基于二维云模型的储罐泄漏风险Bow-tie量化分析  

On the quantitative Bow-tie analysis of the tank leakage risk based on the 2-D cloud model

作　　者：陈樑 金龙[1] 黄兵 曹卫东[3] 王学东 CHEN Liang;JIN Long;HUANG Bing;CAO Wei-dong;WANG Xue-dong(School of Public Safety and Emergency Management,Kunming University of Technology,Kunming 650000,China;School of Environmental Science and Engineering,Kunming University of Technology,Kunming 650500,China;Kunming Yunneng Chemical Co.,Ltd.,Kunming 650300,China)

机构地区：[1]昆明理工大学公共安全与应急管理学院,昆明650000 [2]昆明理工大学环境科学与工程学院,昆明650500 [3]昆明云能化工有限公司,昆明650300

出　　处：《安全与环境学报》2020年第3期809-815,共7页Journal of Safety and Environment

基　　金：国家自然科学基金项目(41261079)。

摘　　要：针对Bow-tie分析无法实现风险量化的问题,引入专家评判与风险矩阵,同时将二维云模型与风险矩阵耦合,利用云模型不确定性推理解决了风险量化过程中定性结论与定量描述之间相互转化、风险矩阵阈值存在模糊性、风险的两个影响因素事故可能性与严重度存在随机性与模糊性的问题,并将风险矩阵与实际评价结果均以二维云图的形式输出。事故可能性与严重度均采用专家打分,将事故严重度一级指标细化为人员伤亡、经济损失、环境污染、社会影响4个二级指标,并采用信任度更高的组合权重对指标赋权。利用云模型的正向云发生器将风险矩阵阈值云模型化,柔化了传统硬性的阈值区间,同时利用云模型的逆向云发生器整合专家对事故可能性与严重度的评价结果,生成相应的云模型参数。将事故可能性与严重度云模型参数代入二维云模型生成实际云与标准云对比,通过计算云的贴近度确定风险等级。以化工储罐泄漏Bow-tie分析为例,证明该模型能够融合专家意见并体现专家集体智慧,将风险以蕴含随机性与模糊性的云分布输出,既能反映风险的不确定性,又能通过计算云分布相似性确定系统的风险等级,实现了将风险从不确定到确定的量化,且分析过程可视化程度高,易于编程实现。Aiming at the problem that the Bow-tie analysis method can’t quantify the likely risks,the paper has first of all introduced the risk matrix and the expert evaluation index. At the same time,we have also coupled a 2-D cloud model with the risk matrix. The method of uncertainty reasoning of the cloud model can be used to solve the problem of mutual transformation between the qualitative conclusion and the quantitative descriptions in the process of risk quantification,though the threshold of the risk matrix may be ambiguous and potential with risks. The probability and severity of the 2 factors may turn out to be in a random and ambiguous form,whereas the risk matrix and the actual evaluation results can be turned out to be in a form of 2-D cloud map,with the possibility and severity of the accidents being scored and judged by the corresponding experts. In so doing,the first-level indicators of the severity of accidents can be divided into 2 second-level indicators,i. e. the casualties,the economic losses,the environmental pollution and the social impact. The combination weights with higher confidence can be used to empower the said indicators. The forward cloud generator of the cloud model can be used to modulate the threshold cloud of the risk matrix, which can help to mitigate the traditional hard threshold intervals. At the same time,the reversed cloud generator of the cloud model can be adopted to integrate the evaluation results of the experts on the possibility and severity of the accidents to produce a corresponding cloud model parameters. Actually,the accident probability and the severity cloud model parameters can be used to substitute the 2-D cloud model to produce the actual cloud and the standard cloud,with the risk level being determined by accounting the cloud closeness. Thus,taking the Bow-tie analysis of the chemical tank leakage as a case study sample,the model can help to integrate the expert ideas and assessments to reflect their collective wisdom and the joint evaluation results of the rando

关 键 词：安全工程 二维云模型 Bow-tie分析 风险矩阵 风险量化 

分 类 号：X913.4[环境科学与工程—安全科学]