Machine Learning Approaches for the Performance Evaluation of Shear Strengthened Reinforced Concrete Beams using BFRP Composites with Different Wrapping Methods

An analytical investigation on the use of basalt reinforced polymer (BFRP) composites for the shear capacity enhancement of reinforced concrete (RC) beams is presented in this study. The focus is on investigating the influence of BFRP composites and different wrapping techniques on RC beam’s shear strengthening. The aim is to address the insufficient shear strength capacity of RC beams while ensuring sufficient flexural moment capacity. To achieve this, the beams are designed with two shear regions: one region with sufficient shear strength and another region with inadequate shear strength. The beams with insufficient shear strength are strengthened using BFRP composites through full wrapping, side and U wrapping techniques. The specimens are subjected to four-point flexural tests to monitor the variations in shear strength and displacement. The analytical investigation using ABAQUS software emphasizes the outcomes with empirical data to evaluate the efficacy of the various configurations strengthening. The outcomes underscore that the side wrapping and U strip wrapping techniques yield superior outcomes among the examined strengthening configurations in relation to shear enhancement. The validated analytical models corroborated these outcomes, providing a pragmatic framework for engineers to execute temporary retrofitting methodologies. Furthermore, this study underscores the paramount importance of integrating machine learning (ML) models in forecasting the performance of shear-strengthened RC beams utilizing basalt-reinforced fiber polymer (BRFP) across diverse wrapping configurations. Model performances were compared with statistical indicators. In addition, SHAP analysis was also performed to assess the relative importance of each input parameter on the output results.