基于人工神经网络的电气设备超高频局部放电模式识别研究  

Research on Pattern Recognition of Ultra-High Frequency Partial Discharge in Electrical Equipment Based on Artificial Neural Network

作　　者：朱瑞丹[1,2] ZHU Ruidan(School of Mechanical and Electrical Engineering,Liuzhou Vocational&Technical College,Liuzhou Guangxi 545006,China;College of Engineering,Shantou University,Shantou Guangdong 515063,China)

机构地区：[1]柳州职业技术学院机电工程学院,广西柳州545006 [2]汕头大学工学院,广东汕头515063

出　　处：《电子器件》2023年第3期831-835,共5页Chinese Journal of Electron Devices

基　　金：2020年度广西高校中青年教师科研基础能力提升项目(2020KY31017)。

摘　　要：为了提高电气设备超高频(UHF)局部放电特征识别能力,提出基于人工神经网络(ANN)的电气设备超高频局部放电模式识别方法。根据电气设备绕组间的短路阻抗特性进行超高频局部放电模型构造,计算电气设备励磁支路增量电感,采用人工神经网络辨识方法进行电气设备超高频局部放电特征提取和模式识别,计算输出的静态电感和增量电感,采用快速傅里叶变换提取电气设备放电脉冲调节,计算一次侧基频电流幅值,根据励磁曲线和励磁电阻的匹配模式,建立电气设备超高频局部放电的参数提取模型,采用人工神经网络实现电气设备超高频局部放电模式识别。仿真分析结果表明,在0.020 s时,所提方法多数值均能检测出局部放电,采用该方法进行电气设备超高频局部放电模式参数识别的准确性较高,收敛性较好,抗干扰能力较强。In order to improve the feature recognition ability for ultra-high frequency(UHF)partial discharge in electrical equipment, a pattern recognition method for UHF partial discharge in electrical equipment based on artificial neural network(ANN)is proposed. The ultra-high frequency partial discharge model is constructed according to the short-circuit impedance characteristics of the electrical equipment windings, and the incremental inductance of the excitation branch of the electrical equipment is calculated. Artificial neural network identification method is used to extract the ultra-high frequency partial discharge feature and perform pattern recognition of electrical equipment. The output static inductance and incremental inductance are calculated, the discharge pulse regulation of electrical equipment is extracted by using fast Fourier transform, and the amplitude of primary fundamental current is calculated. According to the matching mode of excitation curve and excitation resistance, the parameter extraction model for ultra-high frequency partial discharge of electrical equipment is established, the artificial neural network is used to realize the pattern recognition of ultra-high frequency partial discharge of electrical equipment. The simulation results show that most of the values of the proposed method can detect partial discharge in 0.020 s, and the proposed method has high accuracy, good convergence and strong anti-interference ability to identify the mode parameters of ultra-high frequency partial discharge in electrical equipment.

关 键 词：ANN 电气设备 超高频 局部放电模式 识别 

分 类 号：TM41[电气工程—电器]