Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Skin-depth meshing in a two phase fluid model of a cold crucible melting process
Posted 14.01.2014, 19:11 GMT-5 Low-Frequency Electromagnetics, Fluid & Heat, Computational Fluid Dynamics (CFD), Mesh, Modeling Tools & Definitions, Parameters, Variables, & Functions 3 Replies
Please login with a confirmed email address before reporting spam
I'm actually working on a model of a semi-levitating cold crucible melting process. I want to predict the deformation of a metalic mass due to the interaction of EM, hydrodynamic and gravitationnal forces. I assume that the mass inside the crucible is completly melted, so my model comprises only the CFD and EM part.
To deal with mass deformation, I tried first the ALE method. But since the mass can touch the crucible walls, the ALE method is not suitable for me because it cannot handle topological changes.
The best alternative I found, and that actually works well is the Two phase Phase field model. By setting the rigth boundaries and model parameters, I can achieve really good results. The problem is that I'm working at a frequency of 160kHz and the skin depth is very thin at this frequency. Since the mesh resolution is very important in the skin depth, I had to mesh the air + mass region with elements smaller then the skin depth. The simulations takes days but gives good resultats.
I'm trying to find a way to make a dense mesh only at the interface along the skin depth and make it follow the interface displacement, by some kind of remeshing or mesh adaptation trick. This way I can make the simulation faster.
Thank you very much !
Please login with a confirmed email address before reporting spam
I work on the same topic as you! Cold crucible for silicon melting.
I try to calculate the deformation of the liquid silicon due to Lorentz forces, I have given up but I am still interested to make studies like yours. I work with an approximative shape.
For the mesh I think that you can use a size node applicated to the boundaries (You have identity pairs I assume on the interface?), you can define a minimal size element for boundaries. In the time dependent solver you can click on remesh under adaptative mesh. In the remeshing node you can select a Minimum element quality between 0 and 1, remeshing works well.
Best regards
Je suis français, si vous voulez échanger d'avantage d'informations vous pouvez me contacter sur mon adresse email: jgivernaud@emix.fr
Votre travail m'interesse beaucoup. Je peux vous aider sur la partie electromagnetique 3D du creuset froid si vous ne l'avez pas encore fait
Please login with a confirmed email address before reporting spam
I will try your hint and let you know if that works. I'll contact you by email soon.
Merci !
Please login with a confirmed email address before reporting spam
Have you succeed to do your computation with the remeshing Comsol tool?
Best regards
Contactez moi si c'est possible, nous serions intéressé de travailler avec vous