Relaxation factors are a thing with SIMPLE

For piso, if your courant number is <1 you don't need them

Ahh okay - i didnt know that!

But in piso it could be wrong then - then a timestep will be manipulated my relaxtion right? so even thoguh it makes it more stable then its not a true timestep?

Yep

Makes sense thou

some of your variables will be simulated in slow motion

I make sure I always have:

c++
".*Final"    1;

in my relaxation factors, avoids making that mistake

I will - thats a pretty important detail

I think i managed to get it working too. So the changing DDT from backward to euler did it... I used some of the other settings as well afterwards because i believe that is more correct. I still think it is strange that ddt scheme was causing this?

Could try letting it run with Euler, then swap to backward when settled

Could be an option, but i have to run many differnet configuraiton so i think i am satisfied with euler! My i'll try and check the effect of higher order schemes!

Hi all, if you have an example github or paper about free convection inlet and outlet bc for a free convection case, I would greatly appreciate it. Trying to get this working to go away from a proprietary cfd app.

I am trying to get it working with an incompressible solver. It is buoyant flow. Most confusing part is the pressure bc type needed.

It's difficult to setup free convection BCs. Maybe you can find interesting insights there: https://www.cfd-online.com/Forums/openfoam-pre-processing/183661-common-setup-p_rgh-complete-open-domains.html

Hello, So regarding High-Re flow in a pipe with multiple bends - i was experimenting with symmetry and i noticed that it changed the simulation result - My Nu fell with 5%. For most incompressible flows you can apply symmetry if the geometry allows in, but in pipe-bends even though the geometry allows it, there still exist an un-steady behaviour. Is it correct, that in un-steady cases (even with symmetric geometry) secondary flow patterns can still occur, why it symmetry in these cases can adjust the results? Guess it would make kind of sense.

Yeah imposed symmetry affects results , some modes of motion become impossible with the symmetry bc...

Thank you for confirming - i´ll guess the dream of cutting all my meshes in half disappeared..

well, if i am not wrong, when you have turbulence due to its chaotic behavior the symmetry results will differe from non half geoemtry, no? (i dont know in how much, but they will differe...)

Yeah imagine like if it develops a clockwise spiral .. that won't happen under half geometry

here goes your mesh, down the drain... 😄

haha yeah... but it was worth a shot though... But it forced me to update class-blocks at least, haha

If you run a half domain sim vs full one, can you see in residuals convergence that something is off in the half domain simulation?

First picture is with sym. Second picture is a full pipe. The pressure is clearly oscillating in the symmetry case... but im unsure what to interpret from this? https://cdn.discordapp.com/attachments/726377664213942323/1415277752159830047/image.png?ex=68c29f9e&is=68c14e1e&hm=b6b3f499f55947a91a7111718fee3214f39ba3669ec33bf06ea950659680abcb& https://cdn.discordapp.com/attachments/726377664213942323/1415277752562618444/image.png?ex=68c29f9e&is=68c14e1e&hm=9a0566e11de8352162e101b76754fcf3b22082b99dd3b97b714eaca35a329f93&

Hi there, just to ask, any usual practice for fvscheme and fvsolution configuration? I've gone through this documentation https://www.wolfdynamics.com/wiki/tipsandtricks.pdf and found that quite a lot to understand. is it okay to just leave the solver entries for U up to solver, smoother, tolerance and relTol?

Depends on the setup. I go with second order schemes most of the time and i dont touch the fvSolution - Unless my simulation becomes unstable or if it have trouble congering in each step. E.g if the GAMG solver maxes out the number of iteration (50 pr. default) then your simulation will take a long time - and then it can be a good idea to check the settings. There is probably other people with better tips thou 🙂

Thanks for the tips!

I got some luck with these: Inlet p_rgh total Pressure Inlet U pressureInletVelocity Outlet p_rgh zeroGradient Outlet U pressureInletOutletVelocity Needing to also get these working with compressible solver but it diverges

So I posted a Q in #726377692626157579 group