SimScale CAE Forum

CFD Simulation Not Reaching a Solution


#1

Hello,

I am having an issue with the simulation for my FSAE car where the convergence plot does not appear to converge. The forces appear steady for a bit and then increases or decreases to unrealistic values.

I’ve used the same settings previously to mesh and run simulations so I’m unsure what the issue is.

I tried to simplify the geometry and even removed the tube chassis, but that did not appear to be the issue. The front and rear wings have been run in Simscale CFD by themselves as individual bodies, but I seem to run into issues when running a multi body simulation.

I would greatly appreciate assistance with this.

Thanks!

Project link: https://www.simscale.com/projects/traven/fsae_aero_project/


#2

Hi @traven,

From what I can tell, your simulation is diverging due to two possible reasons. I recommend trying fix no 2 and the minor correction to the numerics first and posting about how it goes before trying fix no 1.

The first is that your illegal cell count is still too high. 110 might sound ok, but these 110 might be clustered together and cause problems down the simulation. Reducing this significantly may be difficult however. I’ve done a quick visual check and everything looks good so far. So moving forward, the mesh may actually not be the source of your divergence. If there is a need to indeed resolve this, you will need to run this mesh offline via checkMesh in OpenFOAM and perform “checkMesh -writeAllFields” before post-processing it in ParaView where you can then check where the problematic cells are. Your geometry is already very well simplified with no significant geometrical errors so there is not much else to do there.

The second potential issue may be the “Potential foam initialization” option under Simulation Control. Try disabling it and see if your sim still diverges. Fingers crossed that this resolves it as if not, it would likely be the first issue and as mentioned, not an easy fix at all.

Another potential fix would be to set your gradient schemes to “Cell Limited Gauss Linear”. But do only do this option after you try fix 2 (the potential foam one).

Let me know how it goes.

Cheers.

Regards,
Barry


#3

So I attempted fix 2 first by turning OFF “potential foam initialization”, the simulation diverged.

Then following your advice, I set the gradient schemes to “Cell Limited Gauss Linear” (with the potential foam initialization ON again). The simulation appears to converge this time.

. Is this what should be expected for residuals?

Thanks a lot for the help!


#4

Hi @traven,

looks “good” I would say. Residuals like these are not the only indicator that a simulation is correct (whatever correct means in this case) nor if it shows physical plausible behavior. You can also have a look at a force residual that you set as a result control item and see if this force reaches a steady-state value - if not then you would have to make adaptions to the simulation (simulation time, physical settings etc.).

Cheers,

Jousef