DESIGN OF REINFORCED CONCRETE RECTANGULAR AND SOLID CIRCULAR COLUMNS SUBJECTED TO BIAXIAL BENDINGS USING ARTIFICIAL NEURAL NETWORKS

Abstract

In recent years, important advances in the field of the development of artificial intelligence tools have been obtained in practically all the areas of the scientific knowledge. The systems inspired by biological neural networks have been seen as a promising tool that is being successfully used in the solution of several problems in almost all areas of the technical-scientific knowledge. The paper explores the use of Artificial Neural Networks for design reinforced concrete sections subjected to combined axial load and biaxial bending moments. In a general way, this problem does not have an analytical solution and the computation of reinforcement is often an iterative process. In this context, the paper used Artificial Neural Networks techniques to assess the mapping between variables in reinforced concrete columns design. Feed forwad networks with back propagation training algorithm were used with more than 400 data for each type of cross section studied. Obtained results indicated good performance in real design conditions. Keywords: reinforced concrete columns, artificial neural networks, columns design.