Prediction of bond strength of spliced steel bars in concrete using artificial neural network and fuzzy logic

Abstract

Artificial neural networks (ANNs) and fuzzy logic (FL) models have been used in many areas of civil engineering applications in recent years. The main purpose of this study is to develop an ANN and FL models to predict the bond strength of steel bars in concrete. For this purpose, the experimental data of 179 different splice beam tests were used for training, validating and testing of the models. The models have six inputs including the splice length, the relative rib area, the minimum concrete cover, ratio of the area of longitudinal tension bars to the effective cross section in the splice region, ratio of the cross-sectional area of stirrups to their spacing in the splice region and concrete compressive strength. The bond strength of steel bars in concrete was the output data for both models. The mean absolute percentage error was found to be less than 6.60% for ANN and 6.65% for FL and R2 values to be about 99.50% and 99.45% for ANN and FL for the test sets respectively. The results revealed that the proposed models have good prediction and generalization capacity with acceptable errors. Meanwhile, in this study the proposed ANN is a slightly more accurate than FL.