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
18 juil. 2012, 01:13 UTC−4
Hi
first of all you can always use the default mesh that will work.
For sweep mesh you have some limitations, you cannot sweep over topology changes, what you can do is
1) remove or disable the specific union 1 node in your geometry sequence, then
2) to mesh the internal bloc side i.e. boundaries 25 26 27 (once the union has been removed) with tri or quad and then sweep your domains 2, 3, 4 ,5
Then you need to "convert" all external meshed boundaries of your blocks 2,3,4,5 to tri with a "convert" and add a final Free thet mesh for domain 1
--
Good luck
Ivar
Hi
first of all you can always use the default mesh that will work.
For sweep mesh you have some limitations, you cannot sweep over topology changes, what you can do is
1) remove or disable the specific union 1 node in your geometry sequence, then
2) to mesh the internal bloc side i.e. boundaries 25 26 27 (once the union has been removed) with tri or quad and then sweep your domains 2, 3, 4 ,5
Then you need to "convert" all external meshed boundaries of your blocks 2,3,4,5 to tri with a "convert" and add a final Free thet mesh for domain 1
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
18 juil. 2012, 12:52 UTC−4
Thanks Ivar. I removed the union and switched to the default mesh (I used the physics defined mesh option).
I next wanted to sweep the voltage on the beam from 1 to 5V and added that to the study. When I ran the simulation, I got the following error:
Failed to find a solution for the initial parameter.
In segregated group 3:
Very ill-conditioned preconditioner.
The relative residual is more than 1000 times larger than the relative tolerance.
Returned solution is not converged.
I'm not quite sure what this means or where to go from here.
Thanks Ivar. I removed the union and switched to the default mesh (I used the physics defined mesh option).
I next wanted to sweep the voltage on the beam from 1 to 5V and added that to the study. When I ran the simulation, I got the following error:
Failed to find a solution for the initial parameter.
In segregated group 3:
Very ill-conditioned preconditioner.
The relative residual is more than 1000 times larger than the relative tolerance.
Returned solution is not converged.
I'm not quite sure what this means or where to go from here.
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
19 juil. 2012, 01:25 UTC−4
Hi
1st you should patch your v4.3, check on the main website
You must check your model, several nodes are not defined, or do not have logical (in my sense) settings. This could come from the fact that when you removed the union, it changed the topology, therefore you must recheck your entity definitions. This is particularly important if you have not defined Selection sets from your different domain and BC lists
--
Good luck
Ivar
Hi
1st you should patch your v4.3, check on the main website
You must check your model, several nodes are not defined, or do not have logical (in my sense) settings. This could come from the fact that when you removed the union, it changed the topology, therefore you must recheck your entity definitions. This is particularly important if you have not defined Selection sets from your different domain and BC lists
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
9 nov. 2015, 03:38 UTC−5
Hi Mr Ivar
I am simulating a electromechanics model with a beam covered by an air domain. I want to define the mesh by myself and I meet the same problem as presented one in this post.
I follow your steps :
1. add Convert node in Mesh, chose all boundaries of domain 2, 3, 4 ,5
2. add Free Triangular node, build it with "Remain" option.
Is it correct ?
Hi
first of all you can always use the default mesh that will work.
For sweep mesh you have some limitations, you cannot sweep over topology changes, what you can do is
1) remove or disable the specific union 1 node in your geometry sequence, then
2) to mesh the internal bloc side i.e. boundaries 25 26 27 (once the union has been removed) with tri or quad and then sweep your domains 2, 3, 4 ,5
Then you need to "convert" all external meshed boundaries of your blocks 2,3,4,5 to tri with a "convert" and add a final Free thet mesh for domain 1
--
Good luck
Ivar
Hi Mr Ivar
I am simulating a electromechanics model with a beam covered by an air domain. I want to define the mesh by myself and I meet the same problem as presented one in this post.
I follow your steps :
1. add Convert node in Mesh, chose all boundaries of domain 2, 3, 4 ,5
2. add Free Triangular node, build it with "Remain" option.
Is it correct ?
[QUOTE]
Hi
first of all you can always use the default mesh that will work.
For sweep mesh you have some limitations, you cannot sweep over topology changes, what you can do is
1) remove or disable the specific union 1 node in your geometry sequence, then
2) to mesh the internal bloc side i.e. boundaries 25 26 27 (once the union has been removed) with tri or quad and then sweep your domains 2, 3, 4 ,5
Then you need to "convert" all external meshed boundaries of your blocks 2,3,4,5 to tri with a "convert" and add a final Free thet mesh for domain 1
--
Good luck
Ivar
[/QUOTE]
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:
9 years ago
10 nov. 2015, 04:40 UTC−5
Hi
if you want to Sweep mesh through a set of domains, you need to add internal boundaries to make the geometry look like a cubic drawing. In your case it's rather simple, on Block4 add two Layers (only to the LEFT) of respectively 8 and then 12 um, this cut the last block in plane with the internal items. In this way you can mesh the full cut surface (chose at 8 um from the left inside the block, and then sweep out
That should make the trick.
With sweep mesh the initial surface you have meshed need to be swept WITHOUT change all through the domain.
I'm using triangular mesh, it's simpler and faster, if you want Quads you must work a bit more in manually defining the distributions along the edges
--
Good luck
Ivar
Hi
if you want to Sweep mesh through a set of domains, you need to add internal boundaries to make the geometry look like a cubic drawing. In your case it's rather simple, on Block4 add two Layers (only to the LEFT) of respectively 8 and then 12 um, this cut the last block in plane with the internal items. In this way you can mesh the full cut surface (chose at 8 um from the left inside the block, and then sweep out
That should make the trick.
With sweep mesh the initial surface you have meshed need to be swept WITHOUT change all through the domain.
I'm using triangular mesh, it's simpler and faster, if you want Quads you must work a bit more in manually defining the distributions along the edges
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
11 nov. 2015, 05:08 UTC−5
Dear Mr Ivar
Thank you very much. I see your idea.
Dear Mr Ivar
Thank you very much. I see your idea.