基于改进Fast-ICA的电能质量谐波检测  

Harmonic detection of power quality based on improved Fast-ICA

作　　者：王继[1] 王年[1] 汪炼[1] 沈玲[1] 庄振华[1] 

机构地区：[1]安徽大学教育部电能质量工程研究中心,安徽合肥230039

出　　处：《电力系统保护与控制》2010年第18期126-130,共5页Power System Protection and Control

基　　金：安徽省科技攻关计划重大科技专项项目(08010202034);安徽省教育厅自然科学研究重点项目(KJ2010A007)

摘　　要：独立分量分析是利用信号的高阶统计量快速准确地实现信号分离和恢复。提出一种改进快速分离算法检测电能质量谐波的方法。在介绍了快速分离算法基本原理的基础上对牛顿迭代法进行了改进,减少了迭代次数,提高了收敛速度;依据负熵极大的独立性准则实现谐波信号的盲分离,再进行幅值修正以实现对真实信号的估计;对分离修正后的信号进行FFT分析,得到幅值和频率,进而实现对谐波的检测。仿真实验结果表明了该算法在检测精度和运行效率上都有所提高。Independent component analysis （ICA） separates and restores signal by using the high-order statistics of that signal. A methodofpowerquality＇sharmonicdetectionusingimporvedFast-ICAisproposed. First, the basic pdnciples of Fast-ICA algorithm are introduced and with which Newton＇s iteration is improved, thereby the number ofinterations is reduced and the convergence speed is improved. Harmonic signal is then carried out according to the independency principle of maximizing negentropy, the amplitude of the harmonic signal is revised to obtain the estimation of real signal. FFT is used to calculate the amplitude and frequency of the separative signal, thus the final harmonic＇s detection is completed. The detection accuracy and the operation efficiency have been improved in the simulation tests.

关 键 词：盲源分离 独立分量分析 Fast-ICA 电能质量 谐波 

分 类 号：TM933.4[电气工程—电力电子与电力传动]