Pore-scale convection-conduction heat transfer and fluid flow in open-cell metal foams: A three-dimensional multiple-relaxation time lattice Boltzmann (MRT-LBM) solution

Flow and heat transfer in porous foams are usually simulated based on averaged properties that are far from the actual values and can lead to substantial inaccuracies in the results. The pore-scale simulation could provide accurate and realistic models for such properties. Accordingly, a multiple-relaxation time lattice Boltzmann (MRT-LBM) code is extended for pore-scale direct numerical simulation (DNS) of fluid flow and conductive-convective heat transfer in open-cell metal foams (OCMFs) using the Palabos code, an open-source parallel LB solver. Three foam models with different numbers of pores per inch (PPI) are generated, and through simulations, the permeability, the non-Darcy coefficient beta, the onset of non-Darcy flow, and Nusselt number (Nud) are determined. The results show that by increasing the PPI of the samples from 20 to 40, the flow becomes more complex, and velocity and permeability are reduced by up to 70%. On the other hand, the heat transfer coefficient increases significantly so that in the same Reynolds number (Re), the overall heat transfer coefficient of sample 3 (40 PPI) is on average 110% higher than that of sample 1 (20 PPI). Finally, the global Nusselt number (Nud) is determined as a function of Re, and the related correlations are presented.

Automated summary

Simulation at the pore-scale level provides accurate and realistic models for flow and heat transfer properties in porous foams. Three foam models with different numbers of pores per inch were studied, revealing changes in flow complexity, velocity, permeability, and heat transfer coefficient with increasing PPI. The Nusselt number is also determined as a function of Reynolds number, showing significant variations between different pore densities.