Experimental and numerical characterization of pultruded GFRP structural elements with stiffened Web-Flange junctions

The research presented herein pertains to the experimental characterization and numerical simulation of the behavior of pultruded Glass Fiber-Reinforced Polymer (GFRP) structural elements, with specific attention devoted to enhancing ultimate capacity of the Web-Flange Junctions (WFJs). An experimental campaign, per-formed on as-manufactured I-beam specimens is conducted to serve as a baseline for comparing behavior of pultruded GFRP members with the proposed stiffening strategy. The specimens were produced through the addition of external elements, using bonded L-shaped profiles with variable lengths that were installed in the proximity of the WFJ zone of the I-beams. The experimental tests aimed at highlighting the inherent premature failure that potentially occurs at the WFJ as a result of lack of reinforcement continuity between the flanges and the web that are required to withstand localized radial stress concentration at these locations. To this end, the top flanges of the thin-walled GFRP I-beams were clamped while a localized concentrated load was applied on the bottom flange by means of a steel jaw. Two sets of tests were performed, namely Mid-Point (MP) and End-Point (EP), according to the location of the applied load that caused delamination, and eventually separation of the bottom flange from the web. Failure modes observed in the tests, together with experimental load-displacement curves, are used to investigate the influence of external stiffeners length on the ultimate strength and stiffness of the I-beams. In executing the tests, Digital Image Correlation (DIC) techniques were used to track the spatial distribution of the deformation in the elements. Based on the results of the experimental campaign, a numerical investigation was conducted, to characterize the influence of different stiffeners' arrangements on the me-chanical response of the GFRP elements. The computational model was validated by comparing the observed response, both in terms of measured force-displacement curves, and bi-dimensional displacement maps obtained by means of the DIC technique. Numerical results highlighted the importance and influence of an explicit rep-resentation of the mechanical response of the bonding agent (i.e., the epoxy resin used to attach the stiffeners to the I-beams) to correctly characterize the deformation behavior of the strengthened WFJs.