电工技术学报  2025, Vol. 40 Issue (1): 253-263    DOI: 10.19595/j.cnki.1000-6753.tces.232016
高电压与放电 |
GIS局部放电光信号检测技术及光学图像诊断方法
李泽1,2, 钱勇1, 刘伟3, 臧奕茗1, 盛戈皞1, 江秀臣1
1.上海交通大学电气工程系 上海 200240;
2.国网上海市电力公司电力科学研究院 上海 200437;
3.六氟化硫气体特性分析与净化处理技术国网公司重点实验室 国网安徽省电力有限公司电力科学研究院 合肥 230022
Optical Signal Detection Technology and Optical Image Diagnostic Method of Partial Discharge in GIS
Li Ze1,2, Qian Yong1, Liu Wei3, Zang Yiming1, Sheng Gehao1, Jiang Xiuchen1
1. Department of Electrical Engineering Shanghai Jiao Tong University Shanghai 200240 China;
2. State Grid Shanghai Electric Power Research Institute Shanghai 200437 China;
3. Key Laboratory for Sulfur Hexafluoride Gas Analysis and Purification of SGCC State Grid Anhui Electric Power Research Institute Hefei 230022 China
全文: PDF (2543 KB)   HTML
输出: BibTeX | EndNote (RIS)      
摘要 光测法是一种有效的局部放电检测技术。为研究典型局部放电的光学特征和光学图像诊断方法,首先搭建气体绝缘局部放电光电联合检测实验平台,实现光信号和电信号的同步采集;然后分别采用导光柱法和荧光光纤法对SF6气体中典型放电光信号进行实验测量和统计分析,对比研究两种光学检测方法的放电特征参数,包括起始放电电压、相位分布、脉冲幅值和数量等;最后,利用导光柱采集的典型局部放电信号,提出一种基于金字塔梯度直方图(PHOG)特征的局部放电光学图像故障诊断方法。结果显示该模型诊断准确率达90.4%,验证了基于导光柱的局部放电光学检测与诊断的可靠性。
服务
把本文推荐给朋友
加入我的书架
加入引用管理器
E-mail Alert
RSS
作者相关文章
李泽
钱勇
刘伟
臧奕茗
盛戈皞
江秀臣
关键词 光学局部放电导光柱金字塔梯度直方图天牛须搜索算法支持向量机    
Abstract:Partial discharge (PD) detection and fault diagnosis are of great significance to the stable operation of GIS. The optical detection method is a new method for measuring and analyzing PD signals, which has received wide attention due to its high detection confidence and strong anti-interference. The light guide rod is a kind of optical sensor with high transmittance, high hardness, and easy to seal, which can be used for PD detection in GIS. Currently, there is a lack of research on the characteristics of PD optical pulse in GIS using a light guide rod. In PD fault diagnosis, fewer studies based on optical data, especially optical images. Therefore, in this paper, optical detection and image diagnostic research of PD in GIS is carried out using the light guide rod.
Firstly, a test platform for electrical and optical detection of PD in GIS was built, achieving synchronous collection of optical and current signals. The typical PD defects were designed. Then, the PD optical signals were measured by the light guide rod and the fluorescent fiber. After that, the optical parameters of the two methods were compared and analyzed, including the PD inception voltage, the phase distribution, the maximum amplitude of the optical pulse, and the average amplitude and number of the optical pulse.
For optical image fault diagnosis, the PRPS patterns of optical PDs were constructed based on the data collected by the light guide rod. The pyramid histogram of oriented gradients (PHOG) features of the optical images were extracted and input into the classifier of the beetle antennae search (BAS) and support vector machine (SVM) for fault diagnosis. Finally, the method in this paper was compared with other image feature extractions and support vector machine optimization algorithms.
The conclusions are as follows: (1) The light guide rod can measure the PD in GIS and reflect the signal characteristics of typical PDs. The discharge phase distribution of the optical signals is similar to that of the discharge phase distribution collected by the pulse current method, and the discharge pulse amplitude increases with the increase of the applied voltage. (2) Typical PD signals were collected by the light guide rod and the fluorescent fiber, respectively. The defects were placed in the center of the tank, and the relative distance between the discharge source and the two sensors was kept the same. The results show that the starting detection sensitivity of the light guide rod was higher. For the floating discharge and particle discharge, the optical signals obtained by the two methods are basically the same in phase distribution, amplitude and number due to the large intensity and small number of discharge pulses. For the surface discharge and needle tip discharge, the detection effect of the light guide rod was better. (3) An optical image fault diagnosis method based on PHOG-BAS-SVM is proposed. The PHOG features can well obtain the local and overall feature information of images. The results show that the accuracy of fault diagnosis based on this model reaches 90.4%.
Key wordsOptical partial discharge    light guide rod    pyramid histogram of oriented gradients    beetle antennae search    support vector machine   
收稿日期: 2023-12-04     
PACS: TM85  
基金资助:国家自然科学基金资助项目(62075045)
通讯作者: 钱 勇 男,1977年生,副教授,研究方向为电力设备绝缘状态诊断与在线监测技术。E-mail:qian_yong@sjtu.edu.cn   
作者简介: 李 泽 女,1997年生,博士,研究方向为GIS局部放电光学检测与故障诊断。E-mail:lize123@sjtu.edu.cn
引用本文:   
李泽, 钱勇, 刘伟, 臧奕茗, 盛戈皞, 江秀臣. GIS局部放电光信号检测技术及光学图像诊断方法[J]. 电工技术学报, 2025, 40(1): 253-263. Li Ze, Qian Yong, Liu Wei, Zang Yiming, Sheng Gehao, Jiang Xiuchen. Optical Signal Detection Technology and Optical Image Diagnostic Method of Partial Discharge in GIS. Transactions of China Electrotechnical Society, 2025, 40(1): 253-263.
链接本文:  
https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.232016          https://dgjsxb.ces-transaction.com/CN/Y2025/V40/I1/253