Predicting concrete-steel bond performance at high temperatures: a data-driven approach using AI modelling

In reinforced concrete structures, the interaction between steel and concrete is a complex and diverse phenomenon essential to the structure's performance and design. The mechanical interlock and adhesion between the surrounding concrete matrix and the steel reinforcement bars are referred to as such. Even though the bond is more complex at higher temperatures, a correct estimate is still crucial for design. Therefore, this paper focuses on using data-driven models to explore the performance of the concrete-steel bond under high temperatures using ML modelling, including ANN, RFR, GB and KNN algorithms. To fully understand the impact of elevated temperatures on the concrete-steel bond, these models are designed to replicate the bond performance. The 316-point database generated from laboratory-based studies investigating concrete bond strength at various temperatures and fibre concentrations was utilised to create the dataset for the model. With the help of this study, engineers will be able to predict the performance of concrete-steel bonds at high temperatures and gain a better understanding of the behaviour of these bonds.