[wel]

Greetings!

The SU2 version I'm running is from the feature-AdjTNE2 branch. I'm using TNE2_NAVIER_STOKES. I have two problems:

1. If I don't specify VISC_NUM_METHOD_TNE2 the solver terminates with error "You must define a viscous numerical method for the TNE2 Navier-Stokes equations!!" but when I do it terminates with error "VISC_NUM_METHOD_TNE2: invalid option name".

2. As a work around I installed SU2-master and ran the same case but it returns an error:
" ***Cannot bisect a graph with 0 vertices!
***You are trying to partition a graph into too many parts!
Finished partitioning using METIS. (0 edge cuts).
"

If this is to be corrected in the code, where do I do so?

Regards,
Wel

[hlk]

Originally Posted by wel

Greetings!

The SU2 version I'm running is from the feature-AdjTNE2 branch. I'm using TNE2_NAVIER_STOKES. I have two problems:

1. If I don't specify VISC_NUM_METHOD_TNE2 the solver terminates with error "You must define a viscous numerical method for the TNE2 Navier-Stokes equations!!" but when I do it terminates with error "VISC_NUM_METHOD_TNE2: invalid option name".

2. As a work around I installed SU2-master and ran the same case but it returns an error:
" ***Cannot bisect a graph with 0 vertices!
***You are trying to partition a graph into too many parts!
Finished partitioning using METIS. (0 edge cuts).
"

If this is to be corrected in the code, where do I do so?

Regards,
Wel

For the first problem, since VISC_NUM_METHOD_TNE2 is in config_structure.cpp for that branch, I think this error might mean either that the config_structure.cpp file didn't get updated when you checked out the branch, or that there was an issue with compiling the code after switching branches.
Check to make sure that VISC_NUM_METHOD_TNE2 exists in common/src/config_structure.cpp. It should be, if you have checked out feature_adjTNE2 - if not, try checking out that branch again. You should also make sure the "make clean" before installing the new branch.

For the second problem, that indicates a problem with your mesh. The su2 mesh format is ascii, so you should be able to open it and see if there are any vertices. It could also indicate a problem with getting Metis compiled and linked to SU2.