Convective heat transfer and friction drag in a duct inserted with aluminum foams have been studied experimentally. The combined effects of foam porosity (ε=0.7, 0.8, and 0.95) and flow Reynolds number (1900⩽Re⩽7800) are examined. Frictional drags for flow across the aluminum foam are measured by pressure taps, while interstitial heat transfer coefficients in the aluminum foam are determined using a transient single-blow technique with a thermal non-equilibrium two-equation model. Solid material temperature distribution is further measured for double check of the heat transfer results. To understand the frictional drag mechanisms, smoke-wire flow visualization is conducted in the aluminum-foam ducts. Results show that both the friction factor and the volumetric heat transfer coefficient increase with decreasing the foam porosity at a fixed Reynolds number. In addition, the aluminum foam of ε=0.8 has the best thermal performance under the same pumping power constraint among the three aluminum foams investigated. Finally, empirical correlations for pore Nusselt number are developed in terms of pore Reynolds number under various foam porosities.

This paper presents the results of heat transfer measurement and analysis for two 5×5×1 cm porous channels. The channels were made of sintered bronze beads with two different mean diameters, dp=0.72 and 1.59 mm. The local wall temperature distribution, inlet and outlet pressures and temperatures, and heat transfer coefficients were measured for heat flux of 0.8, 1.6, 2.4, and 3.2 W/cm2 with air velocity ranging from 0.16 to 5 m/s and inlet air pressure of 1~3 atm. The measurement covers the data in both thermal entrance and thermally fully developed regions. The local Nusselt numbers were correlated in the fully developed region. The fully developed Nusselt numbers were analyzed theoretically by using a non-Darcy, two-equation flow model. Heat transfer between the solid and fluid phases was modeled by a relation of the form hloc=ARen. A wall function was introduced to model the transverse thermal dispersion process for the wall effect on the lateral mixing of fluid. The predicted fully developed Nusselt numbers are in good agreement with the measured values.

Combined free and forced convection for steady fully developed laminar flow in horizontal rectangular channels under the thermal boundary conditions of axially uniform wall heat flux and peripherally uniform wall temperature at any axial position is approached by the method of successive overrelaxation. The convergence of the numerical solution is ascertained. Graphical results are presented for streamlines, isotherms, w/w0 versus Re Ra, f Re/(f Re)0 versus Re Ra, and Nu/Nu0 versus Re Ra for the aspect ratios γ = 0.2, 0.5, 1, 2, and 5 and Pr = 0.73. For square channels, velocity and temperature distributions for Pr = 0.73 and heat transfer results for Pr = 7.2 are also presented.

Numerical solutions given by a vorticity-velocity method are presented for combined free and forced laminar convection in the thermal entrance region of a horizontal rectangular channel without the assumptions of large Prandtl number and small Grashof number. The channel wall is heated with a uniform wall heat flux. Typical developments of temperature profile, secondary flow, and axial velocity at various axial positions in the entrance region are presented. Local friction factor and Nusselt number variations are shown for Rayleigh numbers Ra = 104, 3×104, 6×104, and 105 with the Prandtl number as a parameter. The solution for the limiting case of large Prandtl number and small Grashof number obtained from the present study confirms the data of existing literature. It is observed that the large Prandtl number assumption is valid for Pr = 10 when Ra ≤ 3×104 but for a larger Prandtl number when the Rayleigh number is higher.

No abstract