A SYSTEMATIC REVIEW OF SHEAR STRENGTH PREDICTION MODELS FOR REINFORCED CONCRETE BEAM–COLUMN JOINTS UNDER CYCLIC SEISMIC LOADING CONSIDERING JOINT GEOMETRY, REINFORCEMENT DETAILING, AND AXIAL LOAD INTERACTION

Abstract

Reinforced concrete (RC) beam-column joints are critical structural components that significantly influence the seismic performance and overall stability of buildings. The shear strength prediction of these joints under cyclic seismic loading is a complex problem that depends on multiple interconnected parameters including joint geometry, reinforcement detailing, and axial load ratios. This systematic review comprehensively examines existing shear strength prediction models for RC beam-column joints under cyclic seismic loading, with particular emphasis on design code provisions, experimental evidence, and analytical approaches. A critical evaluation of approaches including strut-and-tie models, design code formulations, and empirical equations reveals significant variations in prediction accuracy and safety margins across different methods. Recent advances in machine learning and finite element analysis have demonstrated improved prediction capabilities, with coefficient of determination (R²) values exceeding 0.99 for neural network models compared to 0.83-0.85 for conventional design codes. The analysis of 50+ peer-reviewed studies shows that reinforcement detailing, particularly the provision of transverse reinforcement, is paramount in controlling shear failure modes and enhancing joint ductility. Furthermore, axial load effects introduce nonlinear interactions that current design codes inadequately capture, potentially leading to unconservative designs. This review synthesizes findings on the influence of key parameters including joint aspect ratio, concrete strength, reinforcement ratio, and axial load ratio on joint shear strength. Recommendations are provided for improving design methodologies and implementing performance-based design approaches for seismic applications. The findings emphasize the critical need for unified design criteria and the potential benefits of integrating advanced computational methods with experimental validation.