You're talking about the Broadband Noise Source Model, correct?
Keeping in mind that it is a near-field approach, the model does not predict the sound at receivers far away from the flow field. Otherwise, it is a fairly good prediction.
The model can also be employed to extract diagnostic information on the noise source to check portions of the flow primarily responsible for noise generation.
Actually, I'm talking about general broadband noise. I've seen that the results of the simulation fairly match the experimental outcomes. However, this match compromises with a high error. e.g. The noise reduced in both cases -simulation and experimental- but not with the same percentage.