[martinWatt]

Hello all,

I have an oscillating drag force in a steady state simulation in OpenFoam (external aerodynamics), see picture. The oscillation is always the same, starting with the first iteration. I therefore believe that it has no physical cause. The simulation also converges normally, and the force curve looks normal except for this fluctuation.

Does anyone have experience where this oscillation comes from or how to investigate it further?

Thanks a lot!
Martin

[Alczem]

Hey


According to the scale, the oscillations are less than 0.001 which is already quite small in my opinion. I am not sure you can get even more stable results with a steadystate simulation. Depending on your conditions, the flow can be unsteady and it will be reflected on the values you monitor and the residuals. I believe these oscillations are mainly due to the limitations of a steady state approach and possibly numerical errors, because it looks otherwise decent.

[martinWatt]

Originally Posted by Alczem

Hey


According to the scale, the oscillations are less than 0.001 which is already quite small in my opinion. I am not sure you can get even more stable results with a steadystate simulation. Depending on your conditions, the flow can be unsteady and it will be reflected on the values you monitor and the residuals. I believe these oscillations are mainly due to the limitations of a steady state approach and possibly numerical errors, because it looks otherwise decent.

Thanks for your reply. But do you really think that’s a physical phenomena like flow separation etc.? I would expect a sinus like wave if it’s a physical phenomena not a sawtooth profile…

[Tobermory]

It's not a flow feature - ignore it.

If you really want to go down the rabbit hole, increase the precision of the written data (writePrecision in controlDict), and you will probably find that the variation is even smaller than you reckon. As the earlier poster said - it's just noise, as demonstrated by the "saw tooth" profile.

[martinWatt]

Originally Posted by Tobermory

It's not a flow feature - ignore it.

If you really want to go down the rabbit hole, increase the precision of the written data (writePrecision in controlDict), and you will probably find that the variation is even smaller than you reckon. As the earlier poster said - it's just noise, as demonstrated by the "saw tooth" profile.

Ok, thanks your help guys!