[lakeat]

A small question,

I just performed a cylinder flow with icoFoam, but the result seems to be different from those in ferzifer's book. So i am wondering what might be wrong.

How to set upper and lower boudaries conditions?

wall or cyclic, Any ideas? Thanks a lot.

Because I am afraid if I use wall patch when Re is high, the mesh will be very expensive. I saw some paper used cyclic though in fact it is actually another problem.

Daniel

[mike_jaworski]

Hi Daniel,
This question seems far too general to be answered in any sort of detail. Maybe you could tell us *how* the solution is different and compared to what? An analytical solution? another simulation? experiment?

Also, have you done the potentialFoam tutorial? It's detailed in the Programmer's guide and covers a cylinder in cross-flow. Maybe that would be a good place to start and see how BCs are set there.

Good luck,
Mike J.

[lakeat]

I am making my effort to simulate the case cylinder flow as described in "Computational Methods for Fluid Dynamics" Chapter 8, and I hope to get a better understanding of liftDrag (for my ultimate aim to to simulate drag reduction).

The geometry is the same as Ferziger's.
I meshed with gmsh, and then used gmsh2ToFoam.
Re is 20,40,100,150.
fvSolution and fvScheme are the same as those in icoFoam/cavity, and are not changed.
I lowered the 1st cell hight to satisfy y+=1.
The velocity profile, I used both the parabolic one and the uniform one.
One thing I'm not sure is the upper and lower wall boundary condition, which type should I set?

whatever, the variation of CD on a cylinder is very smooth on the book of Ferziger, however, mine is not smooth at all!

meanwhile, the meanvalue is too small.
Here is Prof. Ferziger's


Hi Mike, thank you,
I noticed in potentialFoam tutorial, the BC is symmetryPlane. But how about cyclic? Will the result be of great defference?

[lakeat]

I'm still very new to CFD.
Could anyone point out what's going wrong here, thanks.