Hello Everyone ,

Hope you are doing fine . I was going through this paper of P. Cosyn and J. Vierendeels on ''Numerical Investigation of Low-Aspect-Ratio Wings at Low Reynolds Numbers'' in which they had done the numerical validation of Torres and Mueller's experimental work in commercial CFD package ANSYS Fluent .

They had done the selection of turbulence model on the basis of 2D airfoil profile (no depth in span direction ) of the same 3D wing . After comparing the results of K-e realizable , K-w SST and SA turbulence model . They concluded that SA gives accurate results when compared with the experimental results among all three turbulence models.

However , at later stage when they tried to do the validation of 3D rectangular flat plate wings at a low Re No of 10e5 they found that at higher angle of attacks SA under-predicts the drag and lift coefficient results . They attributed the under-prediction of results to the incapability of SA model to switch the turbulent length scales to higher values as we have massive amount of flow separation at that higher angle of attacks. So , in short according to them SA model under-predicts the amount or level of separated flow .

Another thing which is to be noticed is that , as the aspect ratio of rectangular flat plate wing increases from 0.5 aspect ratio to 2 aspect ratio the difference between experimental and numerical values increases even further . Why is it so ? Is it because for higher aspect ratios we have a larger surface area of the wing hence increased level of flow separation ?

They postulated that at higher angle of attacks on flat plate rectangular wing there pop ups a 3D massive separation bubble as earlier as at the start of leading edge of the wing, which is not the case for cambered airfoils . Hence for cambered airfoils the difference between the experimental and numerical values is comparatively less than that of flat plate wings .

My last question , is it the limitation of SA model only or overall RANS formulation for all one and two equation turbulence models that it can not accurately predict the flow separation levels at higher angle of attacks for flat plate rectangular wings at Re No of the order of 10e5 ?

Suggestions , recommendations and guidelines to make improvement in numerical results are highly welcome and appreciated in advance .

Regards

Aamir