民机驾驶舱自动化人机失效事故链的探寻及阻断  

Exploring and blocking man-computer failure accident chain for civil aircraft cockpit automation

作　　者：杜红兵[1] 吴森[1] 张庆庆[1] 王健新[1] 

机构地区：[1]中国民航大学安全科学与工程学院

出　　处：《中国安全科学学报》2015年第10期71-77,共7页China Safety Science Journal

基　　金：中央高校基本科研业务费资助项目(SY-1443)

摘　　要：民航飞机驾驶舱自动化应用,面临诸多新的人机交互问题,如错误选用飞行模式、过分依赖自动驾驶等。为探寻其致因机理并找出预防对策,构建民机驾驶舱自动化人机失效的HAFCSTER模型,结合国内外相关文献,研究世界近年发生的典型31起民航事故,找出各事故致因的高频因素及主要事故链。构建数据集合,用JAVA语言模拟影响驾驶舱自动化人机失效的因素组合,结合因素的辨识难度确定最佳预防方案。结果表明,标准操作程序(SOP)存在的错误、企业教育培训不足等12项失效因素最易辨识,完善这些失效因素,阻断所有事故案例的事故链,即能获得预防该类事故的最佳方案。The civil aircraft cockpit automation brings new man-computer interaction problems such as wrongly chosen flight mode,over-reliance on autopilot,etc. To explore causes of the problems and the mechanism through which the causes created the problems and to identify preventive measures. An HAFCS-TER model was built for the civil aircraft cockpit automation man-computer failure accident chain. Data on typical 31 international civil aviation accidents occurring recently were analyzed. Main causes and accident chains were identified for the accidents. Sets of inductive factors were built. Combination of the factors was simulated using the JAVA programming language. The optimal prevention program can be worked out on the basis of factors identification difficulty. The results indicate that 12 inductive factors such as mistakes in the standard operating producers and lack of corporate education and training are most recognizable,and that improving these failures factors can block accidents chain accidents.

关 键 词：驾驶舱自动化 人机失效 HAFCS-TER模型 JAVA语言程序 集合构建 

分 类 号：X912.9[环境科学与工程—安全科学]