low Re (30-170) Karman vortex street

Sponi

Hello to everybody. I am trying to simulate Karman vortex series at low Re number, 30 up to 170. From literature is known critical Re = 47 +- 0.5, whereat is boundary layer periodic break away from cylinder wall and phenomenon Karman vortexreet is created. My 2D case is seting up like this: - isothermal simulation - fluid material is an air - cylinder diameter D=1.07 mm - inspection area has rectangular shape with this dimension: 100*D in front of a cylinder, 150*D below, and width is 200*D. Selectin of this dimensions is coming out from literature, where is noted minimal distance from cylinder to boundary for simulation low Re, 60D to 100D. - height of first cell is 0.001 mm (centroid is at 5e-04 m) - size of grid is about 45 000 (grid independcy was checked up to 0.5 mil cells)

Results from simulation is very good with experimental data in range of Re=60 to 170. Max. relative deviation is 1.2%. I never get simulation with vortex shedding for Re<60.

Solver Settings: - in first step is steady simulation done: 150 iteration 1st order and then Second Order for pressure, QUICK from Momentum and SIMPLEC for press-veloc. coupling. When I get periodic lift and drag coefficinet, then I start unsteady simulation with 2nd order time discetazion and same discret. scheme from steady simulation for press and momentum (2nd order press., QUICK momomentum) - during unsteady simulation are checking all residuals (x,y-veloc and continuity). I set-up 0.5e-04 for continuity - unsteady simulation is convergency in every time step with 3-5 iteration per time step. - legth of time step is derived from known Strouhal number, where is frequency f. I can determine period T=1/f. After that time step is Dt=T/60 (or 6 degree in angle interpretation)

Trouble: How I said, result for Re=60 and higher are very well with present experimental data. But I never achieved vortex shedding for Re<60. Even if sctructure grid is fine enough 45e+03 to 500e+03 cells. Neither Third Order MUSCL for momentum nor shorter time step (Dt=T/120, Dt=T/360) didn't help.

Question: Is there any way, how could I bring near to critical Re=47+-0.5? Thanks a lot for your suggestions! -David Sponiar, Prague, Czech Rep.-