基于核主分量分析的经验模式分解及其工程应用

王 雷;王奉涛;朱 泓;张志新;郭正刚

振动与冲击 ›› 2010, Vol. 29 ›› Issue (2) : 39-41,6.

PDF(1272 KB)
PDF(1272 KB)
振动与冲击 ›› 2010, Vol. 29 ›› Issue (2) : 39-41,6.
论文

基于核主分量分析的经验模式分解及其工程应用

  • 王 雷;王奉涛;朱 泓;张志新;郭正刚

作者信息 +

Application of Empirical Mode Decomposition Based on Kernel Principal Component Analysis

  • Wang Lei;Wang Feng-tao;Zhu Hong;Zhang Zhi-xin;Guo Zheng-gang
Author information +
文章历史 +

摘要

若信号的信噪比较小,经验模式分解不能正确分解出基本模式分量,分量中含有伪分量。根据此种情况,提出一种核主分量分析与经验模式分解相结合的方法。该方法首先建立信号相空间,利用核主分量分析方法提取相空间的核主分量,然后利用投影逆过程将得到的核主分量逆向投影回原相空间,从而重建信号相空间。最后对重建的相空间所对应的信号作经验模式分解。此方法可以有效消除噪声和冗余对经验模式分解的影响,提高经验模式分解的适应能力保证分解的有效性,确保其能够分解出正确的基本模式分量。通过工程实例进一步验证了该方法的可行性。

Abstract


If the signal-noise ratio is smaller, the correct intrinsic mode function(IMF) can not be decomposed by empirical mode decomposition(EMD), which may be a mistake IMF. According to this case, kernel principal component analysis (KPCA) and EMD method would like to be combined. Firstly, signal phase-space is founded , the kernel principal component of signal phase-space is extracted by KPCA, and then the phase-space is reconstructed by using the adopted kernel principal component through the inverse process. Finally, the reconstructed signal is decomposed by EMD method. The noise impact on EMD can be effectively eliminated by this way, so the decomposition ability of EMD is improved to ensure the validity of EMD in all cases and the correct IMF would be broken down. The effectiveness and feasibility of this method is verified further by actual works.

关键词

核主分量分析 / 经验模式分解 / 信号处理

Key words

kernel principal component analysis / empirical mode decomposition / signal processing

引用本文

导出引用
王 雷;王奉涛;朱 泓;张志新;郭正刚. 基于核主分量分析的经验模式分解及其工程应用[J]. 振动与冲击, 2010, 29(2): 39-41,6
Wang Lei;Wang Feng-tao;Zhu Hong;Zhang Zhi-xin;Guo Zheng-gang. Application of Empirical Mode Decomposition Based on Kernel Principal Component Analysis[J]. Journal of Vibration and Shock, 2010, 29(2): 39-41,6

PDF(1272 KB)

2251

Accesses

0

Citation

Detail

段落导航
相关文章

/