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.
Meshing a model with large dimension difference
Posted 29 oct. 2009, 05:22 UTC−4 2 Replies
Please login with a confirmed email address before reporting spam
Hi,
first of all I have to say that I am quite a beginner in modelling with COMSOL.
I am trying to mesh a model of a spring-operated brake. Due to the Maxwell stress tensor method I need a thin gap between the anchor plate and the housing of the coil. And here begins the problem.
I found some similar topics in this forum but I can't mesh the model. I learned that the problem is the large dimension difference and tried to define this gap as a subdomain and use a swept mesh but it didn't work.
Now I'm clueless about how to proceed.
Best regards,
Frederic
PS: I added a simplified model of the brake.
first of all I have to say that I am quite a beginner in modelling with COMSOL.
I am trying to mesh a model of a spring-operated brake. Due to the Maxwell stress tensor method I need a thin gap between the anchor plate and the housing of the coil. And here begins the problem.
I found some similar topics in this forum but I can't mesh the model. I learned that the problem is the large dimension difference and tried to define this gap as a subdomain and use a swept mesh but it didn't work.
Now I'm clueless about how to proceed.
Best regards,
Frederic
PS: I added a simplified model of the brake.
Attachments:
2 Replies Last Post 29 oct. 2009, 08:48 UTC−4
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
Hi
indeed there are several discussions about maxwell stress tensor, on one side and large mesh differences o the other side. Trying to mix both increases the fun !
I can only suggest to try to make several "boxes", cut your thin volumes such that you have interiour boundaries with the same "material" and perhaps subdivide the volume giving you trouble in several layers, as a "russian doll", there is an example somewhere on the forum of a small sphere inside a bigger one, if you add one or two intermediate spheres the system meshes automatically, if you have only the tiny one and the big external one it fails.
Hope this helps, in general meshing is a delicate issue when youhave complex volume with very different dimensions. I always now cut my volumes along symmetry axis, and such to separte "small" and larger volumes. This gives me long lists of domains, but its can be managed by creating groups and then by meshing item per items you finally speed up and get a better control.
All this is really easier in V4, hope it comes soon (even if I fear other surprises)
Good luck
Ivar
indeed there are several discussions about maxwell stress tensor, on one side and large mesh differences o the other side. Trying to mix both increases the fun !
I can only suggest to try to make several "boxes", cut your thin volumes such that you have interiour boundaries with the same "material" and perhaps subdivide the volume giving you trouble in several layers, as a "russian doll", there is an example somewhere on the forum of a small sphere inside a bigger one, if you add one or two intermediate spheres the system meshes automatically, if you have only the tiny one and the big external one it fails.
Hope this helps, in general meshing is a delicate issue when youhave complex volume with very different dimensions. I always now cut my volumes along symmetry axis, and such to separte "small" and larger volumes. This gives me long lists of domains, but its can be managed by creating groups and then by meshing item per items you finally speed up and get a better control.
All this is really easier in V4, hope it comes soon (even if I fear other surprises)
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
2 decades ago
Hi
indeed there are several discussions about maxwell stress tensor, on one side and large mesh differences o the other side. Trying to mix both increases the fun !
I can only suggest to try to make several "boxes", cut your thin volumes such that you have interiour boundaries with the same "material" and perhaps subdivide the volume giving you trouble in several layers, as a "russian doll", there is an example somewhere on the forum of a small sphere inside a bigger one, if you add one or two intermediate spheres the system meshes automatically, if you have only the tiny one and the big external one it fails.
Hope this helps, in general meshing is a delicate issue when youhave complex volume with very different dimensions. I always now cut my volumes along symmetry axis, and such to separte "small" and larger volumes. This gives me long lists of domains, but its can be managed by creating groups and then by meshing item per items you finally speed up and get a better control.
All this is really easier in V4, hope it comes soon (even if I fear other surprises)
Good luck
Ivar
indeed there are several discussions about maxwell stress tensor, on one side and large mesh differences o the other side. Trying to mix both increases the fun !
I can only suggest to try to make several "boxes", cut your thin volumes such that you have interiour boundaries with the same "material" and perhaps subdivide the volume giving you trouble in several layers, as a "russian doll", there is an example somewhere on the forum of a small sphere inside a bigger one, if you add one or two intermediate spheres the system meshes automatically, if you have only the tiny one and the big external one it fails.
Hope this helps, in general meshing is a delicate issue when youhave complex volume with very different dimensions. I always now cut my volumes along symmetry axis, and such to separte "small" and larger volumes. This gives me long lists of domains, but its can be managed by creating groups and then by meshing item per items you finally speed up and get a better control.
All this is really easier in V4, hope it comes soon (even if I fear other surprises)
Good luck
Ivar