Damage Identification Based on Bridge Displacement Response Considering Autocorrelation of the Dynamic Component

Bridges demand efficient and reliable methods of structural health monitoring, as visual inspection is time-consuming and resource-intensive. To address this challenge, this study proposes a damage identification method for bridges based on Bayesian inference, using displacement responses induced by traffic loads. Unlike conventional influence line (IL)–based methods, the proposed method directly employs displacement responses, thereby avoiding explicit IL calculation. The bridge span is divided into multiple regions, and correction factors for bending stiffness are estimated to identify both the location and the severity of damage. In this framework, the dynamic component—treated as the residual between measured and predicted displacement responses—is explicitly modeled through a covariance matrix incorporating its time-domain correlation. To assess its effectiveness, the proposed method is compared against conventional IL-based methods—using both unfiltered and low-pass filtered ILs—in both numerical simulations and experimental applications. Numerical simulations using a simply supported beam model were carried out for three damage scenarios, which were designed based on the actual damage conditions introduced in the model bridge. The results indicated that filtering reduced variability in the estimated values but led to decreased identification accuracy near the bridge ends. In contrast, explicitly modeling the dynamic component with a covariance matrix improves the accuracy of identifying stiffness reductions near the bridge ends, thereby enhancing overall identification performance. In experimental validation, the proposed method accurately identified damaged locations, demonstrating its practical applicability. Furthermore, the estimated correction factors exhibited a clear correlation with actual damage severity, indicating the potential for quantitative damage assessment. These findings demonstrate the effectiveness of the proposed method and its potential to enhance bridge maintenance strategies.