机器视觉算法下配电网电气设备局部放电识别  

Partial Discharge Identification of Electrical Equipment in Distribution Network Based on Machine Vision Algorithm

作　　者：郑鹏 韩鹏程[2] 娄颖 王国栋 ZHENG Peng;HAN Peng-cheng;LOU Ying;WANG Guo-dong(School of Electrical Engineering&Automation,Luoyang Institute of Science and Technology,Luoyang Henan 471023,China;School of Electrical Engineering,Southwest Jiaotong University,Chengdu Sichuan 610031,China)

机构地区：[1]洛阳理工学院电气工程与自动化学院,河南洛阳471023 [2]西南交通大学电气工程学院,四川成都610031

出　　处：《计算机仿真》2025年第1期111-115,共5页Computer Simulation

基　　金：河南省重点研发与推广专项(232102241041)。

摘　　要：当配电网中存在局部放电故障时,会导致设备绝缘性能下降,进而影响电网的电能质量和电力稳定性。为了有效识别电气设备局部放电和合理安排检修方案,提出一种机器视觉算法下配电网电气设备局部放电识别方法。采用小波变换结合高阶偏微分方程,对配电网电气设备局部放电信号去噪。通过经验模态分解(Empirical Mode Decomposition, EMD)分解去噪后的全部放电信号,采用能量门限法对内涵模态分量(Intrinsic Mode Functions, IMF)展开筛选,使用敏感固有模态函数获取敏感IMF,将其和局部放电特征信号对比,完成电气设备局部放电信号特征提取。通过机器视觉算法中的BP神经网络,实现对配电网电气设备局部放电识别。实验结果表明,所提方法对电气设备局部放电识别准确率在98%以上,平均识别时间低于500ms。When there is a partial discharge fault in the distribution network,it can lead to a decline in the insulation performance of equipment,which in turn affects the power quality and stability of the power grid.In order to effectively identify partial discharge of electrical equipment and reasonably arrange maintenance plans,a method for partial discharge identification of electrical equipment in distribution networks based on machine vision algorithms is proposed.Wavelet transform combined with higher order partial differential equations is used to denoise partial discharge signals of electrical equipment in distribution networks.EMD is used to decompose all the noise removed discharge signals,and the energy threshold method is used to filter the IMF.The sensitive intrinsic mode function is used to obtain the sensitive IMF,which is compared with the partial discharge characteristic signal to complete the feature extraction of the electrical equipment partial discharge signal.Through the BP neural network in the machine vision algorithm,the partial discharge identification of electrical equipment in the distribution network is realized.The experimental results show that the accuracy of the proposed method for partial discharge identification of electrical equipment is above 98%,and the average recognition time is less than 500 ms.

关 键 词：机器视觉算法 配电网 电气设备 局部放电 识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]