CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   3-d mesh gambit (cylinder within a cylinder) (http://www.cfd-online.com/Forums/ansys-meshing/42356-3-d-mesh-gambit-cylinder-within-cylinder.html)

 Jason September 21, 2006 13:27

3-d mesh gambit (cylinder within a cylinder)

Any help would be appreciated:

First of all I create two 3-d cylinders in Gambit... one with a height of 5 and a radius1 of 5 and one with a height of 2.5 and a radius1 of 2.5

What I am trying to do is split the volumes so that I can mesh the inner cylinder with a fine mesh and the outer cylinder with a coarse mesh.

Here is the problem I was running into... I can mesh the inner cylinder but when i go to mesh the outer cylinder it lays the mesh over the inner cylinder... How can I get the two cylinders to be connected but independant of each other so i can mesh each seperately... thanks in advance.

-Jason

 zxaar September 21, 2006 16:22

Re: 3-d mesh gambit (cylinder within a cylinder)

. I can mesh the inner cylinder but when i go to mesh the outer cylinder it lays the mesh over the inner cylinder... How can I get the two cylinders to be connected but independant of each other so i can mesh each seperately.

I have meshed these kinds of cylinder many times as you are saying and never got into trouble. After the split operation you should get two volumes, one the inner cylinder and then the rest of outercylinder you should get if you substart inner cylinder from it. In your case it seems instead of getting the second volumes as (Cyl2 - cyl1) you are getting second vol = cyl2. So when you mesh you see that it meshes above cyl1. Probably you have the retain button on. So evenif they are plit, cyl2 original is still there. After split check how many volumes do you have.

 Jason September 25, 2006 15:08

Re: 3-d mesh gambit (cylinder within a cylinder)

Thanks for the response Zxaar, but I am still having trouble getting this to work. I'm sure it is something stupid on my end, but perhaps someone could walk me through a step by step approach to what I am trying to accomplish in my initial post.

Here is what I am doing and it is not working.

1) Create a volume which is a cylinder. It has a height of 5 and a radius1 of 5.

2) Create a 2nd volume which is also a cylinder. It has a height of 2.5 and a radius1 of 2.5

3) Go to the split volume commmand and split the outer cylinder with (volumes real) the inner cylinder. Retain is checked off and bidirectional and connected are checked off. It then says in the command window: "volume split (outercyc) volumes (innercyc) connected bientity.

4. I then go to the mesh tool bar to mesh volumes... the inner cyclinder highlights correctly, but the when i click on the outer cylinder everything is highlight and I can't get independant meshes...

I know this problem is pretty straight forward and simple, but for whatever reason I am missing something (or many things) Any additional help would be appreciated... Thanks in advance.

-Jason

 Jason September 25, 2006 15:09

Re: 3-d mesh gambit (cylinder within a cylinder)

In my last post there was an error on step 3

it should read: Retain is checked off and bidirectional and connected are checked on.

 zxaar September 25, 2006 16:53

Re: 3-d mesh gambit (cylinder within a cylinder)

I am not sure which version of gambit you are using, old versions of gambit could show this kind of error.

 zxaar September 25, 2006 19:54

Re: 3-d mesh gambit (cylinder within a cylinder)

1) Create a volume which is a cylinder. It has a height of 5 and a radius1 of 5.

2) Create a 2nd volume which is also a cylinder. It has a height of 2.5 and a radius1 of 2.5

3) Go to the split volume commmand and split the outer cylinder with (volumes real) the inner cylinder. Retain is checked off and bidirectional and connected are checked off. It then says in the command window: "volume split (outercyc) volumes (innercyc) connected bientity.

I did this and yes, there is not problem. And i used old version of gambit, not the latest one.

 Jason September 26, 2006 07:37

Re: 3-d mesh gambit (cylinder within a cylinder)

First of all, don't use the bi-directional option (turn bi-directional off). This could create extra volumes that you don't want to deal with. If you want the meshes of the two cylinders to be independent, then turn OFF the connected option. This way the two cylinders are no longer sharing a common face. Also, don't try to mesh them at the same time. Mesh the two cylinders separately (one at a time).

The last thing to mention is that you say "everything is highlighted"... that's because the larger cylinder has the geometry of the smaller volume removed from it... it's like an empty box (with thick walls) right now, and when you select it, both the inner and outer walls are being selected...

Hope this helps, and good luck, Jason

 Jason October 2, 2006 15:16

Re: 3-d mesh gambit (cylinder within a cylinder)

Jason,

Thanks for the advice... and I did everything you said, but I am still having problems with the mesh.

I create the 2 cyclinders (one with height and radius of 5, the other with height and radius of 2.5)

After I do this I go to split volumes ... I split the inner volume with the outer volume with only the retain option clicked on.

Next i go to the meshing volumes tool, and when i select the volume i have two options available... the inner cyc and the outer cyc. The inner cyclinder is exactly what I want for that part of the mesh. However, for the outer cyclinder here is what is highlighted:

The whole outer area. When i mesh it this is what i think is happening. It leaves the inner cylinder unmeshed but it is still meshing over the top part of the inner cylinder.

What I want to happen is when i mesh the outer cylinder any part of the whole inner piece is not getting meshed, but that is not what is happening...

What am i doing wrong? Thanks again guys.

 Jason October 3, 2006 07:46

Re: 3-d mesh gambit (cylinder within a cylinder)

Just because it looks like the faces of the inner cylinder are being meshed, doesn't mean they're actually being meshed (this is a common problem when people DO want the meshes to be shared across both volumes). When you split the geometry (with the connected option off), both volumes now have a face that shares the same space (but they each have their OWN face... they're not sharing the faces). Just try meshing the inner volume how you want, and you should see that there's two different meshes at that same face. Or, use the model attributes tool (bottom row... the icon is a red square with a white cube and a series of check boxes on it) to turn off all of the volumes. Then turn back on the inner volume. If there isn't any mesh shown on that volume, then the faces aren't being shared and the inner volume mesh will be independent of the outer volume mesh.

On a side note, you really need to look into what's going on when you split geometry. Each option has a purpose, and if you're not paying attention to what they mean, then you can get things you don't want.

For the split command: Retain means to keep the original volumes (so you'll have the split volumes, and the original volumes used to make the split) Connected means to connect the faces at the interface of the volumes so that both volumes are sharing a common face. Bidirectional means that volume A is going to split volume B and volume B is going to split volume A at the same time...

From your description, you want a larger outer cylinder and a smaller inner cylinder... you want them to have different meshes... therefore, turn off ALL of the options (you don't want the larger cylinder trying to split the smaller one... doesn't make sense, that's why "bidirectional" is off... you only want one big volume with a hole in it and one small volume to fill it, that's why "retain" is off... you don't want the faces to be shared by both volumes, that's why "connected" is off).

If you leave bi-directional on, you may get a weird ghost volume. If you leave retain on, then you'll have extra volumes, and if you leave connected on, the volumes will be sharing faces.

Hope this helps, and good luck, Jason

 brusly October 5, 2006 01:00

Re: 3-d mesh gambit (cylinder within a cylinder)

change the type of mesh in the outer cylinder and try it that may help you (Tri mesh)

 goyalnn May 2, 2011 22:33

Hi

I have created a person (made of different volumes) in GAMBIT. The person is standing inside a big cube which in itself is a separate volume. I want to study the air flow inside the cube. Do I have to first split/subract or do anything like that with outside volume and the volumes inside it (person) or Can I just go ahead and create edge,face and then volume mesh for the whole thing without specifying any such relation between them?? someone told me it is necessary to split but when Im trying to split the person from the outside volume, it gives me an error "coincident face_face_ints with different body vertices". Can anyone help me??
I would greatly appreciate your response.

Thanks!

 alejandro_1980 November 23, 2011 02:33

re

Quote:
 Originally Posted by goyalnn (Post 306014) I have created a person (made of different volumes) in GAMBIT. The person is standing inside a big cube which in itself is a separate volume. I want to study the air flow inside the cube. Do I have to first split/subract or do anything like that with outside volume and the volumes inside it (person) or Can I just go ahead and create edge,face and then volume mesh for the whole thing without specifying any such relation between them?? someone told me it is necessary to split but when Im trying to split the person from the outside volume, it gives me an error "coincident face_face_ints with different body vertices". Can anyone help me?? I would greatly appreciate your response. Thanks!
well, i think it would be a nice idea if you extract the person to the cube, then, you can declare walls BC to the body shaped geometry resulted from the extraction, (it must be as a body shape hole) then, you can mesh with small elements that empty body shaped body, and use T grid volume mesh. Finally you need to declare all the rest conventional BC, as inflow, outflow etc.....

Good luck...

 shk12345 November 27, 2011 13:05

RE

This is a problem which took one month for me so solve lately.

Create a outer cylinder and then inner cylinder.
Apply Boolean operations and subtract the inner cylinder from the outer cylinder.
Create the inner cylinder again.
Mesh it.
Give wall as the boundary to the interface of the both the cylinders.
export to fluent and run it.
It should work fine.

Thanks
SHK

 B N HARSHA December 13, 2012 04:04

3D mesh difficulty

Hi,
This is Harsha actually i have a problem i have three cylinders the radius are 2,4,8m the 2m cylinder is heat generating and 4m cylinder contains air and 8m cylinder is steel i meshed the model making use of split option when i imported to fluent after many iterations i found no heat transfer was taking place was tat due to no connectivity in between the cylinders is it correct to use split opt or should i use subtract opt instead if so how please help me

 ghost82 December 13, 2012 10:46

Quote:
 Originally Posted by B N HARSHA (Post 397296) Hi, This is Harsha actually i have a problem i have three cylinders the radius are 2,4,8m the 2m cylinder is heat generating and 4m cylinder contains air and 8m cylinder is steel i meshed the model making use of split option when i imported to fluent after many iterations i found no heat transfer was taking place was tat due to no connectivity in between the cylinders is it correct to use split opt or should i use subtract opt instead if so how please help me
you have to connect faces between cylinders (you can't move only a cylinder if you try to move it); when imported into fluent, it will create shadow walls; then choose coupled under thermal boundary conditions for each internal wall.

Daniele

 B N HARSHA December 14, 2012 02:40

1 Attachment(s)
thanks Daniele for ur reply but i have meshed volumes how to connect the faces if so which all faces i have attached the jpeg pls go through that and reply

 ghost82 December 14, 2012 05:04

1 Attachment(s)
Quote:
 Originally Posted by B N HARSHA (Post 397474) thanks Daniele for ur reply but i have meshed volumes how to connect the faces if so which all faces i have attached the jpeg pls go through that and reply
Hi,
try this: in gambit select the smaller cylinder and move it somewhere: then move the medium size cylinder and move it somewhere; if your cylinders move then you have to connect faces: lateral faces and bases have to be in common between adiacent cylinders (look at the picture: as an example you may have 2 faces, where indicated, and you have to connect them - geometry ->faces-> connect)
Mesh the domain: note that nodes on faces must be in common for adiacent cylinders.
If you have errors in gambit when you try to move cylinders, then check if in fluent you set coupled walls in thermal boundary conditions.
Daniele

 steven1 June 26, 2013 09:23

hello, i am now conducting a similar projet. and i have a question. is it ture that all the faces which are not difined will be treated as walls? if it is this case, the interface will be treated as a wall or not? thank you in advance.

 -mAx- June 26, 2013 09:41

All boundary surfaces without predefined BC, and all non-connected inner surfaces will be treated as wall.
In your case: if you define a cylinder as volume, and you set one cap as velocity inlet, and the other cap as pressure outlet, then you can export and gambit will create a set as wall which contains the cylinder surface (the one(s) without BC.

Now if you split your cylinder along its axis into 2 volumes and you disconnect both sub-volumes, then the split surfaces (there are 2 since volumes aren't connected) will appear in the wall set. Gambit understands that you have 2 separated volumes and the split surfaces are boundaries (ie: only one side of surface has adjacent fluid region, on the contrary with connected volumes, there is only one split surface, and this surface has 2 adjacent fluid regions).
I hope I am clear
;)

 steven1 June 26, 2013 10:00

thank you very much ; it's clear :)

All times are GMT -4. The time now is 12:58.