[Jose Daniel]

Hello all,

I am currently working in the rotational problem of an airfoil and I am experimenting some problems with it. The outline and the 3D view is attached as "outline.png"

outline.jpg

I have used an steady state problem to initiate the simulation and then it moves to a transient simulation where the circle that can be seen in outline.png rotates periodically between 25º and -25º (the RotationVelocity expression is (pi/180)*25 * 2*pi*1 * cos(2*pi*1*Time))

The issue I am having is that the fluid in the domain is rotating with the domain ("rotation_wrong.png" in attachments)

rotation_wrong.PNG

and I only want to rotate the airfoil profile (wall) and the mesh in the circle, not to insert a rotational velocity in the fluid (since I want to rotate the airfoil). Can you help me?

Thank you very much

PD: I have upload two more images in order to provide more data of my settings. If you need to know something else, do not hesitate to ask for more.

airfoil_1.PNG

Domain_interface.PNG

[ghorrocks]

Try converging to a tighter residuals tolerance. Also try the alternate rotating model in the rotating frames of reference options.

What is the frequency of oscillation? Is the angular acceleration significant? Rotating frames of reference do not take into account angular acceleration.

[Jose Daniel]

It is not about the residuals and I have checked that it cannot be solved with the alternate rotating model (at least not with the configuration that I showed).

The problem is that the fluid solution I am getting is being modified as if the circle was a platform that rotates. And on top of that, the boundary wall (the airfoil profile) is not moving.
The thing that I want to do is to rotate the mesh of that circle (which contains the airfoil) AND do not alter the fluid (except due to the airfoil presence, of course)

The frequency oscillation is 1Hz and the angular velocity is set to a sinusoidal function which its maximum is 2.74rad/s and therefore, its maximum acceleration is 17.22rad/s2.
The formulas that I set for the velocity (the one set on the rotational velocity) of the circular domain is the following:

Velocity = (pi/180*25deg)*2*pi*1Hz*sin(2*pi*1Hz* time)

which comes for the intended position:
Position = (pi/180*25deg)*sin(2*pi*1Hz* time)

[ghorrocks]

What is the free stream velocity and the velocity of the outer edge of the rotating domain at peak velocity?

As I said in post #2, rotating frames of reference assume the rotation is constant velocity so angular accelerations are not accounted for properly. I suspect in your case what you are seeing is this. If this is the case then it will not be possible to remove this effect, and the only way of fixing the problem is by moving to another model, such as making your rotating section a moving mesh (not a rotating frame of reference).

[Albert G]

hi
Daniel
if you want to do transient simulation, please change the option of frame change/mixing models to transient rotor stator

[Jose Daniel]

Thank you for your responses.

In the end I did the rotation with mesh deformation, which is easier than this domain rotation and allows more options and movements.

[ghorrocks]

Moving mesh is much slower than rotating frames of reference, but it does include the rotational terms ignored in the rotating frames of reference model. In your case you obviously need the additional terms so you need to do this.