You are on page 1of 2

XFOIL

Brett Watson
California Polytechnic State University, San Luis Obispo, California, 93405

I. Abstract
XFOIL is an interactive program which utilizes a two-dimensional panel method code with integral boundary
layer theory to analyze airfoils in viscous or inviscid flow fields. The results generated from an XFOIL analysis
of the laminar flow NLF(1)-0115 airfoil are compared with results from Selig, Maughmer, and Somers' paper
"Natural-Laminar Flow Airfoil for General-Aviation Applications." Selig, Maughmer, and Somers use the
existing NACA 23015 airfoil as a departure point for lift comparisons with the NLF(1)-0115. To verify their
results, analysis was done on the NACA 23015 using thin airfoil theory and flat plate boundary layer assumptions.
Upon analysis, XFOIL's generated drag polar and separation points matched closely those presented by Selig,
Maughmer, and Somers leading to the conclusion that for low Reynold's numbers, XFOIL's analysis is relatively
accurate. Problems arise when Reynold's numbers in the millions appear. XFOIL is unable to create a converging
flow field around a blunt leading edge airfoil when the local velocity is too large. The drag polar produced using
lower Reynold's numbers resembles the shape shown in Selig, Maughmer, and Somers' paper. The transition from
laminar to turbulent flow was expected to occur near the leading edge for takeoff and landing conditions, i.e. a
coefficient of lift around 0.1, as predicted by Selig, Maughmer, and Somers. XFOIL predicted the separation at
the leading edge for landing conditions on the upper surface. This result shows the boundary layer formulation
exploited by XFOIL is an idealization of the actual case due to an already separated flow is approaching the
leading edge of the airfoil.

II.Introduction
The NASA NLF(1)-0115 airfoil as described in Somer, Maughmer, and Selig's paper entitled "Natural-Laminar
Flow Airfoil for General-Aviation Applications," is an attempt at creating an airfoil with a high lift range and a low
profile drag. It is a compilation of experience gained through analysis of turbulent flow, high applicable lift range,
airfoils such as the NACA 23015. Somer, Maughmer, and Selig show the NLF(1)-0115 obtains a improved drag
profile due to laminar flow similar to the NACA 23015. The NLF(1)-0115 expects to see laminar flow throughout
the flow field due to a reduction in leading edge contamination inherent in its design.
XFOIL is a program in which it is possible to analyze airfoils in two-dimensional flow fields. The program was
developed by Mark Drela at the Massachusetts Institute of Technology in 1986 and utilizes integral boundary layer
theory in viscid and inviscid flow field solutions. Exploiting Somer, Maughmer, and Selig's results, it is possible to
explore the accuracy and limitations of XFOIL. Analysis of the NLF(1)-0115 in XFOIL can be directly compared to
Somer, Maughmer, and Selig's results, and results
obtained from analysis of the NACA 23015 using a flat 0 . 4

plate boundary layer assumption and thin airfoil theory


are used to verify the results. 0 . 3

0 . 2

III.Results
The NLF(1)-0115 compared to the NACA 23015 0 . 1

revealed the potential and limitations of the XFOIL 0

program. A generated plot in MATLAB of the


coordinates of the NLF(1)-0115 confirm the basic - 0 . 1

geometry. The analysis of the characteristics of the - 0 . 2


airfoil showed the flaws of XFOIL. In an attempt to
compare Selig, Maughmer, and Somers' drag polar data - 0 . 3

to that generated by XFOIL, a Reynold's number of six


million was used, corresponding to the paper's claim for - 0 . 4
0 0 . 1 0 . 2 0 . 3 0 .4 0 .5
x / c
0 . 6 0 . 7 0 . 8 0 . 9 1

standard cruise. XFOIL's results were unable to Figure 1. NASA NLF(1)-0115 airfoil
converge for this case. A Reynold's number of 400,000
generated a converging flow field and created a drag polar
similar to that presented in Selig, Maughmer, and Somers'
1

American Institute of Aeronautics and Astronautics


paper. This leads to the conclusion that XFOIL can only handle a range of freestream velocities at a blunt leading
edge for it to generate a converging solution.
A possible explanation for this is the fact that
XFOIL allows only 500 data points as inputs
for a given airfoil. This may lead to
insufficient data collection for XFOIL's
calculation of panel vortices and creating a
turbulent flow when one may not be present.
Beyond speculation, XFOIL cannot handle
flows in the transonic regime due to shock
effects. It is likely that XFOIL was unable to
converge on a series of solutions for
Reynold's numbers above one million
because of this fact.
XFOIL's prediction of flow separation is
relatively close to that of Selig, Maughmer, and Figure
---- 2. Drag NASA
polar for NACA 23015 and NASA NLF(1)-
NLF(1)-0115
Somer's prediction. At an angle of attack of _____
0115. Data forNACA 23015
NLF(1)-0115 taken from Selig, Maughmer,
negative 2.4 degrees and coefficient of life of 0.1, and Somers. Data for NACA 23015 generated from flat plate
flow separated at the trailing edge on the lower assumptions and thin airfoil thoery.
surface and at about 20 percent chord on the
upper surface. This is accurate to the prediction
of flow separation using flat plate boundary assumptions and thin airfoil theory for the NACA 23015. The fact that
XFOIL's prediction is accurate to results using flat plate boundary layer approximations and thin airfoil theory
further show its limitations. XFOIL uses downstream wake trajectories from inviscid flow solutions to calculate
drag if the angle of attack is specified, as in this analysis. This is an obvious assumption as viscous effects will
degrade lift and increase drag.

IV.Conclusion

It is important to realize the limitations of XFOIL before use. It is a tool to be used when flow separation occurs
downstream of the leading edge and at Reynold's numbers which do not exceed a certain value. XFOIL was
accurate in determining the separation point and predicts drag well for lower Reynold's numbers, so a direct
comparison was unable to be made between its results and those supplied by Somers, Maughmer, and Selig in their
paper "Natural-Laminar Flow Airfoil for General-Aviation Applications." Much can be said regarding XFOIL's
comparison to solutions utilizing flat plate boundary assumptions and thin airfoil theory: XFOIL's prediction of
drag is a huge improvement over these assumptions which produce drag data invariant of angle of attack. XFOIL
may therefore prove to be a useful tool in preliminary airfoil design, but caution must be utilized if greater accuracy
is desired.

References
1
Crowe, C.T., and Elger, D.F., Engineering Fluid Mechanics, Wiley, Hoboken, NJ, 2009, Chaps. 9, 11.
2
Fox, R.W., and Pritchard, P.J., Introduction to Fluid Mechanics, 7nd ed., Wiley,Danvers, Massachusetts, 2009, Chaps. 6, 9.
3
Selig, Michael, Maughmer, Mark, and Somers,Dan, “Natural-Laminar Flow Airfoil for General-Aviation Applications,”
AIAA Journal, Vol. 32, No. 4, 1995.

American Institute of Aeronautics and Astronautics

You might also like