[keepfit]

Suppose a moving solid body in a two-fluid (e.g. Gas + Liquid), the viscosity of the mixture of two-fluid in the fluid region is:

  
mu_fluid = alpha_G* mu_G + alpha_L* mu_L;

and

  
alpha_G + alpha_L = 1.0;

If an artificial viscosity mu_S (e.g. 100* mu_L) is attributed to the solid region, then we can solve the three-phase flow with only ONE equation, where the viscosity and the density in the entire domain is averaged in this way:

  
mu = alpha_G* mu_G  + alpha_L* mu_L   + alpha_S* mu_S;
rho = alpha_G* rho_G  + alpha_L* rho_L  + alpha_S* rho_S;

here

alpha_G + alpha_L + alpha_S =1

in the entire domain.

Since volume fraction field (alpha_S) can be explicitly calculated, before solving the momentum equation and alpha equation (for gas and liquid), we can model these three-phase flow with a modified InterFoam.

The problem remaining is that: how do we effectively modify the fluid properties inside the solid region, if the cells (ID) inside the solid body are known? The first way comes to my mind is rather straight-forward:

for (auto& cID : cellsInSolid) { // fluid-solid boundary cells + inner solid cells  
  mu[cID] = alpha_S* mu_S + alpha_G* mu_G + alpha_L* mu_L;
  rho[cID] = alpha_S* rho_S + alpha_G* rho_G + alpha_L* rho_L;
}

Any better ideas?

Thanks in advance.

[RobertHB]

You can try using a fvOptions source. Take a look at these examples.

[keepfit]

Originally Posted by RobertHB

You can try using a fvOptions source. Take a look at these examples.

Never used the "fvOptions" source before, how do we account for the difference in viscosity and density source for the solid region? Note that there is no boundary defined on the fluid-solid interface. We only know the volume fraction (alpha_f) of the solid region.

Can you explain a bit more in detail?