Predictive modelling of shear capacity of retrofitted RC beams

Fibre-reinforced polymer (FRP) composites have emerged as an effective technique for enhancing the shear performance of reinforced concrete (RC) beams in existing structures. This study proposes a data-driven framework for predicting the total shear capacity (Vtotal) of FRP-retrofitted RC beams using exploratory data analysis and ensemble machine learning (ML) approaches. A comprehensive experimental database comprising rectangular and T-beam specimens was compiled, incorporating geometric properties, material characteristics, reinforcement details and FRP strengthening parameters. Based on the identified influential variables, three ensemble ML models, namely extreme gradient boosting (XGBoost), random forest and AdaBoost regressors, were developed and evaluated. The models were trained using ten-fold cross-validation, while their predictive performance was assessed using the coefficient of determination (R2), adjusted R2 and mean absolute error (MAE). Among the developed models, XGBoost demonstrated the highest prediction accuracy, achieving an R2 value of 0.93 with an MAE below 10 kN. The results highlight the capability of ensemble ML techniques to accurately estimate the shear strength of FRP-strengthened RC beams and demonstrate their potential applicability in structural assessment and strengthening design.