CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Grid adaption (http://www.cfd-online.com/Forums/fluent/71414-grid-adaption.html)

SerSe December 31, 2009 06:19

Grid adaption
 
Hi guys,

I have a capillary pipe with two fluid phases, i.e., water at the left and right end of the pipe separated by an air gap (therefore a unique big bubble which bounds the walls) in the centre of the pipe.
My BC are mass flow inlet (just water) and pressure outlet, interior at the interfaces water/air/water and, of course, walls.
I am using the Volume of Fluid Multi-phase model to get insight into both the motion of the air gap pushed by the inlet water and the shape of the interface.
Should I turn grid adaption on, to model this problem?
and should I also provide as input in Fluent the motion of the interiors?

I made some trials with dynamic mesh but it didn't work. Moreover, dynamic mesh is used when the motion of a certain domain is known, while I don't know the motion law of the bubble.

Many thanks,

SerSe

meb January 14, 2010 10:45

Quote:

Originally Posted by SerSe (Post 241192)
Hi guys,

I have a capillary pipe with two fluid phases, i.e., water at the left and right end of the pipe separated by an air gap (therefore a unique big bubble which bounds the walls) in the centre of the pipe.
My BC are mass flow inlet (just water) and pressure outlet, interior at the interfaces water/air/water and, of course, walls.
I am using the Volume of Fluid Multi-phase model to get insight into both the motion of the air gap pushed by the inlet water and the shape of the interface.
Should I turn grid adaption on, to model this problem?
and should I also provide as input in Fluent the motion of the interiors?

I made some trials with dynamic mesh but it didn't work. Moreover, dynamic mesh is used when the motion of a certain domain is known, while I don't know the motion law of the bubble.

Many thanks,

SerSe

With VOF the boundary movements is not a concern; the ability to carefully track the sepration between phases depends on the mesh size.
You have to turn on dynamic mesh if external boundaries moves.
We are study a similar application (but on a different scale): a flexible pipe partially filled with liquid... In our application you need both VOF and moving mesh.

SerSe February 3, 2010 05:40

Hi Meb,
after some trials I got that I just need to use the VOF model, without using any dynamic mesh, since my boundaries are fixd walls.

By the way, when you talk about refining the mesh to catch the right interface shape, do you refer to the mesh size in the radial direction or in the axial direction? I guess the axial mesh would affect that shape more than the radial one.

Greetings,

Serse

meb February 3, 2010 06:35

Quote:

Originally Posted by SerSe (Post 244732)
Hi Meb,
after some trials I got that I just need to use the VOF model, without using any dynamic mesh, since my boundaries are fixd walls.

By the way, when you talk about refining the mesh to catch the right interface shape, do you refer to the mesh size in the radial direction or in the axial direction? I guess the axial mesh would affect that shape more than the radial one.

Greetings,

Serse

Using VOF you track the interface as a concentration. The ability to resolve the interface using VOF is strongly related to local mesh size.
So the CFD domain part where the interface is moving has to be equally spaced using a fine mesh size (in both directions).

We have used such technique here:

L. Andreassi, R. Baudille, M. E. Biancolini, “ Spew Formation in a single lap joint” International Journal of Adhesion & Adhesives 27 (2007) 458–468.

R. Baudille, M. E. Biancolini, “A general approach for studying the motion of a cantilever beam interacting with a 2D fluid flow” Interaction and Multiscale Mechanics, Vol. 1, No. 4 (2008) 000-000 1

Regards.

Marco

SerSe February 3, 2010 07:58

Hi Marco,

a quick question.. is there any possibility in Fluent to model only half a curved pipe section (let's say a pipe with a 90° curve) and solve it as 2D axisymmetric?

Thanks for the tips on mesh size when using the VOF model!

Ciao,

Serse

meb February 3, 2010 08:21

I don't know, but the mentioned problem is not axis-symmetric!


All times are GMT -4. The time now is 23:54.