Structural health monitoring of exterior beam–column subassemblies through detailed numerical modelling and using various machine learning techniques

Abstract Structural health monitoring of beam–column joints is paramount, as they are critical load-carrying components of reinforced concrete buildings. Evaluating the ultimate joint shear capacity and failure modes of beam–columns, especially in seismic events, is a crucial task, especially in view of life safety concerns. Traditional methods used to determine the joint shear capacity of beam–column joints are often inaccurate and cumbersome owing to improper accounting of governing parameters that influence beam–column joints’ behaviour. In this study, the performance of machine learning-based structural health monitoring techniques are evaluated in predicting the joint shear capacity and the mode of failure for the exterior beam–column joint taking into account their complex structural behaviour through both numerical modelling and various machine learning techniques. The data used to train and test the model was collected from laboratory experiments and other test data available in the literature. The results indicated the superiority of the proposed particle swarm optimized artificial neural network (PSO-ANN) and XGboost over previously used approaches. Hence, the proposed techniques can be efficiently used for monitoring of structural performance by making informed decision regarding condition assessment of RC buildings.