Inviscid Onera
 

I computed the results using both MPI and SU2_CFD, using PARAVIEW output. The non-MPI version does not seem to provide Mach_Number variable in the surface_flow.vtk. The MPI version does output Mach_Number, but it does not appear correct.

The lambda pattern is visible, correct, and identical for Pressure_Coefficients in both MPI and non-MPI versions of surface_flow.vtk.

What could cause these errors? I have tried using Paraview 3.14 (Ubuntu default) and Paraview 3.98-1 compiled manually with no effect.

I have since computed the turbulent onera case in parallel using the default mesh and received the same errors in Paraview when opening surface_flow.vtk

I have computed tutorials 1-5 and 7 without issue, so I opened the flow.vtk (volume mesh) in Paraview for inviscid and turbulent Onera cases, the former computed in serial and parallel, to find that all the information was available and qualitatively correct (M_max 1.53 M_min 0.109), even though Mach_Number is missing from the surface_flow.vtk output. I'm assuming the larger range of Mach number is due to the range of Mach number throughout the volume as opposed to the surface values shown in the tutorial.

I didn't find any part of the configuration file that allows one to select/omit output variables to the Tecplot or Paraview files. Barring that I'm assuming this is a bug in the Paraview output?

Hi

I have received same error for turbulent Onera using fine mesh available in test cases, but only when I run multigrid.... if multigrid is off the Paraview output is good.

Thanks a lot for your interest in SU2,

The new version of SU2 (2.0.2) counts with a new I/O based on Tecplot (binary and ascii), and CGNS (that can be used by paraview users). As far I know VisIt counts also with a Tecplot reader (https://wci.llnl.gov/codes/visit/).

Anyway, using the current structure of the code it is very easy to create a paraview output format from scratch. If you are interested, my recommendation is to use output_WriteTecplotASCII.cpp as baseline and write something similar for paraview. Once you finish it, if you are interested, we can add the code to the official release (with the appropriate credits for the authors).

Best,
Francisco

VisIt is beter than ParaView in reading Tecplot binary format.