Unreliable Torque in CFX
 

Hello All,

For a fluid coupling device that is just a closed rotor-stator system, I have found CFX to show good results for one model (within 5%), BUT for another 2 devices of similar construction it is 3 orders of magnitude off with the same procedure!

I have gone thru the following checks:
1. Check that all meshing and setup for all geometries are the same.
2. Check several mesh sizes from a total of 1-3 million for a single period.
3. Steady state as initial guess with transient to follow once decent convergence is reached. Time step is determined by (2*pi/number of periods)/rotational speed. 10 coefficient loops, second order everything.
4. Convergence is always to at least 5e-6.
5. Axis checked and rechecked to make sure its always rotating about the proper circle center and at the proper lever arm.
6. Boundary conditions same as successful run, physically realistic.
7. Rotation in correct direction (i.e. right hand rule).

I ultimately feel like there is some trick to using the torque feature, or it just plain doesn't work for all cases. Any help would be immensely appreciated.

Thanks

Hi,
Are you checking the torque about you axis of rotation in all cases. See it in post and not in a monitor point ( if you are doing that now )