船舶溢油应急反应人的可靠性分析研究

Human Reliability Analysis on Ship Oil-spill Accident Emergency Response

为减小船舶溢油事故应急反应人员失误概率,提升应急处置效果,在对应急人员可靠性主要影响因素识别的基础上,利用模糊贝叶斯网络(BN)建立船舶溢油应急人的可靠性分析(HRA)模型,将应急反应流程抽象为可进行概率推算的BN,实现单个应急任务和全过程人的可靠性的量化评估,并依据该模型对一起船舶溢油事故进行实例研究。结果表明,首先模糊集值专家评估较好地解决应急人员可靠性情境依赖性强,难以量化评估的问题;其次,BN卓越的概率推理性能实现由单个应急任务向全过程人的可靠性的推算;最后得到应急反应人的可靠性主要取决于关键应急任务的完成情况和相关情境因素的影响。

For sake of reducing human errors and improving emergency response efficiency,a combination model of fuzzy theory and BN based on analysis of relevant influential factors was proposed to estimate human reliability of ship oil-spill emergency response.Human reliability of the single task and the whole process was calculated through formalize actual ship oil-spill emergency response process into a BN with probabilistic inference algorithm,and an application example was studied with this model.The study shows that： first of all,the major problem that human behavior is seriously context-relied and hard to be quantified,is well solved through using HRA based on fuzzy BN;secondly,the change of human reliability calculation from single task to the whole process is achieved by using excellent probabilistic inference algorithm of BN;thirdly,human reliability of the whole emergency response system mainly depends on the completion degree of a series of key tasks and influence of context factors.