电力信息与通信技术2026，Vol.24Issue(5)：13-22,10.DOI:10.16543/j.2095-641x.electric.power.ict.2026.05.02
基于大语言模型构建面向能源数字网络的知识图谱
Knowledge Graph for Energy Digital Network Based on Large Language Model
摘要
Abstract
The energy industry will accumulate a considerable amount of business data internally,and automatically mining the information in the business data will promote the improvement of relevant departments'business capabilities and reduce the huge operation and maintenance costs in the industry.The professional terminology in the energy field is complex and the entity relationships are dynamically changing.Cross domain knowledge fusion needs to solve the problem of unified representation of multimodal data(text,temporal,spatial).In response to the above issues,this paper proposes a framework for constructing an energy knowledge graph based on a large language model,designs a"patiotemporal functional"dual dimensional ontology model,and introduces digital twin technology to achieve dynamic knowledge updates.Domain adaptive prompt function was developed,accurate extraction of professional terms was achieved with Qwen model.A hybrid architecture of graph neural network Transformer was proposed to solve the problem of semantic alignment in multimodal data.The research results can provide a highly reliable knowledge base for energy digital networks and assist in the construction of new type of power system.关键词
能源互联网/稀疏随机投影/卷积神经网络/入侵检测/遗传编程Key words
energy Internet/sparse random projection/convolutional neural network/intrusion detection/genetic programming分类
信息技术与安全科学引用本文复制引用
周爱华,潘森,乔俊峰,朱力鹏,李井泉,张肖杰..基于大语言模型构建面向能源数字网络的知识图谱[J].电力信息与通信技术,2026,24(5):13-22,10.基金项目
国家电网有限公司总部管理科技项目资助"基于数字对象架构的新型能源数字网络模型与机制研究"(5700-202390591A-3-2-ZN). （5700-202390591A-3-2-ZN）
