Using implicit dual-time marching and SS option (10,11) leads to immediate segmentation fault errors

[plaad]

Hi All,

Just a quick concern I have had/wanted to ask about. I currently use UCNS3D on a RedHat HPC system at my university.

1. When conducting a transient Euler equations run, my code works for explicit RK schemes on a new grid, but upon using option 11, the code immediately fails without producing any such Errors.dat file so troubleshooting is tricky. I am able to run the DDES cylinder case fine and see it uses option 11.

I wanted to ask if there are specific parameters/flags (in parameters.f) to be specified to use option 11 as the only other difference besides flow conditions/schemes and grid is the code config of 1 being used in the example whereas mine uses 0. Further, are there inputs in the UCNS3D.DAT file which should be added/removed if trying to use this time-marching option? As in, I included all options needed for implicit time marching (non-zero entries for inner iteration number and pseudo time step size), but maybe there are conflicting flags?

2. For SS simulation, it is the same problem in that the example (SS RANS NACA example) runs fine, but when using option 10 on this working grid, the code immediately fails. I am not sure if certain flags conflict in the input deck, but if someone can tell me what factors can cause/control this behavior it would be appreciated.

I can share my grid here (just rename as .msh) which is simple expansion corner to be run at a supersonic conditions at STP (T=298 K, P= 101325 Pa) conditions; again, unsteady, explicit marching works, but SS and implicit dual-time marching do not for some reason.

The error occurs after the grid is processed as the .case file converts to the three GRID files and then an output file is written, but the solution never proceeds due to the immediate segmentation fault error.

My grid is 3d with 1 cell in the spanwise direction to enforce a 2d flow (with corresponding symmetry conditions) since I do not have a direct way to produce a 2d case file in the given format. If anyone can give some details as to what is going wrong it would help.

grid.txt

[plaad]

I have attached my two input deck for the two options (10, 11) here. I changed the name of the UCNS3D.DAT for each to differentiate between the two input decks for posting on here (now as .txt).
UCNS3D_SS.txt
UCNS3D_ImplicitDualTimestepping.txt