基于故障树-贝叶斯-蝴蝶结分析的复合事故风险模型研究

Study on Composite Accident Model Based on Fault Tree-Bayesian-Bow-tie Analysis

为解决现有单一风险评价方法存在的不足,引入了基于故障树分析、贝叶斯网络与Bow-tie分析(蝴蝶结分析)相结合的复合风险分析模型:首先找出引起故障的基本事件,绘制故障树(FTA)模型;然后以故障树模型为基础,将其转化为贝叶斯网络,计算其中基本事件的后验概率和重要度;最后对重要度等级较高的事故节点进行Bow-tie分析,提出预防事故发生的防护措施和减轻事故后果的控制措施。以粘结漏钢事故为例,应用该模型对粘结漏钢事故进行了系统风险分析,验证了该模型的科学性和合理性,其结果表明:引起粘结漏钢事故占比较大的事故节点为保护渣异常,结合风险分析结果提出了相关预防及控制措施,以期最大程度地降低事故发生概率及减小事故损失的目的。

In order to solve the shortcomings of the existing single risk evaluation method,a composite risk analysis model based on fault tree analysis,Bayesian network and Bow-tie analysis is introduced.First the basic events causing the failure are found out and the fault tree(FTA)model is drawn.Then,based on the fault tree model,it is transformed into a Bayesian network to calculate the posterior probability and importance degree of the basic events.And finally the Bow-tie analysis is carried out on the accident nodes with high importance degree,and the protective measures to prevent the accident and the control measures to reduce the consequences of the accident are put forward.Taking the bonded steel leakage as a case,this model is used to analyze the systematic risk of bonded steel leakage accident,and the scientificity and rationality of this model are verified.The results show that,the node which causes the bonding steel leakage accident is the abnormal protective slag,and combined with the results of risk analysis,relevant preventive and control measures are put forward in order to reduce the probability of accident occurrence and reduce the accident loss to the greatest extent.