Using artificial neural network to optimize the flow and natural heat transfer of a magnetic nanofluid in a square enclosure with a fin on its vertical wall: a lattice Boltzmann simulation

Abstract

The heat transfer of a nanofluid in a square enclosure is numerically simulated using FORTRAN software in this paper, by considering the radiation (Rad) effect. The enclosure is under a magnetic field (MaF) and can also deviate from the horizontal axis. The enclosure has insulated top and bottom walls, as well as hot and cold right and left walls. There is a fin (0.1 × 0.4) with constant thermal conductivity on its hot wall. Entropy generation (EnG) is investigated as one of the important results in this paper. The lattice Boltzmann method is used for the simulations. A sensitivity analysis is performed on three parameters; including Rayleigh (Ra), Hartmann (Ha), and enclosure angle are introduced as the most effective parameters. Finally, based on these parameters, the performance of the cavity was simulated using an artificial neural network. The results show that the Nusselt (Nu) and EnG increase with the Ra and decrease with the Ha. The parameters mentioned above are subjected to optimization, indicating that the maximum Nu which is equal to 22.24 corresponds to the Ra of 992,409, the Ha of 0.004, and the inclination angle of 34.74°.