| A total of 39 GFRP-reinforced and 9 steel-reinforced concrete flexural members were tested under sustained loading for 100 days to determine the effect of several variables on long-term deflections. In addition, existing analytical models to calculate long-term deflections were evaluated, and a new model was developed that allows for better prediction of long-term deflections for both GFRP-reinforced and steel-reinforced concrete flexural members.;The experimental specimens were divided into four sets of 12 specimens with each specimen set used to evaluate a different variable. Set A specimens were used to evaluate the ability of GFRP-compression reinforcement to limit long-term deflections. GFRP compression reinforcement was found to be effective in reducing long-term deflections and a new compression reinforcement index, kr, equation is proposed. Set B specimens were used to evaluate the effect of the service-to-cracking load ratio, Ma/Mcr, and the sustained-to-cracking load ratio, Msus/Mcr on the long-term deflection multiplier. The Msus/M cr ratio had no effect on the long-term deflection multiplier, whereas, the Ma/Mcr ratio significantly affected the multiplier with higher Ma/Mcr ratios causing smaller long-term deflection multipliers.;Set C was used to evaluate the effect of the specimen geometry and reinforcement ratio, rho, on the long-term deflection multiplier by evaluating the multiplier with respect to the gross-to-cracked section moment of inertia ratio I g/Icr. Experimental data shows that as Ig/I cr increases, the long-term deflection multiplier decreases. Set D was used to evaluate specimen depth as it also relates to the span-to-depth ratio, L/d, and elastic neutral axis depth-to-tensile reinforcement depth ratio kd/d on the long-term deflection multiplier. As the specimen depth decreased, the long-term deflection multiplier increased, but this trend was only observed when specimens were tested at low load levels, Ma/Mcr< 1.3, where deflections are usually not a concern.;A semi-empirical analytical model, NOVA model, that considers the effects of Ig/Icr and Ma/Mcr on the long-term deflection multiplier was developed. The model was calibrated to the experimental data in this study and then used to predict experimental deflections from a past research study by Mias, et. al. (2013a, 2013b) for normal and high strength concrete specimens. The NOVA model provides good correlation with Mias's normal strength concrete specimens, but currently over predicts the additional long-term deflection for Mias's high strength concrete specimens with Ma/M cr ratios above those for which the model was calibrated. Further refinement of the analytical model should be performed as additional experimental data becomes available. |
