电测与仪表
電測與儀錶
전측여의표
ELECTRICAL MEASUREMENT & INSTRUMENTATION
2014年
22期
40-43
,共4页
谐波%谐波检测%神经网络%LMS算法
諧波%諧波檢測%神經網絡%LMS算法
해파%해파검측%신경망락%LMS산법
harmonics%harmonic detection%neural network%LMS algorithm
在比较了当今多种谐波检测方法的优缺点基础上，提出了基于线性神经网络的谐波检测方法。构造了谐波检测的网络模型，对利用该方法进行谐波检测的原理做了简单的介绍。文中运用最小均方差算法（ LMS ）调节网络权值W ，使得误差指标达到最小值。最后构造含有奇次谐波的负载电流函数，利用MATLAB软件分别运用加汉宁窗的快速傅里叶变换（ FFT）方法和构造的自适应线性神经网络方法进行仿真实验。通过实验结果表明，基于自适应线性神经网络方法的谐波检测技术具有更好的检测精度。
在比較瞭噹今多種諧波檢測方法的優缺點基礎上，提齣瞭基于線性神經網絡的諧波檢測方法。構造瞭諧波檢測的網絡模型，對利用該方法進行諧波檢測的原理做瞭簡單的介紹。文中運用最小均方差算法（ LMS ）調節網絡權值W ，使得誤差指標達到最小值。最後構造含有奇次諧波的負載電流函數，利用MATLAB軟件分彆運用加漢寧窗的快速傅裏葉變換（ FFT）方法和構造的自適應線性神經網絡方法進行倣真實驗。通過實驗結果錶明，基于自適應線性神經網絡方法的諧波檢測技術具有更好的檢測精度。
재비교료당금다충해파검측방법적우결점기출상，제출료기우선성신경망락적해파검측방법。구조료해파검측적망락모형，대이용해방법진행해파검측적원리주료간단적개소。문중운용최소균방차산법（ LMS ）조절망락권치W ，사득오차지표체도최소치。최후구조함유기차해파적부재전류함수，이용MATLAB연건분별운용가한저창적쾌속부리협변환（ FFT）방법화구조적자괄응선성신경망락방법진행방진실험。통과실험결과표명，기우자괄응선성신경망락방법적해파검측기술구유경호적검측정도。
Based on the comparisons of advantages and disadvantages of the current variety of harmonic detection methods,a harmonic detection method based on linear neural network has been proposed in this paper. The network structure model of harmonic detection has been established and the principle of harmonic detection by using this meth-od has been simply introduced. Least mean square( LMS)algorithm has been adopted to adjust the network weight W so as to minimize the error index. Finally,the load current function with odd harmonics has been constructed,and simulating experiments have been performed using the Hanning window Fast Fourier Transmission( FFT)and linear neural network method separately by MATLAB. The experimental results show that the harmonic detection technology based on linear neural network method has a better detection accuracy .
