Use of neural network and machine learning in optimizing heat transfer and entropy generated in a cavity filled with nanofluid under the influence of magnetic field: A numerical study

Abstract

The current research was conducted with the aim of numerical simulation of the natural convection heat transfer and the radiation heat transfer (RHT) of a nanofluid (NF) in a square enclosure under the effect of a magnetic field (MF). The enclosure can rotate and have various angles. The left-hand wall is warm while the opposite wall is cold. Moreover, the upper and lower walls are insulted and there are two blades on the warm wall. In this study, the Lattice Boltzmann Method (LBM), which was written manually in the Fortran software, were used to solve the partial differential equations and simulate the problem, respectively. The radiation parameter, Rayleigh number (Ra), volume percentage of nanoparticles (VNPs), and enclosure angle were among the parameters studied for both rectangular and triangular blades. By evaluating the results of Nu (Nu), generated entropy (GE), and Bejan number (Be), it was found that the highest value of heat transfer (HTR) and GE is in enclosure 90 with rectangular blades, although the Be is minimal in this case. Finally, the addition of radiation to the enclosure increases HTR for rectangular and triangular blades by 91 and 98%, and decreases GE by 4.2 and 2.6%, respectively.