摘要
危险化学品企业因涉及易燃易爆、有毒有害等物质,本身具有较高固有安全风险,且其风险体系受企业内部运营、外部环境及动态工况等多维度因素交叉影响。因此,针对这一困境,文章提出知识图谱-多元融合方法的风险评估模型,构建危险化学品企业知识图谱,设计“数据层时空对齐+特征层注意力融合”的双层架构,整合企业运营、外部监管及动态监测数据,进行特征层注意力融合,最终构建三级评估模型,以期为危险化学品企业提供更精准、更高效的安全管控工具,为类似研究提供重要理论与实践价值。
Hazardous chemical enterprises inherently face heightened safety risks due to handling flammable,explosive,toxic,and hazardous substances.Their risk management systems are influenced by multiple factors including internal operations,external environments,and dynamic operational conditions.To address these challenges,this paper proposes a knowledge graph-based multi-integration risk assessment model.The framework begins with constructing a knowledge graph for hazardous chemical enterprises,featuring a dual-layer architecture combining“spatiotemporal data alignment at the data layer and attention fusion at the feature layer”.By integrating operational data,regulatory oversight,and real-time monitoring information,the model achieves effective feature-level attention fusion.This three-tier evaluation system provides enterprises with precise and efficient safety management tools,demonstrating significant theoretical and practical value.
作者
彭贵
文蓬涛
赵军
白元
PENG GUI;WEN Pengtao;ZHAO Jun;BAI Yuan(Xinjiang Uygur Autonomous Region Emergency Management Science Research Institute,Urumqi 830000,China)
出处
《化工管理》
2025年第34期90-95,共6页
Chemical Management
基金
2023年新疆维吾尔自治区重点研发计划项目“基于设备工艺数据与专家知识的安全风险评估与操作规范研究”(2023B01031-1)。
关键词
危险化学品
化工企业
知识图谱
多元复杂
融合方法
风险评估
模型研究
hazardous chemicals
chemical enterprises
knowledge graph
multidimensional complexity
integration method
risk assessment
model research
