Natural convection in water-saturated metal foam with a superposed fluid layer.

Abstract

Experimental results are presented for steady state natural convection heat transfer in an enclosure with a horizontal layer of copper foam and an overlying water layer, heated from below. The foam was placed between two parallel plates in a cylindrical enclosure adjacent to either one or both boundaries. The cylindrical copper foam disks were manufactured by ERG Aerospace with 127 mm DIA, with porosity ranging from 0.88 to 0.92, and pore densities of 5 and 10 PPI. The effective thermal conductivity of the foam was measured by running experiments at pre-convection Rayleigh numbers. The permeability and drag coefficient of the foam were determined by measuring the pressure drop across foam samples in a separate facility. With foam on the hot boundary, natural convection heat transfer experiments are run for 5 and 10 PPI foam, Rayleigh numbers, based on the total height of the enclosure and the water properties, from 2 × 106 to 5 × 108, aspect ratios, H/D, from 0.1 to 0.8, and ratios of the total foam layer height to the total height of the apparatus, η = Hp/H, of 0.25, 0.5, and 0.75. With foam on both boundaries, experiments are run for 10 PPI foam, Rayleigh numbers from 1 × 107 to 3 × 108, aspect ratios from 0.3 to 0.8, and η of 0.5 and 0.75. Natural convection heat transfer experiments were also performed for water alone over the Rayleigh number range 9 × 105 ‒ 3 × 108 and for aspect ratios 0.2 to 0.8. Heat transfer results are presented as an enhancement factor, the ratio of heat transfer with foam to that of water alone at the same Rayleigh number. The results with foam on the hot boundary were not a function of η. 10 PPI foam on the hot boundary did not enhance heat transfer with an average enhancement factor of 0.98. 5 PPI foam on the hot boundary had an enhancement factor of 1.1. The increased enhancement of 5 PPI foam over 10 PPI foam is attributed to the greater permeability of the 5 PPI foam. With foam on both boundaries, enhancement tended to increase with Rayleigh number and is a function of η. For η = 0.5 enhancement ranged from 0.93 to 1.33. For η = 0.75 enhancement ranged from 1.1 to 1.29. For the same enclosure filled with copper foam (from Kathare et al. [13]), enhancement was not a strong function of Rayleigh number and ranged from 1.2 to 1.6, though enhancement was typically above 1.35. The decreased enhancement of a partially filled enclosure, compared to a fully filled enclosure, is attributed to a lack of direct conduction path from each boundary, significantly decreasing the apparent conductivity of the enclosure, and decreased advection as the convection cells do not fully penetrate the porous layer(s). Power law correlations relating the fluid Nusselt number to the fluid Rayleigh number were determined for the three foam cases. The data for foam on the hot boundary alone and on both boundaries were also successfully correlated using porous media variables where an apparent conductivity of the composite system, considering the foam and fluid layers in series, was used as the effective conductivity. The porous media Nusselt number is related to the porous media Rayleigh number and a modified porous media Prandtl number.