Dynamic response of the Trinity River Relief Bridge to controlled pile damage: modeling and experimental data analysis comparing Fourier and Hilbert-Huang techniques

This paper presents the implementation of a method for nonlinear, nonstationary data processing, namely the Hilbert-Huang transform (HHT) in traditional vibration-based approaches to characterizing structural damage and shows the frequency signature of local structural damage in nonstationary vibration recordings. In particular, following the review of traditional approaches to characterizing structural damage from nonstationary vibration recordings, this study first offers the justifications of the HHT as an alternative and complementary data process in addressing the nonstationarity of the vibration. With the use of recordings from controlled field vibration tests of substructures in the Trinity River Relief Bridge in Texas in its intact, minor- and severe-damage pile states, this study then shows that the HHT-based approach can single out some natural frequencies of the structure from a mixed frequency content in recordings that also contain the time-dependent excitation and noise frequencies. Subsequently, this study exposes that the frequency downshift for the damaged pile relative to the undamaged one is an indicative index for the damage extent. The above results are also validated by an ANSYS model-based analysis. Finally, a comprehensive HHT-based characterization of structural damage is discussed, and the potential use for cost-effective, efficient structural damage diagnosis procedures and health-monitoring systems is provided.