Prediction Of Reinforced Concrete Beam Depth Using Neural Networks

Abstract

This paper discusses the development and the implementation of a neural network for the depth prediction of singly-reinforced rectangular concrete beams. The procedure of the American Concrete Institute (ACI-318 1995) was used as the basis for the development of the proposed network. A training set of 56 cases was used to train the network. The network adequately learned the training examples with an average training error of 3.0 percent. A testing set of 19 cases was used to validate the network. The network was able to predict the correct beam depth with an average error of 6.8 percent. A case study, where 878 new design cases were considered, was conducted to demonstrate the system's generalization and fault-tolerance properties. The network showed good generalization and fault-tolerance properties since it was able to predict the correct beam depths with an average error of 9.2 percent