摘要
知识图谱是一种利用解释性语言描述研究对象之间关系的知识库,能够增强机器学习能力,挖掘事物间深层及潜藏的关系,为决策者在抉择过程中提供辅助。从知识图谱技术原理出发,结合油气管道特点,提出了以业务需求、本体设计、知识抽取、知识融合、知识存储、知识迭代更新为核心的油气管道行业知识图谱构建流程与方法。依据SY/T 6828—2017《油气管道地质灾害风险管理技术规范》,构建了以风险识别、评价、防治为主体的油气管道地质灾害风险管理知识图谱。以典型黏性土滑坡灾害为例开展应用示范,在明确灾害风险评价及防治逻辑架构的基础上,探讨了知识图谱在油气管道地质灾害智能评估中的决策过程、应用场景及决策结果的推演与可视化,为管道地质灾害风险评价及防治的智能辅助决策提供了可行路径。
As a knowledge base describing the relationship between research objects with explanatory language,knowledge graph can enhance the ability of machine learning and excavate the deep and hidden relationships between things to help deciders make decisions.The process and method of knowledge graph construction in oil and gas pipeline industry,focusing on business requirements,ontology design,knowledge extraction,knowledge fusion,knowledge storage and knowledge iterative updating,were put forward in combination with the characteristics of oil and gas pipelines based on the principle of knowledge graph technology.The knowledge graph of geohazard risk management of oil and gas pipeline based on risk identification,assessment and prevention was constructed according to the Specification for geological hazard risk management of oil and gas pipeline(SY/T 6828—2017)for the frst time,and application demonstration was carried out based on the typical clay landslide hazard.Besides,the deduction and visualization of the decision-making process,application scenarios and decision-making results of the knowledge graph in the intelligent geohazard assessment of oil and gas pipeline was discussed on the basis of defining the logical framework of hazard risk assessment and prevention,which provides a feasible way for the intelligent aided decision-making in geohazard risk assessment and prevention for pipelines.
作者
吴张中
WU Zhangzhong(PipeChina Institute of Science and Technology)
出处
《油气储运》
CAS
北大核心
2023年第3期241-248,共8页
Oil & Gas Storage and Transportation
基金
国家管网集团揭榜挂帅项目“智慧管网理论和技术体系研究”,JCGL202109
国家管网集团揭榜挂帅项目“油气管道线路及站场感知技术研究”,WZXGL202106。
关键词
油气管道
地质灾害
风险管理
知识图谱
本体设计
智能决策
oil and gas pipeline
geohazard
risk management
knowledge graph
ontology design
intelligent decision-making