I need some help or explanations to resolve some issues regarding turblulent modeling in CFX, I am studying an external flow problem and comparing with wind tunnel test. I have tried different schemes in the program : The SST-with high resolution is pretty correct predicting the Lift. K-epsilon with Upwind provides (although still off)the closest to the Drag and The SST with specified blend is in between. I will like to know if anyone knows how the Torque is being calculated in the program, secondly I wonder if the High resolution can be improved. Perhaps some one has some experience with this. Thanks.

Hi,

You might get better accuracy using the specified blend factor spacial differencing rather than high resolution. Use a blend factor of 1.0 to use purely second order differencing. This may give an increase in accuracy over high-resolution differencing.

The torque reported in the output file is around the default coordinate system, that is the point 0,0,0 in your mesh model.

Regards, Glenn Horrocks

Thanks Glenn, Suppose I desire the Torque about a different location, like the centroid of my object in the domain ? I am gradually adjusting the blend factor to see its influence.

Hi,

In CFX-Post you can use the Function Calculator to evaluate the torque on any locator about any global axis, and also about any user Coord Frame axis created in CFX-Post.

Regards,

Johnson