I have a Strange questions?

Let me explain my geometry.

I have inner pipe with 0.05 m radius and then outer pipe with 0.1 m diameter and then another pipe with 0.2 meter diameter.

All of them have same axis of rotation.

Now I am not modeling the flow through smaller pipe. But I want to model flow through the other two outer pipes.

Now my questions whether this case is 2D? Well If I take it as 2D. But then in Fluent when I specify the velocity based upon the area and mass flow rate from 3D and then use that velocity as input to 2d model?

I do not get the same mass flow rate as 3D one. The reason is Fluent calculates the area using a depth of 1 m (this is available in reference values). Of course one solution is to change this depth so that you get specified mass flow rate. But the change can be done only either of tube.

I hope you guys what my question is ? Is there any possibility to model this domain in 2D or not.

Share : Share

11 comments Sort by rating

By ka78• 31 Aug 2009 13:20

Rating: 4/5

The geometry certainly suggests that it is axi-symetric.Perhaps,axi symetrical boundary conditions is not desired for a multitude of reasons such as on uniform velocity profile, walls at distinctive temperatures, gravity in the perpendicular direction of the cylinders, and so on. Though, I would start with an axi-symetrical model and make sure that I have set all parameters correctly.

By the-birdie• 31 Aug 2009 12:52

sameera, when is ur vacation due....... just kidding

By every_mothers_nightmare• 31 Aug 2009 12:10

Rating: 5/5

Well, just select "2D axi-symetrical" in "Model", and make sure the cylinder direction is in X and its center (even if not part of the mesh) at the 0,0 location in Y,Z.

I can google also.......

By Jaweedali• 31 Aug 2009 12:09

Rating: 4/5

This is a good question for, given your description of the geometry, the axis location is outside the computational domain.

Well, just select "2D axi-symetrical" in "Model", and make sure the cylinder direction is in X and its center (even if not part of the mesh) at the 0,0 location in Y,Z.

I (probably naively) believe that the solver should get it right. Check it carefully by reporting the "areas" of a couple of surfaces.

If it is not the case, maybe you should contact your user support, they should know better. Hopefully.

At the worse, you may simply have to mesh (coarsly) the smaller-radius cylinder, so as the axis is actually part of the mesh. If you solve for temperature and you do not want the inner cyclinder to play a role, just freeze the heat exchange at the separation-wall (heat flux 0 at the wall facing the cylinder, and whatever you want for its "shadow" facing the fluid). Does this make sense?