Welcome to the next PoreLab lecture!
Who: Reza Haghani, PhD candidate at the Department of Geosciences / PoreLab at NTNU
Reza’s research is about pore-scale modeling of multiphase fluid flow inside porous media using the lattice Boltzmann method. Various phenomena, including capillary entry pressures, layer flow, snap-off, cooperative pore-body filling, and trapping in multiphase flow, happen in the pore scale and they strongly depend on the wettability of a rock sample. His PhD’s project aims to conduct direct modeling of multiphase flow on segmented micro-CT images of fluid distribution inside a porous medium seen in experiments to find the wettability map of the matrix surface.
When: Wednesday 30 October at 13:00 (Norway time)
Where: The lecture will be in person in the common room (PoreLab Trondheim) and will be streamed in the Kelvin room (PoreLab Oslo) and . From anywhere else, you will be able to join via the following Zoom link:
https://uio.zoom.us/j/65837085049?pwd=WjZianUyN3FJa2liQkxBbzQrOCtGdz09
Title: Multiphase modeling using phase field method
Abstract:
A commonly used relation in the context of the phase-field modeling, i.e., the summation of the order parameter is equal to unity, is widely implemented in numerical modeling. In a recent study, we have shown that this assumption will lead to inaccurate results and it should be evaluated more numerical settings. To investigate the effect of this relation, two different schemes are considered: in the one-equation scheme, one interface-capturing equation is solved and the order parameter of the other fluid is obtained via the summation relation. While in the two-eq scheme, two interface capturing equations are solved for determining the order parameters of the two fluids. These two schemes were investigated in terms of accuracy and computational costs using the LBM as the numerical solvers. It is shown that using the one-eq scheme produces asymmetrical results for a symmetrical flow, and also the results depend on the fluid of choice. To further investigate the differences, we prescribed the velocity field in a way that mimics different difficulties of a real velocity field, i.e., not divergence-free velocity fields and velocity fields with singularities. Results showed that when the velocity field is not divergence-free or it has singularities, the one-eq scheme leads to inaccurate and fluid-dependent results. The problem is alleviated to some extent when the velocity field is totally divergence-free and there are not any singularities. We observe that the one-eq and two-eq schemes give the same results only if the velocity is zero.