unsteady flow around a circular cylinder

Lattice Boltzmann Method LBM: Theory
#1

Dear all,

as a part of my master thesis I’m writing a 2D LB solver. To verify the solver I try to simulate the flow around a circular cylinder and here is my problem: I can calculate the creeping flow for low Re (<4) and I can also calculate the (symmetric) seperation for moderate Re (4<Re<40), but I cant simulate the unsteady instabilities when Re is rising. The solution I get in the unsteady case is simular to the case of 4<Re<40, only the recirculation length rises with Re. In a physical meening the flow seems to be to viscous. I tried to increase Re but this had no effect on the solution, until omega (of the BGK model) was nearly 2 and the solution got instabil. I also used different grid resolutions without sucess. Something to the model I use: its a D2Q9. For BC I use equilibrium at inlet, zero gradient at outlet and the halfway bounce back at walls (most of my code is according to the book of Hänel).
I was looking for mistakes in my code several times but found nothing.
Has somebody an idea what`s the reason for this problem or where a error in my code could be???

Thanks a lot

Andreas

#2

Dear Andy

Is the cylinder located the same distance from both top and bottom walls? Have you tried to located it slightly asymmetrically? Or, otherwise, have you tried to use an asymmetrical velocity profile as inlet boundary condition.
The presence of an asymmetry in the flow domain/structure will trigger the flow instabilities for the Re that you mentioned.

Hope it helps.

Regards

Goncalo Silva

#3

Dear Goncalo Silva,

yes my cylinder was located symmetrically in the fluid domain.
I`ve located it asymmetrically now and it works!!
Thanks for your help

Andreas