Double precision does not converge, Single precision does
 

Hi all,

I'm doing a crosswind analysis of a train and have found the following problem.
When solving with the double precision solver, my monitored coefficients such as lift, drag, ... do converge after some iterations, however my scaled residuals get stuck at 1 e-02 (continuity only).
When solving with the single precision solver, my residuals have no problem dropping below 1e-05, but the converged values of lift etc. differ somewhat from the ones found with double precision.

I know double precision normally is the more accurate, but can I conclude this when my scaled residuals of continuity get stuck at 1 e-02??

Thanks in advance,
jmmichie

PS: I'm using the k-omega turbulence model and wall functions (y+30). I have tried using y+=1, where the scaled residuals where somewhat less but they got stuck at 5 e-03.