Abstract: This paper presents a new output only time domain modal parameter identification method based on Independent Component Analysis（ICA）.Firstly，the method interprets the vibration system modals as virtual sources，and then uses ICA to process the system responses.Finally some single-frequency identification signals can be gotten，thus the multi-DOF system modal identification problem is turned into a single degree of freedom system parameter identification problem.A simply supported beam is tested by finite bandwidth Gaussian white noise excitation，the result shows that the method based on Independent Component Analysis can be used to identify the modal parameters effectively.

Key words: modal analysis, modal parameter, Independent Component Analysis（ICA）, Blind Source Separate（BSS）

摘要： 提出一种新的在只有输出条件下的基于独立分量分析（ICA）的时域模态参数识别方法。该方法首先将振动系统的各阶模态理解为相互独立的虚拟源，然后将系统响应信号进行ICA处理，得到单频可识别的信号，从而将多自由度系统模态识别转化为单自由度系统的参数识别问题，最后用有限带宽的高斯白噪声激励作用下的简支梁进行验证。结果表明基于ICA的模态识别方法可以得到较好的识别效果。

关键词: 模态分析, 模态参数, 独立分量分析, 盲源分离