Parallelization and output

[Alexmine]

Dear OpenLB community.
1.I want to use smagorinsky model to simulate turbulence flow (with curve boundary),the following error occurs if the resolution is insufficient:
[Timer] step=306000; percent=63.75; passedTime=113173; remTime=64353.5; MLUPs=67.7476
[LatticeStatistics] step=306000; t=10.2; uMax=0.01; avEnergy=-nan; avRho=-nan
[getResults] ; drag=-nan; lift=-nan
[getResults] yPlusMax=2.22507e-308
“./tmp/imageData/data/_EuklidNormphysVelocityiT0306000.p” line 21: warning: matrix contains missing or undefined values
“./tmp/imageData/data/_qualityiT0306000.p” line 21: warning: matrix contains missing or undefined values.
The higher the Reynolds number, the higher the resolution required.why?

2.I used parallel computation based on cpu-openmpi to reduce the huge computational time generated by high resolution. At first it works well,however, with the increases of resolution , errors occured as follows:
Primary job terminated normally, but 1 process returned
a non-zero exit code. Per user-direction, the job has been aborted.
————————————————————————–
————————————————————————–
mpirun noticed that process rank 5 with PID 0 on node ALEX-MS-7D17 exited on signal 9 (Killed).
I don’t know where exactly the problem is?

3.The simulation data is stored in the VTK data format and can be further processed with ParaView, however, the output of my simulation is incomplete, every time I use Paraview to process the output of the calculation, only an eighth or a quarter of the calculation field can be display. My simulation results are asymmetrical and I need the entire computational domain.
What should I do?

sincerely,
Alex

• This topic was modified 2 years, 3 months ago by Alexmine.

[mathias]

Dear Alex,

your simulation has diverged. I see that form the line
[LatticeStatistics] step=306000; t=10.2; uMax=0.01; avEnergy=-nan; avRho=-nan
with the “nan”.

You should checkout the last data output and look for a possible cause of the instablity. It may originate from a boundary or the bulk. If the source is the bulk, refine your mesh or increase the smago constant and if the source is the boundary extending the inflow or outflow region or intruducing a fringes might be an option.

Best
Mathias

[Author]

Posts