Statistical modeling of shear strength in RC beams strengthened with EB-FRP: a comparative study with international standards

Abstract

Abstract

Abstract The accurate estimation of shear strength in reinforced concrete beams strengthened with externally bonded Fiber-Reinforced Polymers (EB-FRP) remains a challenge in structural engineering. Current standard formulations, such as ACI 440.2R-17 and fib Bulletins, often exhibit high scatter in results and tend to overestimate shear capacity, frequently neglecting influential parameters such as internal longitudinal reinforcement. This study proposes an improved prediction model grounded in rigorous statistical analyses. To this end, a comprehensive database containing 524 experimental results was compiled and curated. The methodology employed sensitivity analyses and multiple linear regression using backward elimination to identify statistically significant variables. The results culminated in the development of a model with 11 predictors (M11), which demonstrated superior performance compared to the evaluated international standards, yielding a higher coefficient of determination and an approximately 28.17% reduction in Root Mean Square Error (RMSE) relative to the average of the standards. The proposed model highlighted that the inclusion of longitudinal reinforcement characteristics is essential for calculation accuracy, offering a more robust alternative, particularly for beams with shear strength below 400 kN. This research enabled the construction of an experimental database, providing a relevant contribution to future investigations on shear behavior in these elements. Additionally, it allowed for the reduction of the problem's dimensionality, simplifying the modeling of complex relationships between parameters. Regarding the developed models, the M11 model surpassed the accuracy of current standards, providing a practical formulation for straightforward application in engineering design.