电器与能效管理技术 ›› 2026, Vol. 0 ›› Issue (9): 54-64.doi: 10.16628/j.cnki.2095-8188.2026.09.007

• 配网技术与系统 • 上一篇    下一篇

面向变电站倒闸操作的智能防误知识图谱的构建及应用

刘嘉明1, 乔继斌2, 刘东3   

  1. 1 国网辽宁省电力有限公司, 辽宁 沈阳 110055
    2 南瑞集团有限公司 北京科东电力控制系统有限责任公司, 北京 100192
    3 中国电力科学研究院有限公司, 北京 100192
  • 收稿日期:2026-05-15 出版日期:2026-09-30 发布日期:2026-10-16
  • 作者简介:刘嘉明(1983—),男,高级工程师,主要从事电力系统分析及其自动化工作。|乔继斌(1972—),男,工程师,主要从事电力系统分析及其自动化工作。|刘 东(1990—)男,工程师,通信作者,主要从事电力系统分析及其自动化工作。

Construction and Application of Intelligent Anti-Misoperation Knowledge Graph for Substation Switching Operations

Liu Jiaming1, Qiao Jibin2, Liu Dong3   

  1. 1 State Grid Liaoning Electric Power Co., Ltd., Shenyang 110055, China
    2 Nari Group Corporation, Beijing Kedong Electric Power Control System Co.,Ltd., Beijing 100192, China
    3 China Electric Power Research Institute Co., Ltd., Beijing 100192, China
  • Received:2026-05-15 Online:2026-09-30 Published:2026-10-16

摘要:

针对当前操作票缺乏有效的防误校核手段、智能化程度较低等问题,提出了面向变电站倒闸操作智能防误的知识图谱构建方法。首先,利用电网设备专业语料数据,采用深度学习的模型训练方法构建电网设备连接关系图谱、防误规程图谱和历史操作票图谱,并进行自动融合形成智能防误知识图谱。然后,设计实验验证防误模型的有效性,模型的综合评价指标(F1值)达到80%,准确率达到95%。最后,利用Neo4j图数据库对智能防误知识图谱进行高度可视化管理,并给出其对应的全过程应用流程,有效提高了操作人员面对倒闸操作故障的应急处理能力。

关键词: 智能防误, 实体识别, 知识图谱, 辅助校验

Abstract:

To address the problems that current operation tickets lack effective anti-misoperation verification means and suffer from low intelligence level,a knowledge-graph construction method for intelligent anti-misoperation of substation switching operations is proposed.Firstly,using professional corpus data of power grid equipment,a deep learning model training method is adopted to construct a power grid equipment connection relationship graph,an error prevention regulation graph,and a historical operation ticket graph,and automatically fuse them to form an intelligent error prevention knowledge graph.Then,experiments are designed to verify the effectiveness of the error prevention model,with an F1 value of 80% and an accuracy rate of 95%.Finally,the intelligent anti-misoperation knowledge graph is managed with high-level visualization by adopting the Neo4j graph database,and the corresponding whole-process application workflow is provided,which effectively improves operators’ emergency-handling capabilities when encountering switching-operation faults.

Key words: intelligent anti-misoperation, entity recognition, knowledge graph, auxiliary verification

中图分类号: