CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Rotating mesh (http://www.cfd-online.com/Forums/cfx/23667-rotating-mesh.html)

Jesper February 20, 2007 10:07

Rotating mesh
 
Hey CFX-users.

I have the following question:

Is it possible to rotate a mesh in CFX, when the mesh is a structured mesh created in ICEM? And if so, do you have a pointer of how to do this, or where to find info about this function?

I know how to create a moving mesh in translation, so this is not the problem in hand. And my profil is not to be moved.

Thanks in advance...

Regards Jesper

Glenn Horrocks February 20, 2007 16:53

Re: Rotating mesh
 
Hi,

The solver does not care where the mesh comes from. The moving mesh function will work the same on meshes from any meshing package.

So to answer your question, yes a structured mesh can be used in a rotating domain.

Glenn Horrocks

Jesper February 21, 2007 03:32

Re: Rotating mesh
 
Hi again

thanks for that response. But my question still stands:

How to do a rotating mesh or where to read about the procedure? (and with out having to large mesh deformations)

tanks again

Jesper

Patrick February 21, 2007 09:01

Re: Rotating mesh
 
If you have a rotor/stator problem then the frozen rotor concept might be the right for you. That's also the way to do a simple steady state simulation.

Jesper February 21, 2007 10:03

Re: Rotating mesh
 
Im not sure I 100% understand that reply...

I hav a bridge profile centered in a 2D Windtunnel (one element thick), and want to rotate the mesh around the profile. How to do this?

sorry if it was not clear enough at first.

Regards

Jesper

Glenn Horrocks February 21, 2007 17:11

Re: Rotating mesh
 
Hi Jesper,

What are you trying to do? Why do you want to rotate the mesh around the profile?

Glenn Horrocks

Patrick February 22, 2007 05:19

Re: Rotating mesh
 
Then forget about the frozen rotor concept. But I also don't understand your problem. Maybe you should describe your problem a little more.

Johnny February 22, 2007 07:52

Re: Rotating mesh
 
Are you asking if you can define mesh motion in cylindrical coordiates? If so, the sort answer is no, you can only define mesh motion in Cartesian coordinates. But you could easily generate some CEL to apply a rotational mesh motion.

Jesper February 22, 2007 12:05

Re: Rotating mesh
 
Firstly tanks for replying everyone!

The general idea is to rotate the mesh inkluding the bridge profil as a riged body movement (sorry for the misstake earlier) - but to keep the inflow direction constant. In this way the profile will expirience this as a varying angle of attack of the wind alpha = A*sin(omega*t) -and so simulating a marmonicaly movement of the profile with amplitude A and the circular frequency omega.In this way flutter derivatives can be determined.

If further discription is needed just ask again.

Once again thanks everyone.

Regards

Jesper

Glenn Horrocks February 22, 2007 17:44

Re: Rotating mesh
 
Hi,

I can see two approaches being useful here.

If the rigid motion timescales are slow compared to the fluid timescales you may be able to simulate this as a series of steady state simulations with a varying angle of the air at the inlet boundary. This will be a simple simulation and does not need moving mesh, rotating domains or anything.

If the rigid motion timescales are similar to the fluid timescales then you are forced to do a fully coupled transient simulation. I would consider doing this using FSI, but if your bridge motion can be boiled down to a fairly simple equation of motion you might be able to do it entirely in CFX. Don't forget you are likely to have transverse motions as well as rotations so I would forget about rotating domains and just just general mesh motion. An example is the ball valve moving mesh tutorial.

Glenn Horrocks

Jesper February 23, 2007 13:31

Re: Rotating mesh
 
hey Glenn

Thanks for the response and yor great inputs/ideas. I know that transverse movement will occur, and these i know how to acount for - but rotating the mesh harmonically (or the angle of attck) is nessesary to do these calculations in tthe way mentioned earlier on. I have no acces to other programs than CFX, so this will have to do- sorry. Do you think it is possible, and if so - how to go about?

tanks again

regards Jesper

Glenn Horrocks February 25, 2007 17:34

Re: Rotating mesh
 
Hi,

As long as you can write the equation of motion as a CEL expression it will work fine. Also I would only do the motion as a moving mesh simulation, forget about putting the rotating bit in a RFR. As long as the deflections are not huge (and for a bridge I hope they are not!) moving mesh should work fine.

Glenn Horrocks

Jesper February 26, 2007 04:36

Re: Rotating mesh
 
ok thanks Glenn.

I will definetly try that, but are almost certain it will not work, because simulating rater large angles is nessesary to conduct a flutter derivativ analysis.

But you are right - in generalt the deflections are normally small for a bridge, but not when considering flutter - just recall the movements of the Tacoma Narrows bridge - though the motions today are some what smaller.

But thanks for your interesst and inputs/help.

If anyone have other surgestions or experiences with this, do not hesitate to write them here.

regards

Jesper

heatdrive April 27, 2012 13:35

Quote:

Originally Posted by Johnny
;80470
Are you asking if you can define mesh motion in cylindrical coordiates? If so, the sort answer is no, you can only define mesh motion in Cartesian coordinates. But you could easily generate some CEL to apply a rotational mesh motion.

would you please tell me a little bit detail about CEL and how to use it or how it's working...i'm doing my final project of vertical wind turbine in 2d simulation, and i saw in the youtube such simulation which the creator mention about CEL..hope that i can get some advice..thank you

ghorrocks April 27, 2012 17:57

Examples of CEL are included in the CFX tutorials.

This old post is discussing rigid body motion coupled to fluid flow. This has been somewhat superseded by the 6DOF solver in recent releases which means you do not need to do program the basic motion any more, and a much better and more accurate scheme is built-in.


All times are GMT -4. The time now is 23:18.