[Engr07]

Hi all,

I simulated flow over an aerofoil to get the forces and moment using Spalart-Allmaras turbulence model. The grid was done so as to achieve the y-plus values less than 1.0.

At zero degree I was able to keep the y plus less than 1.0 over the complete aerofoil but not the trailing edge-The aerofoil has a blunt trailing edge.

At 10 degree y-plus values at the leading edge were about 3.0 and were also larger than 1.0 at the trailing edge.

Since it is "good practice" to keep the number less than 1.0, I need to know how significant is the issue of not achieving this at the leading and the trailing edges.

Any help is greatly appreciated.

AMal

[rlc113]

Hey AMal, it has been my experience that the y+ values are crucial to drag prediction. I am currently working on a similar case in which my y+ values are << 1 everywhere except the Leading edge. When comparing my case with a case ran with someone with a bit more experience I noticed a significant underprediction on my solution. I have added more fidelity and decreased the delta1 value in that area and I am rerunning the case. I will report back with the difference.

Let me know if you find anything yourself.

[Far]

with the hybrid wall functions (currently implemented in major CFD codes), the yplus values needed are Y+<10. See this reference :

http://num.math.uni-goettingen.de/ba...ings/knopp.pdf


http://num.math.uni-goettingen.de/ba...ngs/alrutz.pdf

http://tau.dlr.de/fileadmin/document...8_englisch.pdf

http://proceedings.fyper.com/eccomas...cuments/64.pdf

[rlc113]

Hey Far, thanks for the info, I haven't checked it all out yet. Just wanted to say that the only solver I use is NASA's FUN3D, which to my opinion doesn't have a wall function. Any insight to that would be interesting.

I have the refined BL case running now. I also have the results of a case in which the inviscid spacing was refined by 40% or so, I will post up all my results when the the results are apparent.

[Martin Hegedus]

For the SA model I usually try to get the y+<1 at about 5% to 10% of the chord. I doubt it matters for less than 5%, if the majority of the airfoil y+ is less than 1. The eddy viscosity at the leading edge is usually very small. It builds up along the chord. Thus, the leading edge flow is laminar, even when using the turbulence model. However, that is dependent on angle of attack. So, look at your eddy viscosity. As for the trailing edge, usually the boundary layer is thick, so I'm not sure why you are having an issue there. Unless you have a recirculation bubble at 10 deg. But, if you are at the very trailing edge tip, I doubt it matters. Again if the majority of the airfoil is y+<1.

If you are concerned, the best thing to do is reduce the off wall grid spacing by 1/2, run the code, and see if you get a difference in the skin friction.