In reinforced concrete (RC) members, the bond behavior is crucial to transfer the reinforcing bar stress to the concrete. However, estimating reinforcing bar bonding failure stress is challenging due to highly nonlinear relationships among components. Moreover, the accuracy of the existing design formulas to estimate reinforcing bar bonding failure stress is low. Hence, precise prediction of the reinforcing bar bonding failure stress is essential for the safe and economical design of the RC members. This study develops the artificial neural network (ANN) and teaching learning-based optimization (TLBO) models for predicting the reinforcing bar bonding failure stress using 394 experimental data points. For practical design, innovative formulas are derived using the developed ANN and TLBO models. The performance of the proposed formulas is compared with the existing design formulas. The comparisons indicate that the proposed ANN-based formula gives the best results, followed by the proposed TLBO-based one. Based on the ANN model, a parametric study is conducted to explore the impact of input parameters on the reinforcing bar bonding failure stress. Finally, a graphical user interface (GUI) is built to apply ANN and TLBO models to predict reinforcing bar bonding failure stress, reducing computational cost and less effort.
Read full abstract