Metal foam regenerators; heat transfer and pressure drop in porous metals

Open pore metal foams with moderate porosity (0.6 – 0.7) may be of interest as regenerators due to their high volumetric heat capacity and large specific surface area. Replication process is a low cost and simple foam manufacturing method which provides moderate porosity metal foams. Due to its simplicity, it provides many opportunities to investigate the effect of porosity, pore size and shape or their combination. In this study, this process was used to manufacture metal foams. A method, called vacuum-gas, was the standard method for manufacturing metal foams in the University of Sheffield Material Science and Engineering department. This method was further investigated and two new methods, gas-only and mechanical infiltration, were introduced. Based on the foams produced by these methods, the gas-only method was adopted due to its repeatability and quality. The method was further investigated by manufacturing eight more samples (1.4-1.7 mm pore size) under various infiltration pressures and the optimum infiltration pressures were found for manufacturing foams with pore size of 1-1.1 mm, 1.4-1.7 mm and 2.0-2.36 mm. A total of nine aluminium metal foams were manufactured for thermal and pressure testing. The manufactured foams had three different pore sizes, 1-1.1 mm (called Small samples), 1.4-1.7 mm (called Mid samples) and 2-2.36 mm (called Large samples). On average foams had porosity in the range of 0.62 – 0.65. Since this type of metal foams never been tested as a regenerator, two extra samples (a packed bed of 10000 2mm ball bearing and a packed bed of 100 layers of wire mesh No. 200) were made to compare with the manufactured foams and the results from other researchers. A test rig was built to test the pressure drop under steady state flow condition from 1 to 6.5 m/s (permeability based Reynolds number from 20 to 175). The extended Darcy-Forchheimer equation and a cubic velocity of Darcy-Forchheimer were used to measure the permeability and form drag of the samples. The results showed that the cubic velocity equation had a better prediction of the permeability and form drag. The Small samples had the lowest permeability and highest form drag coefficient for metal foams. The wire mesh sample had the lowest permeability and lowest form drag among the tested samples. In addition to steady state flow, samples pressure drop was also measured under oscillatory flow. A test rig was built to measure pressure drop and air instant velocity under oscillatory flow (1 to 19 Hz). The results showed that the oscillatory pressure drop was higher than steady state flow except for the Small samples which had higher pressure drop at steady state flow. The pressure drop for the wire mesh sample was measured to compare with other researchers data and a good agreement was observed with some of the published data. Moreover, the instant air velocity was measured by a hot-wire anemometer inside the connecting tube between the sample holder and the compressor. The results showed that the air velocity behaved like a turbulent flow during the acceleration and deceleration period. A single-blow test rig was designed and manufactured to measure thermal performance of the samples. To estimate the average heat transfer coefficient of the samples, several types of the single-blow models were studied and the extended Schumann-Hausen model was implemented for predicting the samples’ outlet air temperature history. Two matching techniques, maximum gradient and direct curve matching were used to match the experimental and modelled outlet temperatures history to estimate samples’ NTU and average heat transfer coefficient. The results showed that NTU increased with decreasing of pore size. Based on mass flow rate Mid samples had the highest h, however the difference between the metal foam samples were insignificant. The foam samples had higher heat transfer coefficient than the ball bearing sample but the wire mesh sample had the highest heat transfer coefficient. The heat transfer results for the wire mesh and ball bearing samples were compared with published data and good agreements were observed.