CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   rotatingWallVelocity vs. SRF vs. MRF vs. AMI/GGI (http://www.cfd-online.com/Forums/openfoam/106847-rotatingwallvelocity-vs-srf-vs-mrf-vs-ami-ggi.html)

Soheyl September 10, 2012 19:14

rotatingWallVelocity vs. SRF vs. MRF vs. AMI/GGI
 
1 Attachment(s)
This is a fundamental question that I need to think it through. I'm trying to model the flow inside the attached geometry. The mill (in white) rotates at a constant speed inside the casing (in red). The casing is a simple cylinder, and the whole system is filled with a single-phase liquid. Also, there is not protrusion on the outer surface of the rotating mill/shaft.

Can I simply model this using rotatingWallVelocity and say, simpleFoam? Would it automatically take care of the coriolis and centrifugal forces? (Notice that I will be modeling this in 3D). Also, considering the holes on the mill surface, would these methods work, or do I need a more advanced dynamic mesh method? Any thought/suggestion is highly appreciated.

kwardle September 11, 2012 13:12

Presumably the casing is stationary? Since you have only one moving domain and there are no baffles on your stationary cylinder (as near as I can tell from the image), there is no need for AMI/GGI. As you suggested, I think rotatingWallVelocity might barf because you have the holes in the rotating part--it is worth a try, but I think it wants to put just a tangential velocity on the surface which is not valid for the interior of your holes. SRF or MRF should have no problem at all. If there is not already a solver available, it is not too tricky to add in MRF capability to simpleFoam or whatever you will be using.
-KW

Attesz September 12, 2012 12:58

I think rotatigWall will give bad results. I would use MRFSimpleFoam, but I would cut the domain into several domains very close to the plates. This would be the most accurate solution - if it is a steady phenomena, of course.

kwardle September 12, 2012 14:19

If there are no baffles on the stationary part, then whether it is one rotating domain or multiple as Attesz has suggested should give the same results.


All times are GMT -4. The time now is 13:44.