Observations from testing of indeterminate FRP-plated steel beams on flexible supports

This paper discusses the fabrication of, failure testing of and test data analysis from multiple two-span continuous FRP-plated steel I-beams. The study incorporated variation between beams of the layout and amount of FRP plating along the positive and negative moment zones, the novel concept of bonding (rather than welding) the steel plate stiffeners to the beams’ webs and flanges, use of longitudinal strain gauges on the bonded and exposed surfaces of key FRP plates, and elastomeric (hence flexible) pads to distribute stresses into the webs at the supports. The bonded steel stiffeners significantly delayed local buckling of the beams. Use of strains from only the exposed surface of any FRP plate produces erroneous magnitude changes and sign reversal of the high interfacial shear stresses near plate curtailment, relative to the prudent approach of using strain averages from the plate’s bonded and exposed surfaces. Differences between bonded and exposed surface strains enabled quantification of interfacial normal stresses. Relaxation of the rigid support assumption in favour of experimentally determined flexible support stiffnesses significantly nudges the predicted moment distributions toward the test distributions. This enables reliable assessment of the potential for end-peel failure arising from offsets between plate curtailment and contraflexure.