Predicting of Punching Shear Capacity of Corroded Reinforced Concrete Slab-column Joints Using Artificial Intelligence Techniques

Abstract

Rebars in reinforced concrete (RC) slab-column structures may corrode under unfavourable conditions, making slab-column joints (SCJs) more susceptible to punching shear (PS) failure. Moreover, PS failure is a common brittle failure, which makes it more difficult to evaluate slab column systems' functioning and failure probability. Thus, the prediction of PS resistance and the related reliability analysis are key factors for building RC slab-column systems. In this study, a highfidelity finite-element model was created using Abaqus. A comprehensive experimental record is compiled for corroded RC slab-column joints subjected to punching shear loading. Then, effective parameters are established by applyingstatistical technique principles. The text then provided a model of artificial intelligence, an artificial neural network (ANN). In addition, it provided guidelines for the future development of design codes by identifying the significance of each variable on strength. In addition, it supplied an expression demonstrating the intricate interdependence of affective variables. The results show that The ACI is the most dependable standard, while the CSA is the least. The ANN model had an average, coefficient of variation (COV), root mean square error (RMSE), and lower 95 % values of 0.93, 12.2 %, 1.8, and 0.82, respectively. As a result, the ANN model was found to be more accurate, reliable, and design-safe than variable uncertainty