[stardust23]

Dear all,
I'm getting crazy because of unrealistic results I got from an incompressible simulation with simpleFoam.
The domain is the one reported in the attached picture, a wall bounded flow over a flat plate (on the domain upper endwall I applied slip condition). I had to check the turbulence decay along the domain to control that the lenght-scale chosen is correct. The problem is that instead of lowering down its value, k and hence the turbulence intensity increases in the X direction.
You can find the case file (ready to be used) in this tar archive: https://www.dropbox.com/s/szkz4vymu9...nline.tar?dl=0
I hope that someone can find what's wrong in my setup because I couldn't...


Thanks in advance!

[stardust23]

Update: I tried to run the simulation with a commercial solver (starccm+), using the same grid, same BCs and, as far as it was possible, the same models.
Find the comparison in terms of TKE streamwise trend (X-direction) in the attached plot.
Actually the result I got from STAR looks way more similar to what I would expect and kind of mines the reliability of OF for the application I'm dealing with. I wrote about the bug on openfoam.com bug section, they told me that will consider the thing but it will take some time.

[alexeym]

Hi,

You do not have convergence criteria in fvSolution:

SIMPLE: no convergence criteria found. Calculations will run for 2500 steps.

so TKE distribution can be quite bizare.

Could you add OpenFOAM version you are using? I have tried running you case with OpenFOAM 5.x and got FPE in turbulence model on the first iteration.

[stardust23]

Originally Posted by alexeym

Hi,

You do not have convergence criteria in fvSolution:

SIMPLE: no convergence criteria found. Calculations will run for 2500 steps.

so TKE distribution can be quite bizare.

Could you add OpenFOAM version you are using? I have tried running you case with OpenFOAM 5.x and got FPE in turbulence model on the first iteration.

Hi alexeym,

I normally do not use stopping criteria in the fvSolution file, but just check the overall residuals trend to decide when to stop.
Anyway, also increasing a lot the iterations count the distribution keeps this rising trend-shape at mid-span (i.e. in the free-stream) which is totally not-realistic.


The OpenFOAM version I'm using is v1912 (https://www.openfoam.com), compiled from source on my computer (OS: debian10 - testing).