contact angle in 3D two compartments Shan-Chen model

[zshi6193]

Dear OpenLB,

I am using two-compartment Shan-Chen model to conduct a benchmark of contact angle in both 2D and 3D. The formula for the effective mass is “PsiEqualRho”. The virtual wall density is used to adjust the contact angle. The simulation can easily converge in 2d case but always failed in 3D cases. I think it is due to the high spurious velocity but not very sure. I am just wondering whether anyone has any suggestions to let the simulation converge in 3D cases.

Many Thanks.

• This topic was modified 4 years, 6 months ago by zshi6193.

[mathias]

Dear zshi6193,

I would check if the values at the boundaries are correctly set. Is the simulation unsatbale or the results are not as expected?

Best
Mathias

[zshi6193]

Dear Dr. Mathias,

Thanks for your quick response. I think that I find a way to solve the problem but want to double-check with you.

In two compartment Shan-Chen model(oil with rho=0.8 inside the water with rho=1), I try different g. When g is 3.5, the simulation can converged in around 40000 steps (the simulation domain is 100*100*60, the initial oil drop size is 40*40*40, periodic boundary in x and y, the top and bottom is set up to be bounce-back solid wall with a density (0.6 for latticeOil, 0.4 for latticeWater)). I increase g for larger repulsive force because I find that there are significant miscibility if g is 3 or less. When g is 3, the volume of formed droplet tends to decrease with the fixed contact angle and it takes very long time to converge. In that case, if the initial volume of droplet is small, the droplet tends to disappear after many iterations.To be concluded, It is very hard to achieve the mass conservation if g is 3 or less. Choosing g of 3.5 may works for setting up the contact angle in 3D.

Thanks for your response.
Best regards,
Simon

[mathias]

That sounds good – a common way to deal with the mass is to project the two avergage densities in the considered domain to a certain values. It is like the pressure in the Navier-Stokes equations which is definded up to a constant. So you basically fix that constant. There is a technical report on how it works for single component flows on our website. Best Mathias

[Author]

Posts