CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   Rotary drum attached with baffles (https://www.cfd-online.com/Forums/cfx/81133-rotary-drum-attached-baffles.html)

 nutow October 18, 2010 01:27

Rotary drum attached with baffles

Dear all

I am a newbie of cfx's users that would like to simulate the behavior of fluid inside the rotary drum attached with baffles. First of all, I create the geometry like a rotary drum attached with baffles and then define the baffle as the internal wall. After that, I write the input for cfx pre program.

In the cfx pre program, I specify the wall and baffle of rotary drum as the rotating wall but I found the problem which is the terminate run in the cfx solver. So I need to change boundary condition from the above condition to specify the wall of rotary drum as the rotating wall and baffles as stationary wall. Then I run the simulation and get result.

In the result of cfx post program, the result is the wall rotated but baffles is stationary.

So, I don't know how to define boundary condition for the baffles and wall are rotating along the z-axis

Pls help me.

Best Regard

http://img171.imageshack.us/img171/1316/qshot0016.jpg

 ghorrocks October 18, 2010 06:16

Your description is not clear but I suspect you have meshed the baffles. Unless you want to model heat flow in the baffles you won't want to do this. Remove the baffles and just leave the walls on the fluid domain left by the baffles. Then this should run fine.

 michael_owen October 18, 2010 07:07

Actually I think the problem is that he needs to put the entire domain in a rotating frame of reference and make all the walls stationary.

 nutow October 18, 2010 10:42

My objective is to study the mixing of fluid in the rotary drum attached with baffles.

Now, I try to run the simulation by define:

Domain Motion : stationary
Mesh Deformation --> Options : regions of motion specified

In Boundary of wall and baffles I define:

Option : No slip condition
Wall Velocity relative to: mesh motion
Mesh Motion --> Option : specified displacement
Displacement : Cylindrical Coordinate

On theta component : r*t*(2*3.14*1/60[s]) -------------> 1 rpm

I found the wall and baffle were rotate in the cfx post. Are these correct?

 nutow October 18, 2010 10:44

Quote:
 Originally Posted by ghorrocks (Post 279584) Your description is not clear but I suspect you have meshed the baffles. Unless you want to model heat flow in the baffles you won't want to do this. Remove the baffles and just leave the walls on the fluid domain left by the baffles. Then this should run fine.
Are these mean as I specified above?

Thanks you

 ghorrocks October 18, 2010 17:55

If the motion is purely rotary then you don't need mesh motion. Get rid of that bit, you want to define a rotating frame of reference as it is much simpler.

I think my first post still comments on an issue - have you meshed the baffles as solid bodies? If so then remove them and just use the baffle cutouts in the fluid domain.

 nutow October 21, 2010 02:23

I changed the mesh like you suggested above.

I meshed the surface of rotary drum by using tri and then extruded them to form prism meshes because I would like to simply my work from 3d to 2d.

After that I did not do anything, I used it in cfx - pre

Is it same that you suggested above?

Thank you very much

http://img230.imageshack.us/img230/1326/qshot0017.jpg

 ghorrocks October 21, 2010 05:02

Have you done the CFX tutorials? I doubt it if you cannot mesh a simple object like this. You really should do the tutorials as they explain all the basics.

 nutow October 21, 2010 05:45

I did the tutorial number 9 which is the 2d flow over bump in cfx pre.

Could you suggest the subject of tutorial to me for studying more?

http://img121.imageshack.us/img121/4615/qshot0018.jpg

 ghorrocks October 21, 2010 06:04

Do as many as you can. Make sure you do the meshing part as well as it appears you do not understand the basics of meshing yet either. The first few tutorials cover meshing better than the later ones.

 nutow October 21, 2010 06:28