Jonathan Florez Giraldo

NUMERICAL STUDY OF BIO-FLUIDS AND MASS TRANSFER PROCESSES THROUGH MEMBRANES

Blood rheology (haemorheology) plays a key role in tissue perfusion and its alteration from physiological conditions is often the main cause of cardiovascular pathologies. Therefore, the study of blood velocity profiles and wall shear stress distribution along micro-vessels is important in the field of cardiovascular diseases research. Recent advances in organ-on-a-chip highlighted the possibility of using artificial lung-on-chips which have been developed to replace the respiratory functions of the human lungs in pharmaceutical tests. We have expanded the study of micro-separation processes through micro-porous membranes by developing a numerical tool able to model the behavior of lung-on-a-chip micro-devices in both two and three dimensional geometries. As this is a multiscale problem, the new code consists in a hybrid LBM-FD (Lattice Boltzmann – finite difference) model on a non-uniform material grid, that models mass transfer processes in non-Newtonian flows. A part from the validation of the code, results obtained include the correlations of the non-dimensional numbers involved in mass transfer processes and the dependence on porosity, and the study of concentration profiles under steady (pipe flow) and the beginning of the study in non-steady (Womersley flow) conditions.
The LBM-FD hybrid model was used to study the mass transport through a hydrophobic micro-porous membrane located in-between a co-current flow passing through rectangular channels, which is similar to the micro-device used in Lung-On-a-Chip research. This code has been used to perform a parametric study to find the empirical correlation between Peclet number in the permeate channel and the mass transfer processes across the membrane which is quantified by mean of Sherwood number. The correlations in the two-dimensional micro-device reproduce correctly the linear scaling law of with the number of pores. The correlations give a power value equal to 1/3 (which characteristic of the Graetz-Leveque problem) for the scaling exponent of the average Sherwood number with Pe. This has been done in 2D and 3D models. In the three-dimensional case, we compared the results obtained using the power-law flow with a shear-thinning degree of n=0.7 against the results obtained using the Newtonian hypothesis (n=1). The non-Newtonian case.

Download: