flow past a missle: how to solve divergence?
 

I am now using Spalart almarmas + 2nd order upwind to do simulations in FLUENT. but it is very fast to diverge after just a few steps of iterations. My mach number is around 2 and Reynolds number in order of 105. So how to solve the divergence problem? How can i get some reasonable initiation field for the iterations to continue?

Thank you very much in advance!