CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [GAMBIT] Gambit mesh around buildings (http://www.cfd-online.com/Forums/ansys-meshing/73329-gambit-mesh-around-buildings.html)

rcerdeira March 4, 2010 14:13

Gambit mesh around buildings
 
1 Attachment(s)
Hi!

I am having some issues to built a mesh around buildings for a street canyon problem. I am following “Gambit tutorial 03 - simulation of wind around buildings”, but I am having some problems to mesh the volume just above the small domain with Cooper method.

I can mesh the small domain, but the top of the buildings are without a mesh and I don't know if this cause the problem. It simple doesn't create the mesh above the small domain... Cananybody help me?

Thank you in advance

-mAx- March 5, 2010 02:19

Try to split your domain along (xy) with y= top of buildings.
If you still have problem, split along (xz)
Meshing complex domain on the fly without any volume decomposition will fail most of time

rcerdeira March 5, 2010 07:25

Gambit mesh around buildings
 
1 Attachment(s)
Thank you for your quick answer!
I have already my domain split, as in the image, but the buildings have different heights and the “small domain” is a little higher than the highest building... I am doing exactly what is told in the Gambit tutorial with some adaptations for my geometry. And worst of all, I am new at this...
Thank you once more

-mAx- March 5, 2010 07:45

Are you also giving the top of each building as source in the volume mesh panel (cooper mesh)?

rcerdeira March 5, 2010 10:32

1 Attachment(s)
Yes... The top of the buildings appear automatically in the face list of sources when I choose the small domain as the volume.

-mAx- March 5, 2010 11:00

are they meshed?
Then either you switch to tet-hexcore, or you want to work with "hexa", but in this case you need to split the domain at each building's height. You will have volume layers, but you should mesh them with cooper.
Else you can provide me your dbs file, and I will take a look at it on monday.

rcerdeira March 5, 2010 11:21

Could you do that really? So nice of you! I want to send you the file but it is too big to go as an attachment through here... Is there any other way? Anyway I will keep on trying, if I manage to make it I will tell you!

Thank you so much for your effort!

rcerdeira March 5, 2010 11:44

Following your advice I think the problem is solved :)! I have changed to tet/hybrid elements with hex core (native) type and the mesh is ok! But on the pave I have quad and on the area above the small domain I have hex with cooper, do you think that can bring any problem?

Thank you so much!!!

-mAx- March 5, 2010 15:31

no problem with mixing hexa-tetra-pyramid-wedge.
Just check the skewness (over 0.95 you can have convergence problem, try to keep it under 0.9)
If you still want to send the file, upload it on a share server like rapidshare

rcerdeira March 8, 2010 05:40

Thank you very much for your help! For now I think it is not necessary to send you the file! But I am almost sure that I will post some more doubts further in time :p...

rcerdeira March 11, 2010 08:24

Here I am again… Well, I have a small domain inside the domain. I have the small domain meshed and I am trying to mesh the pave of the domain. The idea is to have a less denser mesh. And that is the problem. I have made with size function, without, with different growth rate, different size limit and the mesh have always the same size, the size of the pave of the small domain... What can I do? Thank you

-mAx- March 11, 2010 09:34

size function with cooper, won't have a big effect on the volume, because this is just projection of sources along mappable surfaces (so the source faces will be projected (eg: the same), and consequently you cannot have refinement. At the other side, mappable surfaces can only contain mesh grading... but it is not refinement).
If you re-do what you did with tet-hexcore, you will see the difference

rcerdeira March 11, 2010 13:09

I have already done with tet-hexcore, but the problem is that I need a denser mesh near the buildings I think that with tet-hexcore it is not possible, is it? It doesn’t seem like.

-mAx- March 11, 2010 14:34

I don't understand what you want. The logic wants that your refine the mesh around the buildings. :confused:
post a picture or hand-sketch which explains what you want

rcerdeira March 12, 2010 05:31

1 Attachment(s)
Thats what I want, a refined mesh around the buildings. Is it possible with tet-hexcore mesh? Thanks again

-mAx- March 12, 2010 05:52

oh sorry I understood coarser instead of denser...:rolleyes:
Sure it is possible with size function.
Mesh building faces as your wish, then create a size function (type meshed), and let grow the mesh (for instance with factor 1.2) till maximal cell size you want.

jackshao March 27, 2010 07:24

Quote:

Originally Posted by rcerdeira (Post 248609)
Hi!

I am having some issues to built a mesh around buildings for a street canyon problem. I am following “Gambit tutorial 03 - simulation of wind around buildings”, but I am having some problems to mesh the volume just above the small domain with Cooper method.

I can mesh the small domain, but the top of the buildings are without a mesh and I don't know if this cause the problem. It simple doesn't create the mesh above the small domain... Cananybody help me?

Thank you in advance

where can i find “Gambit tutorial 03 - simulation of wind around buildings”? i haven't found this chapter in Gambit 2.2 tutorial.

rcerdeira March 29, 2010 04:40

I can send you, but I think that this is possible only by e-mail. Though the forum they just allow 97.7KB. I found this in another forum, but I don't remember which...

jackshao April 5, 2010 08:20

Rcerdeira,thank you very much! My email box is yunzhonghehit@126.com. Hope for your mail.

teguhtf April 16, 2010 05:34

Hey Guys,,
I have the same problem.. Can you send me “Gambit tutorial 03 - simulation of wind around buildings”? here is my email huget_teguh@ymail.com
Thanks
Teguh


All times are GMT -4. The time now is 17:34.