基于图神经网络的社区户内燃气系统动态风险评估

Dynamic Risk Assessment of Community Indoor Gas System Based on Graph Neural Network

目前对于社区户内燃气系统的风险评估大都为静态分析方法,且一般只对系统中单独组成部分进行建模,忽略了户内燃气系统风险复杂的耦合影响关系。针对现有风险评估方法的局限性,提出了一种基于图神经网络的社区户内燃气系统动态风险评估方法。该方法首先通过梳理社区户内燃气事故、分析社区户内燃气系统的风险源和风险影响因素,构建了社区户内燃气系统动态风险评估指标体系;然后将知识图谱用于社区户内燃气系统的场景构建,基于社区户内燃气系统的构成和指标体系确定知识图谱中的实体类型和实体特征,并在所构建的社区户内燃气系统知识图谱的基础上,提出了一种基于图神经网络的社区户内燃气系统动态风险评估方法;最后,以某一社区为例,对所提出的基于图神经网络的社区户内燃气系统动态风险评估方法进行验证。实例应用分析结果表明:该方法可以针对社区户内燃气系统风险复杂的耦合影响关系,实现社区户内燃气系统的安全动态风险评价,对社区燃气用户的安全管理具有一定的指导和参考意义。

At present, the risk assessment of community indoor gas system is mostly based on the static analysis method, and generally only models the individual components of the system, ignoring the complex coupling effect relationship of indoor gas system risk.Aiming at the limitation of existing risk assessment methods, this paper proposes a dynamic risk assessment method of community indoor gas system based on graph neural network.The method first combines the indoor gas accidents, analyzes the risk sources and risk influencing factors of indoor gas system, and thus constructs the risk assessment index system of indoor gas system.Then it applies the knowledge graph to the scene construction of indoor gas system, and determine the entity types and entity characteristics in the knowledge graph based on the composition and index system of indoor gas system.In addition, based on the knowledge graph of indoor gas system, the paper proposes a dynamic risk assessment method of indoor gas system based on graph neural network.Finally, a community is taken as an example to verify the proposed dynamic risk assessment model.The result indicates that the method can realize the dynamic risk assessment of community indoor gas system in view of the complex coupling relationship of community indoor gas system risk, which has certain guidance and reference significance for the safety management of community gas users.