CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   SU2 (http://www.cfd-online.com/Forums/su2/)
-   -   tutorial to step through a grid adaptation process?? (http://www.cfd-online.com/Forums/su2/112634-tutorial-step-through-grid-adaptation-process.html)

jentink February 1, 2013 11:16

tutorial to step through a grid adaptation process??
 
I am interested in learning how to use grid adaptation and am having problems getting mesh_adaptation.py to run.

Is there a tutorial somewhere that goes through step-by-step a grid adaptation process? I'm mainly interested in adapting to a shock for right now.

Thanks.

copeland February 5, 2013 04:02

Hey Tom --

Thanks for bringing this up! Mesh adaptation is a key feature in the SU2 package and I've been working on a tutorial that showcases the two of the most commonly used adaptation strategies: feature-based (using flow field gradients) and goal-oriented (using the adjoint variables as weighting factors in the adaptation strategy).

The tutorial will be ready very shortly (within the next 24 to 48 hours) and I'll be sure to let the community know as soon as it's available!



Regards,
Sean

copeland February 5, 2013 22:19

Hey Tom --

Just wanted to follow up with you. I've completed the Mesh Adaptation tutorial (currently tutorial #13 on our documentation page) and included a direct link below:

http://adl.stanford.edu/docs/display...esh+Refinement

In the process, I discovered some minor file-processing bugs in the python script. I've corrected those bugs and attached an updated mesh_adaptation.py script to the tutorial page. Please download this file, and overwrite the old mesh_adaptation.py script located in your $SU2_HOME/SU2Py directory and in your $SU2_RUN directory. This updated file will be distributed with our monthly development release scheduled in the near future.

If you have feedback on the tutorial, I'd be happy to hear your comments and suggestions.


Regards,
Sean

jentink February 11, 2013 15:59

I was having computer issues last week. Will try the tutorial this week and let you know.

Thanks!

EMolina February 13, 2013 11:39

1 Attachment(s)
Hi Sean and Trent

Sorry for posting my issues on the wrong topic. So here we go...

I was having problems with this tutorial when I compile the SU2 with all the options (MPI, METIS and CGNS), so Trent advised me to download the update version of SU2v2.0. I downloaded and compiled it with all the same options and the problem persists. See bellow the terminal output:

molina@Desktop-Molina:~/Documentos/Programas/SU2v2.0/TestCases/mesh_adaptation$ python mesh_adaptation.py -f adapt_NACA0012.cfg -p 1 >> output_TECPLOT_MAC_PARALLEL.txt
*** The MPI_Comm_rank() function was called before MPI_INIT was invoked.
*** This is disallowed by the MPI standard.
*** Your MPI job will now abort.
[Desktop-Molina:5070] Abort before MPI_INIT completed successfully; not able to guarantee that all other processes were killed!
Traceback (most recent call last):
File "mesh_adaptation.py", line 387, in <module>
main()
File "mesh_adaptation.py", line 59, in main
options.save_all )
File "mesh_adaptation.py", line 354, in mesh_adaptation
libSU2_run.SU2_MAC( Config_MAC_filename, partitions )
File "/home/molina/Documentos/Programas/SU2v2.0/TestCases/mesh_adaptation/libSU2_run.py", line 37, in SU2_MAC
run_command( the_Command )
File "/home/molina/Documentos/Programas/SU2v2.0/TestCases/mesh_adaptation/libSU2_run.py", line 106, in run_command
raise Exception , "SU2 process returned error '%s'" % return_code

I replaced the original meshadaptation.py file with the one from tutorial's page. After, I decided to compile in another folder a serial version without any options to see what was happening, chossing a SU2_MAC_SERIAL or SU2_MAC changing the libSU2_run.py.

Here is the results:

Configure file with Tecplot - and the SU2_MAC_Serial -> Run ok
Configure file with Paraview - and the SU2_MAC_Serial -> Segmentation fault (core dumped)
Configure file with Tecplot - and the SU2_MAC -> Same problem as above (MPI issues)
Configure file with PARAVIEW - and the SU2_MAC -> Same problem as above (MPI issues)

For me, we have two problems here one with paraview output and one with this strange problem with MPI even when running the code with -p 1 option.

Last question: Why we have this WARNING: unrecognized option in the config. file: GRID_DEFORM_SOLVER.?

Please see the attached files with the output error from terminal.

Thanks in advance

Eduardo

bensciens March 15, 2013 05:17

I am getting the same MPI fault as above. Is there a solution to this?

The error is caused by calling SU2_MAC.


All times are GMT -4. The time now is 00:55.