Flow past cylinder - Lift coefficient

[lauguimel]

Dear Fppimenta
I am interested in calculating the lift coefficient of a cylinder in a channel with a viscoelastic flow, depending on its y-position. I made some computations and got some results and to explain the results obtained, I then checked the Newtonian case for Re=0 (ok, I should have done this in first place !).
Unexpectedly, the lift coefficient for the Newtonian case doesn't remain null when moving the cylinder in the y direction. Even if at the center it is null ($\approx=10^{-14}$), it goes up to $\approx=10^{-4}$, in the upper half part of the channel which is very similar to the results of the elastic case. Based on literature, this lift coefficient should remain null for a Stokes flow wherever the position in the y-axis.
I have doubled checked that the convective term is disabled in fvSchemes, tried multiple convergence test (in time, mesh, etc) and I am not able to figure out what is wrong.

Have you an idea ?
If you want to have a look, I set up a minimum working example here based on your tutorial:
https://cloud.univ-grenoble-alpes.fr/index.php/s/drbj62AKBwZMNBx

Thanks a lot in advance for any suggestion you may have !
Guillaume

[fppimenta]

It's not an issue, but a question related with general fluid mechanics. Please do not use this space for such purpose (there are certainly better places).

Unless I am missing something, what you are getting is the expected behavior. Why should Cl = 0 for an offset cylinder in a Poiseuille flow? It can be << Cd, but not null, as there is loss of symmetry about the transverse direction.

[lauguimel]

Hello Francisco
Thank you for transferring the question to discussion, indeed this is not an issue.
Many papers state that the lift coefficient scale as 0(Re^2), so it should vanish in my case. I also ran the same simulation of Stokes flow with Comsol and the lift coefficient remains close to zero even when the cylinder is not located at the center.
For Navier-Stokes equation, there is no discussion, there is a lift acting but for Stokes flow I am still confused about the results obtained with openFoam...I thus suspected a mistake in my configuration file.

I'll keep digging !
Thank you for your quick answer!

[fppimenta]

Can you share a paper showing 'the expected results'?

[lauguimel]

In Bretherton 1962 it is theoretically shown that there is no lateral force in Stokes flow. It is also mentioned in the introduction of Ho 1974 or in the book of Morris: We saw in Chapters 1 and 2 that a neutrally buoyant spherical particle at zero Reynolds number will follow the motion of the fluid except for a lag owing to the “curvature” of the flow profile associated with the quadratic term (r/R)². Hence, at Re=0 the particle will not move laterally toward either the wall or the center-line – to do so would violate the reversibility property of Stokes flow. (pp201-202)