为解决传统方法在状态多样化的复杂系统可靠性分析中的局限性, 通过贝叶斯网络的推理功能得到系统故障概率与可靠度评估、后验概率及根节点重要度的计算和分析的方法, 充分利用了贝叶斯网络不确定性推理和图形化表达的优势.文章以汽车软关系统为研究对象, 在系统可靠性框图模型的基础上, 应用新方法计算了各 根节点失效概率和重要度并进行了排序, 针对系统的薄弱环节, 提出了相应的改进方案, 为系统的产品和类似系统的可靠性研究提供了理论参考, 具有较强的工程实用价值.

In order to solve the limitations that traditional methods showed in the reliability analysis of complex systems with multiple states, an improved reliability analysis method based on fuzzy multi-state Bayesian networks is proposed taking advantage of the uncertainty inference and graphical expression of Bayesian networks. The analysis process is elaborated to calculate the reliability, the posterior probability and the importance degree through the reasoning function of the Bayesian network. In this paper, taking the integrated closure system as the object, based on the establishment of the reliability index and failure criterion and reliability block diagram model, the method based on of fuzzy Bayesian network is applied to system reliability assessment and analysis. Fuzzy importance degree and critical importance degree based on prior failure probability and current important degree based on current state of some root nodes known are computed for each root node and sorted. According to the results of reliability allocation and analysis, the weaknesses of the integrated closure system are proposed and implemented based on the preliminary design, including the design and debugging. This method is also a good reference for the reliability research of similar systems.