New empirical approach for determining nominal shear capacity of steel fiber reinforced concrete beams

The main objective of this paper is to develop new design formulations for determining shear stress of steel fiber-reinforced concrete (SFRC) beams without stirrups using Gene Expression Programming (GEP) and Artificial Neural Networks (ANNs) based on a large number of test results. The proposed formulations relate the average shear stress to geometrical, and material properties of common reinforced concrete beam (effective depth, ratio of shear span to effective depth, compressive strength of concrete, and longitudinal steel reinforcement) and fiber properties (diameter, length, and volume percentage). In order to verify the validity and reliability of the proposed formulations, a comparative assessment was conducted between measured and calculated average shear stress of beams. The comparative assessment is carried out in terms of common and modified coefficient of determination (R and Rm), root- mean-square error (RMSE), mean absolute percentage error (MAPE), and gradients of regression lines (k and k’). The results obtained for the considered statistical measures and performance criteria reveal that all of the proposed formulations have acceptable ability to calculate average shear stress for a wide range of shear span to effective depth ratios.