|
[Sponsors] |
![]() |
![]() |
#1 |
New Member
Michael Levene
Join Date: Jan 2025
Posts: 6
Rep Power: 2 ![]() |
Hello,
I am very new to CFD in general and SU2 in particular. I have a simple simulation I would like to run on a flat plate in 3D. I have been getting divergence errors with NaN detected before the first iteration. I have reduced my geometry down to just a simple box, with no features other than inlet, outlet, symmetry sides, farfield top and MARKER_HEATFLUX bottom, and still I get this error (compressible RANS simulation). Even if I go to an incompressible Euler simulation with no-slip boundary on the bottom, I get the same error. I have performed enough variations to identify that the problem is definitely the bottom boundary condition. If I make it farfield, I get no error. If I make all sides of the box farfield or symmetry and place an object with no-slip boundaries inside the box, I get no error. It seems to work fine for 2D simulations, but is there something about 3D simulations in SU2 that you cannot have a wall boundary? Thanks for any help, Michael |
|
![]() |
![]() |
![]() |
![]() |
#2 |
Senior Member
bigfoot
Join Date: Dec 2011
Location: Netherlands
Posts: 775
Rep Power: 21 ![]() |
Can you post your mesh and cfg file? This should work so something in your setup must be wrong that is not being picked up by the preprocessor. It can also be that your case simply diverges because your settings are too aggressive. How do you make your mesh?
|
|
![]() |
![]() |
![]() |
![]() |
#3 |
New Member
Michael Levene
Join Date: Jan 2025
Posts: 6
Rep Power: 2 ![]() |
Hi,
I have tried a lot of different options, as well as meshes. I make my geometry and mesh in GMSH (I got an error trying to upload it here). At this point, I don't care about accuracy of results, since ultimately I am trying to simulate something more interesting than a flat plate; I have just reduced my model to a flat plate for debugging purposes. I do not have transfinite mesh definition, or sufficient resolution near the wall, in the latest version but previous versions that did have better resolution gave the same error. Also, the run fails before it completes a single iteration, so I have been thinking it is not likely to be a mesh problem. And using a terrible mesh with a non-slip object inside the box doesn't crash. To make the CFG file, I copied a lot from the ONERA M6 wing tutorial (you can see the header is still present in the file test.txt). Thanks for taking the time to look at this, I really appreciate it. Michael |
|
![]() |
![]() |
![]() |
![]() |
#4 |
New Member
Michael Levene
Join Date: Jan 2025
Posts: 6
Rep Power: 2 ![]() |
Sorry, I realize the test config file I uploaded previously still had the Eler boundary condition on the bottom surface. Here it is the the HEATFLUX boundary.
|
|
![]() |
![]() |
![]() |
![]() |
#5 |
New Member
Michael Levene
Join Date: Jan 2025
Posts: 6
Rep Power: 2 ![]() |
Apologies, here is the correct file.
|
|
![]() |
![]() |
![]() |
![]() |
#6 |
Senior Member
bigfoot
Join Date: Dec 2011
Location: Netherlands
Posts: 775
Rep Power: 21 ![]() |
you have the gmsh .geo file as well?
|
|
![]() |
![]() |
![]() |
![]() |
#7 |
New Member
Michael Levene
Join Date: Jan 2025
Posts: 6
Rep Power: 2 ![]() |
Thanks for your offers of help on this, I found the problem. The setting of REF_AREA = 0, which worked in the Turbulent ONERA M6 tutorial and I understand is supposed to imply an automatic calculation, in this case caused the error. I assume it lead to a divide-by-zero error somehow, but I do not know anything about how the automatic calculation works to really understand why setting it to zero worked in the tutorial, but not for my case. It seems it just used zero as the area instead of doing a calculation.
|
|
![]() |
![]() |
![]() |
![]() |
#8 |
Senior Member
bigfoot
Join Date: Dec 2011
Location: Netherlands
Posts: 775
Rep Power: 21 ![]() |
Glad to hear that you have found the problem.
It might be that ref_area=0 does not work for your geometry because it needs to have a positive nonzero area in a certain plane (2D: y-area, 3D: z-area). So probably for your mesh, the computed reference area or reference length is zero (you can check the SU2 output of the failed case). Can you still share the gmsh file, then I can reproduce the error and add a proper error message to SU2 when this happens. |
|
![]() |
![]() |
![]() |
![]() |
#9 |
New Member
Michael Levene
Join Date: Jan 2025
Posts: 6
Rep Power: 2 ![]() |
I wasn't able to upload the .geo file, but I did upload the GMSH file, although the extension is .txt
Can you use that? |
|
![]() |
![]() |
![]() |
![]() |
#10 |
Senior Member
bigfoot
Join Date: Dec 2011
Location: Netherlands
Posts: 775
Rep Power: 21 ![]() |
Thanks, I could reproduce the error. I'll look into this.
|
|
![]() |
![]() |
![]() |
Tags |
nan error |
Thread Tools | Search this Thread |
Display Modes | |
|
|
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
rhoSimpleFoam High Pressure Cell Crashes Simulation | NorthCFD | OpenFOAM Running, Solving & CFD | 0 | March 3, 2023 05:02 |
laplacianFoam with source term | Herwig | OpenFOAM Running, Solving & CFD | 17 | November 19, 2019 13:47 |
Floating point exception error | lpz_michele | OpenFOAM Running, Solving & CFD | 53 | October 19, 2015 02:50 |
Cannot run the code properly: very large time step continuity error | crst15 | OpenFOAM Running, Solving & CFD | 9 | December 14, 2014 18:17 |
Moving mesh | Niklas Wikstrom (Wikstrom) | OpenFOAM Running, Solving & CFD | 122 | June 15, 2014 06:20 |