PREDICTING THE ULTIMATE BOND STRENGTH OF CFRP BONDED TO CONCRETE USING ARTIFICIAL NEURAL NETWORKS

The application of artificial neural networks (ANNs) to predict the ultimate bond strengths of CFRP bonded toconcrete is investigated. An ANN model is built, trained and tested using 118 sets of test data collected from the technicalliteratures. The data used in the ANN model are arranged in a format of 8 input parameters that cover the compressivestrength, tensile strength andwidth of cube concrete; the thickness, the width, the bond length and the elastic modulus ofthe CFRP plate. 15 sets of experimental data are usedfor predicting. The results predicted by the ANN model are comparedwith those obtained by 4 empirical equations. The results show that ANN is a feasible tool for predicting the ultimate bondstrengths of CFRP bonded to concrete.