Methodology for Rapid Estimation of Deflections in Two-Way Reinforced Concrete Slabs Considering Cracking

Abstract

The codes of practice do not provide a methodology for estimation of deflection in two-way reinforced concrete (RC) slabs. Hence, a methodology has been proposed for immediate estimation of central deflections in two-way RC slabs considering cracking under service load for use in everyday design. The methodology covers the derivation of a closed-form expression for deflections in two-way simply supported RC slabs. The terms of the expression are based on weights, biases, and transfer functions from the developed supervised artificial neural network (ANN). First, a sensitivity analysis has been performed to identify significant input parameters for generating the data sets. The training, validating, and testing data sets for the development of the ANN are produced through an automated process using a programmed Python script and the finite-element software ABAQUS in accordance with the selected significant input parameters. The expression predicts the inelastic deflection (considering cracking) from the elastic deflection (neglecting cracking). In order to gain general applicability, the expression has been verified for a number of slabs that differ in size and reinforcement detailing. The small differences indicate that it can be used for preliminary design purposes. The use of the closed-form expression requires a computational effort that is a fraction of that required in iterative procedures available in the literature. The proposed methodology can be applied under different types of support conditions of the two-way slabs.