可靠度也符合指数分布，故在牵引变电所中选取β因子模型作为共因失效模型，显式分析方法作为共因失效的分析方法。用MATLAB对考虑共因失效的贝叶斯网络进行编程，并绘制相应的可靠度以及瞬时可用度的曲线图，结果表明在考虑共因失效后系统的可靠度有所降低，意味着得出的结果更接近实际。此外，实现牵引变电所定性和定量的评估，计算其可靠性指标，确定系统的薄弱环节。