CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (https://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [GAMBIT] projecting a curve on a face in Gambit (https://www.cfd-online.com/Forums/ansys-meshing/103912-projecting-curve-face-gambit.html)

mazdak June 28, 2012 19:51

projecting a curve on a face in Gambit
 
1 Attachment(s)
Hi.
I want create a circular area on a bend. I draw that circular area but when
i wanna to project that on the bend, for right half-part of that i deal with this error:
Projection of "edge. 27" on "face 1" can not create unique or correct edge.

i can not project edges 27 and 26 on "face 1" (face 1= red face).

edges 24 & 25 are easily projected, although they are similar to edges 27 and
26.

what should i do?

aim of this work is to create a circle with diameter of 25.4 mm on surface of a bend at teta=45 degree.

-mAx- June 29, 2012 02:31

Did you try to project edges 26 & 27 first (before edges 24 & 25)?

mazdak July 3, 2012 18:10

yes i have tried but no results

-mAx- July 4, 2012 00:47

ok
If your volume is still real you can try to split it with plane.
Create first plane (edge 24 & 25 belong to this plane)
Split desired surface with the created surface (plane).
If it is successfull, then reproduce it with the 2 other edges.
If not, you can provide me your dbs, I will take a look

mazdak July 4, 2012 14:55

thank u for ur considering my inquiry. my email address is
mazdak_parsi2000@yahoo.com
plz send me your email address to send u the file.
thanks

mazdak July 4, 2012 17:37

i tried to send you email in a couple of occasion but there is failure to sending.

-mAx- July 5, 2012 02:46

1 Attachment(s)
ok so I managed myself to split as you wanted.
Attachment 14227
Do following:
* project edges 26 & 27 on face1 with split option and with closest distance option
* project edges 24 & 25 on the other face with split option with closest distance option

Now your splits are done, but you have superposed surfaces.
Here is how I did: copy the rounded edge common to the 4 surfaces (the doted one on my picture) with vector (0 0 0)
Delete all the surfaces which should take place at pink or blue suface.
enable shaded mode on, and check if your geometry is hollowed (hollow at pink and blue surface)
Now you reconstruct both surfaces properly, by selecting edges (create face from wireframe and use the copied edge)
Again shaded mode on, and check if hollow is filled.

Do (to be sure):
*connect surfaces All
*connect edges All
*connect vertices All

Voila..
Then you only have surfaces and no volume.
So create the cap surfaces, and create the volume (stitch faces and select all surfaces)

for hexa meshing, sure without those 2 surfaces , it would be very easy to mesh your stuff. But if you need those surfaces, then you need to decompose your volume.
But first do your splits

mazdak July 5, 2012 03:44

max
i don't know how to appreciate your time and help.
1) is it possible for you to send me your dbs files.
2) now, can i generate structured mesh?

-mAx- July 5, 2012 04:06

1-I cannot since you are working with an older version than mine
2-you can try to mesh your volume with hexa on the fly, but it will fail due those 2 surfaces
you need to decompose the geometry

mazdak July 5, 2012 04:09

i have ansys 13

-mAx- July 5, 2012 04:13

i was talking about gambit: (i have 2.4.6)
I received a warning while opening your dbs file (>> older version)

mazdak July 5, 2012 04:20

oh yes. max i really appreciate your help. i hope i could in touch with u and use ur knowledge.

-mAx- July 6, 2012 02:29

5 Attachment(s)
ok very quick generated, here is one method for full hexa decomposition.
I did it only on one half geometry:
Attachment 14232Attachment 14233Attachment 14235
Attachment 14236Attachment 14237

mazdak July 6, 2012 14:44

if i want to find y+ (in the wall-adjacent cell) in bend part what should i do?
cuz sounds that you can not define a curve in r-teta system.

-mAx- July 9, 2012 01:45

Then you have to add a boundary layer, but the split angle between yellow and green volumes will affect the cell-skewness at those jonctions

mazdak July 9, 2012 01:54

no, imagine a regular bend without projection of that circular area.
when i do simulation of fluid flow in the bend i dont know how to find the amount of y+ in the bend part. cuz when i open the "plot" and select a wall of bend,
there is no r-teta coordinate. there are just X,Y,Z that i can put zero for two of them and put 1 for another one.

-mAx- July 9, 2012 02:00

if you select a wall, then you will have all info in that wall
But y+ is talking about distance from wall, so you have to pick info IN your domain

mazdak July 9, 2012 02:19

1 Attachment(s)
so u mean that i can select the curve that is next to the wall and see how much is the y+ along that? if yes how?

imagine that i want to find the pressure distribution along the edge 30 which is located at outer wall of the bend. how can i find that?

-mAx- July 9, 2012 02:27

if you want to PLOT y+ along a curve, I am pretty sure you can create a rake-line in fluent-postprocessing (there are several tool in post-processing like slice etc...)
Once you have the point-line or surface you wanted, you can get all info you need


All times are GMT -4. The time now is 23:51.