基于复调制细化和Adaline网络的谐波检测方法  

Harmonic detection method based on complex modulation refinement and adaptive linear network

作　　者：窦嘉铭 马鸿雁 DOU Jiaming;MA Hongyan(School of Electrical and Information Engineering,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;Institute of Distributed Energy Storage Safety Big Data,Beijing 100044,China;Beijing Key Laboratory of Intelligent Processing for Building Big Data,Beijing 100044,China)

机构地区：[1]北京建筑大学电气与信息工程学院,北京100044 [2]分布式储能安全大数据研究所,北京100044 [3]建筑大数据智能处理方法研究北京市重点实验室,北京100044

出　　处：《中国测试》2022年第5期43-50,共8页China Measurement & Test

基　　金：北京建筑大学博士基金项目(ZF15054)。

摘　　要：针对传统快速傅里叶变换(FFT)与自适应线性(Adaline)神经网络结合的谐波检测方法中,存在难以辨识频率相近谐波的问题,提出一种改进方法。首先,利用复调制频谱细化对快速傅里叶变换后的频谱峰值进行细化,确定谐波的个数和大致频率。其次,将所得谐波个数和大致频率代入Adaline神经网络中进行训练,获得准确的谐波幅值、频率和相位。最后,利用不同谐波算例对所提方法进行仿真验证。所提方法均能准确检测出各算例中的谐波个数、频率、幅值和相位,误差均保持在小数点后三位,且用时较短。结果表明,该方法可以有效解决整数次谐波附近的精细化辨识问题,准确率高且时效性好,并具备一定的自动化程度。In order to address the problem of difficulty in identifying harmonics with similar frequencies in the conventional harmonic detection method combining fast fourier transform and adaptive linear neural network,an improved method is proposed.Firstly,the number of harmonics and the approximate frequency are determined by refining the peak of the spectrum after the fast Fourier transform using the complex modulation spectrum refinement.Secondly,the number of harmonics and the approximate frequencies are substituted into the Adaline neural network for training to obtain the accurate harmonic amplitude,frequency and phase.Finally,different harmonic examples are used to simulate and verify the proposed method.The method can accurately detect the number,frequency,amplitude and phase of the harmonics in each case,and the error is kept to three decimal places,and the time required is short.The results show that the proposed method can effectively solve the problem of refinement of identification near integer harmonics with high accuracy and good timeliness,and has a certain degree of automation.

关 键 词：电能质量 谐波检测 人工神经网络 复调制细化 

分 类 号：TM930[电气工程—电力电子与电力传动] TB9[一般工业技术—计量学]