Problem with vortex shedding frequency
 

Hi friends,
I am currently trying to simulate flow past a circular cylinder for my thesis in OpenFOAM. I am having problem in getting the correct strouhal number. Drag coefficient is comparatively good. The following are the BC’s and details about mesh
Dia=0.034 m; velocity=20 m/s; Re=45,000 ; Solver- PISOFOAM; RANS Model-kOmegaSST


P-freestreamPressure at inlet & outlet, zeroGradient on cylinder (wall), Symmetry on Top & Bottom


U-freeStreamVelocity at inlet & outlet, fixedValue (0) on wall, Symmetry on Top & Bottom


k-fixedValue at inlet (based on k=3/2(UI)^2),zerogradient at outlet, FixedValue (0) on wall and kqrwallfunction, Symmetry on Top & Bottom


omega-fixedvalue at inlet & outlet (based on omega=k^3/2/kcmul) basically took from one journal paper.For wall omega wall function and value =60nu/betaXdely^2. Symmetry for Top & Bottom


nut-fixedvalue(0) at inlet, zerogradient at outlet,nutuspaladingwallfunction on wall, symmetry on Top and bottom.


Mesh details:
Y+=1, Four meshes are tested 50K,65K,80K,and 95K with max aspect ratio less than 22. The domain size is sufficiently long was used.


fvSchmes:
Time- Euler, gradschmes-Gauss linear, Divergence-Gauss limitedlinear,Laplacian-Gauss linear corrected
Interpolation-linear, Sngrad-corrected


Fvsolution:
PCG with DIC as preconditioner for pressure with tolerance of 1e-06, reltol=0
For other terms PBiCG with DILU ,tolerance 1e-08


Ncorrectors=2,northogonalcorrectors=3


Please help me to solve this problem friends, Hope I described all important details

What about the time step size, did you try to use a smaller time step?

Hi taxalian
Thanks for your interest. Intially I used 5e-06 which gave max courant number 0.6-0.8. After your post today, I simulated with the time step of 1e-06 such that the max courant number less than 0.2. I simulated more than 140 non dimensional time units to avoid transient effects and started to monitor the values of Cl and Cd.Still I am getting reasonable drag value but strouhal number is high.I calculated this through calculating the time interval between peaks of the Cl plot as well as FFT of the Cl values.Manual calculated value is 0.25 and FFT value is 0.2464.The actual value for Re 45000 is 0.20.I also attached my case.Please have a look.

https://drive.google.com/file/d/0B8S...it?usp=sharing

Did you solve your problem?

I would recommend using pimpleFoam instead of pisoFoam, to make sure convergence is achieved with a convergence criterion. I made the experience that with pisoFoam no relaxation should be used, but it seems you do not use relaxation.

Isn't Y+=1 rather low, since you are using wallfunctions?

Hi all,
I am solving a similar problem but with reynolds number =100. I got strouhal number same but am unable to get correct Cd(drag co-efficient).

drag co-efficient(Cd) calculated from the forceCoeff.dat through numerical integration doesn't match with that calculated from the formula
Cd=F(drag force)/(0.5*density*diameter*Umax^2) and neither of them confirm with the actual result. Can anyone explain why? Since i am using 2-d simulations so in place of area in the Cd formula i used the diameter of the cylinder over which my flow is taking place.
regards
Subhankar