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 meshing a high aspect ratio geometry

Please login with a confirmed email address before reporting spam

Hello,

I am working on a model of electric field inside a detector for an experiment I am working on. The model is contained in a cylinder ~1 m tall and the smallest parts in the detector are 2 wire grids; one with wire diameter 100 um and one with wire diameter 200 um. I have created a meshing sequence and managed to mesh everything but the grid with 100 um wire diameter. If I decrease the number of wires in a grid down to 10 (instead of 100), COMSOL manages to mesh it. Any suggestions of how to mesh the entire model please?

The model is available here: drive.google.com/file/d/0B_rFjLlSlYF-Nm45Q21NWlFHa0k/view?usp=sharing

Thank you for suggestions,

Lucie

4 Replies Last Post 24 mars 2015, 18:51 UTC−4
Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 23 mars 2015, 14:25 UTC−4
Lucie,

my first question is: do you really need to resolve the grids or could you replace them by potential planes. If you really need to resolve the grid wires you could try:

- Put structured meshes to the wires first and then go ahead
- You might consider replacing the explicite wire geometry by lines, i.e. infinitely thin wires.

Meshing this will be very expensive in terms of memory and computation time. You might consider to look at the wire grids in a seperate model on a smaller scale.

Cheers
Edgar

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Lucie, my first question is: do you really need to resolve the grids or could you replace them by potential planes. If you really need to resolve the grid wires you could try: - Put structured meshes to the wires first and then go ahead - You might consider replacing the explicite wire geometry by lines, i.e. infinitely thin wires. Meshing this will be very expensive in terms of memory and computation time. You might consider to look at the wire grids in a seperate model on a smaller scale. Cheers Edgar -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 23 mars 2015, 16:21 UTC−4
Dear Edgar,

Thank you for your suggestions. I have replaced all the grid for which it was possible by planes already. I would like to know how the grids influence the field inside the detector so solving for the grids in a different model won't really help. I have already tried replacing the wires by lines, but unfortunately the shape of the field seems to be too different with infinitely thin lines compared to the wires.

What do you mean by "putting a structured meshes to the wires first"?

Thank you,

Lucie
Dear Edgar, Thank you for your suggestions. I have replaced all the grid for which it was possible by planes already. I would like to know how the grids influence the field inside the detector so solving for the grids in a different model won't really help. I have already tried replacing the wires by lines, but unfortunately the shape of the field seems to be too different with infinitely thin lines compared to the wires. What do you mean by "putting a structured meshes to the wires first"? Thank you, Lucie

Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 24 mars 2015, 05:48 UTC−4
Lucie,

here is an example of a meshing sequence with a mapped mesh on the wire surface. There are many more option to generate controlled, structured meshes instead of using the free meshing. This is often needed in cases like yours.
In the meshing sequence these controlled steps on small entities should be the first steps to allow the free mesher that you need at a certain point can connect the free mesh to the small structures.
You may have to spend some time to play with all the available options.
I recommend to try that in a smaller setup but still containing the essential features.

Cheers
Edgar

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Lucie, here is an example of a meshing sequence with a mapped mesh on the wire surface. There are many more option to generate controlled, structured meshes instead of using the free meshing. This is often needed in cases like yours. In the meshing sequence these controlled steps on small entities should be the first steps to allow the free mesher that you need at a certain point can connect the free mesh to the small structures. You may have to spend some time to play with all the available options. I recommend to try that in a smaller setup but still containing the essential features. Cheers Edgar -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com


Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 24 mars 2015, 18:51 UTC−4
------citation start
Dear Edgar,

Thank you very much for your suggestion. However even in the model you sent me only the boundaries get meshed and when I try to mesh the domain outside the wire (Domain 1) using the Free Tetrahedral Mesh I get an error saying that the "Boundary mesh must consist of triangular elements only." How do I mesh the domains in the model using your method please?

Thank you for your answer,

Lucie
------citation end

Lucie,

you can add a conversion node to allow connecting a tetrahedral mesh. Please keep the subject in the forum and don't send pm, so you have a better chance to receive more contributions.
And as I said, explore all meshing options.

Cheers
Edgar

--
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
------citation start Dear Edgar, Thank you very much for your suggestion. However even in the model you sent me only the boundaries get meshed and when I try to mesh the domain outside the wire (Domain 1) using the Free Tetrahedral Mesh I get an error saying that the "Boundary mesh must consist of triangular elements only." How do I mesh the domains in the model using your method please? Thank you for your answer, Lucie ------citation end Lucie, you can add a conversion node to allow connecting a tetrahedral mesh. Please keep the subject in the forum and don't send pm, so you have a better chance to receive more contributions. And as I said, explore all meshing options. Cheers Edgar -- Edgar J. Kaiser emPhys Physical Technology http://www.emphys.com

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.