BULETINUL INSTITUTULUI POLITEHNIC DIN IAŞI Publicat de Universitatea Tehnică „Gheorghe Asachi” din Iaşi Volumul 63 (67), Numărul 1, 2017 Secţia CONSTRUCŢII. ARHITECTURĂ VARIATION OF INTERNAL FORCES USING ARTIFICIAL NEURONAL NETWORK BY ALEXANDRINA-ELENA PANDELEA * , MIHAI BUDESCU and LUCIAN SOVEJA “Gheorghe Asachi” Technical University of Iaşi, Faculty of Civil Engineering and Building Services Received: February 2, 2017 Accepted for publication: March 2, 2017 Abstract. Color maps for normal and tangential tensions, for a masonry wall under a dynamic action, are used as input dates into artificial neuronal network (ANN). The advanced modeling software with finite element used for this model is ANSYS. The mansory wall has been discretized with the steps:10 cm, 7.5 cm, 5 cm, 2.5 cm, 1.25 cm and 1 cm. The numerical value of internal are used to obtained the way how they varies. Keywords: FEM; meshing step; internal forces; artificial neuronal network; back-propagation algorithm. 1. Introduction In the last years, the numerical modeling has become an area highly studied for researchers with the development of modern computing technology, that allow programs using parameters and networks of mesh as fine to get results closer to reality . Mathematical modeling of a structure subjected to dynamic actions requires a description of the system in relation to a small number of parameters. * Corresponding author: e-mail: pandelea.alexandrina@gmail.com