燃气管道泄漏知识图谱构建与应用  

Construction and Application of Knowledge Graph of Gas Pipeline Leakage

作　　者：张季娜 王凡[2] 周宏健 ZHANG Jina;WANG Fan;ZHOU Hongjian

机构地区：[1]中石油昆仑燃气有限公司燃气技术研究院,黑龙江哈尔滨150010 [2]中国石油天然气股份有限公司天然气销售分公司(昆仑能源有限公司),北京100101

出　　处：《煤气与热力》2025年第3期I0034-I0041,I0046,共9页Gas & Heat

基　　金：中国石油前瞻性基础性战略性技术攻关项目“油气管道建设与完整性管理技术研究”(2021DJ2806)。

摘　　要：针对燃气管道领域特点,分析燃气管道领域知识图谱技术架构和多源数据,选取Bi-LSTM-CRF模型作为知识抽取模型,对燃气管道泄漏相关的资料进行知识抽取,定义关系类型,明确实体间关系,并利用Neo4j图数据库存储,构建燃气管道泄漏知识图谱。通过分析燃气管道泄漏知识图谱的信息查询方式及燃气管道泄漏演化路径预测分析及处理措施推理,验证了知识图谱在燃气管道领域应用的有效性。对下一步的应用进行展望。According to the characteristics of gas pipeline field,the technical framework and multi-source data of knowledge graph in gas pipeline field are analyzed.The Bi-LSTM-CRF model is selected as the knowledge extraction model to extract knowledge from data related to gas pipeline leakage,define the relationship types,clarify the relationship between entities,and use the Neo4j graph database to store and construct the knowledge graph of gas pipeline leakage.By analyzing the information query method of gas pipeline leakage knowledge graph,predicting and analyzing the evolution path of gas pipeline leakage,and inferring the treatment measures,the effectiveness of the application of knowledge graph in gas pipeline field is verified.The next steps of application are prospected.

关 键 词：燃气管道泄漏 双向长短期记忆网络条件随机场模型 知识图谱 

分 类 号：TU996.6[建筑科学—供热、供燃气、通风及空调工程]