摘要
为了改善经验模态分解(EMD)过程中产生的端点效应,本文提出一种基于支持向量机和镜像延拓相结合的新方法对短时间序列进行延拓。首先应用支持向量机(SVM)对原始信号两端分别延拓一个极大值和一个极小值,再用带镜像延拓程序的EMD方法对延拓后的信号进行边分解边延拓,逐渐抛弃受"污染"的点,得到具有原始信号长度的固有模态函数(IMF)。本文将该方法应用于电力系统的谐波分析中,仿真结果表明该方法能有效抑制EMD方法的端点效应,可以得到效果更好的单分量谐波信号。
A new method, a combination of support vector regression machines and mirrorizing extension, is proposed in this paper in order to restrain the end effects of Empirical Mode Decomposion （EMD）. Firstly, the support vector regression machines are used to predict one maximum and one minimum in each side of the signal. Secondly, EMD method which includes mirrorizing extension procedure is applied in the decomposition of the extended signal. Finally, IMFs （intrinsicmode functions） are obtained with the same length of the original signal. This new method is applied to harmonic analysis, and simulation results show that this method is effective to restrain the end effects of EMD method. Accurate harmonic components are obtain with this method.
出处
《电工电能新技术》
CSCD
北大核心
2008年第2期33-37,共5页
Advanced Technology of Electrical Engineering and Energy
基金
教育部霍英东青年教师基金资助项目(101060)
四川省杰出青年基金项目(07JQ0075)
关键词
EMD
端点效应
支持向量回归机
镜像延拓
谐波分析
EMD
support vector regression machines
mirrorizing extension
end.effect
harmonic analysis
