Files

Abstract

Open-pore aluminium foams are of interest in thermal management of power electronics due to their large specific surface area, low weight and relatively high thermal conductivity. The replication process is a low-cost and simple foam manufacturing method that offers many opportunities for the investigation of the influence exerted by the microstructure (porosity, pore size) and geometry or their combination in the performance of metal foams in heat-exchange applications. In this study, we explored replicated aluminium foams as a part of an integrated heat sink manufactured in one step that seamlessly combines metal and ceramic phases. Replicated aluminium foams were manufactured and tested first under natural and then under forced convection. A dedicated test apparatus was built to test these samples under both natural and forced convection, using the guarding ring technique. In the case of natural convection, the aluminium foams had different pore sizes (125-180 µm, 400-450 µm and 5 mm), porosities (78 to 86%) and geometry (disk, cylinder and finned structures). In all cases, replicated aluminium foams were found to dissipate the same amount of heat as their bulk aluminium (alloy) equivalents. The thermal heat exchange performance of replicated microcellular aluminium was also measured under forced convection. The test rig was modified to measure the pressure drop and volumetric air flow in a cylindrically symmetric jet impingement configuration. Foams of different pore size (125-180, 400-450 and 900-1300 µm), porosity (between 75 and 85 %), height (1 and 2.8 cm) and infiltrated at various pressures (between 2.7 and 70 bars) were manufactured for thermal testing, all having a hollow central channel, through which air is injected at a range of flow rates. The results show that within the covered parameter range foams of higher porosities, larger pore sizes and lower infiltration pressures dissipate more heat at lower cost in terms of pumping power. Shorter heights also dissipate more heat for a given volumetric flow but do so at the cost of an increased pressure drop and hence greater pumping power expenditure. A numerical model, developed and written by Dr. D. Ingram, using the finite-volumes technique, was employed to elucidate how the various parameters influence the heat transfer. The model solves the Darcy-Forchheimer formulation of fluid flow in porous media, which is then coupled to a convective/conductive heat transfer model of the same structure. These simulations were used to perform parametric studies on the influence of height and pore size with regard to the thermal behaviour of replicated foams. Furthermore, the model was benchmarked against the experimental results; overall, despite some discrepancies with measured temperature profiles, the model captures data trends well, and provides insight on the physical phenomena that underlie the experimental data and observations. In particular, the presence and influence of recirculating air flow along the inner portion of the heat exchanger structures was identified. Finally, we successfully fabricated the proposed integrated multimaterial structure and tested it under forced convection, using the same conditions as the replicated aluminium foams. Results demonstrate that this type of structures can be effectively used to cool power electronics.

Details

PDF