orion
R.I.P.
This is an interesting discussion excerpt that i got this morning from one of my associates. It was written by Mark Drela:
Even so, there are some trends related to experimental data that I think we should look at. Comparing the Clmax results between Xfoil and NACA data ("Summary of Airfoil Data" by Abbott and Doenhoff) it seems to me that the dispersion between them increases for airfoils with small LE radius. Some examples:
Naca 0006 - Xfoil Clmax=1.57 / Tunnel Clmax=0.90
Naca 1408 - Xfoil Clmax=1.83 / Tunnel Clmax=1.35
Naca 0010-34 - Xfoil Clmax=1.40 / Tunnel Clmax=0.80
Naca 0012 - Xfoil Clmax=1.86 / Tunnel Clmax=1.60
Naca 2412 - Xfoil Clmax=1.94 / Tunnel Clmax=1.70
Naca 23012 - Xfoil Clmax=1.89 / Tunnel Clmax=1.80
Before blaming the wind tunnel data, maybe it would be interesting to think about whether there are any kind of simplifications or assumptions in the Xfoil code that under these conditions might lead to incorrect values.
Very few people seem to realize that _some_ of the NACA airfoil data is very strongly affected by compressibility. For the NACA 0006 at Re = 9M, these are the CLmax values I get versus Mach number:
Mach CLmax
0.00 1.56
0.05 1.53
0.10 1.46
0.15 1.33
0.20 1.21
0.25 1.06
0.30 0.94
The reason it's so sensitive is that with the very small LE radius the flow at the LE can easily become supersonic at large AoA's, even if the freestream Mach is low and the bulk of the flow is effectively incompressible. But the supersonic or nearly-supersonic flow at the LE strongly degrades the separation resistance of the downstream boundary layer, and thus decreases CLmax.
So the big question is what's the Mach number of the test data. There's no way to know, because it was never documented in the original NACA reports. The main report on the pressure tunnel which was used for these tests is here: http://naca.central.cranfield.ac.uk/reports/1947/naca-tn-1283.pdf The models had a 2 ft chord, so the 9M data was at 4.5M/foot. The chart in Figure 10 then shows that the Mach number could have been anywhere from 0.10 to 0.30, depending on the pressurization of the tunnel. So an Xfoil calculated CLmax number could be anywhere from 1.46 to 0.94 -- pick one.
The bottom line is that one should not take the Abbott & Doenhoff CLmax numbers verbatim without considering the possible Mach number effects, especially if they have low camber and/or pointy leading edges. For airfoils with larger LE radii, like the NACA 0012, the CLmax is much less sensitive freestream Mach, but the effect is still noticeable.
Even so, there are some trends related to experimental data that I think we should look at. Comparing the Clmax results between Xfoil and NACA data ("Summary of Airfoil Data" by Abbott and Doenhoff) it seems to me that the dispersion between them increases for airfoils with small LE radius. Some examples:
Naca 0006 - Xfoil Clmax=1.57 / Tunnel Clmax=0.90
Naca 1408 - Xfoil Clmax=1.83 / Tunnel Clmax=1.35
Naca 0010-34 - Xfoil Clmax=1.40 / Tunnel Clmax=0.80
Naca 0012 - Xfoil Clmax=1.86 / Tunnel Clmax=1.60
Naca 2412 - Xfoil Clmax=1.94 / Tunnel Clmax=1.70
Naca 23012 - Xfoil Clmax=1.89 / Tunnel Clmax=1.80
Before blaming the wind tunnel data, maybe it would be interesting to think about whether there are any kind of simplifications or assumptions in the Xfoil code that under these conditions might lead to incorrect values.
Very few people seem to realize that _some_ of the NACA airfoil data is very strongly affected by compressibility. For the NACA 0006 at Re = 9M, these are the CLmax values I get versus Mach number:
Mach CLmax
0.00 1.56
0.05 1.53
0.10 1.46
0.15 1.33
0.20 1.21
0.25 1.06
0.30 0.94
The reason it's so sensitive is that with the very small LE radius the flow at the LE can easily become supersonic at large AoA's, even if the freestream Mach is low and the bulk of the flow is effectively incompressible. But the supersonic or nearly-supersonic flow at the LE strongly degrades the separation resistance of the downstream boundary layer, and thus decreases CLmax.
So the big question is what's the Mach number of the test data. There's no way to know, because it was never documented in the original NACA reports. The main report on the pressure tunnel which was used for these tests is here: http://naca.central.cranfield.ac.uk/reports/1947/naca-tn-1283.pdf The models had a 2 ft chord, so the 9M data was at 4.5M/foot. The chart in Figure 10 then shows that the Mach number could have been anywhere from 0.10 to 0.30, depending on the pressurization of the tunnel. So an Xfoil calculated CLmax number could be anywhere from 1.46 to 0.94 -- pick one.
The bottom line is that one should not take the Abbott & Doenhoff CLmax numbers verbatim without considering the possible Mach number effects, especially if they have low camber and/or pointy leading edges. For airfoils with larger LE radii, like the NACA 0012, the CLmax is much less sensitive freestream Mach, but the effect is still noticeable.
