摘要
实测信号中的噪声,以及模型阶次的不确定性给模态参数的准确识别带来困难。以提高模态参数识别精度为目标,提出基于模型定阶和信噪分离的复指数模态参数识别方法。该方法借助奇异值分解技术确定模型阶次,采用结构低秩逼近方法进行信噪分离。在此基础上,利用复指数法进行模态参数识别。分别选取一维的悬臂梁模型和二维的悬挂板模型进行物理模型实验,结果表明:该方法提高模态参数的识别精度,尤其是阻尼比的识别精度,具有较好的工程应用前景。
Modal parameters can not be identified easily due to the noise in measured data and uncertainty of the model order. A modal parameter estimation approach based on model order determination and signal-noise separation is proposed to improve the estimation accuracy. First, the model order is determined by the singular value decomposition (SVD). Then, the structured low rank approximation (SLRA) is ap- plied to get the filtered data. Finally, the modal parameters are identified using the complex exponential method. Experimental studies include a one- dimensional cantilever beam and a two-dimensional hung plate. The results indicate the proposed method can improve the accuracy of modal paremeters identification, especially in estimating damping ratios, and has good prospects for engineering applications.
出处
《中国海洋大学学报(自然科学版)》
CAS
CSCD
北大核心
2011年第1期155-160,164,共7页
Periodical of Ocean University of China
基金
国家高技术研究发展计划项目(2008AA09Z101)
国家自然科学基金项目(51079134)
山东省自然科学基金项目(ZR2009FQ007)资助
关键词
复指数法
实测信号
降噪技术
结构低秩逼近
complex exponential method
measured signal
noise elimination
structured low rank approximation