Dynamic damage analysis of tunnel structure across multiple fault fracture surfaces based on Hilbert transform coupled with improved wavelet packet transform

Abstract

An improved wavelet packet threshold (IWPT) function combined with the Ensemble Empirical Mode Decomposition (EEMD) denoising method is proposed. Coupling with Hilbert transform, seismic damage dynamic identification on tunnel structure through multiple fracture surfaces is analyzed, based on the data obtained by shaking table test. The results reveal that the IWPT method significantly enhancing signal denoising performance compared to traditional threshold function methods. The IMFs optimal method based on the Composite Multiscale Permutation Entropy (CMPE) markedly improves accuracy of signal decomposition and reduces noise interference. The proposed damage identification method enables precise analysis of degree and opportunity of damage on tunnel structure crossing multiple fracture surfaces, surpassing traditional analytical methods. By analyzing the Hilbert marginal spectrum, instantaneous power, and instantaneous energy spectrum of tunnel structure, it is concluded that seismic damage is most severe near the main fault fracture surface (FFS) on hanging wall, followed by the main FFS on footwall, with the least damage occurring in the central of fault fracture zone. The method provides valuable reference for the seismic technology and damage identification analysis of tunnel engineering in high-intensity earthquake regions.