[Bekka]

Hi,

I'm undertaking a study at the moment on the drag and lift forces of a vehicle. All my residuals/mesh/pressures etc seem to be what I expected. However, my drag and lift forces and coefficients keep drifting downwards and I can't work out why.

Does any body have any ideas?

Thanks!

[scipy]

What do you mean by "drifting downwards"? If your solution has not converged yet, then just give it more time/iterations until force/coefficient monitors level out.

If on the other hand you mean the values are too small, did you properly set reference values for coefficient calculation (such as vehicle's projected frontal surface area)?

[Bekka]

Yes, the values are much lower than expected and the more iterations completed, the lower they seem to get.

Yes, all reference values e.g. density/frontal area/relative velocity have been filled in correctly.... Completely at a loss with what could be wrong

[scipy]

Can you post a screenshot of a convergence graph for Cd/Cl or something? If they're asymptotically getting closer to some "value" then let it run until the values level out. However, if they're just obviously wrong then I don't know.

Are you using symmetry? Did you set only the half of the frontal area as the reference value? Did you select all the walls of the vehicle in the force calculation (maybe you missed some of them)?

[Bekka]

No, I havent used symmetry due to the kart i'm using not being symmetrical. I've been through my set up numerous times and can't find anything wrong.

I've attached screen shots of the force plots and residuals

Drag Plot.PNG

Lift Froce.PNG

Residuals.PNG

[scipy]

This all looks pretty normal/standard to me. I don't know how large your mesh is, but I had convergence graphs that looked exactly like this. Some playing with under-relaxation factors might help to "calm" things down a bit (I'm assuming a steady-state simulation) and more iterations may be needed to get the forces to level out (again, depends on the solver.. but even using a Coupled version of the solver I needed ~1000 or more to converge cases of 8-12 million elements).

[fluid23]

I would argue that 650 iterations is probably no where near enough to call this converged. In a lot of external aero cases I find that my residuals sometimes appear to converge before the engineering values (lift, drag, mass flow rate) sometimes level out. I would AT A MINIMUM, run another 650 iterations before making any judgement about stability of your values (i.e. their drift).

As for your forces being low I would suggest checking that your mesh is appropriate. Make sure wall y+ is appropriate. That you don't have any large volume changes or skewness issues. I would also verify that you have the dimensions you think you do. Sometimes CAD imports or improper unit system selection can cause you to have bodies that are quiet a bit smaller or larger than what you intended. You may be putting the right constants into the report, but you should make sure that those match your model. Also, make sure you are refining the wake otherwise you aren't really capturing what is going on with your drag values.

[GeorgeH25]

Does anyone know how to fix a unable to compute error showing up in force coefficient report?

I have everything as it should in the coordinates, pressure and part regions but it keeps saying unable to compute and no data

the frontal area reports on the same parts so I'm not sure what is going on?

Thank you