Leonardo Fernando Ferreira

Abstract

The computational fluid dynamics (CFD) is a field of study of great interest and with several applications, and that, due to the complexity and extensiveness of the calculations involved, may require a large computational time to perform. On the other hand, Artificial Neural Networks (NN) is a field that has been shown to be useful in several areas of knowledge, including fluid dynamics. In order to explore a relationship between these areas, and investigate possible improvements in the CFD branch, single-phase flow simulations were created using the Lattice Boltzmann Method, and then NN techniques were used to recreate the simulations, considering only the first frame of the dynamics sequence produced by the simulator. The results obtained from it are consistent with the expected results, and the generalization of the method presents promising results. Thus, the use of NN techniques proves to be a good proof of concept for their CFD applications

Advising

Advisor: Anjos, G. R.; Leite, K.T.F; Mangiavacchi, N.