为提高结构 自振频率识别精度和识别工作的 自动化水平,提出基于希尔伯特-黄变换(HHT)和随机减量技术(RDT)的结构模态参数识别方法.首先利用傅里叶变换、巴特沃斯滤波获得目标频段动力响应,接着用经验模态分解(EMD)获得系列固有模态函数(IMF),采用RDT提取每个IMF的自由衰减振动信号,通过Hilbert变换,获得相位曲线和振幅曲线,最后拟合曲线斜率,获得自振频率和阻尼比.研究发现,无论是线性平稳信号还是非平稳信号,提出的方法在鲁棒性和准确性上均较传统的快速傅里叶变换(FFT)和HHT有优势,且无须人工介入判断模态真伪.对于非线性短信号,为获得更好的分析效果,推荐截取阈值取1.2σ,自由衰减时长取75 s.
An Operational Modal Parameters Identification Algorithm for Structures Based on HHT and RDT
An algorithm,utilizing the Hilbert-Huang Transform(HHT)and the random decrement tech-nique(RDT),was proposed to enhance the precision and automation of identifying the modal parameters of structures under operational conditions.The algorithm involves the use of Fourier transform,Butterworth fil-ter,empirical mode decomposition(EMD),random decrement technique(RDT),and Hilbert transform.Fourier transform and Butterworth filter were applied to obtain the dynamic responses of the target frequency range(DRTFR).EMD was employed to decompose the DRTFR into multiple intrinsic mode functions(IMFs).RDT was utilized to derive the free decay response signal of each IMFs.By subjecting the free decay response signals to Hilbert transform,phase curves and amplitude curves were generated.The slops of the phase curves and the amplitude curves were the frequencies and damping,respectively.The results showed that the proposed algorithm was suitable for processing both stationary linear signals and non-station-ary non-linear signals.This algorithm exhibited superior accuracy and robustness compared to the fast Fou-rier transform(FFT)and HHT.It is recommended to use the intercept threshold of 1.2σ and the free de-cay duration of 75 seconds for processing the non-stationary non-linear signals.
NETL
NSTL
万方数据
