Dependence Inlet conditions : k and eps

PK

Hi,

I had a look about all related post but didn't find a proper answer. I am trying to simulate the flow around some simple bluff body , like cylinder a sphere in 3D. I am using the keps realizable, steady, y+<3, fluid water. first oder in pressure , second order in M, SIMPLE.

Inlet Turbulence Specification Method : K and epsilon according to 6.2.2 in the fluent user guide.

I am using the 2 formulas to get k and eps (6.2.4 and 6.2.5 in the guide): k=1,5 x (Uavg x I)^2 AND eps = f(Cu,k,l).

1-if I used I=0,1 and l=0,8m I get some value of k and eps and get approx the right value for the Drag (Re=25000).

2-If I used I=0,01 (smaller) my drag in than 20% less.

I have used finner grid and it seems to be the same result. Does the converged solution should be independent of the intial values ? which value of I and so k should be choosen ?.

Thanks for your help

PK