Neural Network Modeling for Assessing Punching Shear Strength in GFRP-Reinforced Concrete Slabs

Corrosion is a major problem that occurs in steel reinforcement concrete members. To counteract corrosion, glass fiber-reinforced polymer (GFRP) bars are gaining prime importance in the construction industry as flexural reinforcement members. Hence, this modeling research aims to optimize the shear behavior of the GFRP embedded concrete slabs. Different approaches have been used to forecast the punching shear behavior of GFRP-reinforced concrete slabs. However, well-established traditional design equations require refinement due to their poor accuracy and broad dispersion. This study adopted an artificial neural network (ANN) modeling to predict punching shear strength of GFRP embedded concrete slabs with the training of an experimental dataset. The ANN model’s efficiency was determined using coefficient of determination (R2), mean absolute error (MAE), and root mean square error (RMSE) metrics. The ANN model performance demonstrates that it can efficiently forecast the punching shear strength of concrete slabs, with an R2 value of 0.92. It also found that punching perimeter is the most important factor influencing the punching shear strength in GFRP-reinforced concrete slabs.