CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   SU2 Shape Design (http://www.cfd-online.com/Forums/su2-shape-design/)
-   -   Problem running ONERAM6 analysis on finer grids (http://www.cfd-online.com/Forums/su2-shape-design/119784-problem-running-oneram6-analysis-finer-grids.html)

Akash C June 25, 2013 04:15

Problem running ONERAM6 analysis on finer grids
 
Hi,

I am using SU2 v2.0 for completing the tutorials. When running the oneram6 tutorial, the analysis fails to run on any grid other than the coarse hexa grid. There are three other grids provided viz., hybrid coarse, hybrid fine and hexa fine. But on all these grids the analysis throws the following error after some iterations

Code:

Negative pressure at point: 0.876234 -0.985696 0.00872704               
Negative pressure at point: 0.887721 -0.957623 0.00686946               
Negative pressure at point: 0.876234 -0.985696 0.00872704               
Negative pressure at point: 0.887721 -0.957623 0.00686946               
CSysSolve::modGramSchmidt: w[i+1] = NaN                                 
terminate called after throwing an instance of 'int'                   
Aborted (core dumped)
[akash.chaudhari@TSPLDEV-11 SU2Py]$

I ran into a similar problem with the low mach no. euler ONERAM6 tutorial but ignored it at that time.
Can someone please help resolve this issue?
Thanks.

economon June 28, 2013 18:35

Hi Akash,

It looks as if the calculation is going unstable (after getting negative pressures, the implicit linear solve breaks down). Have you tried adjusting the CFL and other options, like multigrid, down to more conservative levels?

Also, are you using the most recent version of the code? Developer version 2.0.4 contained some notable improvements in the stability of the code, especially when using the Krylov linear solvers.

Hope this helps,
Tom

Akash C June 29, 2013 02:14

Thanks for your reply Tom. I sure will try out the suggestions you mentioned with my current version i.e. v2.0.

For the developer versions, I just merged/overwrote the trunk folder provided in the developer versions(v2.0.5) with the trunk folder of v2.0 and then compiled the code using python script as described in the manual. Hope I did it right and please correct me if I am wrong.

The manual being old(v1.1 I guess) and also the python build script, it does not compile the SU2_SOL module and neither could I get any info as to how it would affect my analysis. And I also could not compile the SOL module separately as the process stopped after an error. I hope this module isn't required for any crucial function in the analysis. I will try and post the error as soon as possible.

Besides this the code really seems great and I am happy with way it has been converging in no time for my cases.

Thanks,
Akash

shirazbj June 29, 2013 02:57

Hi,

I can't compile the sol module on win7 with gcc for win either.

I think it is only needed for parallel computation. After finish simulation, sol module sort of Combining the result.

Regards,

Cean


All times are GMT -4. The time now is 00:35.