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.

Rotating Mode Stirrer in Room excited by log-periodic antenna

Please login with a confirmed email address before reporting spam

Dear Sir/Madame,

I am looking into the effect of adding a mode stirrer to a room excited by a Log-periodic antenna on the electric field in the room. I would like to see how the electric field would change between two stirrer positions. However I am experiencing difficulties adding the mode stirrer and getting two solutions (I start with turning the mode stirrer 180 degrees).

I created the geometry of the mode stirrer by defining a workplane and specifying rectangles in this workplane which I then manipulate such that they form the form of the mode stirrer. This stirrer is then considered infinitely thin to ease the computational requirements. I then try to add boundary conditions to this geometry (PEC/Impedance) but they are not applicable. I added a screenshot for clarification.

If I use very thin rectangular blocks instead of the planes I can apply the boundary conditions however the mesh becomes incredibly fine because of the thinness of the blocks causing very large computation times which I want to avoid as much as possible.

Also I would like to model the electric field in the room for two different stirrer rotations. I have had no succes in finding a way to do this. The most similar problem I found was a microwave heating problem with a rotating dish on the bottom which used identity pairs and a rotating domain (https://www.comsol.com/forum/thread/234941/applying-moving-mesh-rotating-domain-in-microwave-heating). Though I dont think I can use this as my mode stirrer does not have overlapping faces that form identity pairs. The ugly solution would be to save the system in different files for every stirrer position and solve the models seperately. Though comparing the solutions would be more tiresome.

Any help with the boundary conditions and the rotation of the mode stirrer would be much appreciated. Thanks in advance, Gijs Mast



4 Replies Last Post 2020年6月5日 GMT-4 20:58
Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 4 years ago 2020年6月4日 GMT-4 15:10

From your description alone, I already know what is going wrong. Don't model the stirrer as infinitely thin and don't try to model (mesh) the interior volume of the stirrer material either. Instead, create a stirrer with a reasonable thickness (fairly thin is ok) and subtract (using boolean difference operation) the stirrer from the chamber. This will result in (default) perfect electric conductor (PEC) boundaries on the surfaces of your stirrer. You won't be meshing the interior volume of the stirrer at all, since that is no longer part of the computational space. Now the only finely-meshed part will be the edge-surfaces of the stirrer, which shouldn't be too serious. Now, if you want to consider loss on the stirrer's surfaces, you can then define its boundaries as impedance boundaries (instead of the default PEC), and you can also assign material properties (such as a conductivity value) to those surfaces, where you normally specify materials. I hope that makes sense.

-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
From your description alone, I already know what is going wrong. Don't model the stirrer as infinitely thin *and* don't try to model (mesh) the *interior* volume of the stirrer material either. Instead, create a stirrer with a reasonable thickness (fairly thin is ok) and subtract (using boolean difference operation) the stirrer from the chamber. This will result in (default) perfect electric conductor (PEC) boundaries on the surfaces of your stirrer. You won't be meshing the interior volume of the stirrer at all, since that is no longer part of the computational space. Now the only finely-meshed part will be the edge-surfaces of the stirrer, which shouldn't be too serious. Now, if you want to consider loss on the stirrer's surfaces, you can then define its boundaries as impedance boundaries (instead of the default PEC), and you can also assign material properties (such as a conductivity value) to those surfaces, where you normally specify materials. I hope that makes sense.

Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 4 years ago 2020年6月4日 GMT-4 15:14
Updated: 4 years ago 2020年6月4日 GMT-4 15:15

In regard to the mode stirrer rotation, you can add a rotation operation in your geometry specification, to rotate it by some angle parameter (e.g., a quantity like "mytheta"). Then you can do a parametric study where you cause the simulation to step through various values of mytheta (which should force the model to remesh and re-solve) at each value. If you only have a few angular positions of interest, you should probably due it all manually, since that way you can (and should) inspect the mesh for good quality along the way. But once you are sure it is meshing and working well, and if you have many angles of interest, a parametric approach would be appropriate.

-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
In regard to the mode stirrer rotation, you can add a rotation operation in your geometry specification, to rotate it by some angle parameter (e.g., a quantity like "mytheta"). Then you can do a parametric study where you cause the simulation to step through various values of mytheta (which should force the model to remesh and re-solve) at each value. If you only have a few angular positions of interest, you should probably due it all manually, since that way you can (and should) inspect the mesh for good quality along the way. But once you are sure it is meshing and working well, and if you have many angles of interest, a parametric approach would be appropriate.

Please login with a confirmed email address before reporting spam

Posted: 4 years ago 2020年6月5日 GMT-4 03:50

Dear Robert Koslover,

Thank you for your fast response. I indeed came to the conclusion that using the difference operator solved my problem, but good to know that using an infinitely thin mode stirrer will not work. The suggestion to use the parametric study helps a lot, thank you. Even though I probably will not use too many angles it will still improve my understanding of Comsol.

Dear Robert Koslover, Thank you for your fast response. I indeed came to the conclusion that using the difference operator solved my problem, but good to know that using an infinitely thin mode stirrer will not work. The suggestion to use the parametric study helps a lot, thank you. Even though I probably will not use too many angles it will still improve my understanding of Comsol.

Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 4 years ago 2020年6月5日 GMT-4 20:58
Updated: 4 years ago 2020年6月5日 GMT-4 21:00

You are welcome. I should add that, in principle, you can use an infinitely thin surface, but to set it as a PEC, you will need to create a separate PEC boundary condition for it. E.g., see the attached figure for how to do that. Personally, I prefer to represent most real-world 3D objects by 3D geometries, when working in a 3D model. But again, you don't necessarily have to do that.

-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
You are welcome. I should add that, in principle, you *can* use an infinitely thin surface, but to set it as a PEC, you will need to create a separate PEC boundary condition for it. E.g., see the attached figure for how to do that. Personally, I prefer to represent most real-world 3D objects by 3D geometries, when working in a 3D model. But again, you don't necessarily have to do that.

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.