基于经验模态分解和奇异值分解的振动声调制信号分析方法研究

Research on the Analysis Method of Vibro-acoustic Modulation Signal Based on EMD and SVD

针对振动声调制特征信号被强噪声淹没无法有效提取的问题,提出一种基于经验模态分解与奇异值分解相结合的振动声调制信号分析方法。先对振动声调制信号进行经验模态分解,选取imf分量,然后将imf分量进行奇异值分解降噪,得到非线性特征信号,最后对特征信号进行Kolmogorov熵计算。将该算法应用于实际碳纤维复合材料的检测,利用Kolmogorov熵进行损伤评估。该方法成功提取了特征信号,实现了损伤诊断和定量评估,而且具有较强的自适应能力。

In order to solve the problem that due to the influence caused by strong noises the characteristic signal could not be extracted effectively, the analysis method of vibro-acoustic modulation signal based on empirical mode decomposition and singular value decomposition is proposed. The signal is decomposed by empirical mode decomposition for selecting the imf component, then singular value decomposition is conducted on the imf component for obtaining the nonlinear characteristic signal. Kolmogorov entropy of the characteristic signal was calculated. Secondly, the algorithm was applied to detect the actual carbon fiber composites. Finally, Kolmogorov entropy is used to assess damages. The method is successfully extracted characteristic signal, achieving diagnosis and quantitative assessment of the damages and the algorithm has strong adaptive capacity.