CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   SU2 (http://www.cfd-online.com/Forums/su2/)
-   -   Problems on mesh adaptation (http://www.cfd-online.com/Forums/su2/124195-problems-mesh-adaptation.html)

ervasph September 29, 2013 11:57

Problems on mesh adaptation
 
I had tried the tutorial 13 for mesh adaptation and found some problems.

1. For feature based adaptation, the results of SU2-v2.0.0 are similar to tutorial, but the results of v2.0.6 are not correct. I guess the reason is that the restart files are different. The "restart_flow.dat" file of v2.0.0 has 12 columns, while v2.0.6 has only 8 columns and includes additional file head. I delete the file head of v2.0.6 and get the correct result.

2. For goal oriented adaptation, the same problem appears. I try to delete the file head of v2.0.6's adjoint restart file, but the results are not still compatible with v2.0.0.

3. Now I use version 2.0.0 to finish goal oriented adaptation. In tutorial, only upper grids at shockwave are refined and the refinement area is large, but in my run, the grids below airfoil are refined too and the range of refinement at shockwave is small.

thanks for any help.

fpalacios October 3, 2013 14:27

Quote:

Originally Posted by ervasph (Post 454147)
I had tried the tutorial 13 for mesh adaptation and found some problems.

1. For feature based adaptation, the results of SU2-v2.0.0 are similar to tutorial, but the results of v2.0.6 are not correct. I guess the reason is that the restart files are different. The "restart_flow.dat" file of v2.0.0 has 12 columns, while v2.0.6 has only 8 columns and includes additional file head. I delete the file head of v2.0.6 and get the correct result.

2. For goal oriented adaptation, the same problem appears. I try to delete the file head of v2.0.6's adjoint restart file, but the results are not still compatible with v2.0.0.

3. Now I use version 2.0.0 to finish goal oriented adaptation. In tutorial, only upper grids at shockwave are refined and the refinement area is large, but in my run, the grids below airfoil are refined too and the range of refinement at shockwave is small.

thanks for any help.

Thanks a lot for letting us know, as you have notice we have introduce some changes in the restart formats (mainly because of the new unsteady capability). Please stay tuned for new updates, and fork the latest version of the code https://github.com/su2code. We'll try to fix the problem today or tomorrow.

Best,
Francisco

ervasph October 11, 2013 10:04

2 Attachment(s)
hello developers,

I had tried on v2.0.8, the featured-based adaptation now work correctly, but the goal-oriented adaptation still has some problems.

The figure 1 is the result of v2.0.0, the figure 2 is the result of v2.0.8. 2.0.0's result looks reasonable, but 2.0.8's result is strange.

thanks for any help

ervasph October 12, 2013 11:34

The problem had been resolved. WRT_RESIDUALS should be set YES for goal oriented adaptation.


All times are GMT -4. The time now is 19:37.