Artificial neural network for the prediction of the steel–concrete bond behaviour

Abstract

This paper aimed to show possible applicability of artificial neural networks (ANN) to predict the performance of the bond between reinforcement and concrete. An ANN model is constructed, trained and tested using the available test data of 117 different pull-out cylindrical concrete specimens with an embedded reinforcing bar. The data used in ANN model are arranged in a format of four input parameters that cover the concrete compressive strength, cover thickness, embedment length and related rib area. The ANN model, which performs in Matlab software, predicts the bond strength of anchoring capacity of the reinforcement in the concrete. The results showed that ANNs have strong potential as a feasible tool for predicting bond strength. Comparisons with empirical formula and experimental results of several different researchers show an acceptable accuracy of the proposed ANN model.