ANN Model for Predicting Ultimate Shear Strength of Reinforced Concrete Corbels

Abstract

ABSTRACT The artificial neural network (ANN) model was developed using previous experimental data on Reinforced concrete (RC) corbels to simulate the behavior of RC corbels. The neural network model has six input parameters representing the concrete compressive strength ( ), shear span (a), effective depth (d), corbel width (b), area of main reinforcement (As), area of secondary reinforcement (Ah), one output parameter representing the ultimate shear load (Vu). A back propagation neural network (BPNN) with the log-sigmoid activation function is adopted due to its accuracy of prediction. The ANN model is constructed using the experimental data from the literature. The ANN predicted ultimate shear load which compared with those calculated by ACI318-08 code Formula and Russo model. The neural network model is to predict the shear load of RC corbel more accurate than the ACI318-08 code Formula, and Russo model. Through the parametric studies using the ANN model, the effects of various parameters such as ( , a, d, b, As, and Ah) on the behavior of RC corbel were shown. The results reveal that the proposed network model captures the RC corbel underlying shear behavior very well.