基于特高频局放技术与CG-BP算法的GIS绝缘故障类型识别方法
    点此下载全文
引用本文:柯拥勤,潘亦斌.基于特高频局放技术与CG-BP算法的GIS绝缘故障类型识别方法[J].计算技术与自动化,2022,(3):159-164
摘要点击次数: 144
全文下载次数: 0
作者单位
柯拥勤,潘亦斌 (国网福建省电力有限公司莆田供电公司福建 莆田 351100) 
中文摘要:针对封闭式气体绝缘开关装置(Gas Insulated Switch Gear,GIS)由于生产、运输安装和运行环境等因素,引发的局部放电(Partial discharge,PD)现象而造成的绝缘故障类型问题,提出一种基于特高频(UHF)局放技术与CG-BP算法的GIS绝缘故障类型识别方法。首先,对GIS内的PD产生机理,以及UHF局放技术的实现过程进行描述,并在此基础上,分析了GIS内部4种常见的PD信号的UHF传播特性。其次,利用共轭梯度法(Conjugate Gradient, CG )来优化BP算法的初始阈值和权值,进而提高BP算法学习效率和GIS故障类型识别率。最后,仿真结果表明:相比基于传统BP算法的GIS故障类型识别方法,改进的CG-BP识别算法可有效减少训练次数,且识别准确率可达91%,提高了10%。
中文关键词:GIS  特高频  局部放电  CG-BP
 
GIS Insulation Fault Type Identification Method Based on UHF Local Discharge Technology and CG-BP Algorithm
Abstract:The insulation failure of gas insulated switch gear (GIS) is caused by partial discharge (PD) caused by production, transportation, installation and operation environment. A GIS insulation fault type identification method based on UHF local discharge technology and CG-BP algorithm is proposed. Firstly, the mechanism of PD generation in GIS and the realization process of UHF local emission technology are described, and on this basis, the UHF propagation characteristics of four common PD signals in GIS are analyzed. Secondly, conjugate gradient (CG) method is used to optimize the initial threshold and weight of BP algorithm, so as to improve the learning efficiency of BP algorithm and GIS fault recognition rate. Finally, the simulation results show that compared with the GIS intelligent PD fault identification method based on the traditional BP algorithm, the improved CG-BP intelligent identification algorithm has a higher accuracy of 91%, an increase of 10%.
keywords:GIS  UHF  PD  CG-BP
查看全文   查看/发表评论   下载pdf阅读器