convergence in rsm
 

hi guys ,
i am doing a problem of flow over a sphere for a reynolds number from 1.2*10^5 to 2.42*10^5 i am planning to use the rsm model
after many trail and errors i found that i was going nowhere
so i am in desperate need of some expert advice please help me!! i would dearly appreciate it

ok so i have the the following questions
1)i have a y+ of less than one everywhere on the boundary so i am using rsm i.e
the stress omega version with low re correction on is it ok for my flow which is unsteady with flow separation
2) is it requried that the flow be converged at every time step?
if so then what changes should i make regarding
a)setting of initial conditions i.e for turbulent intensity and
other parameters
b)it was suggested that i use the PISO scheme for Pressure-
velocity Coupling rather than SIMPLE is it justified?
c)what should the quality of the mesh be in terms of the skewness and aspect ratio .Is a skewness of 0.8 allright ?and an aspect ratio nearly 200 and 350000 cells
3)is there a need for me to adapt the mesh?
4)many people suggested that i first run the simulation for a steady state k epsilon to get the boundary conditions but they give very bad results and do not converge as my flow is unsteady
5)the transition takes place at an re of 3.8*10^5 so should i be looking at transition models? how well does rsm work for transition flows?
6)i am using turbulence intensity of 1% at both inlet and outlet and hydraullic diameter as my sphere diameter for both inlet and outlet as i read that for external flows the turbulent intensity is less than 1
7) the domain i used for my problem is a spherical domain which is ten times bigger than my sphere but i am suspecting that to capture the wake i need another type of domain what is your opnion ?
:confused: