Abstract
The goal of this numerical investigation is to explore and compare the hydrothermal characteristics of free, mixed, and forced convection of water-copper nanofluid in a cavity using single-phase (SPM) and two-phase (TPM) methods. The top and bottom walls of the cavity are insulated, and the left and right walls are kept at a constant temperature so that the temperature of the left wall is higher than that of the right wall. The impact of volume fraction of nanoparticles (phi), Rayleigh number (Ra), Richardson number (Ri) and Reynolds number (Re) on the performance features is assessed. In addition, the artificial neural network (ANN) is employed to develop a predictive model of average Nusselt number of nanofluid. The results showed that the TPM was more accurate than the SPM. Additionally, it was observed that increasing the phi, Ra, Ri and Re leads to increasing the average Nusselt number of nanofluid. The maximum heat transfer improvement for the mixed convection and forced convection modes was 28% and 32%, respectively. Furthermore, the ANN modeling revealed the nanoparticle concentration has a negligible role on the results of free convection, while the opposite is true of forced convection.