一种引入动量项的自适应谐波检测算法  

An Adaptive Harmonic Detection Algorithm with Momentum Term

作　　者：马立新[1] 王玉珠[1] 王月晓[1] 彭华坤[1] 

机构地区：[1]上海理工大学电气工程系,上海200093

出　　处：《控制工程》2013年第6期1077-1080,1083,共5页Control Engineering of China

基　　金：国家自然科学基金(F050304);国家科技部政府间科技合作项目(2009014)

摘　　要：基于自适应噪声对消原理的谐波电流检测方法,得到了广泛研究并且已被验证是可行的。针对传统定步长最小均方自适应算法无法解决收敛速度和稳态误差这一矛盾,在研究最小均方算法基础上,提出一种引入动量项的变步长自适应谐波检测算法,采用动量项加快权值收敛,并且采用当前误差信号和上一次误差信号的自相关估算来调整步长迭代,使得谐波检测在实时性和精度上有显著提高。该算法计算量比较小,易于实现,具有良好的工程应用前景。仿真结果验证了该算法的有效性。The harmonic current detection method based on adaptive noise cancellation principle has been researched extensively and proved to be feasible. Invariable stepsize Least Mean Square adaptive algorithm can＇t solve the contradiction between convergence speed and steady state error,this paper presents a variable stepsize adaptive harmonic detection algorithm with momentum term,which uses the momentum term to speed up the convergence of weights and uses the autocorrelation estimates of the eurret and previous error signal to adjust stepsize itera tion. This algorithm has significant enhancement of harmonic detection ：in real time andaccuracy, and good engineering application prospects because of theJsmall calculation quantity and easy realization. The simulation results verify the effectiveness of the algorithm.

关 键 词：动量项 自适应滤波 谐波检测 变步长 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]