陈孝信,周童浩,刘江明,戴鹏飞,刘延琦,李文栋,张冠军
CHEN Xiaoxin,ZHOU Tonghao,LIU Jiangming,DAI Pengfei,LIU Yanqi,LI Wendong,ZHANG Guanjun

• 1:国网浙江省电力有限公司电力科学研究院
• 2:国网浙江省电力有限公司检修分公司
• 3:西安交通大学电工材料电气绝缘国家重点实验室

摘要(Abstract)：

研究有效的特征提取及优选算法是提高GIS（气体绝缘封闭组合电器）局部放电检测准确性的有效手段。为了有效利用放电谱图的相位信息，提高GIS局部放电检测的精度和可靠性，提出了一种基于余弦相似度的局部放电模式识别神经网络。通过分析每种放电类型的谱图，总结出相位特征，并将每种放电类型的特征与待判断的放电谱进行比较，得到图像的相位参考量，再引入网络结构中进行训练。研究结果表明，引入余弦相似度计算后，神经网络的正确率提升了3.9%，且提升效果与谱图的相位特征有关，采集谱图的相位特征越明显，余弦相似度模块对精度提高的效果就越显著。该方法可以提高局部放电故障识别的精度，为GIS提供预警和故障评估功能。
The study of effective feature extraction and optimization algorithms is crucial for improving the accuracy of partial discharge(PD) detection in gas-insulated switchgear(GIS). To enhance the precision and reliability of GIS partial discharge detection by utilizing the phase information in discharge spectra, this paper proposes a neural network-based PD pattern recognition method that incorporates cosine similarity. By analyzing the spectra of various discharge types, phase features are summarized, and the features of each discharge type are compared with the spectrum of the discharge under evaluation. The resulting phase reference values are incorporated into the network structure for training. The research findings demonstrate that introducing cosine similarity calculations enhances the neural network's accuracy by 3.9%, and that the improvement is closely related to the spectral phase features—greater phase feature clarity leads to more substantial accuracy gains from the cosine similarity module. This method significantly improves PD fault recognition accuracy, equipping GIS with enhanced warning and fault evaluation capabilities.

关键词(KeyWords)： 气体绝缘封闭组合电器;局部放电;神经网络;余弦相似度计算;放电相位谱图
GIS;PD;neural network;cosine similarity calculation;discharge phase spectrum

Abstract:

Keywords:

基金项目(Foundation): 国网浙江省电力有限公司科技项目（B311DS23000K）

作者(Author): 陈孝信,周童浩,刘江明,戴鹏飞,刘延琦,李文栋,张冠军
CHEN Xiaoxin,ZHOU Tonghao,LIU Jiangming,DAI Pengfei,LIU Yanqi,LI Wendong,ZHANG Guanjun

DOI: 10.19585/j.zjdl.202502012

参考文献(References)：

• [1]董玉林，曾鹏，朱磊，等.绝缘子气隙缺陷下的局部放电发展过程研究[J].宁夏电力，2024(3):44-50.DONG Yulin,ZENG Peng,ZHU Lei,et al.Research on the development process of partial discharge in GIS due to insulator gap defects[J].Ningxia Electric Power,2024(3):44-50.
• [2]刘航斌，林厚飞，褚静，等.基于CWGAN-div和Mi-CNN的GIS局部放电图谱识别[J].浙江电力，2023,42(8):75-83.LIU Hangbin,LIN Houfei,CHU Jing,et al. Recognition of partial discharge patterns of GIS based on CWGAN-div and Mi-CNN[J].Zhejiang Electric Power,2023,42(8):75-83.
• [3]王桐，刘畅，王雍会.基于分数阶噪声滤波器的特高频局部放电检测抗噪声分析[J].河北电力技术，2024,43(1):61-65.WANG Tong,LIU Chang,WANG Yonghui. Anti noise processing for UHF partial discharge detection based on fractional order noise filter[J]. Hebei Electric Power,2024,43(1):61-65.
• [4]王其林，巩俊强，张文，等.“L”形测试截面下GIS绝缘间隙UHF信号传播特性[J].高压电器，2023,59(11):193-200.WANG Qilin,GONG Junqiang,ZHANG Wen,et al.UHF signal propagation characteristics of GIS insulation gap under “L shaped” test section[J]. High Voltage Apparatus, 2023, 59(11):193-200.
• [5]张国治，田晗绿，张磊，等.具备局部放电超声波感知功能的PZT基特高频传感技术[J].电工技术学报，2024,39(19):6215-6227.ZHANG Guozhi,TIAN Hanlü，ZHANG Lei,et al. Research on PZT-based ultrasonic-ultra high frequency composite partial discharge sensing technology[J]. Transactions of China Electrotechnical Society,2024,39(19):6215-6227.
• [6]苏志雄，孙康，丁浩，等.基于脉冲电流和紫外脉冲的多源局部放电诊断方法[J].绝缘材料，2024,57(1):101-108.SU Zhixiong,SUN Kang,DING Hao,et al.Multi-source partial discharge detection method based on pulse current method and ultraviolet pulse method[J].Insulating Materials,2024,57(1):101-108.
• [7] SHI H N,TIAN Y,WANG Y R,et al. Study on discharge mode and characteristics of insulation defects in GIS equipment[C]//2022 9th International Conference on Condition Monitoring and Diagnosis(CMD).November 13-18,2022. Kitakyushu,Japan:IEEE,2022:429–432.
• [8] WANG Y X,YAN J,WANG Z B,et al.Multi-source partial discharge diagnosis in gas-insulated switchgear via zero-shot learning[J].Measurement,2023,217:113033.
• [9]闫泽玉，杨洋，刘云鹏，等.基于神经监督决策树算法的多感知GIS局部放电识别[J].中国电机工程学报，2024,44(14):5821-5832.YAN Zeyu,YANG Yang,LIU Yunpeng,et al. Multiaware GIS partial discharge identification based on neural supervision decision tree[J]. Proceedings of the CSEE,2024,44(14):5821-5832.
• [10]彭曼，史钰潮.基于多域特征融合及概率神经网络的GIS绝缘故障诊断[J].电工技术，2024(1):55-59.PENG Man,SHI Yuchao. GIS insulation fault diagnosis based on multi-domain feature fusion and probabilistic neural network[J].Electric Engineering,2024(1):55-59.
• [11] YAN J,WANG Y X,YANG Z,et al.Few-shot mechanical fault diagnosis for a high-voltage circuit breaker via a transformer-convolutional neural network and metric metalearning[J]. IEEE Transactions on Instrumentation and Measurement,2023,72:1-11.
• [12] WANG L H,HOU K,TAN L Q.Research of GIS partial discharge type evaluation based on convolutional neural network[J].AIP Advances,2020,10(8):085305.
• [13] SONG H,DAI J J,SHENG G H,et al.GIS partial discharge pattern recognition via deep convolutional neural network under complex data source[J]. IEEE Transactions on Dielectrics and Electrical Insulation,2018,25(2):678-685.
• [14]冯旗，邵振华，余祉宏.基于PRPD图谱的气体绝缘开关柜多缺陷局部放电模式识别[J].电气应用，2023,42(6):48-54.FENG Qi,SHAO Zhenhua,YU Zhihong. Partial discharge pattern recognition of multiple defects in gas insulated switchgear based on PRPD spectrum[J].Electrotechnical Application,2023,42(6):48-54.
• [15]吴闽，蒋伟，罗颖婷，等.基于改进SSD的GIS多源局放模式识别[J].高电压技术，2023,49(2):812-821.WU Min,JIANG Wei,LUO Yingting,et al.Multi-source partial discharge pattern recognition in GIS based on improved SSD[J].High Voltage Engineering,2023,49(2):812-821.
• [16]许辰航，陈继明，刘伟楠，等.基于深度残差网络的GIS局部放电PRPD谱图模式识别[J].高电压技术，2022,48(3):1113-1123.XU Chenhang,CHEN Jiming,LIU Weinan,et al.Pattern recognition of partial discharge PRPD spectrum in GIS based on deep residual network[J]. High Voltage Engineering,2022,48(3):1113-1123.