Predicting Shear Strength of Reinforced Concrete Beam s ibers W ith and W ithout F iber s U sing A rtificial N eural N etw orks

Abstract

Concrete generally is brittle material and having low tensile strength. These undesirable properties have greater effect on the shear behavior of concrete. The shear behavior, shear strength and shear failure mechanism of nonFIBER and FIBER reinforced concrete Beams without shear reinforcement (RC and FRC Beams) were investigated. The use of FIBER in concrete modifies shear behavior to be more ductile and enhances shear strength by resisting formation and growth of cracks. Shear strength is a quite complex system to be predicted and analyzed accurately as there are several factors affecting it. And yet, there is no agreed rational procedure to design of shear of RC and FRC Beams, although, several empirical and theoretical models have been proposed to attempt acquiring adequate equations with good accuracy for designing engineers. Artificial Neural Networks (ANNs) modeling technique was used in this research to predict shear strength of RC and FRC Beams. One ANN Model was built of three layers feed-forward with back propagation system and consists of nine input nodes, nine hidden layer nodes and one output node. The ANN Model was developed by the Optimization Modeling System "Solver" in the Microsoft Office Excel (2007) and using 177 set of actual and reliable data collected from previous studies. The developed ANN Model gave better performance when was evaluated and compared with: (1) current design codes equations (ACI 318-08, EC2-04, BS8110-97 and Spanish EHE-99) and (2) The proposed empirical/theoretical equations for shear strength of FRC Beams (Sharma-86, Narayanan and Darwish- 87, Ashour et al -92, Imam et al.-97 and Khuntia et al.- 99). The developed ANN Model was also used to evaluate the effect of the parameters governing shear strength of the RC and FRC Beams.