Kernel Extreme Learning Machine Application in Prediction of Bond Strength Between EBR FRP and Concrete Substrate

In this study, the Kernel Extreme Learning Machine (KELM) is applied to develop a new accurate and efficient prediction model. For this purpose, 342 datasets related to the effective parameters involved in the debonding process are gathered from different sources in literature. A Gamma Test (GT) is used to determine the most effective parameters. Mechanical properties of the FRP plates and the bond length are determined as the most important parameters. Furthermore, the performance of the KELM model is investigated against the developed ANN and ANFIS models and the existing most common design equations.