基于EMD-WPT-SVD和指数加权平均的MOA阻性电流去干扰方法  

Method for Eliminating MOA Resistive Current Interference Based on EMD-WPT-SVD and Exponential Weighted Average

作　　者：王星 阮莹 陈敏维 舒勤[1] WANG Xing;RUAN Ying;CHEN Minwei;SHU Qin(College of Electreical Engineering,Sichuan University,Chengdu 610065,China;Electric Power Research Institute of State Grid Fujian Electric Power Corporation,Fuzhou 350007,China)

机构地区：[1]四川大学电气工程学院,成都610065 [2]国网福建省电力公司电力科学研究院,福州350007

出　　处：《高电压技术》2021年第10期3664-3673,共10页High Voltage Engineering

基　　金：国网福建省电力有限公司科技项目(521304190001T)。

摘　　要：金属氧化物避雷器(metal oxide arrester,MOA)阻性电流测量受环境干扰、高频噪声干扰和脉冲干扰等因素的影响,阻性电流实测信号必然含有多种强干扰成分。针对传统阻性电流去干扰方法在多种并发干扰情形下难以适用的难题,提出了一种基于EMD-WPT-SVD和指数加权平均的MOA阻性电流去干扰方法。首先,依据阻性电流中周期干扰的基波频率、采样频率和特征频率是否被间谐波浸没的特点,选取最佳分解层数和最优小波基;其次,对MOA阻性电流进行经验模态分解(empirical mode decomposition,EMD),将含有特征频率的模态分量再使用小波包变换(wavelet packet transform,WPT),使每个子信号最多含有一个特征频率;再采用奇异值分解(singular valuedecomposition,SVD)准确提取所有子信号中的周期分量;最后,从原信号中准确消除阻性电流受到的周期干扰,然后通过指数加权平均算法得到能真实反映MOA工况的阻性电流变化趋势。将模拟信号和工程实测数据与多种经典方法进行比较,结果证明:所提方法能消除阻性电流的多种干扰,得到能真实反映MOA实时工况的阻性电流,其结果满足实际工程测量的要求。The measurement of resistive current for metal oxide arrester(MOA) is affected by environmental interference,high-frequency noise interference, and pulse interference. The measured resistive current signal inevitably contain a variety of strong interference components. In order to cope with the difficulty of traditional methods for eliminating resistive current’s interference in a combination of many interference, this paper proposes a new method for eliminating this interference which can affect the resistive current of MOA based on EMD-WPT-SVD and exponentially weighted average.Firstly, according to whether the fundamental frequency, sampling frequency and characteristic frequency of resistive current are immersed by inter-harmonic, the optimal numbers of decomposition layers and wavelet basis function are selected. Secondly, the MOA’s resistive current is performed by EMD, and then the wavelet packet transform is used to make each sub-signal to only contain one characteristic frequency, so that the characteristic components of all sub-signals can be accurately extracted by SVD. Finally, when the periodic interference of the resistive current is eliminated accurately from the original signal, the exponential weighted average algorithm is used to obtain the changing trend of resistive current that can truly reflect the MOA working condition. The method of this paper is compared with many classical methods in simulation and instance, respectively. It is proved that the proposed method can eliminate many kinds of interference and reflect accurately the real-time condition of MOA. The processing result meets the requirements of practical engineering measurement.

关 键 词：金属氧化物避雷器 阻性电流 经验模态分解 奇异值分解 小波包变换 指数加权平均 

分 类 号：TM862[电气工程—高电压与绝缘技术]