[stroccel]

Dear all,

I am trying to simulate an APC propeller (low reynolds). I used the two domain method and the Frozen Rotor as indicated in many tutorials. The problem is that I found a thrust one half of the one measured experimentally. I do not understand the boundary condition on the wall of the propeller inside the inner cylindrical volume. Does the inner volume have a rotational velocity opposite to the propeller velocity and the wall fixed respect to the stationary domain?

Thanks a lot,

Regards
Stefano

[ghorrocks]

Your general question is an FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

Can you post an image showing what you are talking about with the wall boundary? Also, you might find it useful to view the velocity vectors (both velocity and velcoity in stationary frame) on the wall boundaries so you get a feel for what the different rotation states on a wall give you.

[stroccel]

Thank you for your replay. I will update the picture. When I speak about wall I intend the surfaces of propeller blades. In fact I build a cylindrical volume with the enclouser function around the propeller and then I subtracted the solid propeller. Then I assign a rotational velocity to the domain but I do not understand in which direction the volume has to rotate (the direction of the propeller or the opposite one?). In addition the wall representing propeller blades surfaces has to rotate togheter with the volume or not?

Regards,

Stefano

[ghorrocks]

The propeller domain rotates in the direction of rotation of the propeller. By default walls are assumed to move with their frame of reference (stationary in stationary domains, rotating with the domain in rotating domains). A wall which is stationary but is located in a rotating domain should be set to counter rotating, and a wall which is rotating but is located in a stationary domain should be set to rotating.

[stroccel]

ok, I tried again with a rotating domain and blades surfaces. I also inserted an inflation boundary with 5 layer of total thikness 2 mm around propeller surfaces. The rotating domain is a cilindrical enclosure around the propeller. The results are not satisfying because the thrust is still 50% lower than the real one and the streamlines in the stationary reference behind the propeller are almost straight and do not roll up. as soon as possible I update pictures.

Thanks a lot

Stefano

[ghorrocks]

THis sounds like an FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

[stroccel]

refining the mesh and enlarging the stationary domain I got 10.2 N against the 14.5 measured experimentally. The strange thing is that streamlines seem to roll up in the opposite direction respect to the propeller rotation but the direction of thrust and of the flow seems to be right. Next to the oulet (that I defined as opening entrainment) there is a big vortex surrounding the internal flow that has straight streamlines. I am not able to upload pictures. How can I do?

Thank you very much
Regards
Stefano

[ghorrocks]

No need for pictures, I think I understand what is happening.

If you have run a finer mesh and a bigger domain and got answers significantly closer to expected that says you are on the right track - you need an even finer mesh and/or a bigger domain and you will get closer still.