Studying the Effect of RC Slab Corrosion on Punching Behavior Using Artificial Neural Networks

In this paper, the Punching Shear (PS) behavior of RC Slab-Column Joints (SCJs) exposed to rebar corrosion is modeled using an Artificial Neural Network (ANN). A total of 629 experimental and numerical datasets were used to develop the ANN model. Eight influencing parameters were considered as the input variables in the network namely, column cross-sectional area, effective depth of the slab, compressive strength of concrete, span-to-depth ratio, reinforcement ratio, column dimension, yield strength of steel, and corrosion degree. The punching shear capacity and the ultimate deflection were considered as the output variables. A graphical user interface was developed as a practical tool for predicting the PS behavior of corroded RC slab-column joints. The developed ANN model was compared with two empirical models from the literature. The results proved the efficiency of the proposed ANN model in predicting the PS behavior of corroded RC SCJs for different slab and column geometries, material properties, reinforcement ratios, and corrosion ratios. Additionally, the proposed ANN model was compared with the design equations of two codes, the latter yielded unsafe predictions for the PS capacity of RC slab-column joints in the event of corrosion. Furthermore, the proposed ANN model was utilized in carrying out a parametric study to assess the effect of the different parameters on the PS behavior of corroded RC SCJs. The ANN model proved to have the advantage of its simplicity in application compared with conventional methods such as experimental tests and finite element modeling, which are cumbersome and expensive.