Punching Shear Strengths of RC Slab-Column Connections: Prediction and Reliability

Based on a database of 342 experimental results with a wide range of influencing parameters, a new empirical model was developed using Multiple Linear Regression (MLR) for predicting the concentric punching shear strength of RC slab-column connections, without shear reinforcement, and was presented in this paper. An Artificial Neural Network (ANN) and international codes of practice as well as a state-of-the-art approach were also employed for comparison purpose. To avoid the over-fitting problem, the ten-fold cross-validation method was used to evaluate the performances in prediction. Finally, by using reliability analysis, the safety of these formulations evaluated in terms of safety indexes was also conducted in this comparison. The results indicated that the authors’ approach developed using MLR provided the best combination of accuracy, simplicity, and safety for computation of the punching shear strength.