Menu

Discussion Forum

New Discussion
Filter

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 The information provided may be out of date.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.

About the upside, downside of an interior boundary, and others related

Posted Feb 6, 2011, 9:41 p.m. EST Geometry, Parameters, Variables, & Functions Version 4.0 5 Replies

Xiaoyi Yu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

details please see the attachment, because I draw a figure to help description, thank you!

Attachments:

5 Replies Last Post May 13, 2016, 5:03 a.m. EDT
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 7, 2011, 2:42 a.m. EST
Hi

To make it clear, I believe (without being 100% sure) that you are right.

But I can state that the direction, in particular of interiours boundaries, is not unique, it all depends on how you build your geometry. The best is to check with the new arrow direction plots of un and dn.

In assembly mode it's easier, as then you have the domain and you can point, on a boundary, outwards (up) and inwards w.r.t the domain. and the boundaries are dedoubled between the assembled domains. Which means that you have two sets of "up / down", that are mostly opposed.

Now I seldom need to us the up/down normals, and can 99% of the time rely on COMSOL to look after this, and that is done without problems

--
Good luck
Ivar
Hi To make it clear, I believe (without being 100% sure) that you are right. But I can state that the direction, in particular of interiours boundaries, is not unique, it all depends on how you build your geometry. The best is to check with the new arrow direction plots of un and dn. In assembly mode it's easier, as then you have the domain and you can point, on a boundary, outwards (up) and inwards w.r.t the domain. and the boundaries are dedoubled between the assembled domains. Which means that you have two sets of "up / down", that are mostly opposed. Now I seldom need to us the up/down normals, and can 99% of the time rely on COMSOL to look after this, and that is done without problems -- Good luck Ivar
Xiaoyi Yu
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 7, 2011, 9:42 a.m. EST

Hi

To make it clear, I believe (without being 100% sure) that you are right.

But I can state that the direction, in particular of interiours boundaries, is not unique, it all depends on how you build your geometry. The best is to check with the new arrow direction plots of un and dn.

In assembly mode it's easier, as then you have the domain and you can point, on a boundary, outwards (up) and inwards w.r.t the domain. and the boundaries are dedoubled between the assembled domains. Which means that you have two sets of "up / down", that are mostly opposed.

Now I seldom need to us the up/down normals, and can 99% of the time rely on COMSOL to look after this, and that is done without problems

--
Good luck
Ivar


Thank you, Ivar
You said "The best is to check with the new arrow direction plots of un and dn", how to plot? My way is Right click Results --- 2D Plot Group, choose Arrow Surface, but I cannot find "un" or "dn" in the Repalce expression.
Thank you!
[QUOTE] Hi To make it clear, I believe (without being 100% sure) that you are right. But I can state that the direction, in particular of interiours boundaries, is not unique, it all depends on how you build your geometry. The best is to check with the new arrow direction plots of un and dn. In assembly mode it's easier, as then you have the domain and you can point, on a boundary, outwards (up) and inwards w.r.t the domain. and the boundaries are dedoubled between the assembled domains. Which means that you have two sets of "up / down", that are mostly opposed. Now I seldom need to us the up/down normals, and can 99% of the time rely on COMSOL to look after this, and that is done without problems -- Good luck Ivar [/QUOTE] Thank you, Ivar You said "The best is to check with the new arrow direction plots of un and dn", how to plot? My way is Right click Results --- 2D Plot Group, choose Arrow Surface, but I cannot find "un" or "dn" in the Repalce expression. Thank you!
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Feb 7, 2011, 10:03 a.m. EST
Hi

I do not have access to COMSOL just now, but I believe the normal vectors are not in the physics submenu but the coordinates ones, the names are slighly more complex than just "dn", my shortcut for "downwards normal"

--
Good luck
Ivar
Hi I do not have access to COMSOL just now, but I believe the normal vectors are not in the physics submenu but the coordinates ones, the names are slighly more complex than just "dn", my shortcut for "downwards normal" -- Good luck Ivar
Evgeni Sergeev
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 8 years ago May 3, 2016, 1:01 a.m. EDT
For those searching about up/down direction, the definitive part in the documentation is the Tangent and normal variables section in Comsol Reference Manual.

In 2D, go to Definitions > View 1 > Show edge direction arrows, then the up side is to the left of the boundary if the direction is taken as forward. Normals are usually outward normals everywhere.
For those searching about up/down direction, the definitive part in the documentation is the [b]Tangent and normal variables[/b] section in Comsol Reference Manual. In 2D, go to [i]Definitions > View 1 > Show edge direction arrows[/i], then the [i]up[/i] side is to the left of the boundary if the direction is taken as forward. Normals are usually outward normals everywhere.
Evgeni Sergeev
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 8 years ago May 13, 2016, 5:03 a.m. EDT
By the way, here's a way of not having to worry about which way is up and which is down, in case of getting accurate fluxes (dflux and uflux, typically only one of which is defined on an external boundary).

if(tx*hzeq.ny-ty*hzeq.nx > 0, dflux.Ve, uflux.Ve)

The variables tx and ty is the tangential vector. And the normal on an external boundary (with respect to a physics interface) is always directed outwards.

In the above "hzeq" is the tag of the physics interface, and "Ve" is the dependent variable. The conditional part is the perp product (cross product for vectors in the plane, omitting the zero entries).
By the way, here's a way of not having to worry about which way is up and which is down, in case of getting accurate fluxes (dflux and uflux, typically only one of which is defined on an external boundary). if(tx*hzeq.ny-ty*hzeq.nx > 0, dflux.Ve, uflux.Ve) The variables tx and ty is the tangential vector. And the normal on an external boundary (with respect to a physics interface) is always directed outwards. In the above "hzeq" is the tag of the physics interface, and "Ve" is the dependent variable. The conditional part is the perp product (cross product for vectors in the plane, omitting the zero entries).

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.

Suggested Content