meshing only the big domain covering the small one
Posted May 5, 2025, 2:05 a.m. EDT Mesh 5 Replies
Please login with a confirmed email address before reporting spam
There are two domains in a model . The smaller domain is a real subset of the bigger. In other words, the big domain entirely covers the small one. The two domains form a union. Is there any way to mesh the whole model so that the mesh only considers the big domain as if there were no small domain? In other words, there is no need to refine the mesh at the boundary between the two domains.
Please login with a confirmed email address before reporting spam
Maybe you could send a sketch of what you would like to achieve? If you do not need the smaller domain for meshing, the question is why do you include it at all? If you simply want to disregard the small domain, why do you need to get a union? An assembly would be just fine to mesh one or the other domains (or both of them) with completely separate meshes, that you can reconnect if needed with non-local coupling variables at the physics level). A sketch would help to understand your need. Take care, Eric Favre
-------------------Eric Favre
Renaissance Fusion
Please login with a confirmed email address before reporting spam
Is there any way to mesh the whole model so that the mesh only considers the big domain as if there were no small domain?
In Mesh there is 'Delete Entities' and in Geometry there is 'Virtual Operations' to remove all kind of entities.
With Mesh Plot and Filter or a Clip Plane you can see the inner result.
Nice Greetings
Please login with a confirmed email address before reporting spam
Maybe you could send a sketch of what you would like to achieve? If you do not need the smaller domain for meshing, the question is why do you include it at all? If you simply want to disregard the small domain, why do you need to get a union? An assembly would be just fine to mesh one or the other domains (or both of them) with completely separate meshes, that you can reconnect if needed with non-local coupling variables at the physics level). A sketch would help to understand your need. Take care, Eric Favre
Hi Eric thanks for your reply. It could be just a rectangle enclosed in a bigger rectangle but they have different material properties.
Please login with a confirmed email address before reporting spam
In Mesh there is 'Delete Entities' and in Geometry there is 'Virtual Operations'
Hi Sebastian, thanks so much. But I haven't found any simple online examples for me to understand how to use these two features for my purpose. Do you happen to have any links that can help me? Thank you.
Please login with a confirmed email address before reporting spam
You mention that there are two regions that have "different material properties". Typically, this would be represented by assigning different material specifications to those two different regions. And in Comsol, those would be called different "domains." Now, those two different domains (i.e., (1) the inner rectangle and (2) the space around it) are separated by a boundary. And that boundary becomes a boundary to the mesh, not just the geometry & material specs. So, it seems to me that you are presenting an unusual objective, if I understand you correctly, of not wanting to allow the mesh to conform to the boundary between your two different material/geometric domains. (In most physics problems of the type I work on, this should not be necessary, but you are likely doing something different.) Anyway, there is at least one way I know to accomplish this: You don't have to assign material properties on the basis of domains. Set the entire problem to being just one domain (i.e., do not even draw that inner rectangle at all). Second, specify the material properties of the (now) single domain to be spatially dependent by means of equations or Boolean conditions. E.g., a material property can be set to some value if abs(x)<2 but set to another value if abs(x)>=2. Anyway, I hope you get the idea. From a meshing perspective, there is then only one domain (but with spatially dependent properties) and it can be meshed by whatever means and rules you wish, without causing the mesh to conform to any "boundaries" where the material properties change (since you have eliminated those boundaries). I hope that helps. Bear in mind that you will have mesh elements that, in effect, span two different material properties. This may result in some ragged-looking plots and may or may not impact the quality or correctness or convergence of your calculations.
-------------------Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
Reply
Please read the discussion forum rules before posting.
Please log in to post a reply.