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.
Problems with mesh definition
Posted 26 févr. 2014, 09:54 UTC−5 1 Reply
Please login with a confirmed email address before reporting spam
Hi,
I want to know if it is normal results change with the size of the mesh?
I am using RF module, and implemented one of Comsol examples the Branch-line Coupler, when I change the size of the mesh (Course, Normal, Fine, Finer, etc) the final results (computed values) keeps constant.
But when I try to implement a simple microstrip line, and change the size of the mesh the final results also change.
It's normal? If not, where's my mistake? How do I find the optimal size of my Mesh?
Greats,
Daniel Santos
I want to know if it is normal results change with the size of the mesh?
I am using RF module, and implemented one of Comsol examples the Branch-line Coupler, when I change the size of the mesh (Course, Normal, Fine, Finer, etc) the final results (computed values) keeps constant.
But when I try to implement a simple microstrip line, and change the size of the mesh the final results also change.
It's normal? If not, where's my mistake? How do I find the optimal size of my Mesh?
Greats,
Daniel Santos
1 Reply Last Post 27 févr. 2014, 02:17 UTC−5
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Yes, the result depends on the mesh size! But this is clear you must
resolve all features of the solution in the mesh.
The promise of the finite element method is that, if the longest edge
in the mesh tends to zero the FEM-solution tends to the solution of
the differntial equation. But this is only true for infinite precision arithmetics.
On a computer you have round-off errors and a finer mesh can be worse.
For my problems I check the solution for steep gradients and refine if necessary.
But for other cases different rules could be found.
Regards
Jens
resolve all features of the solution in the mesh.
The promise of the finite element method is that, if the longest edge
in the mesh tends to zero the FEM-solution tends to the solution of
the differntial equation. But this is only true for infinite precision arithmetics.
On a computer you have round-off errors and a finer mesh can be worse.
For my problems I check the solution for steep gradients and refine if necessary.
But for other cases different rules could be found.
Regards
Jens