CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Questions about STAGE interface (http://www.cfd-online.com/Forums/cfx/75480-questions-about-stage-interface.html)

pad April 27, 2010 09:13

Questions about STAGE interface
 
Hello all,

I am calculating a impeller-volute design in which I am using the stage interface which I have some difficulties with.

I have done simulations on similar geometries in which the stage interface seems to work OK but on some design the run reports an overflow after 20-30 iterations. I have tried tuning different parameters but I have not succeeded.

Apart from the fact that the stage interface isn't working properly in off-design flow points (according to the manual) are there any restriction in the use of the stage interface? E.g. the distance from the interface to rotating / stationary walls? Should the flow be mostly radial, i.e. perpendicular to the interface?

Any comments appreciated, thanks!

Kind regards,
PAD

ghorrocks April 27, 2010 18:35

There are few restrictions on where to put interfaces and it can handle flow in any direction. Having said that, if possible you put interfaces away from critical regions of the flow if you can help it - but it is no great disaster if that is unavoidable, they are generally quite stable and accurate.

If you are getting divergence then there is a fault in your simulation setup - If it seems to be related to the interface I would suggest you mesh is too coarse.

pad April 28, 2010 04:33

Thanks for the reply.

I did manage to get a converged solution using the frozen rotor approach. However this will not get me the correct efficiency of the design as this is dependent on the impeller blade position relative to the tongue of the volute.

As mentioned I did other simulations that worked with similar meshes. How would you suggest I refine the mesh? Is there a direction which is more important than others, e.g. perpendicular to the interface?

Regards
PAD

ghorrocks April 28, 2010 09:08

That is problem dependent. Assuming you are using a hybrid prism/tet grid then refine the grid isotropically.


All times are GMT -4. The time now is 05:14.