面向数字孪生工程的水利知识图谱构建及应用  

Building and Application of Water Conservancy Knowledge Graph for Digital Twin Project

作　　者：张军珲 黄希扬[3] 桂明宇 昝红英 张坤丽[3] ZHANG Junhui;HUANG Xiyang;GUI Mingyu;ZAN Hongying;ZHANG Kunli(Yellow River Engineering Consulting Co.,Ltd.,Zhengzhou 450003,China;Yunhe(Henan)Information Technology Co.,Ltd.,Zhengzhou 450003,China;School of Computer and Artificial Intelligence,Zhengzhou University,Zhengzhou 450001,China)

机构地区：[1]黄河勘测规划设计研究院有限公司,河南郑州450003 [2]云河(河南)信息科技有限公司,河南郑州450003 [3]郑州大学计算机与人工智能学院,河南郑州450001

出　　处：《人民黄河》2024年第4期121-124,130,共5页Yellow River

基　　金：河南省科技攻关项目(232102211033,232102211039)。

摘　　要：构建水利领域的知识图谱可为数字孪生水利工程建设提供结构化的知识支持。基于黄河小浪底、万家寨水利枢纽预报调度、工程安全、业务规则数据,使用实体识别和关系抽取模型构建面向数字孪生工程的水利知识图谱(KG4DT),并对图谱进行可视化展示。以水利枢纽为核心确定实体分类体系及关系分类体系,构建KG4DT的模式层。使用二次开发的标注平台对非结构化文本进行半自动标注和人工校对,构建KG4DT的数据层。以黄河小浪底和万家寨水利枢纽防洪、减淤、供水、发电业务为主线,结合AC自动机等技术设计知识检索、预案自动生成应用,可为工程运行提供智能决策支撑。The building of a knowledge graph within the water conservancy sector can offer structured knowledge support for digital twin water conservancy projects.Based on the forecast and scheduling,engineering safety and business rule data of Xiaolangdi and Wanjiazhai key water control projects on the Yellow River,an entity recognition and relationship extraction model was used to build a water conservancy knowledge graph for digital twin engineering(KG4DT),and the graph was visualized and displayed.It determined the entity classification system and relationship classification system with the key water control project as the core,and built the pattern layer of KG4DT.It used a secondary de-velopment annotation platform to perform semi-automatic annotation and manual proofreading on unstructured text and built the data layer of KG4DT.Taking the flood control,sedimentation reduction,water supply and power generation businesses of the Xiaolangdi and Wanjiazhai key water control projects on the Yellow River as the main line,it combined the technologies such as AC automaton to design knowledge re-trieval and automatic generation of contingency plans,which could provide intelligent decision-making support for engineering operations.

关 键 词：水利知识图谱 数字孪生 知识描述体系 应用 

分 类 号：TV61[水利工程—水利水电工程] TP391[自动化与计算机技术—计算机应用技术]