摘要
以危险品航空运输事故树为基础,建立了贝叶斯网络(BN)。运用贝叶斯原理以及推理计算,找出了危险品航空运输事故的主要致因,并分析了改进措施对危险品航空运输系统的影响程度。结果表明,引进改进措施后,事故一发生就已经超出可控范围的后验概率最大,将成为预防危险品航空运输事故的工作重点。研究表明,风险评价领域中的贝叶斯网络分析法是对事故树方法的有益改进。
The paper is intended to present our researeh results of some key accident-prone factors, such as the failure of training, the threat brought by the new products likely to cause transportation haz- ards yet unknown or not enough known to the airline crew members. In addition, it is also found that the fault-tree approach also has its own limitation for the hazard assessment, which pushed us to map up a fault tree of air transport of hazardous goods by putting the fault-tree information onto a Bayesian network, which includes a conditional probability distribution table for the need of assessing the system work under the inference of the Bayesian network. In doing so, we have analyzed the Bayesian network quantitatively while using the inference algorithm of Bayesian network through analyzing the data calculation. Next, we have put forward the improved measures such as ensuring the service quality of the crew members by training according to the service demands of CCAR - 276. 159, ensuring the operation regula- tions establishment concerning the air transport of hazardous goods and improving the specific qualification system for the air transport of hazardous goods. Then, we have done inspections on the effective- ness of related measures with the help of Bayesian networks. Thus, it can be seen that the Bayesian network can help us to ensure a com- pletely new hazard assessment in air transport of dangerous goods. In doing so, successful adoption of Bayesian network can greatly con- tribute to the effectiveness of the related improved measures, which can reduce the possibility of air transport of hazardous goods. The last but not the least is that the effectiveness of the improved measures taken helps to reduce the potential of accidents. In a word, an exam- ple on air transport of hazardous goods illustrates that the Bayesian network approach is a good substitute of the fault tree approach for risk assessment, and the perspective of using Bayesian network to make hazard assessment will find a wide application.
出处
《安全与环境学报》
CAS
CSCD
北大核心
2010年第5期163-166,共4页
Journal of Safety and Environment
