Prediction of punching shear strength of two-way slabs

Abstract

The punching shear strength of two way slabs without shear reinforcement and without unbalanced moment transfer is estimated using both neural networks and new simplified punching shear equations. An artificial neural network (ANN) was used to predict the punching shear strength of internal slab–column connections. Neural network analysis is conducted using 244 test data available in the literature and experiments conducted by the authors to evaluate the influence of concrete strength, reinforcement ratio and slab effective depth on punching shear strength. A wide range of slab thicknesses (up to 500 mm) and reinforcement ratios were used. In general, the results obtained from the neural network are very close to the experimental data available. The test results were used to develop two new simplified practical punching shear equations. The equations also showed a very good match with available experimental data. Four equations for the punching shear strength prescribed in well-known specifications were evaluated based on the available experimental results. This paper includes a discussion of the parameters of punching shear strength in the American, Canadian, British and European specifications.