Artificial Neural Network and Reliability-Based Design of Concrete Beams Reinforced by FRP Bars

Hau Tran (1) Laboratory for Computational Civil Engineering, Institute for Computational Science and Artificial Intelligence, Van Lang University, Ho Chi Minh City 70000, Vietnam. 2) Faculty of Civil Engineering, Van Lang School of Technology, Van Lang Univers)
Trung Nguyen-Thoi (3) Laboratory for Applied and Industrial Mathematics, Institute for Computational Science and Artificial Intelligence, Van Lang University, Ho Chi Minh City 70000, Vietnam. 4) Faculty of Mechanical, Electrical, and Computer Engineering, Van Lang Scho)
Quang-Thien-Buu Nguyen (Department of Civil and Environmental Engineering, Seoul National University, Seoul)

Civil Engineering Journal
Vol. 12 No. 5 (2026): May

FRP bars have been utilized widely to replace steel bars in concrete beams due to their excellent corrosion resistance. Therefore, this paper aims to propose an efficient procedure based on artificial neural network (ANN) and reliability analysis to predict the moment capacity, the failure modes, and the resistance reduction factor for the design of concrete beams reinforced by FRP bars. In particular, 200 FRP RC beams are collected to train and verify the ANN model. In addition, a source code based on the Monte Carlo method is developed in MATLAB for the reliability analysis. The ANN model and the Matlab code are integrated to determine the failure probability, the reliability index, and the resistance reduction factor of FRP RC beams by rigorously considering the uncertainty of numerous variables. According to the findings of this study, ANN can be applied to predict the ultimate moment of FRP RC beams well since the mean and CoV of the model error are only 0.98 and 0.12, respectively, which are better than those obtained from ACI 440.1R. Furthermore, the resistance reduction factors for the design of FRP RC beams by ANN can be taken as 0.65 corresponding to the target reliability index of 4.0.