mesh of a 3D bump

92 views
Skip to first unread message

Giulia Deolmi

unread,
Oct 2, 2014, 9:43:15 AM10/2/14
to dea...@googlegroups.com
Dear deal.ii User Group,

I would like to implement a mesh corresponding to a 3D bump (the domain I have in mind is sketched in the attachment... it is the complementary of the bump in the parallelepiped...)
Is there someone who knows how I can do it? Can you please help me?

Thanks a lot in advance!
Best wishes

Giulia
domain.png

Bärbel Janssen

unread,
Oct 2, 2014, 10:06:24 AM10/2/14
to dea...@googlegroups.com
Hi Giulia,

I would write a geometry description in a file and then read this file in your deal code with the GrindIn class. An easy example for auch a file is given here:
http://numerik.uni-hd.de/~lehre/SS10/dealcompact/ucd_format.pdf

Make sure that you "color" your curved boundaries. You have to tell deal that one of you boundaries is part of a sphere. This is described in the manual for boundary descriptions
http://www.dealii.org/developer/doxygen/deal.II/group__boundary.html
If you did this correctly. new points will be moved to the boundary if you refine your cells.

If you need further help, let me know.

Good luck,
Bärbel
 

--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dealii+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Jean-Paul Pelteret

unread,
Oct 2, 2014, 10:10:48 AM10/2/14
to dea...@googlegroups.com
Dear Giulia,

You haven't mentioned whether you are wanting to produce the mesh using the deal.II grid generator or an external mesher. As an alternative to Barbel's suggestion, the basic strategy that I would use to mesh this domain would be break up the geometry into 5 subregions by projecting the top 4 corners of the domain down onto the upper region of the bump. This picture, which might represent a cross-section of your domain (mirrored around the lower plane), conveniently illustrates the point: http://montjoie.gforge.inria.fr/img/MaillageSphereCube.png

Kind regards,
Jean-Paul

Giulia Deolmi

unread,
Oct 3, 2014, 6:14:08 AM10/3/14
to dea...@googlegroups.com
Hi Baerbel! thanks a lot! I will have a look at it! I hope you are doing fine :)
Thanks Jean-Paul, if it's possible i would prefer to use the deal.ii's grid generator... Do you think it's possible? for example, if I generate a parallelepiped and half sphere, can I tell deal that the domain is the parallelepiped minus the sphere? is the grid of the link produced with deal.ii?
Thanks a lot!
have a nice day
Giulia

--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to a topic in the Google Groups "deal.II User Group" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dealii/YnODAhMMgs0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dealii+un...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.



--
  Giulia Deolmi

  RWTH - Aachen 
  Institut fuer Geometrie und Praktische Mathematik
  Templergraben 55, 52056 Aachen - Germany

Wolfgang Bangerth

unread,
Oct 10, 2014, 8:51:48 PM10/10/14
to dea...@googlegroups.com
On 10/03/2014 05:13 AM, Giulia Deolmi wrote:
> Hi Baerbel! thanks a lot! I will have a look at it! I hope you are doing fine :)
> Thanks Jean-Paul, if it's possible i would prefer to use the deal.ii's grid
> generator... Do you think it's possible? for example, if I generate a
> parallelepiped and half sphere, can I tell deal that the domain is the
> parallelepiped minus the sphere? is the grid of the link produced with deal.ii?

Not, it's not that simple. You can piece together meshes in deal.II, but you
can't just describe a geometry and let deal.II find a mesh for this geometry.
For this you really need a mesh generator, which we don't have.

Did you look at step-49?

Best
W.


--
------------------------------------------------------------------------
Wolfgang Bangerth email: bang...@math.tamu.edu
www: http://www.math.tamu.edu/~bangerth/

Jean-Paul Pelteret

unread,
Oct 13, 2014, 10:27:15 AM10/13/14
to dea...@googlegroups.com
Dear Giulia,

I'm not sure how far you got on this, but I've attached for you a more or less complete example of how one might use deal.II's built in grid generator to produce the geometry you have described. Basically one has to piece-wise construct the geometry from the primitives that deal.II provides and then use a hyperball refinement boundary to produce the curved region of the domain. If you run this example (based off of step-49) you will find that there's still a minor issue in projecting some of the vertices to the curved boundary - I can't recall right now how to fix this little issue so I'll leave it to you or someone else to figure out. Once you've figured that out, you'll likely need to smooth the mesh somehow to reduce the degree of distortion of some cells with vertices on the curved boundary (look at cutplanes through the origin with a normal of (1,1,0) and (1,-1,0) ).

Kind regards,
J-P
3d_bump_geo.cc
CMakeLists.txt

Giulia Deolmi

unread,
Oct 14, 2014, 8:37:59 AM10/14/14
to dea...@googlegroups.com
Thanks a lot! :)
I'm gonna have a look at it right now
have a nice day
Giulia
Reply all
Reply to author
Forward
0 new messages