Experimental and Numerical Investigations of a Full-Scale Steel- and GFRP- Reinforced Concrete Bridge Deck under Pulsating and Rolling Load Fatigue

Abstract

A full-scale concrete bridge deck (15.24 m x 3.89 m) was tested under rolling load fatigue using Canada’s only Rolling Load Simulator (ROLLS) at Queen’s University. The deck, supported by steel girders spaced at 3.05 m, was reinforced with various materials to assess fatigue behavior. Shear studs ensured composite action, and cross braces provided lateral support, in compliance with Canadian Highway Bridge Design Code. The deck was divided into four longitudinal sections: two end sections reinforced with glass fibre-reinforced polymer (GFRP) rebar and subjected to stationary pulsating (P)-load and rolling (R)-load, respectively, and two middle sections—one with GFRP stay-in-place (SIP) form and top rebar, and the other with conventional steel rebar—both under R-load. Each section underwent 3 million (3M) fatigue cycles, except the steel-reinforced section, which experienced 6M due to its proximity to the travel path. The fatigue damage resulted by P-load and R-load is compared, and the fatigue performance of bridge decks with various reinforcement is assessed with R-load. R-load resulted in greater fatigue damage than P-load, with a 71% and 54% reduction in stiffness (k/ko) for the GFRP-reinforced sections, indicating that one R-cycle equates to 120 P-cycles. Significant cracking and concrete pitting were observed at sections under R-load. A conversion factor of 0.59 at 3M cycles, projected to 0.5 at 10M cycles, was established to translate stiffness degradation from P-load to R-load for concrete bridge decks reinforced with GFRP rebar. GFRP SIP section performed similarly to GFRP rebar in terms of stiffness degradation, with 60% lower residual deflection at 3M cycles. Nearly 90% of stiffness reduction occurred within the first 0.4M cycles. GFRP-reinforced sections experienced very similar reduction in stiffness to that of the steel-reinforced section of 69%, at 3M R-loading cycles. Despite differing stiffness degradation, both GFRP rebar reinforced sections exhibited similar residual punching shear strength (Vu), with only a 4% difference. Vu under 2-half axles spaced at 1.2 m increases by only 34% compared to a single load. A nonlinear finite element model, validated by experimental data, accurately simulated deck behavior under P-load and R-load. A parametric study examined the impact of moving direction, load location, and load level.