Discussion Forum

Hi, a moving mesh implementation of the Navier Stokes equations which includes surface tension effects is now available in the Microfluidics module which was released with V4.2:

www.comsol.com/products/microfluidics/

Trying to implement this yourself would be practically impossible.

Dear Sumit,

I have the same problem, could you guide me through this for start?

Is it possible to use the level set method with deformed geometry option together ? how did you define the two phase with ALE method ? where can I start ?

level set and phase field methods that are based on the fixed mesh are not good for my two phase problem too.

best

Dear Daniel,

Is it possible to have the level set with deformed geometry together in COMSOL ?

I am using V4.2 but I do not have the Microfluidics module . I have large deformation in my problem which can not be done by level set method.

you mean in V4,2 Microfluidics module we can trace the interface of two phase flow using Moving Mesh ? ( Instead of level set and phase field which are based on the fixed mesh)

best

Hi, a moving mesh implementation of the Navier Stokes equations which includes surface tension effects is now available in the Microfluidics module which was released with V4.2:

www.comsol.com/products/microfluidics/

Trying to implement this yourself would be practically impossible.

The above quote is from Mr. Smith of Comsol. Is this really true, that setting up surface tension stresses on a moving mesh boundary is practically impossible before v.4.2? We have access to tangential derivatives via the dtang(f,x) function. It seems we could take the divergence of the normal vector to get surface curvature, and take the gradient of the surface tension to get tangential stress components.

So far I have been unsuccessful with v.4.1. I may have misunderstood the formulas. If anyone has set up a general stress boundary condition like this, I would appreciate a note about your method.

Thanks,
Dave...

Hi David, the difficulty is not only in implementing the force on the boundary due to surface tension, it is in applying the correct reaction force on the points (or edges in 3D) which intersect the wall and fluid-fluid interface boundaries. I would also say that using the dtang() operator on the normal vector is not recommended, neither is using the built-in variables for curvature, which were introduced in V4.3b. A better way of implementing the boundary force due to surface tension is to use a surface divergence theorem, which avoids having to use tangential derivatives. This approach is more complicated, but much more numerically stable.

We were actually able to further improve the implementation in the Microfluidics Module in V5.0, which you can see if you go to www.comsol.com/release/5.0 then Release Details>Fluid>Microfluidics Module.

Hi Daniel,

Thanks for the rapid reply on this question. At present I am limited to v.4.1. Perhaps there is a description of the improved scheme in the Theory Guide for 5.0 - I'll take a look at that.

Meanwhile, would you be willing to elaborate a bit on the surface divergence theorem treatment you mentioned? In the Jet Instability model I find a body force term added to the momentum equations (I suppose only for cells at the free surface):

div(sigma*(I - (n)transpose(n))
This seems to be a volume term that depends on a quantity that's defined only on a surface, so I'm not clear how to interpret it.

If the code admits it, I might try to get a net force at each free surface face by summing the attractive surface tension forces exerted by surrounding faces. Or we might apply attractive body forces between centroids of neighboring free surface cells. Any thoughts on the likelihood of success for these strategies?

Thanks,
Dave...

Hi,
i simulated the deformation of a drop in time in a multiphase flow level set. Now, i would plot on bidimensional graph maximum radius/time and minimum radius/time. Is it possible? If yes, how can i do?

Help me, please!