Estimating the Behavior of FRP-Strengthened RC Structural Members Using Artificial Neural Networks

AbstractFiber reinforced polymers (FRPs) have found increasingly wide applications in structural engineering due to their high strength-to-weight ratio and high corrosion resistance. One important application of FRP-composites is as a confining material for concrete. In the investigation, new approach based on an intelligent predicting system (artificial neural network) is developed using experimental data in order to simulate the behavior of FRP-retrofitted reinforced concrete members. With known combinations of input and output data, the neural network can be trained to extract the underlying characteristics and relationships from the data. Then, when a separate set of input data is fed to the trained network, it will produce an approximate but reasonable output. Neural networks are highly nonlinear and can capture complex interactions among input and output variables in a system without any prior knowledge about the nature of these interactions. A large number of tests have been reported in the literature on the axial compressive strength of concrete members confined by FRP. Having parameters used as input nodes in ANN modeling such as dimensions of member, material properties of FRP and material properties of concrete, the target node was confined compressive strength for structural member. In hidden layers of the networks, Log-sigmoid, Tansigmoid and pure-linear transfer functions were utilized. After constructing networks with constant input neurons but with different number of hidden-layer neurons, the best network by considering the mean squared error, regression value and overall performance of the system was selected. Then the results simulated by idealized neural network were compared with experimental data which showed a good and reasonable agreement