摘要
针对各厂家继电保护定值信息语义不标准,严重影响信息审核和应用的问题,提出基于知识图谱和文本相似度计算的继电保护定值信息智能识别模型。建立语义预处理、语义智能识别、定值信息知识图谱等模块。基于正则表达式处理字母、数字以及特殊符号不标准表述;通过文本相似度算法,计算待规范信息和规范信息间余弦值,判定两者间的相似程度;基于自然语言处理技术,建立继电保护标准语义信息知识网络。通过某地区继电保护定值信息验证,模型能够准确地标准化各厂家的定值信息。
Aiming at the problem that the non-standard semantics of relay protection setting information of various manufacturers seriously afiects the information review and application,an intelligent identification model of relay protection setting information based on knowledge graph and text similarity calculation is proposed.It establishes semantic preprocessing,semantic intelligent recognition,fixed-value information knowledge graph and other modules,handles non-standard expressions of letters,numbers and special symbols based on regular expressions,uses the text similarity algorithm to calculate the cosine value between the standard information and the standard information to determine the degree of similarity.The relay protection setting information knowledge graph module is based on natural language processing technology to establish a semantic information knowledge network of relay protection standards.Through the verification of relay protection setting information in a certain area,the model can accurately standardize the setting information of various manufacturers.
作者
熊学海
万春竹
杨波
赵凌
齐雪雯
XIONG Xue-hai;WAN Chun-zhu;YANG Bo;ZHAO Ling;QI Xue-wen(Guizhou Power Grid Co.,Ltd.Power Dispatching Control Center,Guiyang 550000 China;Anshun Power Supply Bureau of Guizhou Power Grid Co.,Ltd.,Anshun 561000 China)
出处
《自动化技术与应用》
2023年第5期128-132,共5页
Techniques of Automation and Applications
基金
贵州电网有限责任公司科技项目(GZKJXM20190394)。
关键词
定值信息模型
文本识别
文本相似度计算
setting information model
text identification
text similarity calculation