基于SONS的结构模态参数识别研究

A Study of Structure Modal Parameter Identification Based on Second Order Non-stationary Source Separation

仅利用结构的振动响应信号,提出了使用2阶非平稳源(SONS)盲分离方法识别结构模态参数的方法。在介绍SONS算法和采用盲源分离方法识别模态参数流程的基础上,以简支梁锤击振动响应信号进行了实验,并分析了响应信号含噪声的情况。实验表明,SONS方法准确地提取了简支梁结构的低阶模态振型、固有频率和阻尼比,抗噪性能好,且分离结果较二阶盲辨识(SOBI)算法准确、稳定。针对高阶模态直接盲分离识别效果不好的现象,采用先滤波再盲源分离的方法,成功识别了高阶模态固有频率。

A method of identifying the structure＇s modal parameters by second order non-stationary source separation（SONS） arithmetic was proposed only using structure＇s vibration response signals.On the basis of introducing the process of SONS algorithm and the steps of using BSS method to identify modal parameters,an experiment was carried out using the hammering signals of simply supported beam,and the response signals with noise were analyzed.The experiment demonstrates that the structure＇s low order mode shape,natural frequency and damping ratio are picked up by SONS method accurately.The SONS algorithm has good noise immunity,and its separation results are better and more steady than SOBI（second order blind identification） algorithm.Aiming at the phenomenon where high order modal parameters can not be identified very well using direct BSS method,a method of doing BSS after signal filtering was proposed and high order natural frequency was identified by this method successfully.