Review on seismic performance of reinforced concrete interior and exterior beam-column joints

Abstract

Beam-column joints represent critical structural components of a reinforced concrete structure. They play a pivotal role in maintaining the structural integrity of an infrastructure. Damage to the beam-column joint could result in partial collapse and even complete structural failure. Joint shear strength emerges as the paramount indicator for evaluating the structural integrity and efficacy of joint design, to ensure the robustness and reliability of the overall structural framework. This study aims to identify the best performing joint shear strength model that is capable of producing predicted joint shear strength with the greatest accuracy. Additionally, the study seeks to utilize the most appropriate machine learning model to estimate the damage index of the joint based on displacement. A comprehensive database of beam-column joint data collated over 33 research articles are utilized in the analysis of joint shear strength. The dataset includes essential parameters such as concrete compressive strength, yield strength of various reinforcement types, and the experimental joint shear strength that is used in the evaluation of predicted joint shear strength. Several machine learning methods were incorporated to estimate the damage index. The results were substantiated by two main error estimators RMSE and MAE which solidified Kassem’s joint shear strength model as the most accurate and reliable model alongside identifying Gradient Boosting as the optimal machine learning technique for forecasting damage index.