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] 3d wing from 2d airfoil-volume sweep (https://www.cfd-online.com/Forums/ansys-meshing/78548-3d-wing-2d-airfoil-volume-sweep.html)

ibzyuk July 25, 2010 16:13

3d wing from 2d airfoil-volume sweep
 
Hi all,

I am trying to create a 3D wing in GAMBIT. I am using a 2D airfoil in a Cgrid. When i use Form Volume -Sweep Face, with path vector. I get an error :

ACIS error 95034 :Attempt to sweep a non-manifold face.
Sweeping or revolving face circl1 failed.

The two rectangle faces are made to 3D but the circle face with the airfoil fails.

Could anyone guide me how to resolve this please. Could it be due to the two airfoil edges ending at the same point (trailing edge ending at one point)?

Thanks in advance.

-mAx- July 26, 2010 02:15

you can try a workaround: copy your face with the desired vector, and then create (wireframe) the lateral surfaces.
And finally create (wireframe) the volume

ibzyuk July 26, 2010 03:56

Thanks Max for that , I am trying it out.

The semi-circle face with the airfoil , I have copied it with the required vector.

when you say create wireframe the lateral surfaces. Do I create the edges first to join the two semi-circle faces together then make them into a new surface?

-mAx- July 26, 2010 04:00

yes you need to "close" the lateral surfaces by joining your 2 semi-circles.

ibzyuk July 27, 2010 19:59

Hi Max,

I managed to create the 3D volume of the semi-circle. Many thanks!

I have meshed both sides of the 3D cgrid and have created the volume mesh for the two cubiod parts.

When I try to create a volume mesh for the semi-circle I am getting the below error:

"volume circle3d cannot be meshed because two sets of source faces which form natural starts for the projections could not be found"

I am trying to do a hex mesh type Cooper and giving the two semi-circles as the sources.

Do you know what is causing this and what I can do to fix this?

-mAx- July 28, 2010 06:35

a picture may help to understand... :)
you can try to split the volume into 2 simpliest volumes

ibzyuk July 28, 2010 06:55

sorry about that Max :D, I did think of adding my Gambit files , but it was 17Mb. Is there a place I can upload it. Or would a tiff image or even generate mesh do.
I will not be able to do send it untill later on this evening.

-mAx- July 28, 2010 07:49

just a picture of the volume giving you trouble

ibzyuk July 28, 2010 17:56

1 Attachment(s)
Attached is the image. The semi-circle volume part is what is giving me a problem.

-mAx- July 29, 2010 02:55

ok.
I suppose you have an airfoil or something like this at the center of semi-circles.
I would split the volume along the airfoil geometry, like this : Attachment 3013 (see http://www.cfd-online.com/Forums/flu...e-meshing.html)

ibzyuk July 29, 2010 03:56

Yes in the centre is the airfoil. I will give it ago and let you know how I get on.
I just relised that the wing I am modeling does not have the same chord lenght accross the wing and has rounded tips. I will create a CAD drawing then import into Gambit and start again.

Thanks Max for all your help!

nvtrieu August 4, 2010 01:55

Hello ibzyuk,

I've generated mesh for 3D wing in GAMBIT. The first time I also got the same error with yours. Then I solved it by let the gap between two surfaces of the foil at the trailing edge (I used NACA 0012). Now the mesh works well!

Good luck!

ibzyuk August 5, 2010 03:31

Hi Nvtrieu,

Thanks for that. I did initially try this leaving a gap at the trailing edge, but ran into a few problems. The airfoil I am using is the FX63-137 which is not symmetrical on the x-axis. I took 0.001 of each of the airfoil edge , but when trying to set the surfaces they did not line up. I was following this tutorial:

http://web.missouri.edu/~solbrekkeng/TherMMEC/pubs/thesis_logsdon.pdf

Do you have any ideas how I can overcome the airfoil not being symmetrical?

nvtrieu August 5, 2010 04:02

Hi ibrahim,

I also followed that thesis. First step, I create two curves that will be used to make the upper and lower surface later, then use a straight like to connect them at the trailing edge. Then save it in iges file. Second, import the iges file into GAMBIT. After create the faces, the volume is created by using Sweep face function. So in the first step, if we donnt let the gap at the trailing edge, we cannot create the volume for the semi-circle face!
Okay, I haven't meshed on non-symmetric foil. However, I think they should be meshed in the same process. So, I just suggest you try to do it one more time then we'll see what happen.
Good luck!

Trieu.

wind.cfd May 2, 2012 20:52

Hi guys,
I have a same problem about sweeping the ellipse edge, when I use sweep edge to create the surface, I receive the following error:
the edge being swept has moved from its original location. Try changing the default GEOMETRY.VOLUME.SWEEP_PATH_ALIGNMENT to 1 or try aligning the path(edge/vector) to start from edge being swept.
ERROR: sweeping or revolving edge edge.3 failed
I would really appreciated if you let me know what the problem is?!
Thanks

asmita April 10, 2013 21:03

Hello,
I have the same error ACIS 95034-attempt to sweep a non-manifold.
I am trying to sweep a face in Gambit to create a volume.The face contains two blades of a turbine with rectangle surrounding it.I want to create separate volumes for the turbine and the rectangle.Can anyone help me please..?

-mAx- April 11, 2013 04:18

post pictures

asmita April 11, 2013 12:15

1 Attachment(s)
Hello, I have attached the photo.I tried using split face option and split faced the rectangle and turbine. I also want to mesh the face before sweeping it. PLease can you help me

asmita April 11, 2013 19:37

please help me as soon as possible.
Thank You in advance

-mAx- April 12, 2013 08:13

sweep the blades for getting volume
sweep (or create your outer box as volume)
substract blades volume from outer box

asmita April 15, 2013 22:14

Hello,
I tried doing as you said. N it did created volumes.but my total mesh is in 100 KB. which should not be the case.I want mesh in 100 MB's and more. Also once when I tried to import that mesh in Fluent I couldn't see the rectangle.Only turbine was imported.This means my geometry wasn't good.

-mAx- April 16, 2013 03:54

That means your volume has not been imported.
Your turbine should appear as surface and not as a volume.
Did you receive any warning while exporting mesh from Gambit (like warn unmeshed entity are not exported)

asmita April 16, 2013 11:24

No. It did not gave me any warning. N when u say that turbine should be surface and not volume, what exactly do u mean ? Because when I sweep the face of turbine it creates volume.And i do want volume for both turbine and rectangle.
Do Split -Face option help? My friend suggested to split-face turbine from rectangle, then mesh and sweep it. But that is not helping me either.

-mAx- April 16, 2013 13:37

You have 2 volumes: your outer domain and the impeller.
If you subtract the outer domain with the impeller, then you get the outer domain with impeller as hollow.
That means your impeller is only represented by its outside surfaces.
And the volume represents the fluid domain around the impeller.

asmita May 2, 2013 18:05

Hii, thanks for the advice.. I got my geometry done. But now i have to restrict the motion of turbine in all directions i.e. translational and rotational except for rotation in x-direction. How do i do that..? Do i use the 'specify continuum types' in zones option in Gambit ? Do i define turbine as fluid or solid type for that?
Please help me

-mAx- May 3, 2013 09:50

do you want moving mesh, Single or Multiple Reference..?
check here:
http://aerojet.engr.ucdavis.edu/flue...ug/node412.htm
http://aerojet.engr.ucdavis.edu/flue...ug/node436.htm

asmita May 4, 2013 16:02

Moving mesh...

-mAx- May 6, 2013 03:39

then you have to disconnect rotor zone to stator zone.
Doing this, you will able to set interfaces between both zones (one interface for rotor, and one for stator).
Then fluent will be able to move rotor zone, and data will be interpolated on interfaces (which you have to define "again" in fluent /define/grid-interfaces)

asmita May 7, 2013 00:05

How to disconnect rotor from the rectangle ?

-mAx- May 7, 2013 11:01

do a split to distinguish rotor from stator.
If the split is well done, then both domains are connected, which is not wanted.
So copy rotor zone with any translations-vector.
Delete original rotor zone.
Now you can see your stator with an hollow.
Define the interface face (stator side)
Define the interface face (copy of rotor side)
Move back the rotor zone.
That's it (Now both interfaces are stacked)

asmita May 29, 2013 19:31

Hello,
What do you exactly mean by copy rotor zone with any translations-vector ? I did not understand that where should I copy and what rotor zone are you talking about.. What should I do before deleting the rotor?

-mAx- June 3, 2013 02:20

if you copy your volume, it will be localised at it original place, and you won't see it
If you give a translation's vector, it will be copied and translated, so that you will be able to delete the original volume
http://aerojet.engr.ucdavis.edu/gamb..._volm_movecopy

asmita June 4, 2013 20:17

Hii, You said to define the interface on rotor and stator.. Is that in Gambit or in Fluent ? and how do you define it? Also you said that - move back the rotor zone.. I did not get it. I will translate the rotor volume- delete original. Then I did not get the interface defining part and again moving back the rotor zone part...

-mAx- June 5, 2013 02:51

5 Attachment(s)
You have to do that in gambit.
I give you an example:
-you have a square 100x100
-in the middle you create a rotor zone with radius 10 (just create a circle R10)
Attachment 22437
So you split your square (surface) with the circle. It gives you 2 connected zones: the rotor surface (disk R10) and the rest (stator).
Attachment 22438
If you mesh both surfaces, they are connected and you don't have any chance to move the rotor zone.
For sliding mesh, you need to disconnect both zones and the data interpolation will be done through interfaces. In this case the interfaces will be circles (you need an interface for the stator, and one for the rotor).
Now both surfaces are still connected. Copy and translate the disk anywhere and delete the original. Doing that you disconnect both zones and you can assign interfaces easily.
Attachment 22439 Attachment 22440
Just move back the copy of rotor.
Attachment 22441
That's over: you have stator and rotor disconnected with interfaces

Then in fluent grid/interfaces...

asmita June 6, 2013 18:45

Hii, I tried as you said.. N that was really well explained. Thank you for that. I have a doubt... When i copied the rotor and deleted the original, it did not got deleted in real coz I had used split option between outer box and rotor. So even when i deleted it,I could see the rotor as a whole volume with the box. Is that a problem ?

And after the interfaces are separated, how do i restrict the motion of rotor in all directions except in only one direction..? is that done in gambit itself or I have to use fluent?

-mAx- June 7, 2013 02:40

Quote:

Originally Posted by asmita (Post 432532)
When i copied the rotor and deleted the original, it did not got deleted in real coz I had used split option between outer box and rotor. So even when i deleted it,I could see the rotor as a whole volume with the box. Is that a problem ?

pictures please

Quote:

Originally Posted by asmita (Post 432532)
And after the interfaces are separated, how do i restrict the motion of rotor in all directions except in only one direction..? is that done in gambit itself or I have to use fluent?

once stator and rotor zones are defined (with interfaces), then the rest of setup occures in fluent (define/dynamic mesh)

yoooodarulez July 3, 2013 19:17

1 Attachment(s)
Good evening to everybody.

Im new to the cfd world so my questions may seem a bit newbish but i really cant help it...so here it goes!First of all id like to say that i am studying the airflow over a wing,namely a s809 3d airfoil,and my goal is to compare the results of this analysis with the experimental ones for various angles of attack as well as various turbulence models (SA,k-e realizable,k-w sst),as far as cl,cd,cm are concerned.Im modelling the wing in gambit and processing with fluent 6.3.My problem came up in 2nd order k-e realizable where i either get the error "turbulent viscosity ratio limited" or i do not get convergence.That is for most of the AoA anyway since ive come across some convergence in some others.My first question then is about the mesh itself.The worst element is in a critical region of the flow and i cant seem to get rid of it.Secondly id like some insight on how to mesh the flowfield next to the wing,meaning from where the wing ends till the boundary.I would be glad if some1 gave me some piece of information!


All times are GMT -4. The time now is 00:06.