CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   transient analysis of a free surface (http://www.cfd-online.com/Forums/cfx/24064-transient-analysis-free-surface.html)

Juan Martin Catelén June 1, 2007 10:06

transient analysis of a free surface
 
What will happen if I set an adaptive mesh on a transient running? Will it last forever? I imagine the mesh adapting in each zone where the surface is flowing. What should I do to have a good result? Should I make a fine mesh in all the zones of the geometry?

Joe June 1, 2007 11:55

Re: transient analysis of a free surface
 
CFX doesnt allow mesh adaption in transient analyses.

Alex June 1, 2007 12:59

Re: transient analysis of a free surface
 
This may not be the group to ask, but I will anyways since I have been wondering about this topic. Anyone have any idea why CFX doesn't do this?

On the one hand, I can see requiring the steady state simulation for "mesh convergence," but at the same time you can use the same procedure and gradients in a transient analysis to converge the mesh for each time step in a dual time stepping scheme. I could use this feature myself in the same application: a transient free surface bubble column simulation. Also, I have seen other solvers that do utilize mesh adaption of transient solutions (Craftech/ Sandy Dash touts this as one of their competitive advantages.)

A very smart CFD engineer with many many years of experience once cautioned me against automatic mesh adaption altogether on the principle that if you begin with a coarse grid and get a poor solution, then refine the mesh based on that poor solution, you may be forcing the final result towards something that is incorrect or unphysical. I open the floor for comments...

-Alex


brunoc June 1, 2007 13:08

Re: transient analysis of a free surface
 
Well, actually CFX does it, but the GUI won't let you turn it on. It has to be done through CCL. I believe this is not allowed because in mesh adaption variables are linearly interpolated between old and new meshes. This does not guarantee conservation. On steady state it doesn't really matter since intermediate results may not be conservative, but in a transient simulation it can be a big problem. I believe this is the same reason why CFX does not re-mesh (as Fluent).

Joe June 1, 2007 13:52

Re: transient analysis of a free surface
 
Fluent does offer MA for transient flows. CFX, annoyingly, does not.

Lee Axon (Point Zero Consulting Ltd) June 4, 2007 17:19

Re: transient analysis of a free surface
 
Mesh adaption is available for CFX transient calculations. I have a document from support regarding this, drop me a mail if you want a copy.

Cheers,

Lee.



All times are GMT -4. The time now is 11:00.