基于动态贝叶斯的危化品泄漏应急决策模型  

Emergency Decision Model for Hazardous Chemicals Leakage Based on Dynamic Bayesian

作　　者：张思晨 方欢 ZHANG Sichen;FANG Huan(School of Mathematics and Big Data,Anhui University of Science and Technology,Huainan Anhui 232001,China;Anhui Province Engineering Laboratory for Big Data Analysis and Early Warning Technology of Coal Mine Safety,Huainan Anhui232001,China)

机构地区：[1]安徽理工大学数学与大数据学院,安徽淮南232001 [2]安徽省煤矿安全大数据分析与预警技术工程实验室,安徽淮南232001

出　　处：《兰州工业学院学报》2024年第4期94-99,共6页Journal of Lanzhou Institute of Technology

基　　金：国家自然科学基金项目(61902002)。

摘　　要：基于危化品泄漏事故,以双乙烯酮泄漏为案例,构建了一种基于动态贝叶斯网络的危化品泄漏事故应急管理全过程演练专家决策模型,准确计算了各状态转换之间的概率,并添加各状态之间转换的代价,实现了对泄漏事故的全面模拟以及应急决策。通过整合专家知识和历史案例以及各项影响因素,有效地降低了转换概率对专家打分的依赖性,提高了模型的客观性和准确性。在模型验证阶段,通过模拟实验和双乙烯酮泄漏事故场景演练,展现了较为出色的决策支持能力,为危化品泄漏事故的应对提供了创新性解决方案。In view of the leakage accident of hazardous chemicals,taking diketene as an example,an expert decision-making model for the full-process of emergency management drill of the leakage accident of hazardous chemicals based on Dynamic Bayesian Network is constructed.The model accurately calculates the probability between the state transitions,and adds the conversion costs between the different states,so as to realize the comprehensive simulation and decision-making of the leakage accident of hazardous chemicals.By integrating expert knowledge,historical cases and various influencing factors,the dependence of conversion probability on expert scoring is effectively reduced,and the objectivity and accuracy of the model are improved.In the model validation phase,excellent decision support capability is demonstrated through simulation experiments and diketene leakage accident scenario drills.The decision-making model provides an innovative solution for the response to hazardous chemical leakage accidents.

关 键 词：危化品泄漏 专家决策 动态贝叶斯网络 应急管理 

分 类 号：X937[环境科学与工程—安全科学]