|September 21, 2010, 04:14||
Join Date: Mar 2009
Posts: 12Rep Power: 9
You said above "My main concern is the CFdesign code does not maintain the steady state such as not maintain equal in and out flow rates (at inlet and outlet)." If you've not been able to fix that I would ignore the CFdesign results. However, assuming you have, and you've checked you've got a good mass balance in both codes I'd look carefully at the differences in the solutions to see if either exhibit any non-physical behaviour based on your engineering knowledge. Beyond that I'd coarsen the mesh in both cases and look to see how the results change to get a feel for which code is giving you results that are closest to mesh independent. You may be a long way off with first order on a tet mesh. I'd be inclined to go with the result that's the least mesh dependent. You could look at overall pressure drop as a measure of mesh independence. Hope that helps,
Dr John Parry, CEng CITP
Electronics Industy Manager
Mecahnical Analysis Division
Mentor Graphics Corporation
|February 9, 2012, 17:49||
Join Date: Feb 2012
Posts: 1Rep Power: 0
What I think is behind differences between CFX and Cfdesign is that Cfdesign uses the SIMPLE or SIMPLER algorithms and segregated solvers, which were state of the art in the late 80s. CFX on the other hand uses multigrid and a powerful coupled solver developed in the mid 90s. In brief I believe that the result is better converged equations.
What is especially uncomfortable about Cfdesign is the criteria it uses for convergence -it looks at plateauing of residuals as opposed to how many orders of magnitude their size has been reduced. I don't think this correct.
best of wishes,
|February 9, 2012, 20:49||
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 12,638Rep Power: 98
The coupled solver in CFX is frequently a good thing for steady state runs where it allows convergence in far less iterations than an uncoupled solver. But for transient runs an uncoupled solver is often faster as the small time steps required by a transient run mean the weaker p-v coupling in an uncoupled solver is more than compensated by the much faster time per coefficient loop.
But there are exceptions to these general principles as well - really you need to compare it for your specific case as it is case by case as to which is best.
|Thread||Thread Starter||Forum||Replies||Last Post|
|Pros and Cons for CFX, CFdesign, COMSOL||Val||Main CFD Forum||3||June 10, 2011 02:20|
|FloEFD CFDesign or CFX||Bigga||CFX||4||January 4, 2010 07:41|
|CFX or CFDesign||Ed Chavez||Main CFD Forum||8||October 18, 2007 04:26|
|PhD using CFX||Rui||CFX||9||May 28, 2007 05:59|
|CFDesign > CFX or ICEM CFD||Endlos||CFX||0||June 14, 2005 02:55|