Abstract
Engineered surfaces are increasingly used to cool electronic devices with dielectric fluids in two-phase change. A common type of such surfaces is the open-cell metal foam. In this work, our goal was two-fold, to assess the influence of the thickness of the foam and which models of foam-finned efficiency can be most suitably used to predict their thermal efficiency. Numerical simulations of heat conduction in open-cell metal foams were carried out in foam-extend-4.0. Copper and nickel foams with three different thicknesses, 3 mm, 2 mm, and 1 mm, were used in the simulations under convective boundary conditions obtained from pool boiling experiments with two dielectric fluids, HFE-7100 and Ethanol, at saturation conditions and atmospheric pressure. The geometries of the foams were segmented from μCT images, converted to the stereolithography (STL) format, and used to build the computational meshes. The numerical extended surface efficiency was compared with classical analytical models and other correlations from the literature. The foam efficiency increased as the thickness decreased, i.e., the thinnest foams showed better efficiency than the other ones. Metal foam under pool boiling of Ethanol had lower efficiency than HFE-7100. The pin fin model with adiabatic tip exhibited the lowest mean absolute error, lower than 10% for copper foams and 20% for nickel foams; the copper foam with 1 mm and HFE-7100 had the lowest error, approximately 1%. Therefore, this model could be used as a guide during the design step to determine the metal foam characteristics of a cooling system. On the other hand, the models that consider a three-dimensional matrix presented errors higher than 30%.
NSTL
Elsevier