• HOME
  • About Journal
    • Historical evolution
    • Journal Honors
  • Editorial Board
    • Members of Committee
    • Director of the Committee
  • Submission Guide
    • Instructions for Authors
    • Manuscript Processing Flow
    • Model Text
    • Procedures for Submission
  • Academic Influence
  • Open Access
  • Ethics&Policies
    • Publication Ethics Statement
    • Peer Review Process
    • Academic Misconduct Identification and Treatment
    • Advertising and Marketing
    • Correction and Retraction
    • Conflict of Interest
    • Authorship & Copyright
  • Contact Us
  • Chinese
Site search        
文章摘要
基于预训练语言模型的电力领域设备缺陷检测
Device defect detection in power field based on pre-trained language model
Received:December 08, 2021  Revised:December 12, 2021
DOI:10.19753/j.issn1001-1390.2022.05.024
中文关键词: 缺陷检测  预训练语言模型  缺陷报告  事件三元组
英文关键词: defect detection, pre-trained model, defect report, event triplet
基金项目:国家电网有限公司大数据中心科技项目(项目名称:多模态认知图谱基础能力研究,项目编号:SGSJ0000FXJS2100099 )
Author NameAffiliationE-mail
Wang Honggang Big Data Center, State Grid Corporation of China, Beijing 100031, China hgwang@sgcc.com.cn 
Ji Xin 1.Big Data Center, State Grid Corporation of China, Beijing 100031, China. 2.Beihang University, Beijing 100191, China xin-ji@sgcc.com.cn 
Wu Tongxin Big Data Center, State Grid Corporation of China, Beijing 100031, China tongxin-wu@sgcc.com.cn 
Yang Zhiwei* Big Data Center, State Grid Corporation of China, Beijing 100031, China hsyangzhiwei@163.com 
He Yude Big Data Center, State Grid Corporation of China, Beijing 100031, China yude-he@sgcc.com.cn 
Hits: 2142
Download times: 546
中文摘要:
      电力设备缺陷种类繁多,部分缺陷会引发设备故障,及时检测电力设备存在的缺陷是防止发生设备故障的重要手段。设备缺陷检测旨在从文本中识别触发词并且将文本划分对应的设备缺陷类型。针对电力领域缺陷数据集标注不足,以及由于文本中包含大量专业术语造成语义理解难等问题,研究基于深度学习的设备缺陷检测方法,设计电力领域设备缺陷检测预训练语言模型,利用事件三元组知识。文中,构建一个电力设备缺陷检测数据集,在模型进行缺陷检测任务之前,通过事件三元组预训练的方式提高语言模型的表征能力。实验表明,基于现场设备案例数据经过预训练的模型在缺陷检测任务上具有更好的表现效果,可以有效实现对电力领域缺陷报告文本的缺陷检测。
英文摘要:
      There are many kinds of defects in electrical equipment, which may lead to electrical equipment faults. Timely detection of defects is an essential means to prevent equipment failure. It aims to identify triggers from the text and divide the text into corresponding equipment defect types. Aiming at the problems of insufficient annotation of defect data set and difficulty in semantic understanding due to numerous professional terms in the text, we study the device defect detection method based on deep learning, and design a pre-trained language model for device defect detection in the power field by utilizing event triplet knowledge. In addition, we construct an electrical equipment defect detection data set. The representation ability of the language model is improved by event triplet pre-trained before the defect detection task of the model. Experimental results show that the pre-trained model based on field equipment case data performs better in the event detection task and can effectively realize the defect detection of power field defect report text.
View Full Text   View/Add Comment  Download reader
Close
  • Home
  • About Journal
    • Historical evolution
    • Journal Honors
  • Editorial Board
    • Members of Committee
    • Director of the Committee
  • Submission Guide
    • Instructions for Authors
    • Manuscript Processing Flow
    • Model Text
    • Procedures for Submission
  • Academic Influence
  • Open Access
  • Ethics&Policies
    • Publication Ethics Statement
    • Peer Review Process
    • Academic Misconduct Identification and Treatment
    • Advertising and Marketing
    • Correction and Retraction
    • Conflict of Interest
    • Authorship & Copyright
  • Contact Us
  • 中文页面
Address: No.2000, Chuangxin Road, Songbei District, Harbin, China    Zip code: 150028
E-mail: dcyb@vip.163.com    Telephone: 0451-86611021
© 2012 Electrical Measurement & Instrumentation
黑ICP备11006624号-1
Support:Beijing Qinyun Technology Development Co., Ltd