I have used Buossinesq hypothesis and k epsilon method, with excellent results.

For accademic reasons I have to do the following things. -Check the spalart-Allmaras model to estimate the gain when using the k epsilon model. Which parameter should i check at the end to make comparison??

-Use RSM to validate the result when using Boussinesq mode. Can I use it? My simulation are all steady... How can i understand if i don-t have enough power calculation or when the RSM is not suitable for my case?

I hope you can help me. Regards Sergio

I guess that you should check the spalart-allmaras model on the same parameters you say you had excellent results with k-e...anyway, in RANS they usually are global parameters like a force coefficient, a reattachment lenght, a flow rate through some area or something similar depending on the problem you're solving.

In theory, if you're using a Boussinesque-hypotesis, more than ever you can use a RSM. Both of them are actually steady models because RANS are steady equations and the trivial activation of the time derivative to get an URANS model is actually not that trivial, not at all.

However, several issue arise in the modeling of the terms which form the RSM so it is not sure that you'll get better results just because you're using a more complex model.

In 3D RSM has 6 independent components + 2 more equations (if i remember well), so the memory requirement for any 3D problem has 2 times the number of variables of a k-e model. The time per iteration is maybe lower than double but it's better to check by yourself