Learning Forum

[josue]

[Rob]

Can you check what the surface type is that bounds the rotor? There's a cylinder just a little bigger than the spinning part - look at the black line just outside the red parts of the contour.

Are you using Fluent or CFX?

 

[josue]

The black line is the projection of the output. the medium is just air. I use fluent.

I have tried another case without a wrapper, and the results are similar to visually normal behavior. but I have seen that the majority use the wrapper, and the same behavior is not achieved.

 

with wrap

 

without wrapper

[Rob]

Middle image in the first post. There's a very visible jump in the contour. However, can you plot the same image in Fluent? CFD Post doesn't always read the moving zone/refrence frame so your data may be correct but you're displaying the velocity in the wrong reference frame. 

[josue]

Here you can see something clearer, however it is assumed that the external area should have speed due to the rotor rotating, which is only seen to be concentrated inside the envelope.

[Rob]

You may need to do some more post processing to see what's going on - flow is going about where we'd expect but that may be masked by the higher core velocity. Angular velocity is a good one to review along with axial.  

How well converged is the model?

[josue]

This last image is the rotor without casing. At first glance, the fluid behavior seems more real. That's why I was wondering if it is normal for the wrappers to have that behavior. Or should they be the same as this last image?
I always match the mesh with share of the spaceclaim. Should it have some interface?
The first image is the convergence curve.

[Rob]

What casing, and what wrappers? For mrf I'd expect an interior, for sliding mesh an interface pair.