Estudio experimental y simulación del comportamiento inelástico de paneles compuestos usando redes neuronales artificiales

Abstract

The analysis of complex structures, such as panels composed of various materials, is difficult to model due to the variability in the mechanical properties of the materials. The foregoing, coupled with non-linearity in the behavior of materials, makes the application of traditional numerical methods difficult and highly demanding in computational time. The present work introduces a less conventional technique like the artificial neural networks (ANN) for the modeling of the permanent deformation and damages in a compose slab subjected to flexion. 400 ANN models were trained and verified, which were able to model the non-linearity of the structural element, successfully reproduce the damages due to cracking and buckling of the panel, as well as reproduce the global permanent deformation of the element.