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 The information provided may be out of date.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.

Mesh Size in Mapped mesh

Please login with a confirmed email address before reporting spam

Hi,

Does any one knows what is the maximum number of element comsol 4.2 can use on an edge. I am trying to carry out the mesh independent analysis but when I go beyound a certain number of element comsol always gives error (cannot find a solution). The edge is only 0.5mm and if I use any thing more than 50 elements on this edge the Comsol gives error. Is there any way to increse the limit in comsol 4.2.

Thanks

4 Replies Last Post 2 mai 2012, 16:16 UTC−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 avr. 2012, 13:53 UTC−4
Hi

from my knowledge there is no limit on the number of elements per edge (apart that it must be ersolved by the binary number representation, but normally one woul run out of RAM before hitting that limit ;)

I suspect rather that its related with a mesh size ration or some of the settings. Have you tried to play with the "custom" mesh Size settings ?

--
Good luck
Ivar
Hi from my knowledge there is no limit on the number of elements per edge (apart that it must be ersolved by the binary number representation, but normally one woul run out of RAM before hitting that limit ;) I suspect rather that its related with a mesh size ration or some of the settings. Have you tried to play with the "custom" mesh Size settings ? -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2 mai 2012, 00:08 UTC−4
Hi Dr. Ivar

Thank u very much for the reply. Actually my domain is an axisymmetric annulus with square ribs on inner surface hence it can be modeled in 2D. With little modifications with the geometry we can use mapped mesh with quadriletral elements which is rather a good choice for simple geometries without holes or curvatures. If we restrict the number of elements on an edge by applying distribution node then we cannot use custom mesh sizes. Apart from this problem some times the simulation takes huge time to reach a moderate convergence tolerance of 1X10^-4 if quadriletral mapped mesh is used. Cannot really figure out whats going wrong. Is there any way to increase the convergence time for the simulations in comsol 4.2?
Hi Dr. Ivar Thank u very much for the reply. Actually my domain is an axisymmetric annulus with square ribs on inner surface hence it can be modeled in 2D. With little modifications with the geometry we can use mapped mesh with quadriletral elements which is rather a good choice for simple geometries without holes or curvatures. If we restrict the number of elements on an edge by applying distribution node then we cannot use custom mesh sizes. Apart from this problem some times the simulation takes huge time to reach a moderate convergence tolerance of 1X10^-4 if quadriletral mapped mesh is used. Cannot really figure out whats going wrong. Is there any way to increase the convergence time for the simulations in comsol 4.2?

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 2 mai 2012, 06:36 UTC−4
Hi

there are many buttons one might tweak in the solvers nodes, but it's not easy to say you should this or that one. Take a closer look (and it the doc), try it outon simpler models

--
Good luck
Ivar
Hi there are many buttons one might tweak in the solvers nodes, but it's not easy to say you should this or that one. Take a closer look (and it the doc), try it outon simpler models -- Good luck Ivar

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2 mai 2012, 16:16 UTC−4
Usually a finer mesh is more accurate and converges better. However, if the model has singularities a very fine mesh may work against you. This applies for example if you have point loads in a structural problem. Also, having 50 elements along an edge is not that much.

Nagi Elabbasi
Veryst Engineering
Usually a finer mesh is more accurate and converges better. However, if the model has singularities a very fine mesh may work against you. This applies for example if you have point loads in a structural problem. Also, having 50 elements along an edge is not that much. Nagi Elabbasi Veryst Engineering

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.