摘要
在模态参数识别中,模态阻尼比往往比固有频率、振型参数更难被准确识别。为了提高频域分解模态识别方法的阻尼识别精度,分析了算法中产生误差的主要因素,提出了采用自回归模型谱估计代替原有周期图谱估计的改进方法。开展了仿真算例研究,以白噪声激励下结构的振动响应为输入条件,使用频域空间域分解法进行模态参数识别。算例结果验证了采用基于自回归模型的谱估计代替原有的周期图法谱估计,既保留了对密频模态的分辨能力,又提高了阻尼参数的识别精度,且在采样数据长度较短的情况下具有优越性。
Modal damping ratios are more difficult to be accurately identified compared with nature frequencies and mode shapes in modal parameters identification.To improve accuracy of the damping identification using frequency domain decomposition method,an error analysis is conducted and a spectrum estimation method based on autoregressive model is presented instead of the previous periodogram method.Two simulation examples are implemented in which the measurement response data of structure under white noise excitation is adopted as input to identify the modal parameters using the frequency and spatial domain decomposition.Results indicate that with the proposed method the ability of distinguishing closely spaced modes is retained,and identified damping ratios can be more accurate.Moreover,the proposed method is superior to periodogram method when the sampling data is not long enough.
出处
《振动工程学报》
EI
CSCD
北大核心
2015年第4期518-524,共7页
Journal of Vibration Engineering
基金
国家自然科学基金资助项目(10902024)
教育部新世纪优秀人才支持计划资助项目(NCET-11-0086)
航空科学基金资助项目(20090869009)
江苏省普通高校研究生科研创新计划资助项目(CXZZ13_0084)
关键词
参数识别
阻尼识别
频域分解法
谱估计
parameters identification
damping identification
FDD
spectral estimation
