The determination of gas density in free-surface examples.

[aaron]

Hello,

Regarding the LBM-VOF implementation, please refer to examples/freesurface/breakdingDam3D.

In this implementation, three types of cells are involved: fluid, interface, and gas. Population communication occurs only between the fluid and interface cells, although the mass field propagates among all of them.

This raises a question: does the gas density matter in LBM-VOF, or is it numerically unstable to consider the equations of state (EOS) for both the fluid and the gas?

The first possibility is that the gas density does not matter because the gas is assumed to have zero pressure, except near the interface where the surface tension and fluid pressure require a local gas pressure for force balance. Elsewhere, since no population advection/diffusion of gas occurs, no density is needed.

The second possibility is that both densities (or pressures) must be tracked during the LBM iteration. However, incorporating two different EOS within a single LBM framework could cause spurious dynamics and even crash the simulation. (I understand that some multi-components/phases models like Shan-Chen may fail to cope with case of high density ratio).

In my view, if one is trying to recover the Navier–Stokes equations from LBM, it would be natural to compare the performance of LBM with other methods, such as the FVM, and at least discuss whether the material properties and EOS are comparable.

This brings us back to the issue of gas handling, because FVM requires a defined density and EOS for both the gas and the water.

Thank you for your attention.

Best regards,
Aaron

[mathias]

This method is refferred to as “Free Surface” in the context of LBM. It does model surface tension but not he gas phase. If you want to also model the gas dynamics, you should use a multi-phase/component method as “pseudo potential”, “free engergy” or “phase field”.

[Author]

Posts