Effect of steel fibres on the torsional behaviour of concrete elements: unified model using Artificial Neural Networks

In the present study, the effect of steel fibres on the torsional behaviour of concrete elements has been studied by performing experiments on cylindrical beams. The percentages of steel fibres used were 0.5%, 0.75%, and 1% by volume of concrete. Three cylindrical concrete beam specimens of each percentage of fibre along with three specimens of plain concrete were cast and tested under torsion after 28 days of curing. Simultaneously, 150 mm cubes were also cast from each mix and were tested under a 200-ton compression testing machine, and the stress–strain curves were obtained. Addition of steel fibres increased the ultimate torsional strength of concrete beams and improved the ductility. The percentage increase in the ultimate torsional strength for 0.5%, 0.75% and 1.0% fibre content by volume in concrete over plain concrete elements was 15.6%, 23.12% and 38.15%, respectively. Apart from the experimental work, two different unified models using Artificial Neural Network (ANN) have also been developed in this study, namely, ANN-1 and ANN-2 models. The ANN-1 model predicts the stress–strain curve, while the ANN-2 model predicts the torque-twist curve. These ANN models are trained using the experimental data collected in this study. The performances of developed ANN models were evaluated, and the results were compared with the experimental results. It was found that the predictions by the developed ANN models are in good agreement with experimental results.