|March 24, 2017, 13:36||
Unexpected restart solutions
Join Date: Dec 2015
Posts: 14Rep Power: 3
I'm using SU2/5.0.0 and I'm trying to get a wing polar at different AoAs. I assumed that once a calculation for a given AoA1 (say 8) was finished I could accelerate the calculation for AoA2 = 9 by simply restarting the solution and changing the angle in the cfg file from 8 to 9.
However this leads to a very quickly convergence .... to the values of AoA = 8!
I tried running AoA = 9 from ground zero and it yields very different results than when I try restarting based on AoA = 8.
I was able to do this kind of procedure before but now I can't.
I'm thinking that it has to do with my convergence criteria (which is relatively strict). It makes me think that starting from a previously converged state makes it so that it takes too long for the new information (change of AoA) on the farfield to reach the cells around the wing... which "converge" faster than the new information can arrive.
btw, I use a Cauchy criteria on the CL for 4000k iterations
Has anybody come across this? Any advice?
|Thread||Thread Starter||Forum||Replies||Last Post|
|FLUENT installation on UBUNTU 12.04 (LTS)||teymourj||FLUENT||2||March 1, 2017 23:24|
|Restart to adjoint for Unsteady Turbulence Cylinder flow||yweng3||SU2||0||October 7, 2016 14:50|
|ANSYS Workbench on "Uncertified" Linux Distros||hsr||CFX||262||September 7, 2016 21:05|
|error 262144 and unexpected disconnect when run on 2 nodes||Emily1412||SU2 Installation||1||April 10, 2014 17:08|
|Problem with multiframe restart of two-way fsi coupled problem||Dimone||CFX||26||December 20, 2011 11:10|