基于贝叶斯网络的多状态机载系统安全性评估方法

Research on Safety Analysis Approach of Airborne System Based on Bayesian Network

传统的安全性评估方法不适用于具有多状态属性的现代机载系统。根据系统与系统组成单元之间的状态关系,构建了基于贝叶斯网络的机载系统多状态安全性模型;运用通用生成函数给出了贝叶斯网络非根节点的条件概率表,基于变量消元法提出了系统失效状态发生概率计算方法,推导了系统组成单元重要度算法。结合某型飞机副翼控制系统给出了应用实例。结果表明,方法为解决多状态机载系统的安全性评估问题提供了一种简洁直观的方法,能够有效评估机载系统的安全性水平,确定各单元对系统安全性中的影响。

Traditional safety analysis methods are not suitable for modern airborne system with multi-state probability .In this paper ,a Bayesian Network based multi-state safety model of airborne system is built according to the state relationship between the system and its constituting components .Universal genera-ting function is used to build the conditional probability table of the non-root nodes .Based on the varia-ble elimination,a method of computing the system failure state′s probability is proposed .The algorithm of the importance of each component in the system is inferred .Combined with the aileron control system of an aircraft an example is given .The result shows that this method in this paper provides a simple and in-tuitive measure to deal with the safety analysis of airborne system with multi-state property .The airborne system safety level can be evaluated and the affection of each component to the system safety can be con-firmed effectively .