基于知识图谱的电力通信设备故障智能诊断方法  

Intelligent Diagnosis Method of Power Communication Equipment Faults Based on Knowledge Graph

作　　者：崔鸣石 邬雪阳 朱宏伟 王冲 粘中元 王蕾 杨莎 CUI Mingshi;WU Xueyang;ZHU Hongwei;WANG Chong;NIAN Zhongyuan;WANG Lei;YANG Sha(State Grid Inner Mongolia East Electric Power Co.,Hohhot 010020,China;Computer College,Northeast Electric Power University,Jilin 132012,Jilin,China)

机构地区：[1]国网内蒙古东部电力有限公司,呼和浩特010020 [2]东北电力大学计算机学院,吉林吉林132012

出　　处：《科技和产业》2024年第5期212-221,共10页Science Technology and Industry

基　　金：国网内蒙古东部电力有限公司科技项目(SGMDXT00JSJS2200133)。

摘　　要：国家电网公司信息化建设飞速发展,电力通信设备运行与故障处理日益复杂。因此,提出基于知识图谱的电力通信设备故障智能诊断方法。首先,利用BERT-BiGRU-CRF(基于转换器的双向编码器-双向门控递归单元-条件随机场)模型及人工规则提取故障信息实体与关系。在构建电力通信设备知识图谱后,采用WBLA(加权双边拉普拉斯算法)模型判断故障严重性并规划处理顺序。随后,通过TFIDF-cos(词频-逆文档频率余弦相似性)模型获取高可信度的故障处理结果,实现电力通信设备故障的智能诊断。实验证明该方法准确度分别达到98.4%和97.5%,验证了在实现电力通信设备故障智能诊断高效化方面的可行性。As the informationization construction of the State Grid developes rapidly,and the operation and troubleshooting of power communication equipment are becoming increasingly complex.Therefore an intelligent diagnosis method for power communication equipment faults based on knowledge graph is proposed.Firstly,the BERT-BiGRU-CRF(bidirectional encoder representations from transformers-bidirectional gated recurrent unit-conditional random fields)model and artificial rules are utilized to extract fault information entities and relationships.After constructing the knowledge graph of power communication equipment,the WBLA model is used to determine the fault severity and plan the processing sequence.Subsequently,the TFIDF-cos(term frequency-inverse document frequency cosine similarity)method is used to obtain high-confidence fault processing results and realize the intelligent diagnosis of power communication equipment faults.Experiments demonstrate that the accuracy of the method reaches 98.4%and 97.5%respectively,which verifies the feasibility in realizing efficient intelligent diagnosis of power communication equipment faults.

关 键 词：知识图谱 电力通信设备 故障诊断 智能处理 

分 类 号：TM73[电气工程—电力系统及自动化] TP391.1[自动化与计算机技术—计算机应用技术]