论文部分内容阅读
针对事故树分析法(FTA)在风险评价中的局限性,在可控飞行撞地(CFIT)事故树的基础上,建立贝叶斯网络(BN)。运用推理运算对贝叶斯网络进行定量分析,通过分析计算数据,寻找主要事故致因,并提出对应的改进措施。再将改进措施引入到贝叶斯网络中,评价相关措施的有效性。结果表明,改进措施后,高度设置错误的后验概率最大,将成为预防CFIT的工作重点。最后指出贝叶斯网络方法是对传统的基于故障树分析的风险评价方法的有益改进。
Aiming at the limitation of FTA in risk assessment, a Bayesian network (BN) is established based on the accident tree of CFIT. The Bayesian network is quantitatively analyzed by reasoning operation. The main cause of the accident is found by analyzing and calculating the data, and the corresponding improvement measures are put forward. Then introduce the improvement measures into the Bayesian network to evaluate the effectiveness of the relevant measures. The results show that after the improvement measures, the highest posterior probability of height mis-setting will become the focus of CFIT prevention. Finally, it is pointed out that the Bayesian network method is a useful improvement on the traditional risk assessment method based on fault tree analysis.