A new bond-slip model for NSM FRP systems using cement-based adhesives through artificial neural networks (ANN)

Abstract

This paper introduced a novel Artificial Neural Networks (ANN)-based bond–slip model for the Near-surface mounted (NSM) FRP system using cement-based adhesives, as an alternative to epoxy adhesives due to their high-temperature resistance and moisture-durability problems, employing experimental data. Therefore, closed-form formulas were presented for key components of the bond-slip law, including maximum bond stress, corresponding slip, fracture energy, and post-peak branch, while taking important factors into account. Compared to available bond-slip laws, this innovative model demonstrates promising potential in predicting the bond behaviour, thereby enabling more efficient and reliable designs for the NSM FRP strengthening applications using cement-based adhesives.