Hi,

I'm trying to run a 2D simulation on a pitching airfoil (sinusoidally varying angle of attack) in a wind tunnel. I've figured out how to move the blade in the normal cartesian coordinates using mesh deformation, but my problem involves pitching the blade, or rotating it about a specified point, and having the mesh move with it. Is this possible? If so, how would I go about doing this?

Note: I have already tried solving this problem using two domains: a rotating domain (transient rotor-stator) inside a rectangular stationary domain, and specifying the rotational motion of the rotating domain with a sinusoidal CEL expression based on time. However, when the angular velocity changed, CFX superimposed velocity vectors into the rotating domain, on top of the already existing velocities. So I decided to try using the mesh deformation technique I mentioned above.

Any help would be much appreciated. This problem is likely very similar to a rotating butterfly valve, so if anyone knows how to model that I would appreciate their feedback.

Regards,

Andy

Hi,

Your TRS approach is probably the best way. When you post-process it look at the "velocity in stationary frame" rather than the "velocity" variables to see a consistent velocity field.

Glenn Horrocks

Thanks for the reply Glen,

That's what I thought, that TRS would be the best way, however, I already am looking at Velocity in STN frame. The odd results are due to my accelerating rotating frame of reference (the motion of the rotating domain follows a sinusoidal variation in angular velocity to simulate a blade pitching). With constant angular velocity this isn't an issue, but when the rotation speeds up and slows down, then changes direction, I've got problems. I believe it may be due to the added Coriolis force terms when running a transient simulation with rotating reference frame, but have yet to figure this out. Surely someone has encountered a similar problem when pitching a blade? I know I certainly am not the first to be running blade pitching experiments...

Any thoughts? Anyone?

Regards

Hi,

You may well be right in that TRS does not account for rotational accelerations. I am not sure, I have not checked.

If that is the case then make it a moving mesh region and define the mesh motion to be your rotation function. That way all the acceleration terms should be in there. Still connect the rotating bit to the outside bit with a GGI so you don't get excessive mesh deofrmation. You may need to use a fortran subroutine to define the mesh motion at each node directly rather than using the default internal smoothing. This will keep the mesh fixed so the mesh does not distort just rotates.

Glenn Horrocks

[euzinho]

Hello Andy,
I take this oportunity to ask you for some help...
I trying to study a transient flow through a butterfly valve during its openning process
so that, I would like to rotate a butterfly valve 45º (from close stage to open stage)
Can you explain me how to do that?

do I have to write a UDF to generate the move?
because I don´t have any idea (even after reading the udf manual)

isn´t there any standard procedure to get it done?

thanks in advance

[ghorrocks]

The standard procedure is to post the question on the forum relevant to the software you are using. Try the Fluent forum.

[euzinho]

sorry about that!
I am a new user...
thanks anyway

see ya