hi all, I come from Taiwan. My English is not very good. So... my model: a axial fan in the middle of very long channel. how to make the channel's wall is stationary and the fan is rotation? I use the method of tutorial 7.4. But the result is inner(omega=300) velocity is very large and outer(omega=0) is suddenly become zero.It is not continuous in the interface. Even I use Boundary couple, it's the same. How to define the model in steady flow with implicit MRF or single rotating frame? help .

hello i use MRF for radial ventilator and also get huge values for velocity at some regions but get OK values for pressure. at many parts the velocity is ok though. have you decreased the residual ( i used 0.001 and 0.0005 and 1000 iterations) for speed in radial case: c (absolute) = u (rotation) + w (relative) the u component is due to rotation and this is zero after blade ends therefore at the interface the c is not continuous. am i wrong? greetings

hi sunlight: Thank for your Response.But I still think my result is wrong.Because my blade is have a little distance with inner channel,but the velocity in ineer channel is very big and in outer channel "suddenly" become zero. All of the region in outer channel is zeor,too. I guess the interface in inner and outer region is not couple even I use the boundary couple in it's interface.In fact,the momentum in inner region is not translation to outer region.

Before looking at the results, you remembered to do a "REFRAME STATIC", right? Otherwise you're looking at computational values, and you will see discontinuities.

Thanks for Brian:

Can you tell me how to define the static "REFRAME STATIC"? My fan is steady state model.Can it define the "REFRAME STATIC"?

It's a command - REFR. See HELP REFR for details, but basically it adds the rotation speed to the velocities in the regions that are rotating. This is a post-processing function, it has no effect on your solution.

The static reference frame is from a viewpoint that is not moving, so you will see high rotation speeds in the rotating areas in this frame.

The computational frame (default, REFR COMP) will show very little rotation, as the viewpoint is rotating with the rotating part, but stationary for the non-rotating parts.

I hope that makes sense - it may be easier to try it and think about what you're seeing in each case. I am sure you will get it.

Thank for Brian.I already slove my problem.My result is considered resonable now.Thank for your help.