Modal identification of high-rise buildings by combined scheme of improved empirical wavelet transform and Hilbert transform techniques

Recently, the empirical wavelet transform technique has attracted much attention due to its advantage of dealing with non-stationary and nonlinear signals. For empirical wavelet transform-based methods, their performance is heavily dependent on the accuracy of Fourier spectral segmentation. Structural response signals of high-rise buildings under ambient excitations often include high-level noises, which may lead to inaccurate spectral estimation, thereby introducing errors in modal decomposition. To this end, an improved empirical wavelet transform method is presented in this paper to analyze the structural response signals with high-level noises, which employs the projection information of Fourier spectra to improve spectral representation. Then, the improved empirical wavelet transform method is combined with the natural excitation technique and Hilbert transform for identifying modal properties of super-tall buildings under ambient excitations. Through a numerical simulation study, the validity and accuracy of the combined scheme are verified even when structural response signals were embedded with high levels of noise. Furthermore, the proposed method is adopted for modal identification of a 600 m supertall building for evaluating its applicability in field measurements. It is demonstrated that the combined scheme can effectively and accurately evaluate the modal properties of the skyscraper under ambient vibrations with low amplitudes.