Hi guys, I’m working on a flow simulation of a front wing for a Formula SAE racecar. I’ve tried following some of the tutorials for other wings, but I’m not completely sure that the numerical parameters I’ve input are correct.
For this simulation, I’m trying to get numbers for a coefficient of drag, coefficient of lift, amount of down force, and center of pressure on the wing. I’m not really sure if I’ve set up the simulation parameters correctly to find these values, but their seems to be a larger problem.
After I ran the simulation and opened it up in the post processor, I wasn’t able to see any values, and the whole mesh box is a single color. This is the first CFD simulation I’ve run, so I honestly have no idea what I’m doing. Any help would be greatly appreciated.
Checking it, give me some time to see what causes this issue.
Update: Can you please fix the units of the wall/inlet? You have the same numeric value but one is inch per second and the other one meter per second which might cause the unplausible values in the post-processor. Once that is fixed check out if putting a slice through your wing gives you a nice velocity/pressure distribution.
I changed the value to 15.65 m/s, ran the simulation again and ended up getting an error at around 32 core hours in. I can’t really make sense of the solver log.
Your simulation is hitting absurd force values and started diverging near the start. The simulation ended when your divergence reached unacceptable levels as stated by the numerics.
So some initial observations. Your mesh is quite ok, no big problems from what I can see. Boundary condition wise, similar observations, everything seems in order. The numerics are where the problems potentially lie. First, try setting the numerics to default and try not use leastsquares as it may be the cause of the divergence. Secondly, if its possible (not sure if the option is greyed out) enable automatic relaxation factors. That will save you some pain in tuning it to get a sim properly converged quickly. Lastly, increase the tolerence to 1e-6 at least, yours is 1e-8 atm and you would see the simulation continue to run even if its diverged significantly which will waste your core hours.
Hi,
I followed your advice on resetting the numerics to default, and the simulation ran. I was able to get some visualizations for the pressure distribution across the wing. Here’s what the convergance plot looks like for the run, does this seem acceptable? I’d really like to ensure that this simulation is at least somewhat sound before I start using it to iterate my designs, and eventually begin doing full car sims.
Thats good news! I just checked your simulation and it does look good indeed.
Being able to judge whether a sim is acceptable or not solely via the convergence plot is insufficient. You can tell whether the simulation is having errors via the convergence plot but you still have to check additional parameters to determine if the sim is usable even if the convergence plots does not show any errors. You’ve already done so in way by having a result control monitoring the force values and I’ve pasted the plot below for your reference but you can always find it in the simulation. As you can see, after the initial period where the simulation is settling down, you get realistic force values. On top of this, the results have started to reach a steady-state and we can say your simulation is converged.
All in all, yes, you can proceed to base your data off this simulation and continue optimizing your geometry/designs from here on out.
Do you want to view if after the simulation is running or before? If its before, it simple, just go to the side panel to hide the faces by clicking on the eye. Another way is to right click the faces you want to hide and select Hide that particular face. Attached screenshots below for reference.
If you’re talking about after simulation, that would be post-processing. SimScale does have a new post-processor, but due to how new it is, I have little experience in it. Instead, what I do for my post-processing is to download the data and process it offline via ParaView. You will need to look through some tutorials and videos to understand how to use ParaView but if you’re considering doing more CFD, this will be a very useful skill.
