[pbo]

Hi everyone,

running icoFoam on a 2D, corrugated, airfoil-like geometry, I obtained the following history for the vertical force:



Freestream Re is 10^4.
linearUpwind cellLimited leastSquares 1 was used for U, Laplacian schemes are linear corrected (mean non-Orthogonality of the mesh is 5 deg), Euler was used as temporal discretization. Max Co was always around 0.6-ish. Plotting the contours of p or U at different times does not highlight any noticeable unsteadiness. So, what do I observe, a truly unsteady signal or simply numerical noise (the oscillation is about 3% of the mean value). It's still there whether I reduce or increase the time step by one order of magnitude. I ran simpleFoam, starting from the latest time computed by icoFoam (t=100s), I still observe those oscillations but the magnitude of the mean force is 7% higher than with icoFoam.
I used the laminarForceAndTorque library from Hrv to compute the forces.

[ngj]

Hi Patrick

For a circular cylinder exposed to a uniform flow, the r.m.s. value of the lift coefficient is different from zero and a function of Re, [1]. This might be the same thing you are observing. Try replacing your geometry with a cylinder and see if you get an value in the order of magnitude of 0.4 in the r.m.s value of the lift coefficient for Re_D=1e4.

Best regards,
Niels

[1] Hydrodynamics Around Cylindrical Structures, B. Mutlu Sumer and Jørgen Fredsøe, Advanced Series on Coastal Enginnering, World Scientific, 1999