CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   problem in modeling flow over a 3d airfoil (http://www.cfd-online.com/Forums/main/73750-problem-modeling-flow-over-3d-airfoil.html)

guess March 16, 2010 07:32

problem in modeling flow over a 3d airfoil
 
Hello everyone,
I am trying to model fluid flow over a 3d airfoil in gambit-fluent. For this, I am making an airfoil and a surrounding rectangular domain, within which I am putting in more cells (fine mesh). Then there is a bigger rectangular domain (enclosing the airfoil and smaller rectangular domain) where, i am putting less number of cells (coarser mesh). I am adding this bigger domain so as to capture the effect of flow in free air where the flow is not affected by domain walls.
But, even though i am specifying Boundary Condition as Interior for the smaller rectangular domain, fluent is giving the error message as: Cannot change interior_wall to interior because there is only one adjacent cell thread.

Can anyone help me with this problem.

Jonny6001 March 16, 2010 13:10

Hello, I would hazard a guess that you must have subtracted a face or volume from another, then re-created that face/volume for use in the other zone.

The best way I found to perform an opperation like this is to use the "Retain" check-box in the bolean opperation box. That way you know that face or volume is exactly the same in both defined zones.

Hope that is clear.

guess March 16, 2010 13:30

I had checked the retain check box while subtracting volumes. What i exactly did was:

1) Created a 3d airfoil

2) Created the smaller rectangular domain.

3) Created the bigger rectangular domain.

4) Subtract airfoil volume from smaller recatngular domain ( retaining the airfoil volume).

5) Subtract airfoil volume from the bigger rectangular domain ( not retaining the volume so as to create a cavity in the domain in the shape of airfoil )

Am I soing anything wrong here ?

Jonny6001 March 16, 2010 19:17

Yes, I think you are doing something wrong.

You can't subtract the airfoil volume from the large one as it will leave the volume inside the airfoil section.
Try this:

1)Subtract the small rectangular domain from the large one (but retain the small volume)

2)Subtract the airfoil from the small volume

guess March 18, 2010 16:02

Thanks Jonny6001 for your help, but it is still showing the same error in fluent.

karananand March 19, 2010 12:26

I don't know much about this, but this may help: create the airfoil, create the larger rectangular domain, subtract the airfoil, from the larger domain. While meshing, put the ratio of grid density such that you have course mesh near the edges and high intensity mesh near the blade. Hope this helps.

guess March 19, 2010 12:31

@Karananand, I have tried this approach. The problem with it is that either a proper mesh is not obtained or a large number of cells are being created which is forcing the computer to run out of memory. So I am looking for alternate ways.

Jonny6001 March 20, 2010 05:01

Really sorry, I couldn't quite see what you were trying to do.

Rather than substract faces/volumes, you should try and split the volume if your case allows, split it with the smaller volume.

karananand March 20, 2010 14:54

Well, if you are performing laminar flow analysis, you need not have that big a mesh size where you have memory problems (assuming you have a "moderate computer") . If you are using a simple turbulence model (1 equation or 2 equation) you do not need that big a mesh as you would need for LES or v^2-f equation. If it is supersonic flow with laminar B.L conditions on the airfoil, you again do not need that big a mesh size. What are the flow conditions and what turbulence model will you use? Also if you consider making the NS equation non dimensional it might help for convergence and help you to work with reduced mesh sizes. Non dimensional NS is easier for incompressible flow, if you are in subsonic region. You may want to consider all these and start with a really coarse mesh (start with 500 x 500, for example), where you don't have memory problems while meshing.
You will have to play around with it for a while. But this method should work. I remember doing a 3D simulation for a ~Boeing 747 wing (size and airfoil shape were similar).


All times are GMT -4. The time now is 21:27.