Numerical investigation and optimization of natural convection and entropy generation of alumina/H2O nanofluid in a rectangular cavity in the presence of a magnetic field with artificial neural networks

Abstract

The SIMPLE algorithm and the volume control technique are used to investigate the natural convection of alumina/H2O nanofluids (NFs) flow in a cavity. The cavity consists of two hot triangular blades on the bottom wall. The two sidewalls are insulated and the upper wall is at a low-temperature. The cavity can have different angles. Brownian nanoparticle motion is also considered to model their thermal conductivity and viscosity. Finally, the values ​​of Nusselt number (Nu), Bejan number (Be) entropy generation rate (Sgen) are estimated by changing the cavity angle (γ), the Hartmann number (Ha), and the Rayleigh number (Ra). It is found that an increment in the Ra enhances the amount of Nu and Sgen, but the amount of Be is reduced. An enhancement in the Ha does not effect on Sgen and Nu at low values of Ra, while for high amounts of Ra, Sgen and Nu are reduced. The maximum value of Sgen and Nu and the minimum amount of Be, which are 10.36, 5.21, and 0.34, respectively, occur when γ = 30°. Also, the minim values of Nu and Sgen occur for strong free convection in the horizontal and vertical cavity, respectively.