Meshing two regions with (very) different densities.

Please login with a confirmed email address before reporting spam

Hi all,

I'm partitioning a square with side length 300 um into an inner and outer square (see attached model). Domain A is the (inner) square with side length 299.7 um, and domain B is the remainig part of the geometry.

I need to mesh domain A extremely finely (max. element size 20 nm). For that, I use the "Mapped" node, which yields a regular structure mesh with the desired element size.

I don't really care about the mesh in domain B (it can be super coarse).

My problem is that once domain A is (successfully) meshed, domain B doesn't mesh, pretty much regardless of the parameters (triangular mesh, quadrilateral mesh, structured mesh, different element sizes or growth rates) for that domain. Here, "doesn't mesh" means that the mesh either takes super long and COMSOL crashes, or I get error messages like "there are too many elements to mesh".

Do you have any suggestions that include keeping domain A as it is (i.e. meshed with e = 20 nm), but enabling the meshing of domain B (which can be very coarse) ?

The relevant mesh in the attached model is named "Mesh - 8 domains".



3 Replies Last Post Feb 19, 2024, 5:56 a.m. EST
Acculution ApS Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 2 months ago Feb 19, 2024, 4:10 a.m. EST

Looks like you are applying contraints on boundaries only, and that is what drives the generation of domain meshes. Look at tetrahedral instead of triangular, for example.

-------------------
René Christensen, PhD
Acculution ApS
www.acculution.com
info@acculution.com
Looks like you are applying contraints on boundaries only, and that is what drives the generation of domain meshes. Look at tetrahedral instead of triangular, for example.

Please login with a confirmed email address before reporting spam

Posted: 2 months ago Feb 19, 2024, 4:49 a.m. EST

Hi,

thanks for the swift response. It is true that I apply constraints on the boundary only, as my goal is to first mesh the surface (2D mesh) and then sweep along the third dimension.

Building a full 3D mesh always took longer in previous models.

Why would a 3D mesh be faster here? Thanks for your answer.

Hi, thanks for the swift response. It is true that I apply constraints on the boundary only, as my goal is to first mesh the surface (2D mesh) and then sweep along the third dimension. Building a full 3D mesh always took longer in previous models. Why would a 3D mesh be faster here? Thanks for your answer.

Acculution ApS Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 2 months ago Feb 19, 2024, 5:56 a.m. EST

Oh sorry, yes it is 3D. I don't have a quick fix, but I start out with perhaps a quarter of the geometry first, and see how rough the Mapping has to be for this to mesh properly, and take it from there.

-------------------
René Christensen, PhD
Acculution ApS
www.acculution.com
info@acculution.com
Oh sorry, yes it is 3D. I don't have a quick fix, but I start out with perhaps a quarter of the geometry first, and see how rough the Mapping has to be for this to mesh properly, and take it from there.

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.