Modeling of confined circular RC columns using artificial neural network and finite element method

A Finite Element (FE) and an Artificial Neural Network (ANN) model were developed to investigate the effect of different parameters, such as the space between stirrups and the diameter of longitudinal steel, on the behavior of circular Reinforced-Concrete (RC) columns confined with fiber reinforced polymer (FRP) sheets using the actual hoop rupture strain of the FRP. The FE model was accomplished using the FE software ABAQUS 6.13 which incorporates the nonlinear behavior of concrete material, the bilinear stress-strain curve of steel, and the linear elastic behavior of FRP. It was found that there is a good agreement between the FE model (FEM) results and the experiments. The ANNs were trained and tested on an experimental database from the literature. The database contains the experimental ultimate FRP strain, and the ultimate load results of 92 FRP confined circular RC columns. The ANN results agreed well with the FEM results. The neural network results were carried out to develop empirical equations to predict the effective rupture FRP strain and the ultimate load for circular confined RC columns with R2 of 0.912 and 0.932, respectively. The proposed equations estimate the ultimate strain from experimental works of the FRP with a small error up to 20%. In addition, the predicted results from the proposed equations exhibited good accuracy compared with previous guidelines and experimental and FE results from the literature, and it can be easily used in engineering designs as well.

Automated summary

A finite element and artificial neural network model were developed to investigate the behavior of circular reinforced-concrete columns confined with fiber reinforced polymer sheets. The models showed good agreement with experimental results, and empirical equations were proposed to predict column behavior accurately.