中文摘要：

在采用人工激励进行模态实验时,不可测的自然环境激励会引发过程噪声,降低频响函数估计的准确性。为此,提出了一种大噪声环境下提高振动结构频响函数测量精度的新方法。与现有方法的最大不同之处在于：设计了分别采用cos和sin线性扫频信号进行激励的模态实验,构造了＂虚拟＂的复数实验信号,避免了实数信号在分数阶傅里叶域内易产生信号成分混叠的不足,进一步提高了去噪效果。仿真和应用实例表明,新方法可显著改善信号品质,提高频响函数的估计测量精度。

英文摘要：

To decrease the heavy noise caused by natural excitation,an enhanced approach is proposed for measurement of Frequency Response Function（FRF）in operation conditions.Different from previous work,a’virtual’ complex experiment data instead of real data is used for Fractional Fourier Transform（FRFT）.The benefit of complex data is that the cross-term in FRFT domain existed in real data case will be removed.The corresponding modal test design is also presented to construct’virtual’ complex data by using cos and sin sweep excitation separately.Finally,the advantage of the enhanced method is illustrated by means of simulated and real cantilever beam test data.