基于压缩感知的微网谐波分析方法  

Harmonic Analysis Method in Microgrid Based on Compressed Sensing

作　　者：杨挺[1] 徐明玉[1] 袁博[1] 

机构地区：[1]天津大学电气与自动化工程学院,天津300072

出　　处：《天津大学学报（自然科学与工程技术版）》2016年第5期480-484,共5页Journal of Tianjin University：Science and Technology

基　　金：国际科技合作专项资助项目(2013DFA11040);国家自然科学基金资助项目(61172014);天津市自然科学基金资助项目(12JCZDJC21300)

摘　　要：多类型分布式电源和电力电子装置的运用使微网中谐波分量更加复杂,针对目前Nyquist采样框架下前段谐波采样数据量大、压缩复杂度高的问题,基于压缩感知理论提出了一种微网谐波分析和同步检测方法.首先理论证明了电网谐波信号在DFT基下的稀疏性满足压缩感知必备条件;随后依据国标要求,选定采用稀疏测量压缩采样方法和谱投影梯度恢复算法的压缩感知过程,使采样端存储空间降低为传统Nyquist采样的M/N,且降低传统稠密测量复杂度.实验结果表明:谱线插值修正算法可有效提升检测精度,新方法对微网谐波的频率、幅值和相位的检测误差分别降低到0.000,1%,、0.053%,和0.05°以内,对间谐波的检测误差分别在0.002%,、0.15%,和0.2°以内.Distributed generations and power electronic devices make the harmonic environment of microgrid more and more complex. To solve the shortages of large volumes of stored data and high complexity of compression in sampling side with the framework of the Nyquist sampling theory,the compressed sensing techniques were integrated into harmonic analysis in this paper and then a novel method of microgrid harmonic analysis and synchronous detection was proposed. Firstly,it was theoretically proven that the sparsity of the microgrid harmonic signals in the base of DFT,which met the necessary conditions of compressed sensing technology. Then,to satisfy the Chinese standard,the sparse measurement and the spectral projected gradient algorithm were used in the processes of compressing and sensing. The new method reduced the storage space of sampling side to original M/N,and decreased the sampling complexity compared with the traditional dense measurement. The experimental results show that the spectrum interpolation correction algorithm can effectively improve the detecting accuracy,with which the detecting errors of microgrid harmonic＇s frequency,amplitude and phase have been reduced to lower than 0.000,1%,,0.053%,,and 0.05°respectively;the detecting error of simple harmonic＇s frequency,amplitude and phase are respectively within 0.002%,,0.15 %, and 0.2°,respectively.

关 键 词：压缩感知 谐波分析方法 谱线插值修正 稀疏测量 

分 类 号：TM73[电气工程—电力系统及自动化] TP393[自动化与计算机技术—计算机应用技术]