Dear Ivelin,

Depends what you need to predict. Could you tell me what kind of cannel do you use ?

Regards

Abel Rouboa

K-ε will do. See N.Markatos Heat Mass and Momentum Transfer across wavy interfaces ,Comp.Meths.Appl.Mechanics Engineering,14,3,pp323-376,1978 and refs therein. Of course Reynolds-Stress and LES is better but not by much. It depends on what you need. For engineering purposes, for your problem k-e is OK.

HI Ivelin,

I do believe that the k-epsilon will give you good results, as Markatos says, if the waves are generated by the flow. But if the flow velocity decrease to zero this is different, this could be happen in you example.

Also see: Turbulent air flow over water waves ,Trans.Inst.Chem.Engrs,Vol 58,pp251-259,1981