I am making computation or non rotating circular cylinder to obtain von karman street. diameter of that cylinder is 1 cm.fluid is air , velocity 100 m/s. I am using k eps RNG model. vortex sheedind freq should be 500 hz in fluent manual is written that optimal no of iterations for unsteady flow is about 20. 500 hz means (i am not sure)that vorticies occurs 500 tines a second . i also know tat delta t should be at least 20 times smaller that period of vorticies 500hz is 0,002s so delta t is 0,0001

i tried with 100, 200,and more iteration but resusts are funny i thing that mesh is good. Y plus near wall is about 7 what else should i take into account or there is something very important that i do not know

please advise me A

don't use k-epsilon as it tends to "stick" to much to the surface. Use a k-omega based SST instead. Have a look at www.pbase.com/michaelbo/cfd - Michael Bo

can you tell me if it is important to have very long chanel before cylinder or can I compute turbulent velocity profile in another completly empty channel and then paste it into that one with cylinder. second think that I dont know is how to execute unsteady computation. Do I have to at initial run set it as steady flow and after let say 700 iteration set it back to unsteady flow (to have initial velo profile) or at very begining at 0 iteration set as unsteady? what is procedure for unsteady computation?