In-situ assessment of structural steel: Correlating mechanical properties using multi-modal hardness and Leeb rebound testing

Abstract

The in-situ diagnostic assessment of existing carbon steel infrastructure often relies on non-destructive testing (NDT) to estimate structural capacity without compromising member integrity. This study investigates the reliability of multi-modal surface hardness testing as a proxy for macroscopic tensile properties in structural carbon steels. Specifically, Brinell, Rockwell, Vickers, and Leeb hardness measurements are correlated with the 0.2% offset yield strength (YS) and ultimate tensile strength (UTS) of Grade 300 Universal Beams and Grade 350 Square Hollow Sections. To account for manufacturing-induced heterogeneities, 36 localized coupons were systematically extracted from varying spatial zones, including the web and flanges of hot-rolled profiles. Destructive uniaxial tensile benchmarking confirmed significant spatial variation driven by differential cooling rates during manufacturing, with the thinner, faster-cooling web elements demonstrating an average yield strength nearly 100 MPa higher than the corresponding thicker flange elements. This variation is consistent with the expected differences in thermo-mechanical processing and section geometry. Bivariate linear regression analyses revealed that predictive performance depends on the testing scale and measurement modality. For yield strength prediction, the macro-scale Leeb rebound method showed strong correlation (R² = 0.947, MAE = 10.49 MPa, MPE = −0.09%). In contrast, the Vickers micro-indentation method demonstrated greater predictive accuracy for UTS (R² = 0.847), indicating greater sensitivity to localized plastic deformation behaviour at higher strain levels. A key observation of this study shows that the reliability of dynamic rebound measurements is strongly influenced by testing boundary conditions. It was observed that uncoupled testing on thin structural elements (< 5 mm nominal thickness) induces kinetic energy dissipation via localized flexure, resulting in significant predictive variance. Consequently, operational guidelines are proposed for field assessments, indicating that rigid coupling for thin-walled elements, or the isolation of uncoupled testing to high-rigidity zones (such as flat flange elements), is necessary to maintain measurement accuracy. Ultimately, this research provides a preliminary framework suggesting that the portable Leeb hardness method can serve as a supplementary diagnostic tool for the yield assessment of these specific steel grades, provided appropriate testing boundary conditions are maintained.