Adjoint solver with INC_SOLVER and periodic boundaries

[hareshramn]

Hello,
I'm trying to run a shape optimization simulation for a 2D external flow over a hydrofoil using INC_RANS and Discrete adjoint solver. I have periodic boundary conditions in one direction.
The flow solver works fine and I get a converged solution.
I get a segmentation fault error when I run the adjoint solver using SU2_CFD_AD.
I have attached a screenshot of the error.

I also find that the error goes away when I use Outlet BCs instead of the periodic BCs.

Any help will be appreciated.
Thanks!

[TobiKattmann]

Hi, can you please attach the case with mesh if possible?

You can also take a look here which is a INC_RANS with periodic case where a working Discrete adjoint is available. At the bottom there is a link to a 2D CHT case where a README.md explains a bit more. I.e. to my knowledge there is no problem per se with INC_RANS + periodic +DiscAdj

In case you compiled the code yourself. Please attach --buildtype=debug to the meson.py ... command. That way you will get much better information where the segfault occurs in the code

[hareshramn]

Thank you for the reply. I have attached my mesh and cfg file.
planar_2D.zip

I was able to run the shape optimization script without the segmentation fault when I updated to version 7.2.1 and SA turbulence model. Have not really found the source of my issue though.

I will compile the code with the debug flag on and check.

Thanks for the info!

[TobiKattmann]

👍 If you cannot reproduce the segfault on the latest version and you are happy with the solution, I guess we could call it day :D
I am not aware of any fixes to the periodic bc lately ... but lets wait for your results