Discussion Groups

Simulation CFD

New Member
Posts: 1
Registered: ‎02-26-2013

How to improve rotating region anlysys !!!!

435 Views, 3 Replies
02-26-2013 10:54 PM

Dear sin sqad,

At the beginning sorry for my English. I have a problem with rotating region analysis. I have an electric motor and simulates his fan . How to set simulation parameters to get   most convergent results (meshing, solver). I did laboratory veryfication, the simulation results are less than 20 to 50% compared to the test results. What can I change in the settings to have better convergence results in the analysis of this case. How many iterations, etc. I set the soft start for 50 iterations, the time blade passage. In solver already used to set ADV5 and low re kepsilon, it got better convergence results but diference to the laboratory verification 50% ;(.........please help

I attached screen with my model.

Please use plain text.
Product Support
Posts: 957
Registered: ‎08-25-2011

Re: How to improve rotating region anlysys !!!!

02-27-2013 02:32 AM in reply to: Bart84



There are a couple of things to try here:


  1. Have you also run it out for longer with just 3deg/timestep? This often helps hone in on the result.
  2. You likely need a longer outlet, we need to ensure we are properly capturing everything that is occurring and have no recirculation over a boundary - if you extended the boundary length by about 4x, that might help
  3. The final thing is probably mesh. Are the blades nicely refined? Is there a uniform mesh on the external surfaces of the Rotating Region? How good is the mesh overall, is there enough through the smaller outlet?



Jon Wilde
Please use plain text.
Product Support
Posts: 307
Registered: ‎08-31-2011

Re: How to improve rotating region anlysys !!!!

02-27-2013 12:36 PM in reply to: wildej

I would agree with what Jon listed as the Pressure and axial (Vx) velocity has not fully stabiliezed.

Aside from the mesh, you may want to run out another 100iterations  to ensure that the axial velocity stabilizes, then to dial in the results further as Jon mentioned run ~and addition 2-5 revolution with a timetep size or 3º/timestep.


On average rotating regin models will require between 5-20revolutions worth of simulated time to get the flow solution evolved/stable. The smaller timestep is used to capture the more detailed flow intricacies once the solution is steady.


Please use plain text.
Posts: 5
Registered: ‎01-05-2013

Re: How to improve rotating region anlysys !!!!

05-21-2013 01:48 AM in reply to: apolo.vanderberg
i was working on this problem and i solved that jon list.

Please use plain text.