Modal Parameter Identification Of High-rise Building Under Non-stationary Environmental Excitation

A key issue in the health monitoring,control,and condition assessment of building structures is the estimation of modal parameters(structural damping,natural frequency etc.)based on measured response data of structure under environmental excitation.Traditional modal parameter identification methods under environmental excitation often need to consider the excitation signal as white noise or stationary signal,However,when high-rise building is subjected to wind loads,the excitation is non-stationary,the non-stationaryity is stronger under the strong typhoon,then these traditional identification methods are not applicable.How to deal with non-stationary signals has become a key issue in the modal parameter identification of high-rise building under environmental excitation.To solve this problem,in this paper,the latest nonlinear and non-stationary signal analysis methods,such as variational mode decomposition(VMD),empirical wavelet transform based on multiple signal classification algorithm(MUSIC-EWT),and adaptive non-parametric empirical wavelet transform(APEWT),were introduced into the modal parameter identification of engineering structure,combined with the traditional modal parameter identification methods,We proposed three methods for modal parameter identification of structure under non-stationary excitation.In the proposed methods,the structural response data was decomposed by the above three time-frequency analysis methods to obtain a series of single modal signals.Then,the free vibration response of each mono component was obtained by using NExT,and finally the frequency solution methods and curve fitting were used to estimate natural frequency and damping ratio of the structure.A numerical simulation of a three-story frame structure was employed to discuss the performance of the three parameter identification methods proposed in this paper.Firstly,the acceleration responses of different floors were analyzed to investigate the stability of these methods,and then the acceleration responses with different noise levels were analyzed to evaluate the noise immunity of these methods.Finally,the acceleration responses of structure under non-stationary excitation were analyzed to validate the effectiveness and reliability of these methods for estimating structural modal parameters.Furthermore,the measured responses of CITIC Plaza during Typhoon Damrey were analyzed based on the approachs.The accuracy of the developed methods were further verified by comparing the identified modal parameters of this super-tall building with those determined by EEMD method.Since the high-rise building is the structure with closely spaced modes,modal aliasing occurs during the decomposition of the measured response data.To solve this problem,in this paper,the SET was introduced into denoising of single component signal,exploiting the characteristic of SET that only retains the time-frequency information from the STFT results most related to the time-varying features of the target signal to eliminate frequency aliasing and noise.Firstly,the measured acceleration responses of CITIC Plaza were processed by MUSIC-EWT to obtain a series of single modal signals.Then,the single modal responses were denoised by SET method.Finally,the modal parameters of the structure were obtained by NExT,INHT and curve fitting.the ability to eliminate frequency aliasing and noise was validated by comparing the identified result with the value estimated by VMD-NExT.At the same time,an improved normalized Hilbert transform(INHT)and direct interpolation(DI)method were proposed to solve the frequency fluctuation and the ending effect of Hilbert transform(HT).A numerical simulation of a three-story frame structure and the analysis of measured data of CITIC Plaza were employed to verify the accuracy and stability of the two methods for solving instantaneous frequency proposed in this paper.