Wavelet Packet Singular Entropy-Based Method for Damage Identification in Curved Continuous Girder Bridges under Seismic Excitations

Abstract

Curved continuous girder bridges (CCGBs) have been widely adopted in the civil engineering field in recent decades for complex interchanges and city viaducts. Unfortunately, compared to straight bridges, this type of bridge with horizontal curvature is relatively vulnerable to earthquakes characterized by large energy and short duration. Seismic damage can degrade the performance of CCGBs, threatening their normal operation and even resulting in collapse. Detection of seismic damage in CCGBs is thus significantly important but is still not well resolved. To this end, a new method based on wavelet packet singular entropy (WPSE) is proposed to identify seismic damage by analyzing the dynamic responses of CCGBs to seismic excitation. This WPSE-based approach features characterizing damage using synergistic advantage of the wavelet packet transform, singular value decomposition, and information entropy. To testify the algorithm, a finite element model of a typical CCGB with two types of seismic damage is built, in which the seismic damage is individually modeled by stiffness reductions at the bottom of piers and at pier-girder connections. The displacement responses of the model to El Centro seismic excitation is used to identify the damage. The results show that damage indices in the WPSE-based approach can correctly locate the seismic damage in CCGBs. Furthermore, the WPSE-based method is competent to identify damage with higher accuracy in comparison with the wavelet packet energy based method, and has a strong immunity to noise revealed by robustness analysis. An array of responses used in this approach paves the way of developing practical technologies for detecting seismic damage using advanced distributed sensing techniques, typically the optical sensors.

Marginal Hilbert spectrum and instantaneous phase difference as total damage indicators in bridges under operational traffic loads

“…Actually, recent advances in sensor and information technologies have improved the performance in big data (Entezami et al, 2020), monitoring (Spiridonakos et al, 2016), and damage identification (Delgadillo & Casas, 2020;Moughty & Casas, 2017) in bridges. The main dynamic signal processing techniques are fast Fourier transform (FFT) (Diez et al, 2016), wavelet transform (WT) (Li et al, 2019) and Hilbert-Huang transform (HHT) (Delgadillo & Casas, 2020;Chen et al, 2014). The Hilbert-Huang transform allows energy-time-frequency resolution for analysing data from nonstationary and nonlinear processes.…”

Quantifying Information without Entropy: Identifying Intermittent Disturbances in Dynamical Systems

“…Of the four SE extensions, the IIF is most similar to wavelet singular entropy (WSE), which uses the singular values from a singular value decomposition (SVD) of a wavelet time-frequency decomposition in the Boltzmann equation to express an entropy [ 43 ]. WSE operates on the premise that the singular values correlate to dynamical states.…”

Performance of Damage Identification Based on Directional Wavelet Transforms and Entopic Weights Using Experimental Shearographic Testing Results

“…Ren and Sun [ 51 ] used wavelet entropy (WE) for the identification of damage in beams, determining entropy from the measured time histories at the measurement points. A similar approach using WPT was presented in [ 52 , 53 ]. The improvement of this approach was proposed in [ 54 ], where the authors defined two entropy measures using probability analysis for the identification and quantification of cracks in beams.…”