Elmer Discussion Forum

Good evening dear community,
My problem is analyzing the flowing of a zinc film along a moving wall. Atop the zinc film, there is air. I am trying to include the Mesh Update Solver in my case.sif, in which I have already included the level set solver. I would like to see the mesh evolution in the overall domain; however, it looks like not solving the mesh update solver when running the case. I attached my case.sif.
Thanks in advance for your answer and help.
Andrea

I think along the free surface you need something like
Mesh Update 2 = Equals FreeSurface

https://github.com/ElmerCSC/elmerfem/tr ... s/freesurf

Hi, kevinarden,
Thanks for your reply, the problem is that I do not use the solver "FreeSurface", so my variable does not call FreeSurface.
Regards,
Andrea

Hi Andrea,

Problems involving shape can be solved typically with two different strategies. Either you have an Eulerian strategy with fixed mesh where the shape is described by some field. Levet-set falls into this category. Or you have Lagrangian strategy where the mesh follows the movement of the interfaces. Here you need some strategy to extend the mesh movement beyond the interface and this is the MeshUpdate solver.

So usually you would not combine these two strategies. Maybe you can take a step back and describe what kind of problem you have? I see you have a Navier-Stokes equation but I don't understand what is being convected and how it would change the shape. Does the shape change radically? Maybe you would need AdvectionDiffision equation + some reaction terms at interface + FreeSurfaceSolver + MeshUpdate?

-Peter

It may help understand the problem if you can share the mesh files or the geometry.

Hi
My colleagues and I would like to use the combination of levelset with meshupdate to avoid using a finer mesh everywhere. We want to refine the domain part of interest, namely where the gradients are stronger or there is a mayor levelset curvature. I attach the geometry.
Regards,
Andrea

Hi

Ok, this is understandable. However, this would mean that all relevant solvers should have ALE formulation. FlowSolver does, HeatSolver does, but notably there is no mesh velocity considered in the advection of the levelset function itself. So this would perhaps not be my 1st strategy to speed up things.

-Peter

Thanks so much, Peter. Recently, I have been proving to put the adaptive mesh command in the Navier-Stockes solver removing the mesh-update solver, nevertheless, anyway, it gives a problem with level set solver. I attached the case.
Regards,
Andrea