摘要
由可控飞行撞地造成的死亡人数在民航运输飞行事故中占第1位,因而研究可控飞行撞地的原因并采取预防措施对保障飞行安全有重要意义。介绍了事故树模型和贝叶斯网络之间的内在转换关系,采用事故树分析了可控飞行撞地事故发生的主要原因是人为因素和技术因素,其中人为因素包括机组、管制员和维修人员的失误,技术因素包括燃油重量、风速风向、能见度、飞机速度、偏离航线等,最后将事故树转化为贝叶斯网络。以人为因素中的机组人员失误为例,用贝叶斯网络模型对失误原因进行了分析,并计算出机组人员失误概率。
The paper takes its target to analyze the causes of accidents with controlled flight bumping into terrain and ways to take pre- ventive measures so as to ensure the flight safety particularly in civilian aviations. As is known, according to the death toll of air crashes, accidents due to the controlled flight bumping into terrain rank No. 1 in the death toll of air crashes. It is just for this need that we have established a fault tree analysis model hoping to make quantitative and qualitative analyses of the causes of such flight accidents. But there are still some assumptions saying that fault tree analysis, for instance, bi-state assumption, is not reliable in analyzing human factors in the accidents. However the characteristic features of Bayesian net- work model can supply a gap of fault tree analysis, such as description of multi-state information of the basic event and of indefinite logic relation among the events. In addition, the transformation principle has also been introduced from the fault tree analysis to that of the Bayesian network. In so doing, the fault tree analysis can be applied to analyze the accidents of controlled flight bumping into terrain only to find human factors and technical factors are the main causes, including airerew errors, air controller errors and the ground airflight maintenance errors, whereas technical factor errors may cover air fuel-supply weight, wind speed and orientation-piloting, visibility, air speed and airway deviation, etc. Then the data concerned can be put into Bayesian network. In this way, the airerew errors can be used to be analyzed along with the Bayesian network model while the proba- bility of aircrew errors is calculated. Thus, our research work can provide a new approach to the cause analysis and accident-prevention of such controlled flight accidents.
出处
《安全与环境学报》
CAS
CSCD
北大核心
2009年第5期136-139,共4页
Journal of Safety and Environment
基金
中国民航大学科研基金项目(05QD08S)
关键词
安全管理工程
可控飞行撞地
贝叶斯网络
事故分析
概率计算
safety management engineering
controlled flight into terrain
Bayesian network
accident cause analysis
probability calculation
