Hi,

I am trying to simulate flow in a hydro-cyclone. But irrespective of what turbulence model I use, I continuously obtain forced vortex rotation. (Tangential velocity keeps increasing away the center axis with radius r and becomes zero at the wall at a sudden). In reality, I should get the forced vortex trend only in the central core region and then free vortex should start. But I am not able to get it so far. I have tried k-e, RSM etc models.

To plot tangential velocity at a particular axial height, I create one line in the cyclone's cross section (say horizontal plane XY plane) passing through the center and along the X axis (Y=0). And then I plot 'Velocity v' to read the tangential velocity (swirl velocity). Is my procedure of plotting is correct?

Please help !

Thanks a lot, KM

Hi,

This is a classic test of the k-eps turbulence model, and it is well known that it predicts the flow you describe and fails to correctly predict the central vortex core. RSM models can predict the vortex core but are limited by the RANS assumptions - generally the flow cannot be "averaged" into turbulent and bulk flow components.

You need to go to LES models to get good accuracy for this type of model.

Glenn Horrocks

Hi KM,

You method is correct. You can also get the tangential velocity by clicking on the turbo tab in Post and setting the axis of rotation to align with your cyclone, then select "Calculate Velocity Components" to get the axial, tangential and radial velocity components (they will be presented as scalar variables).

As for getting a free cyclone, verify that your mesh is fine enough. If your mesh is unstructured, consider slicing the cyclone near the top so you can sweep a mesh from top to bottom. The sweep method (in the workbench meshing application) will allow you to get more resolution in the transverse direction while reducing keeping a larger element spacing in the vertical direction.

-CycLone

Or SAS-SST, or DES.

Thank you very much Glenn and CycLone. Your answers help me a lot ! KM