Airfoils

Design process of new airfoils for the KR-2S wing

Ashok Gopalarathnam. First created: 22 February 98, Last update: 7 April 98

links edited by Mark Langford, January 2003, added template download link

Results from preliminary design studies - 22 February 98, modified 2 March 98

In this page I have compared the predicted performance of several candidate airfoils for the KR-2S. Among them are the RAF48, the NLF(1)0115 and four new airfoils. The predicted performance for all the airfoils has been obtained by analyzing them using XFOIL, a program written by Prof. Mark Drela of MIT. The four new airfoils have been chosen as the best of several that I have designed over the past few weeks. These new airfoils were designed using PROFOIL (a limited version of PROFOIL is available on the web as PROFOIL-WWW), an inverse airfoil design program written by Prof. Michael Selig and MFOIL, a MATLAB graphical user interface that I created for designing airfoils with PROFOIL.
Along with the airfoil performance, I have also included some of the results from my analyses of the effect of these airfoils on the aircraft performance. Assumptions have been made for the drag characteristics of the baseline airplane (with the RAF48 airfoil for the wing) and the engine and propeller performance. With these assumptions, the effect of changing the wing airfoil on the power characteristics, rate of climb, angle of climb, top speed, range and endurance has then been studied.

Comparison of the RAF48 and the NLF(1)0115 airfoils

Both these airfoils have a max. t/c ratio of 15 %. The RAF48 is being used as the wing airfoil for the stock KR-2S, and the NLF(1)0115 airfoil was being considered as a candidate for a redesigned KR-2S wing.
The following figure shows the predicted performance for the airfoils using XFOIL. It must be mentioned that the coordinates for RAF48 are the high resolution coordinates from Mark Langford's page. I had a hard time getting XFOIL to converge for this airfoil, and the results (as seen in Fig. 1) are not smooth. A big reason for this difficulty is the waviness in the coordinate listing. Therefore, one must keep in mind that these predictions for the RAF48 may be erroneous. (I experienced even greater difficulties getting XFOIL to converge with the RAF48 coordinates from the UIUC airfoil database).
The airfoils were analyzed for a reduced Reynolds number of 2 million. By analyzing at such a "reduced" Re, the effect of decreasing Re with increasing values of C_l (owing to decreasing flight speeds) is automatically taken into consideration. More on reduced Re is available here.
For comparison, a wing loading of 12.85 psf results in a reduced Re of 1.38 million for a 25" chord and a reduced Re of 2.76 million for a 50" chord at standard sea level conditions. Therefore, I felt that a reduced Re of 2 million will serve as a good choice for the current study.

Fig.1 Predicted performance for the RAF48 and the NLF(1)0115 airfoils at a reduced Re of 2 million.

As seen from the drag polars, the NLF(1)0115 airfoil has a significantly lower drag at a C_l of 0.3, but it has higher drag than the RAF48 a higher C_l values. The predicted C_lmax for the NLF(1)0115 is nearly 1.6 and for the RAF48 is 1.5.
How does one decide whether it is worthwhile trading off lower drag at high C_l conditions for lower drag at the low C_l conditions? Put another way, by switching from the RAF48 to the NLF(1)0115, it is quite obvious that lower drag can be expected at cruise conditions (low C_l), resulting in higher range for a given fuel quantity. But will this switch result in a significant loss in the rate of climb and angle of climb at low speeds (high C_l)? Perhaps the best way to answer this question is by examining the effect of these drag polars on the aircraft performance.

Effect on aircraft performance

Throughout this page, in all the estimates of the aircraft performance, I have concentrated on determining the effect of the airfoil drag on the aircraft performance rather than the actual performance of the aircraft. With such an objective, I have used rather simple and approximate methods to determine the "baseline" aircraft performance with the RAF48 airfoil. Since, at this time, I am not interested in the actual performance of the airplane, such simple estimates are sufficient.

Assumptions made in calculating the baseline performance with the RAF48 airfoil for the wing

Standard sea level
W = 1000 lb (airplane weight)
S = 79.27 sq. ft. (wing area)
b = 23 ft. (wing span)
f = 1.2 sq. ft. (from fig. 5.17 of Ref. 1) = equivalent parasite area
K = 1.4 = (1/e), where e is Oswald efficiency

q = 0.5*rho*(V^2) = dynamic pressure; where rho = air density, V = flight speed

D = Drag = parasite drag + induced drag = f.q + (K/pi)*((W/b)^2)/q = baseline drag
(in appropriate units)

Change in performance due to change in airfoil

D = baseline drag + DC_d*q*S,
where DC_d = (C_d)_new
airfoil - (C_d)_RAF48

Power required = DV

Other assumptions
Engine = 80 hp,
prop efficiency varies from 45% at 50 mph to 70% at 200 mph (cruise prop), resulting in the power available curve shown in Fig . 2(a)

Engine SFC = 0.5 (lb of fuel consumed/hr)/BHP
1 U.S. gallon of aviation gasoline weighs 6 lb
Fuel capacity = 30 US gallons
Range and endurance estimates are approximate; these calculations do not take into account the decrease in the airplane weight as fuel is consumed. The weight is assumed to be constant at 1000lb. (This assumption results in quicker estimates).

Fig. 2(a) Estimates of the power available and the effect of the airfoil drag on the power required

To determine what happens to the performance when I change the wing airfoil from the RAF48 to the NLF(1)0115, I have taken the drag difference between the two polars from the XFOIL predictions, and translated that difference into a difference in the power required for level flight. As expected, the NLF(1)0115 performs better than the RAF48 at high speeds, but does not perform too badly in comparison with the RAF48 at the low-speed end. The reason, of course, is that the induced drag plays a much larger contribution in determining the low-speed performance than does parasite drag. This illustration shows that while designing a new airfoil to improve on the RAF48, it is important to concentrate on decreasing the drag at low C_l values even at the expense of slightly higher drag at high C_l conditions. To further examine the effects of the airfoil on the aircraft performance, the climb and cruise performance with the two airfoils are shown in Figs. 2(b) and 2(c).

Fig. 2(b) Estimates of the effect of the airfoil drag on the climb performance

The climb rate and the climb angle estimates with the RAF48 and the NLF(1)0115 airfoils can be seen in Fig. 2(b). As seen from the figure, the lower drag for the NLF(1)0115 at higher speeds results in better climb rates at those speeds. The best rate of climb has not changed significantly by using the NLF(1)0115 instead of the RAF48. The RAF48 does have a small advantage at the lower speeds, but the advantage is minor for both climb rate and climb angle. The figure also provides information on the effect on the maximum cruise speed. Maximum cruise speed corresponds to the speed at which the R/C becomes zero. From the figure, the maximum cruise speed is about 2 mph higher for the NLF(1)0115 than for the RAF48. While this result may be somewhat surprising, the reason for this result is that the C_d for the two airfoils are nearly identical at a C_l value of 0.15, which corresponds to the top speed.

Fig. 2(c) Estimates of the effect of the airfoil drag on the cruise performance

As mentioned earlier, the estimates for range and endurance do not take into account that the airplane weight decreases with flight time. The reason for this omission is to simplify the calculations. Again, the emphasis here is on the effect of the airfoil on the performance rather than on the absolute performance of the airplane.
As seen from Fig.2(c), there is a significant improvement in the range at high speeds with marginal loss of endurance at low speeds. Given the role of this airplane, the loss in endurance by using the NLF(1)0115 instead of the RAF48 is considered inconsequential.

Comparison of the new 15% GA19980222A and the NLF(1)0115 airfoils

A new 15% thick airfoil, temporarily designated the GA19980222A was designed to match/excel the performance of the NLF(1)0115 while satisfying the pitching moment design constraint that Cm > -0.055.
A note about the temporary designation: GA stands for "general aviation," "19980222" refers to the date on which this airfoil was designed (22 Feb. 1998) and "A" denotes that this was the first airfoil designed on that day. In the process of designing airfoil(s) for any project, I go through several design iterations and I use this numbering system (adapted from the system used by Prof. Selig) to keep track of all the iterations.

Fig. 3 Predicted performance for the GA19980222A and the NLF(1)0115 airfoils at a reduced Re of 2 million.
The predicted performance using XFOIL for this airfoil is compared with that for the NLF(1)0115 in Fig. 3. As seen from the figure, the pitching moments for the two airfoils are nearly identical, demonstrating that the constraint has been satisfied. The drag polar shows that the 22A has an improved performance at the low-C_l end (high-speed end), with the bottom corner of the polar lower than the bottom corner of the NLF(1)0115 polar by a difference of about 0.1 in C_l. There is no degradation in performance at any other condition. The predicted C_lmax for the 22A is also about 0.05 higher than for the NLF(1)0115.

Effect on aircraft performance

Fig. 4(a) The effect of the airfoils RAF48, NLF0115 and the GA22A on the power required

Fig. 4(b) The effect of the airfoils RAF48, NLF0115 and the GA22A on the aircraft climb performance

Fig. 4(c) The effect of the airfoils RAF48, NLF0115 and the GA22A on the aircraft cruise performance

The effect of the 22A and the NLF(1)0115 airfoils on the aircraft performance is compared with that of the RAF48 (baseline) in Figs. 4(a)-(c). These figures are similar to Figs. 2(a)-(c)
From the three figures, it can be seen that the aircraft predicted performance is equal or better everywhere with the 22A airfoil than with the NLF(1)0115 airfoil. In particular, from Fig. 4(b), the maximum cruise speed (the speed at which the R/C goes to zero) for the new airfoil 22A is significantly better than for the NLF(1)0115 and the RAF48 airfoils; about 10 mph higher than with the RAF48 and about 8 mph higher than with the NLF(1)0115.

Effect of the pitching moment constraint

To answer the question of whether any performance improvement can be obtained by allowing a more negative pitching moment, the performance of the 22A is compared with that of the 15% thick GA19980220D. The 20D and the 22A share common design philosophies, with the exception that the 20D was designed for a Cm of -0.075, instead of the Cm of -0.055 used for the design of the 22A.

Fig. 5 Predicted performance for the GA19980220D and the GA19980222A airfoils at a reduced Re of 2 million.

As seen from Fig. 5, the 20D airfoil has a more negative pitching moment than the 22A. Notice that for both the airfoils, the lower corner of the polar has not changed. The reason is that the C_l value for the lower corner of the polar was a design specification; this C_l value was fixed at 0.1 for both the airfoils. Except for a small reduction of the low-drag range for the 20D, the polars for the two airfoils are nearly the same. The biggest effect of allowing a more negative (nose down) value for the pitching moment is an increase in the C_lmax. By going from a Cm of -0.055 to -0.075, the predicted C_lmax has increased from about 1.62 to about 1.65.
The effect of the 20D airfoil on the aircraft performance has not been included as the curves for the 20D are nearly identical to those for the 22A.

Comparison of the new 18% GA19980222B, the 15% 22A and the NLF(1)0115 airfoils

An 18% thick airfoil was designed to examine the effect of thickness. The pitching moment was constrained to C_m > -0.055. The predicted performance characteristics for the 18% 22B is compared in Fig. 6 with those for the 15% airfoils 22A and the NLF(1)0115.

Fig. 6 Predicted performance for the GA19980222B, the 22A and the NLF(1)0115 airfoils at a reduced Re of 2 million.

As seen from Fig. 6, the 22B and the 22A both have the lower corner of the drag polar at a C_l of 0.1. Also both these airfoils have nearly the same pitching moment values. As explained earlier, these are both common design requirements for the two airfoils. The most noticeable characteristic of the 18% 22B is the much wider low-drag range, stretching from a C_l of 0.1 to 0.8, as opposed to a range from Cl of 0.1 to 0.55 for the 15% airfoils. However, at high Cl values (of about 1.3), the 22B has a much larger drag than the 15% airfoils. Also the predicted C_lmax for the 22B is about 0.1 less than that for the thinner 22A.

Fig. 7(a) The effect of the airfoils GA22B and the GA22A on the power required

Fig. 7(b) The effect of the airfoils GA22B and the GA22A on the aircraft climb performance

Fig. 7(c) The effect of the airfoils GA22B and the GA22A on the aircraft cruise performance

As we saw earlier, the aircraft climb and cruise performance is relatively insensitive to higher drag for the wing airfoil at high values of C_l. In the Figs. 7(a)-7(c), the high-drag behavior of the 22B at high Cl values does not even show up, as this condition corresponds to speeds of less than about 65 mph. It should be realized, however, that this high-drag behavior for the 22B at high C_l is caused by flow separation at the trailing edge of the upper surface. Examination of the boundary layer characteristics for the 22B, as predicted by XFOIL, shows that the flow begins to separate at the trailing edge of the upper surface at a C_l of about 1.0, and very gradually moves forward until at a C_l of about 1.5, the aft 30%-40% of the airfoil is separated. While this results in an extremely soft stall, with ample warning, such an airfoil, when used in the outer portions of the wing may results in somewhat ineffective ailerons at high alphas. However, when used in the root regions of the wing in combination with the 15% 22A airfoil at the tip, it is possible that the gentle stall of the 22B can be utilized to good advantage. The soft stall and the associated trailing edge flow separation for the 22B at high alpha can provide ample stall warning, while the thinner airfoils with a higher C_lmax at the tip can continue to provide good aileron effectiveness and stall margin.
It may be worthwhile trying to design another 18% airfoil which does not have such a high-drag behavior at high values of C_l. One candidate is the 18% GA19980211D, which has the same design philosophy as for the 22B, except that the 11D has a more negative pitching moment of Cm = -0.075.

Comparison of the new 18% GA19980211D, the 18% 22B and the 15% 22A airfoils

Fig. 8 Predicted performance for the GA19980211D, the 22B and the 22A airfoils at a reduced Re of 2 million.

As seen from Fig. 8, the 18% 11D has a more negative pitching moment as well as a slightly higher C_lmax and lower drag at high values of C_l than the 22B airfoil.
I am planning to work on the design of more 18% thick airfoils in the next few days. During this design effort, I'll be focusing on reducing the trailing-edge flow separation at high angles of attack.

Airfoil geometries

Fig. 9 Geometries for the RAF48, the NLF(1)0115 and the four new airfoils 22A, 20D, 22B and the 11D.

Inviscid velocity distributions

(a)	(b)
(c)	(d)
(e)	(f)

Fig.10 Inviscid velocity distributions for the six airfoils

Summary

An effort is underway to design a custom airfoil to replace the RAF48 for the KR-2S wing. This page describes the status of the effort. The predicted performance for the RAF48, the NLF(1)0115 and four new airfoils were presented and discussed. Also presented were the effect of these airfoil characteristics on the changes in the aircraft performance from an assumed baseline.

Upcoming design activities

Design more 18% airfoils without the high-drag characteristics of the 22B at high C_l values.
Examine the effect of leading-edge roughness
Examine the effect of plain flap deflections to determine aileron effectiveness

References

1. Stinton, D., "The Design of the Aeroplane," available in the USA and Canada from the American Institute of Aeronautics and Astronautics, Washington DC, 1995.

Characteristics of one of the final airfoils - the 15% AS5045 - added 7 April 98

The Ashok Gopalarathnam/Selig 15% airfoil AS5045 was arrived at after considerable tweaking of the GA19980222A. The design philosophy included prescribing laminar and turbulent boundary characteristics over different portions of the airfoil at their design operating conditions in a multipoint fashion.
The following figures compare the predicted aerodynamic characteristics of the AS5045 with those for (1) the GA19980222A and (2) the RAF48. As done earlier, all of the airfoil predictions have been made for a reduced Re of 2 million.

Fig. 11 Predicted performance for the AS5045 and the GA19980222A airfoils at a reduced Re of 2 million.

Fig. 12 Predicted performance for the AS5045 and the RAF48 airfoils at a reduced Re of 2 million.

As seen from Fig. 11, the AS5045 is an improvement over the GA22A at both the low-C_lend and the high-C_lends of the polar. There is a small increase in the pitching moment (more negative) for the AS5045 at the higher values of C_l. The slight "kink" in the drag polar for the GA22A at the upper end of the drag-bucket (C_lof 0.5) has also been smoothed out.

Effect on aircraft performance

The effect of using the AS5045 on the aircraft performance is compared with that obtained using the GA22A in Figs. 13 (a)-(c)). The geometry of the AS5045 is compared with the other 15% airfoils in Fig. 14.

Fig. 13(a) The effect of the airfoils AS5045 and the GA22A on the power required

Fig. 13(b) The effect of the airfoils AS5045 and the GA22A on the aircraft climb performance

Fig. 13(c) The effect of the airfoils AS5045 and the GA22A on the cruise performance

Fig. 14 Geometries for the RAF48, the NLF(1)0115, the GA19980222A and the new AS5045 airfoils.

As seen from the geometries, the new airfoil has lesser camber than all the other 15% airfoils in the figure. Also the trailing edge region of the airfoil has been made thinner. The AS5045 has a blunt trailing edge (1/8" thick for a 50" chord) to make it easy to build.

Preliminary results from the wind tunnel tests for the AS5045 and AS5048 airfoils

Ashok Gopalarathnam and Michael Selig, 7 July 98

Summary of the main design goals for the AS5045 and AS5048 airfoils Table 1 The main design goals for the two airfoils

	AS5045	AS5048
t/c	15%	18%
C_d at cruise (C_l = 0.1, Re = 5.6 million)	<0.005	<0.005
C_lmax (Re = 1.5-2 million)	1.3	1.25
C_lmax	not dependent on extensive laminar flow (not highly sensitive to roughness)	not dependent on extensive laminar flow (not highly sensitive to roughness)
Cmo	-0.055	-0.06

Brief description of the experimental setup The UIUC 3x4 ft. open-return, low-speed wind tunnel was used for the experiments. The high-quality airfoil models with excellent surface finish and smoothness were made by Steve Eberhart. The 18" chord models were mounted vertically on to the tunnel balance on the floor. Lift and pitching moment on the model were measured using the balance. Drag was measured using a wake rake.

The models have not yet been digitized. We expect to digitize the models using a coordinate measuring machine in September 1998. No wind-tunnel corrections have been applied to the data yet.

The key focus of the tests were to obtain experimental verification of the C_lmax, stall characteristics and the effect of roughness on the stall behavior. Owing to limitations on the maximum tunnel speed, the maximum Re that could be achieved with an 18" chord model in this tunnel is around 1.8 million. For the airplane wing, this Re corresponds to the C_lmax condition. Re values for cruise condition for the wing in flight are much higher than 1.8 million, and cannot be measured in the tunnel.

Acknowledgments Ashok would like to gratefully acknowledge the invaluable assistance provided by Sam Lee and Andy Broeren during these tests. Without Sam's help in adapting his data acquisition code to suit the current experimental needs, the data presented in this page would not exist. Andy spent several hours of his time helping Ashok get up to speed with the flow-viz and model installation. Ashok would also like to thank B.J.Jasinski and Philippe Giguere for their help with installing and changing the models. Results for the AS5045 (15%) airfoil Figure 1 (repeat_1m.gif) shows the lift, drag and moment characteristics for the AS5045 airfoil at a Re of 1 million. (The data shown is from two different runs to check repeatability). Figures 2 (repeat_1-5m.gif) and 3 (repeat_1-8m.gif) show the data for a Re of 1.5 million and 1.8 million. As seen from the figures, the data is repeatable.

The following observations can be made from the figures:

Table 2 Summary of the AS5045 airfoil performance from the tests

Re (million)	C_lmax	a_max (deg)	Cm (nominal)	Stall type
1.0	1.2	~14	-0.05	gentle
1.5	1.25	~14	-0.05	gentle
1.8	1.28	~14	-0.05	gentle

Note that although the data is shown for C_l values from -0.4 to stall and beyond, the only data relevant for the airplane is the data close to stall. In other words, the C_d values from the data at low values of C_l are quite irrelevant. This is because at the speeds corresponding to the airplane climb and cruise conditions, the airfoils on the wing operate at Reynolds numbers much higher than 1.8 million. (See the web page on reduced Re for information on the variation of Re with flight speed and C_L for the KR-2 type airplanes). The limitations on the tunnel speed, however, prevented us from testing the airfoils at Reynolds numbers greater than 1.8 million. The airfoil C_d at flight Re corresponding to cruise and climb should be significantly lower than those shown in the above data.

Figure 4 (test_res1.gif) compares the data for Re of 1, 1.5 and 1.8 million with the results predicted by XFOIL for Re of 1, 1.5 and 2 million. It is seen that the C_lmax is much less than that predicted by XFOIL. This behavior was expected. It was known even before the design of these airfoils that XFOIL consistently over predicts C_lmax (at least for airfoils of this class). For example, Fig. 5 (compare_nlf0215_3m.gif) compares the XFOIL predictions for the NASA NLF(1)0215 airfoil with the experimental results from tests done at NASA. While the XFOIL predictions for the low C_l agrees quite well with the NASA data, the XFOIL predicted value of C_lmax is about 0.25 higher than that obtained from experiments. Other comparisons (not shown here) show that the XFOIL-predicted C_lmax is sometimes up to 0.3-0.4 higher than that obtained from experiments. During the design of the airfoil, this tendency of XFOIL to over predict C_lmax was taken into account and it was expected that the airfoil C_lmax would be around 1.3-1.4. The more important design goals related to stall were that the stall be gentle and that C_lmax of this airfoil be not less than that of the RAF-48 and the NLF(1)0115. Because experimental results for the RAF-48 and the NLF(1)0115 are not available, it is not possible to verify this goal. It is expected that although XFOIL over predicts C_lmax, the differences in C_lmax from one airfoil to another should still be correctly predicted. As shown below, these goals have been satisfied (based on XFOIL predictions for the AS5045, the NLF(1)0115 and the RAF48 airfoils).

Effect of leading-edge trips on the AS5045 performance
To test the effect of leading-edge roughness on stall characteristics, the performance of the airfoil was tested with leading-edge trips. Three types of tape were used:

Table 3 Types of trip tape used for the current study

Trip number	Trip type	Trip height	Trip width	Trip location (aft edge of tape)
1	4 layers of ultracoat	0.009"	0.2"-0.25"	2% x/c on upper surface 5% x/c on lower surface
2	1 layer of "DYMO" adhesive labeling tape	0.009"	0.38"	2% x/c on upper surface 5% x/c on lower surface
3	2 layers of "DYMO" adhesive labeling tape	0.018"	0.38"	2% x/c on upper surface 5% x/c on lower surface

Figure 6 (trip1_2_re1-8m.gif) compares the clean performance of the AS5045 airfoil with the performance with trip configurations 1 and 2 at a Re of 1.8 million. It is seen that with the smaller trip (Trip 1), the C_lmax of the airfoil drops by about 0.06 compared with the clean case. With the wider trip (Trip 2), the drop in C_lmax is about 0.15. Figure 7 (trip1_2_3_re1-8m.gif) compares the effect of the large trip (Trip 3) with the clean case and the smaller trips for the same Re of 1.8 million. With the larger trip, the drop in C_lmax compared with the clean case is about 0.2. Also seen is that for all the trip configurations, there is a significant increase in drag. This increase in drag is a consequence of the loss in laminar flow resulting from the forced transition. A part of the drag increase can also be attributed to the "device-drag" of the trip itself.

These results emphasize the need to keep the wing (the leading edge in particular) free of bugs and contamination. While it is hoped that small roughness elements will not significantly decrease C_lmax, there is a decrease in C_lmax with increasing size of the leading edge roughness. This decrease in C_lmax with increasing size of the roughness element is expected. The design philosophy used for the AS5045 and the AS5048 minimize this decrease in C_lmax with increasing roughness size. For this reason, it can be expected that the stall speed will increase with large roughness (in rain, for example). For example a decrease in C_lmax of 0.2 results in a stall-speed increase of roughly 5 mph. Therefore, airspeeds should be increased by at least 5-10 mph when flying close to stall in rain or with leading edge roughness.

Results for the AS5048 (18%) airfoil The AS5048 airfoil was tested at Re of 1.0, 1.5 and 1.7 million. This airfoils could not be tested at Re values beyond 1.7 million because of problems arising out of interference between the model and the tunnel floor at high loads. Only the results for Re of 1.5 million and 1.7 million are presented here.
Figure 8 (as5048_1-5m.gif) and Fig. 9 (as5048_re1-7m_repeat.gif) show the performance of the AS5048 at Re of 1.5 million and 1.7 million respectively.
The following observation can be made from the figures:

Table 4 Summary of the AS5048 airfoil performance from the tests

Re (million)	C_lmax (drag rise)	a_max (deg) (drag rise)	Cm (nominal)	C_lmax (peak lift)	a_max (deg) (peak lift)	Stall type
1.5	1.15	~13	-0.045	1.28	~16	gentle
1.7	1.15	~13	-0.045	1.25	~15	gentle

A peculiar feature that can be seen from the reults for the AS5048 is that although the C_lmax corresponding to the maximum in the lift curve is around 1.25-1.28, the drag rises to significantly high values at a C_l of around 1.15. It is felt that this behavior is due to large trailing-edge separation at C_l of 1.15, causing high drag, but the lift continues to increase with increasing a till around 1.25. It is felt that the actual C_lmax for this airfoil (as "felt" by the pilot) at these Re values will be around 1.15-1.2.
Figure 10 (as5045_48_re1-5m.gif) compares the experimental results for the AS5045 and the AS5048 with XFOIL results for the airfoils at Re of 1.5 million.
Effect of leading-edge trips on the AS5048 performance
Owing to lack of time, the effect of trips for this airfoil were investigated only for one trip size (the large trip # 3). Figure 11 (as5048_re1-7m_trip.gif) shows the effect of using the large trip (Trip 3) on the performance of the AS5048 airfoil at Re of 1.7 million. It is seen that the drop in C_lmax is around 0.25, which is a larger drop when compared with that for the AS5045.

Recommendations

The following recommendations are made regarding the use of the AS5045 and the AS5048 airfoils on the modified KR-2S airplanes:

The gentle stall characteristics of the 15% AS5045 make it an ideal choice as an airfoil for the entire wing.
If it is desired that a higher thickness section be used at the root for structural considerations, the AS5048 at the root in combination with the AS5045 outboard form a good combination. The earlier stall and trailing-edge separation of the AS5048 when compared with the AS5045 can prove beneficial in providing the pilot with good stall warning.

However, the early trailing-edge separation at high C_l can also result in high interference drag penalties. This fact must be kept in mind.

It is felt that the AS5048 may not be a good choice for the outboard wing portions, where early trailing-edge separation may reduce aileron effectiveness.

KR2S Airfoil Templates

Full-size airfoil templates were drawn up on CAD for two of the new airfoils using coordinates resulting from the design effort.. They are shown fitted to KR2 spars using standard spar construction methods. The AS5045 was too thin at the main spar to fit the "stock" KR main spar, so the very similar AS5046 was created. These templates have been output to Adobe Acrobat PDF files and donated to public domain for free download, at http://www.krnet.org/AS504x/templates.html .

KRNet would like to thank Dr. Ashok Gopalarathnam, Dr Michael Selig, and the staff of UIUC Applied Aerodynamics Department for their outstanding efforts during the creation of these new KR2S airfoils.