Problem of torque prediction in CFX
 

Now I'm simulating an outdoor unit fan of HVAC. And I calculated it at three different speed 500RPM, 700RPM and 950RPM. The pressure rise of all three speed is reasonable matching the test data. As to the torque of blade, it is good at 500RPM. But at 700RPM the value of torque is about 15% over predicted. At 950RPM the torque is about 20-30% over predicted. The whole case use the same mesh and the turbulence model is k-epsilon. I checked the Yplus in Post and found it is about 12-27 at 500RPM, 15-38 at 700RPM and 20-50 at 950RPM. Anyone who have ever met this problem and how could I got a more accurate torque? Thanks.

Re: Problem of torque prediction in CFX
 

Hi,

Have you done all the normal checks? I mean Mesh size, convergence, physical models, timestep etc.

Glenn Horrocks

Re: Problem of torque prediction in CFX
 

Glenn, I've checked all the things you mentioned. And I make sure the convergence is good. I do not take the mesh-independence test and I just gave the mesh size according my experience so I do not know if the case is mesh independent.Is it necessary to make some more mesh to test this? I'm wondering since the low speed case could give the right toruqe why the high speed case could not give the right toruqe? Is that related to the mesh or the Yplus or other things?

Re: Problem of torque prediction in CFX
 

Hi,

It is typical for a low speed case to run OK on a coarse mesh but a higher speed case requires a finer mesh. This is because the gradients get larger, boundary layers smaller etc. Sounds like a mesh sensitivity check on the high speed run is required.

Glenn Horrocks

Re: Problem of torque prediction in CFX
 

Patrick,

How are you determining the "right" torque?

-CycLone

Hi,
We have the same issue of torque prediction. Did you conclude anything from your study. I am also simulating HVAC outdoor unit fan.
Thanks in advance for your help.
Please reply at trilok99@gmail.com

We have now written an FAQ on this: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F