windPlantSolver.ALM solver blow-up about sowfa HOT 4 OPEN

Hello,

I am replying to this for future references and people who have this same issue.
I had this same issue and spent several weeks trying to solve it. I had a refined mesh, even I used snappyHexMesh tool. The mesh had no errors, still, the run blew up after a few iterations in parallel run. I noticed that if do the simulation in serial mode, everything is fine and that was weird somehow. Finally, I solved the issue by changing a few parameters of my mesh (e.g. cell size and refinement parameters). So, just change your mesh a little bit in terms of refinement areas and cell size and see if it works for you.

Regrads,
Reza

from sowfa.

Dear sonddol,
I had the same issue one year ago, but I was not able to solve it yet.
If you are facing my same problem, I think you are running the simulation with the wind not aligned with the principal axes (not aligned with the mesh) and you performed mesh refinements when running the windPlantSolver.
The issue for me is that the velocity shows a large somewhere on the borders of the refinement boxes and the whole simulation blows up.
Let me know if these are your settings.
Cheers
Paolo

from sowfa.

Hi Sonddol,

Paolo makes a good point. Your Courant number is too high. What does your mesh look like, and what is your timestep? Do you have any flow field plots that you can show?

Cheers,
Bart

from sowfa.

Hello,

I met the same problem when I used windPlantSolver.ALM. And the above reply say that it's about timestep and refinement.I used ABL.neutral as my precursorDir,and the Domain size and number of cells are 3km3km1.02km and 300300100.
In ALM's setUp ,I set 3km3km1.02km and 250250100; in controlDict, the deltaT is 0.04;in topoSetDict.local.2 and 1 the sourceInfo are:



And the turbines baselocation ars:(1500.0 870.0 0.0),(1500.0 1500.0 0.0),(1500.0 2130.0 0.0) and NacYaw is 270.
The solver run well between 20000s to 20175s(perhaps，I didnot save the log) then it failed cause mpirun noticed that process rank 42 with PID 0 on node ubuntu exited on signal 8 (Floating point exception).
Then I change the timestep(delta T) to 0.05 but it failed at 20097.25s.
Now I would like to ask how much timestep and refinement are appropriate, and how should the parameters be determined.

Thank you! @Bartdoekemeijer @reza2031
windPlantSolverALMlog.txt

from sowfa.