Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

Mirror plane in 3D

Please login with a confirmed email address before reporting spam

I am working on calculating the E-field distribution for a rectangular 3D geometry. I require a very fine mesh and therefore I run into memory limitations of my computer. I therefore would like to take advantage of the symmetries of my geometry i.e. it is symmetric about the x-y plane. So if I could define a mirror plane at y=0, I would only need to calculate the field distribution for half the computational cell. Is this possible for a 3D structure in comsol?

I have looked a bit a 2D axial symmetry but the symmetry does not seem to be extruded from a line to a plane when extruding the geometry to 3D.

Cheers,

Claus

6 Replies Last Post Apr 16, 2014, 4:32 a.m. EDT
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 Oct 11, 2011, 8:41 a.m. EDT
Hi

in 3D you have symmetry planes (or antisymmetry), and in 2D too. In 2D axi, a point becomes a loop, and a line a surface in 3D, once revolved (2D-axi revolves by default the results into 3D in v4.2

try to use the right click on your physics (v4)

--
Good luck
Ivar
Hi in 3D you have symmetry planes (or antisymmetry), and in 2D too. In 2D axi, a point becomes a loop, and a line a surface in 3D, once revolved (2D-axi revolves by default the results into 3D in v4.2 try to use the right click on your physics (v4) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 11, 2011, 9:41 a.m. EDT
Thanks for your speedy reponse.
I am not sure I understand what you are saying. I should say I am much more familiar with v3.5a than v4.
If I have drawn a cube, how do I impose a symmetry plane down the center?
Thanks for your speedy reponse. I am not sure I understand what you are saying. I should say I am much more familiar with v3.5a than v4. If I have drawn a cube, how do I impose a symmetry plane down the center?

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 Oct 11, 2011, 10:15 a.m. EDT
Hi

you need an external boundary to set a symmetry condition,

that is you must cut your full domain along the symmetry plane and delete 1/2 of the the geometry (delete entity so it's reversible), then you can apply a symmetry boundary condition to this plane

--
Good luck
Ivar
Hi you need an external boundary to set a symmetry condition, that is you must cut your full domain along the symmetry plane and delete 1/2 of the the geometry (delete entity so it's reversible), then you can apply a symmetry boundary condition to this plane -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 11, 2011, 11:43 a.m. EDT
In v3.5a I have cut the geometry in two. I saw an micro oven model file where they do a symmetry cut. For the symmetry BCs they use perfect magnetic conductors when applying TE waves. I should mention that I employ hybrid mode waves so that does not apply in my case.

I guess a mirror plane can be represented by periodic BC's with anti-periodicity. If I do that I get an error 5043: "Only one group of parellel boundaries could be found". I guess the periodic BCs come in pairs. Any suggestions?



In v3.5a I have cut the geometry in two. I saw an micro oven model file where they do a symmetry cut. For the symmetry BCs they use perfect magnetic conductors when applying TE waves. I should mention that I employ hybrid mode waves so that does not apply in my case. I guess a mirror plane can be represented by periodic BC's with anti-periodicity. If I do that I get an error 5043: "Only one group of parellel boundaries could be found". I guess the periodic BCs come in pairs. Any suggestions?

Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Oct 13, 2011, 4:25 p.m. EDT
My thoughts:

A plane of "mirror symmetry" in an RF problem is generally either fully-equivalent to a PEC or a PMC, depending on the details of the problem. That said, I wonder if you may be, at least in part, confusing symmetry in geometry with symmetry in the EM fields. You must have symmetry in the fields themselves in order to employ a mirror symmetry condition.

Also, I do not know of any case where periodic boundary conditions would be appropriate to handle mirror symmetries.

My thoughts: A plane of "mirror symmetry" in an RF problem is generally either fully-equivalent to a PEC or a PMC, depending on the details of the problem. That said, I wonder if you may be, at least in part, confusing symmetry in geometry with symmetry in the EM fields. You must have symmetry in the fields themselves in order to employ a mirror symmetry condition. Also, I do not know of any case where periodic boundary conditions would be appropriate to handle mirror symmetries.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 16, 2014, 4:32 a.m. EDT
Hi,
I have been looking through the problem of applying symmetry to the waveguide. (across the height of the waveguide, z-axis) to reduce the computation time . I came across PEC and PMC BCs. I try to apply PMC boundary condition on the top of the waveguide (i draw it one half of its original height) . But i cannot get the electric fields to simulate at resonant frequencies( i can see the field strength in the legend though at certain frequencies but cannot plot the Electric fields, they appear not to be travelling visually through the waveguide)). Also, if i do apply symmetry, do i need to half the power input to get the same results (Electric field strength). Would appreciate your response.
Hi, I have been looking through the problem of applying symmetry to the waveguide. (across the height of the waveguide, z-axis) to reduce the computation time . I came across PEC and PMC BCs. I try to apply PMC boundary condition on the top of the waveguide (i draw it one half of its original height) . But i cannot get the electric fields to simulate at resonant frequencies( i can see the field strength in the legend though at certain frequencies but cannot plot the Electric fields, they appear not to be travelling visually through the waveguide)). Also, if i do apply symmetry, do i need to half the power input to get the same results (Electric field strength). Would appreciate your response.

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.