Damage detection in reinforced concrete slabs of underground chambers using impulse response testing and statistical feature recognition

Accurate condition assessment is crucial for infrastructure safety, operational reliability, and cost-effective maintenance. This task is particularly challenging for buried structures, given their limited accessibility and complex boundary conditions. This study introduces a rapid non-destructive testing (NDT) methodology based on Impulse-Response test (IRT), enhanced with statistical feature selection and probabilistic analysis, enabling efficient diagnostics with a minimal number of symmetrically placed test points. A reinforced concrete (RC) slab extracted from a decommissioned underground utility chamber was evaluated using IRT at four symmetric locations. Frequency Response Functions (FRFs) were generated for each test point, and Dynamic Time Warping (DTW) was applied to quantify deviations from pairwise FRF comparisons. Statistical dispersion among DTW distances was further characterized using the Gini coefficient (G). The distances were then standardized into Z-scores to derive probabilities reflecting the likelihood of relative local damage. Points with the highest deviation exhibited defect probabilities exceeding 70%, which were subsequently confirmed by ultrasonic shear-wave tomography. To examine robustness, a finite element (FE) model of the slab was developed and validated, then used for parametric analyses of shallow (60 mm) and deep (240 mm) delaminations, both with and without a 150 mm viscoelastic asphalt overlay. Analyses were performed by progressively reducing defect stiffness and mass density to 1%, 5%, 10%, 30%, and 50% of intact properties. Across all scenarios, the framework preserved stable localization and consistent ranking, with defect probabilities reaching up to 81%, demonstrating sensitivity to defect depth and severity as well as resilience to surface damping effects. These results indicate that the proposed baseline-free, symmetry-based framework provides reliable localization of subsurface anomalies in RC slabs, including buried and asphalt-covered configurations.