EMD和小波变换在结构模态参数辨识中的应用

针对希尔伯特-黄变换方法在振动信号处理中不能有效地进行模态解耦的缺点，提出一种将经验模式分解和小波变换相结合的结构模态参数辨识方法．首先利用经验模式分解方法对结构响应信号进行分解，获得多个本征模函数以完成结构模态的筛选过程，并根据模态筛选得到的本征模函数进行信号重构；然后对重构信号进行Morlet小波变换和模态解耦，获得信号小波变换系数的瞬时幅值和瞬时相位拟合曲线，并在此基础上计算出结构的各阶模态频率、阻尼和振型．结果表明，利用小波变换对本征模函数的叠加信号进行参数辨识代替对单个本征模函数进行希尔伯特变换，更能有效地进行模态解耦，从而获得更为准确的结构模态参数信息．

Application of empirical mode decomposition and wavelet transform in structural modal parameter identification

To overcome the limitation of hilbert-huang transform in decoupling the structural modes in vibration signal processing,a novel structural modal parameter identification method is proposed by combining empirical mode decomposition(EMD) with Morlet wavelet transform.First,the dynamical response signal is decomposed by using the EMD method to achieve intrinsic mode functions and implement the sifting processing of modes,on the basis of which the signal is reconstructed.Then,the variation of the transient amplitude and phase of wavelet transform coefficients with time is given by using the Morlet wavelet and the least squares fit,on the basis of which corresponding modal frequencies and modal damping ratios and mode shapes are achieved.Simulation results show that the method is effective in modal decoupling and accurate in modal parameter identification,which can not be implemented effectively by using the Hilbert transform of the individual intrinsic mode function due to modal mixture.