基于反向模糊Petri网的航空事故致因分析

Causative analysis for the aviation accidents based on the reverse fuzzy Petri net

加强航空运输安全隐患的预防控制,能够降低航空事故发生的概率。以257份航空事故调查报告为基础,基于复杂网络理论,利用节点的度中心性、中介中心性、接近度中心性,结合逼近理想解排序法对45个致因因素进行节点重要度排序,识别"人-机-环-管"中的关键因素;在21个关键致因的基础上,运用模糊Petri网的故障诊断特性,提出了相应的逻辑规则,并构建了航空事故模糊Petri网,利用推理算法对相关子系统进行有效提取;进而,以结果命题"飞行员操作失误"为例,根据该命题的真实度,通过构建反向模糊Petri网,反向推理得出各相关命题的可信度。结果表明,恶劣天气相关类别的可信度最高。最后,利用正向模糊Petri网推理算法验证了该方法的可行性。

The given paper is intended to strengthen the prevention and control of the potential hazards in the air traffic and transportation so as to reduce the probability of the aviation accidents.For the said purpose,first of all,we have calculated 45 causative factors in utilizing the degree centrality,the betweenness centrality,as well as the closeness centrality of nodes based on the investigation reports of the 257 aviation accidents and the theory of the complex network by analyzing 45 causative factors as the aviation accident causative network nodes.And,then,in combination with the TOPSIS method(short for the technology for order preference by similarity to the ideal solution),we have sorted out the main causal factors based on its importance and identified 21 key factors by classifying the impacts of the factors,like those of “the human,the equipment,the environment and the management”.And,then,based on the 21 key causative factors,the corresponding logic rules have been proposed via the fault diagnosis features of the fuzzy Petri net.And,next,we have also set up the aviation accident fuzzy petri nets,so as to draw out the relevant subsystems based on the reverse reasoning algorithm.And,later then,we have taken the result proposition of “the pilot operation errors” and its reliability as example so as to gain the credibility of each related proposition in terms of the reverse reasoning.The results of our proposition and examination indicate that the credibility of the low visibility should be the highest one responsible for the accidents,which means that low visibility should be mostly responsible for the pilot operation irregularity.And,finally,the feasibility of this method can also be verified by the forward fuzzy Petri net algorithm.Thus,the above case samples can help us to verify that the method for analyzing the cause analysis of the aviation accidents based on the reverse fuzzy Petri net proves to be justifiable.And,judging and determining whether the factor is the most important factor for causing the accident,it would be possible for us to reduce and even eliminate the likely aviation accidents.At the same time,it can also help us to formulate and predict other kinds of research methods to be responsible for such accidents.