论文部分内容阅读
为研究危险化学品道路运输事故中各影响因素的结构关系及其重要程度,以2006—2012年162起事故数据为样本,借助贝叶斯网络进行事故影响因素间关系分析及概率推理。网络结构由D-S证据理论综合专家意见得出,并依据条件独立检验实施了修正;表征影响因素对事故影响程度的后验概率由EM算法学习并进行推理得到。研究结果描述了事故影响因素间复杂的作用关系,推理分析表明,最重要的3个直接因素依次为人的因素(0.558)、运输车辆和设备(0.318)、危险化学品包装及装卸(0.212),且目前降低运输事故的应对措施重点应放在对人的失误及运输企业的管理上。
In order to study the structural relationship and importance of various influencing factors in road transportation accidents of dangerous chemicals, 162 incident data from 2006 to 2012 were taken as samples to analyze the relationship between accident factors and probability reasoning using Bayesian network. The network structure is derived from the comprehensive expert opinion of D-S evidence theory, and is amended according to the condition independent test. The posterior probability characterizing the influence degree of the influencing factors on the accident is learned and deduced by the EM algorithm. The results of the study described the complex interaction between the factors affecting the accident. Inference analysis showed that the three most important direct factors were human factors (0.558), transportation vehicles and equipment (0.318), hazardous chemicals packaging and handling (0.212) At present, the measures to reduce transportation accidents should focus on human error and the management of transportation enterprises.