基于改进的互补因散经验模式分解法的谐波检测法  

A improved method of harmonic detection based on the CEEMD algorithm

作　　者：吴衍 马碧芳 李立耀 陈国钦 Wu Yan;Ma Bifang;Li Liyao;Chen Guoqin(School of Electrical and Information Engineering,Fuqing Branch of Fujian Normal University,Fuqing 350300)

机构地区：[1]福建师范大学福清分校电子与信息工程学院,福清350300

出　　处：《高技术通讯》2019年第5期462-466,共5页Chinese High Technology Letters

基　　金：福建省教育厅中青年教师教育科研(JAT160570)资助项目

摘　　要：提出了基于改进的互补因散经验模式分解(CEEMD)算法的谐波电流检测法。该方法能将电流信号分解成内在模式函数(IMF),并创新地在分解过程中加入正负成对的高斯白噪声,抵消噪声余量,抑制了模式混叠问题,同时在筛分过程中加入平滑处理这个改进措施。使用新方法设计谐波检测电路,并做了与EEMD算法的对比仿真实验,结果表明:两个算法都可以分解出电流信号的谐波和基波分量,但是改进的CEEMD算法抑制了模式混叠问题,分解出来的基波分量与原信号基本吻合,两者的相关系数(CORR)为0.997,相对均方根误差(RRMSE)为0.00411,说明该法能够准确有效地分解谐波电流信号,同时该算法做了平滑处理的改进,可满足有源电力滤波器(APF)的需要。An improved method of harmonic detection based on the complementary ensemble empirical mode decomposition (CEEMD) algorithm is proposed. The method decomposes the signal into the intrinsic mode functions (IMFs), adds positive and negative Gauss white noise in the decomposition process to overcome mode mixing problem, decreases the residue of added white noises, and adds smoothing procedure. The harmonic detection circuit is designed by using the new method, and the simulation experiments in comparison with the EMD algorithm indicate that the two methods can detect harmonic current effectively and accurately. However, the improved CEEMD algorithm can eliminate the mode mixing. The decomposition of the fundamental wave is basically consistent with the original signal. The correlation coefficient (CORR) of the fundamental component and the original signal is 0.997, and its relative root mean square error (RRMSE) is only 0.00411. This algorithm can accurately and effectively decompose harmonic current signals and add smoothing procedure, which meets the need of accurate and fast detection of harmonic currents in active power filter (APF).

关 键 词：改进的互补因散经验模式分解(CEEMD)算法 模式混叠 谐波检测法 有源电力滤波器(APF) 电力系统 

分 类 号：TM935[电气工程—电力电子与电力传动]