correct prediction with massive separation

efb

I have used k-omega SST to model low speed airfoil. I have used yplus less than 1.0 and the convergence criteria with residuals less than 0.0001 togheter with steady behaviour of lift.. The results are almost perfectly in agreement with experimental before "numerical" prediction of stall (alfa = 18°), but fluent give stall when exp continue to rise for other 5-6 degree (stall to 23°).... Is possible that for the strong separation that occur after 15° (almost 50% profile separated) k omega cannot capture correctly shedding? is normal this kind of behaviour or not?..

thank you