融合经验模态分解与时频分析的单通道振动信号分离研究

摘要

The separation method of single-channel vibration signal was explored to accomplish the extraction of mechanical vibration source signals. The ensemble empirical mode decomposition technology was utilized to construct the pseudo multi-channel measurement signals. The principal component analysis method was used to determine the number of vibration source signals by the distribution of eigenvalues. Then, the blind source separation algorithm based on time-frequency analysis was implemented to acquire the vibration source signals. The experimental results show that the single-channel vibration signal can be effectively separated. The proposed method is feasible and practical.%通过单通道振动信号分离研究探讨机械振动源信号提取问题.采用集合经验模态分解方法将单通道信号构造成多通道信号,再通过主分量分析方法得到多通道构造信号的特征值分布情况并以此进行源信号数目估计,进而利用基于时频分析的盲源分离技术获取振动源信号.实验表明,该方法能有效实现单通道振动信号分离,具有较强实际应用价值.