Logo image
Evaluation of FE Analysis of Hybrid GFRP-Steel RC Beams Integrating an ANN-Based Bond Behaviour Model
Conference paper   Peer reviewed

Evaluation of FE Analysis of Hybrid GFRP-Steel RC Beams Integrating an ANN-Based Bond Behaviour Model

Rajeev Devaraj, Ayodele Olofinjana and Christophe Gerber
12th International Conference on FRP Composites in Civil Engineering (CICE 2025) Volume 2, pp.1020-1029
International Conference on Fibre-Reinforced Polymer (FRP) Composites in Civil Engineering, 12th (Lisbon, Portugal, 14-Jul-2025–16-Jul-2025)
Lecture Notes in Civil Engineering, 778, Springer Nature
2026

Abstract

artifical neural network bond modeling FE analysis hybrid GFRP-Steel structural analysis
Hybrid GFRP-steel reinforced concrete beams combine the durability of fibre-reinforced polymers with the ductility of steel reinforcement. However, accurately simulating their behaviour in finite element (FE) environments remains challenging due to the lack of explicit bond interface modelling. Conventional FE approaches typically assume perfect bond conditions between reinforcement and concrete, which often results in inaccurate predictions. To address this limitation, the present study introduces a computational framework that integrates an artificial neural network (ANN)-based bond model into FE analysis. The ANN, trained on experimental pull-out test data, predicts bond-slip relationships based on key parameters such as bar diameter, concrete strength, and cover thickness. This predicted behaviour is implemented in the FE model using COMBIN39 spring elements to represent the bond interface. Validation against three-point bending tests confirms that the ANN-integrated FE model accurately simulates the quasi-linear behaviour of hybrid GFRP-steel beams, achieving a coefficient of determination (R2) of 0.98. The results demonstrate clear advantages over traditional perfect-bond assumptions. The proposed framework provides a robust and data-driven approach for optimising hybrid GFRP-steel beam designs, enhancing the accuracy and reliability of structural performance predictions. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2026.

Details

Metrics

1 Record Views
Logo image