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.

Problem in meshing small structures

Please login with a confirmed email address before reporting spam

Hi,

I have waveguide whose length is 70 um, textured with periodic gratings (square pattern) having height gradually varying from one amstrong to one nanometer. Apparently, this is causing problems for the mesh generator and it always gives me the error in advancing front mesher. Any tips on how to get around this problem?

Ideally, I would like a way out in which mesh density is increased only around the grating and not the whole computational field of view so as to keep the computational burden from unnecessarily increasing.

Thanks

2 Replies Last Post 03.05.2010, 05:33 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 30.04.2010, 14:47 GMT-4
Hi

first with the new release v4 we sould start to remind in which version we are working, I assume 3.5 is your case, then I can suggest to check a few things:

1) the relative resolution for geometry import (if you are importing your volume) this is by default set to 10um or 10E-5, any feature smaller will dissapear, so when importing a small geometry, always check the file open "OPTIONS" sub-tab and adapt the repair tolerance to the smalles dimension of your geometry
2) normally I use as rule of thumb to not exceed a 1:40'000 in mesh size reatio for the same subdomain, in your case it seems to be more than the double. This leads mostly to irregular shaped elements, and hence easily inverted elements durng solving, particularly if you have deformations involved.

One way around is to surround the smlles features by some "boxes" so that for each subdomain / boundary you can respect the 1:10-40'000 rule. Or better, when you have such large differences in size, one can often reduce the model by using equations for the physical settings on some boundaries with very small features, but this require often quite some equation writing. Still an interesting and often a good model simplification approach.

Have fun Comsoling
Ivar
Hi first with the new release v4 we sould start to remind in which version we are working, I assume 3.5 is your case, then I can suggest to check a few things: 1) the relative resolution for geometry import (if you are importing your volume) this is by default set to 10um or 10E-5, any feature smaller will dissapear, so when importing a small geometry, always check the file open "OPTIONS" sub-tab and adapt the repair tolerance to the smalles dimension of your geometry 2) normally I use as rule of thumb to not exceed a 1:40'000 in mesh size reatio for the same subdomain, in your case it seems to be more than the double. This leads mostly to irregular shaped elements, and hence easily inverted elements durng solving, particularly if you have deformations involved. One way around is to surround the smlles features by some "boxes" so that for each subdomain / boundary you can respect the 1:10-40'000 rule. Or better, when you have such large differences in size, one can often reduce the model by using equations for the physical settings on some boundaries with very small features, but this require often quite some equation writing. Still an interesting and often a good model simplification approach. Have fun Comsoling Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 03.05.2010, 05:33 GMT-4
Thanks Ivar for the response, I cannot appreciate more.
Thanks Ivar for the response, I cannot appreciate more.

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.