Prediction of FRP shear contribution for wrapped shear deficient RC beams using adaptive neuro-fuzzy inference system (ANFIS)

This study aims to predict the shear contribution of fiber-reinforced polymers (FRP) for the wrapped reinforced concrete (RC) beams using an adaptive neuro-fuzzy inference system (ANFIS). The training and testing data for ANFIS model development were prepared using 119 existing sets of data derived from different published literature. The input parameters of the ANFIS model includes effective depth of RC beams, shear span to effective depth ratio, compressive concrete strength, internal transverse steel ratio, and FRP properties like thickness, number of layers, ultimate strain, orientation of principal fiber, width and center to center distance of strips, whereas the shear contribution of FRP is considered as the output parameter. A head-to-head comparison between the predictions of the ANFIS model and the experimental results has shown that the ANFIS results are in good correlation with the experimental results. In addition, the performance of the present ANFIS model was compared with the predictions made by seven widely used design guidelines. Based on the comparison of different performance evaluators, it can be concluded that the present ANFIS model has better performance in the prediction of the shear contribution of FRP. A parametric study was also conducted to study the effect of individual parameters on the shear contribution of FRP composites.