2010(电工技术学报)基于修正遗忘因子RLS算法的谐波电流检测新方法_刘国海.pdf

2010 年 1 月 电 工 技 术 学 报 Vol.25 No. 1 第 25 卷第 1 期 TRANSACTIONS OF CHINA ELECTROTECHNICAL SOCIETY Jan. 2010 基于修正遗忘因子RLS 算法的 谐波电流检测新方法 刘国海 吕汉闻 陈兆岭 刘 慧 （江苏大学电气信息工程学院 镇江 212013 ） 摘要 为提高电能质量，需对电网谐波进行补偿与抑制，其效果取决于谐波检测的精度和动 态响应特性。本文考虑到有源滤波器（APF ）的低信噪比特性，提出了一种基于自适应对消技术 的修正遗忘因子RLS 谐波电流检测算法，该算法从研究权值收敛方式出发，通过设置阈值来判断 突变，使得检测过程既有较快的动态响应又有较高的检测精度。最后通过Matlab 仿真分析和DSP 实验，验证了本文的修正算法既克服了LMS 算法初始收敛慢、精确度低，又克服了传统RLS 算 法的动态响应过慢的不足，证实了该方法是一种非常有效的谐波电流检测方法。 关键词：谐波检测 自适应滤波 RLS 算法 动态响应 中图分类号：TM48 A New Method for Harmonic Current Detection Based on Self-Adjustable Forgetting Factor RLS Algorithm Liu Guohai Lü Hanwen Chen Zhaoling Liu Hui （Jiangsu University Zhenjiang 212013 China ） Abstract To improve the quality of electric energy, the effects of restraining and compensating for the harmonics of power grid are determined by precision of detection and its dynamic response characteristics. Given the low signal-to-noise ratio of APF, a self-adjustable exponential forgetting factor RLS algorithm of harmonic current detection is introduced, based on adaptive noise canceling technology. Studied the convergence means of weights, the mutation of current can be judged by setting threshold. A fast dynamic response can be provided by ensuring high detecting precision. Finally, by analyzing the Matlab simulation and the DSP experiment, the RLS algorithm with self-adjustable exponential forgetting factor has a faster initial convergent speed and a higher precision than LMS, as well as a faster dynamic response than that of the normalized RLS algorithm. Therefore, it is concluded th