Natural convection of water-based nanofluids in three-dimensional enclosures

The single-phase approach, in which nanofluids are treated as pure fluids assuming that the solid and liquid phases are in local thermal and hydrodynamic equilibrium, has been used with good results to simulate forced convection flows. Conversely, its extension to natural convection has revealed to be more problematic, as the obtained numerical results are substantially different from the experimental data, mainly due to the effects of the slip motion occurring between the suspended nanoparticles and the base liquid, whose consequent nonuniform distribution of the solid phase concentration can significantly affect both heat and momentum transfer. In the present article, a two-phase model based on a double-diffusive approach is proposed to study the natural convection flows of water-based metal oxide nanofluids inside cylindrical and rectangular enclosures heated and cooled at their opposite sides with the scope to reproduce a number of experimental data-sets available in the literature. Given the assumption of local thermal equilibrium between nanoparticles and host liquid, and considering the Brownian and thermophoretic diffusion as responsible for causing significant relative velocity between the solid and liquid phases, the governing equations are solved by a control-volume numerical method implemented using the open-source platform OpenFOAM (Open Field Operation and Manipulation). It has been found that the nanoparticle diffusion gives the major contribution to the decrease of the heat transfer rate red usually observed experimentally in natural convection flows of nanofluids inside differentially-heated enclosures. This means that a two-phase model, in conjunction with proper correlations for the evaluation of the nanofluid effective physical properties, must be necessarily enforced to obtain reliable results for many engineering applications.