Optimization of structural elements in highly seismic areas using neural networks

Abstract

The aim of this research is to use Artificial Neural Networks (ANN) to dimension structural elements in regular 6-storey buildings. The necessary data for the training of the algorithm was elaborated manually with the help of the ETABS software, these were 30 buildings of reinforced concrete with a system of structural walls. The configuration and training of the neural network was carried out in the MATLAB software. The validation was carried out in an additional analyzed building in which the concrete savings were calculated, and the requirements of the current regulations were verified. Finally, the dimensioning obtained with the neural network generated a reduction of more than 10% in the total volume of concrete used in a 6-level building and establishes that the algorithm used provides effective results for an optimal design.