基于FastICA改进EMD的算法研究

经验模态分解(EMD)算法在非线性、非稳态的信号处理上具有显著的优势,但EMD在实际应用过程中存在着一些缺陷,其中以模态混叠和虚假模态现象最为突出。模态混叠现象可以简单地概述为在1个本征模函数 (IMF)含有多于一阶的结构固有模态分量;虚假模态现象则是指不该有的频率组分对结构模态参数识别精度的严重影响。针对这一问题,对EMD中存在的以上两大缺陷展开研究,提出了利用频带滤波和独立分量分析算法(ICA)中的快速ICA算法(FastICA)相结合改进的EMD 算法。利用希尔伯特变换(HT)识别结构频率,并通过Benchmark 结构验证了所提算法的有效性。

Research on Improved EMD Algorithm Based on FastICA

The Empirical mode decomposition (EMD) has significant advantages in nonlinear and non steady response signal processing algorithms, but in the actual application process, the EMD existed some defects, in which the modal aliasing and false mode phenomenon is most prominent, the modal aliasing phenomenon can be simple in the 1 intrinsic mode function (IMF) containing more than one order structure of intrinsic mode components. The false mode phenomenon is the frequency component that should not be existed, which has a serious influence on the identification accuracy of structural modal parameters. To solve this problem, this paper studies the above two defects in EMD, and proposes an improved EMD algorithm which combines the fast ICA algorithm (FastICA) in frequency band filtering and independent component analysis (ICA) algorithm. The structure frequency is identified by hilbert transform (HT), and the effectiveness of the proposed algorithm is verified by the the Benchmark structure.