[ludvigum]

Hello!

I am conducting a control volume analysis of a car simulation, but for some reason the forces don't add up.

My process: I have set up a force report with all the car parts (from regions) and set the vector direction in the drag direction. I set up a similar force report with just walls, ground, inlet and outlet boundaries from regions as well. Same was done for downforce, just in the downforce direction.

My problem: The drag force and downforce should equal the force reports of the domain in the same directions, but this is not the case. I get an error of approximately 10% for drag and 2% for downforce.

Does anyone have a clue why? I'm thinking it might be the mesh near the wall, but with some refined mesh I still don't get much lower errors.

This is for my master's thesis so it's important that the CV analysis is good for me to be able to trust the results.

[FliegenderZirkus]

Are you sure that the same surfaces are selected in both of the force reports? It's hard to advise without seeing the simulation. BTW any reason why you're calling this a control volume analysis? If I get it right, you're just comparing two ways of selecting surfaces in a force report - either from parts or from regions.

[ludvigum]

Thank you for replying!
I think my question was a little unclear. I am only using boundaries from regions, as this is where all data is located.

I am attempting to see if Newtons second law (sum of all forces in a direction equals zero) is achieved.

So I set one force report that reports the forces on the car, and then a separate force report where I input the domain boundaries which should provide an equal force in the opposite direction.

Hope that clarified what I am asking. I can’t show the simulation as it is confidential.

[FliegenderZirkus]

Ah ok, I understood the question differently.

So you're evaluating a force report on inlet and outlet boundaries. I'm not quite sure what that means to be honest. I've always evaluated forces on walls only. But that's not to say you're doing something wrong, let's just see if someone else replies.

[f-w]

If your walls are not part of the free stream, then you can't group them with the inlet and outlet boundaries. That is, your boundary force balance should be: car + walls = inlet + outlet. Also, make sure your simulation is converged, and your force reports are configured to include forces due to pressure and shear (which should be the default).

[FliegenderZirkus]

So this is a common thing to do? What would be the physical interpretation of a force evaluated on an inlet and outlet boundaries? Assuming the solver just integrates dot product of the user-supplied direction vector with (static pressure * face normal vector), you'd get something resembling the static pressure drop of the domain...? I'm guessing there's no shear contribution at inlet and outlet. Just thinking out loud here. I'm curious to hear from the experts.

[ludvigum]

Originally Posted by f-w

If your walls are not part of the free stream, then you can't group them with the inlet and outlet boundaries. That is, your boundary force balance should be: car + walls = inlet + outlet. Also, make sure your simulation is converged, and your force reports are configured to include forces due to pressure and shear (which should be the default).

That is exactly what I have done this far, and I still get an error of 10%. The simulation is converged and forces use Pressure+Shear (as default).

[f-w]

Originally Posted by ludvigum

That is exactly what I have done this far, and I still get an error of 10%. The simulation is converged and forces use Pressure+Shear (as default).

If you say so, but your first post described it differently. You said you were comparing forces on [domain] to [car parts]. What I suggested is to compare forces on [domain, comprised of only inlet and outlet] to [everything else in your simulation]. If you are indeed doing this, then I can't think of any other solutions. As a sanity check, see if your inlet and outlet mass flows are equal in magnitude.