P.

Giguere
Graduate Research Assistant.
New Airfoils for Small Horizontal
e-mail: [email protected]
Axis Wind Turbines
M.S. Selig
Assistant Professor. In a continuing effort to enhance the performance of small wind energy systems, one
e-mail: [email protected] root airfoil and three primary airfoils were specifically designed for small horizontal
axis wind turbines. These airfoils are intended primarily for 1-5 kW variable-speed
Department of Aeronautical and wind turbines for both conventional (tapered/twisted) or pultruded blades. The four
Astronautical Engineering. airfoils were wind-tunnel tested at Reynolds numbers between 100,000 and 500,?0~.
University of Illinois at Urbana-Champaign. Tests with simulated leading-edge roughness were also conducted. The results zndl-
306 Talbot Laboratory. cate that small variable-speed wind turbines should benefit from the use of the new
104 S. Wright Street. airfoils which provide enhanced lift-to-drag ratio performance as compared with
Urbana. IL 61801-2935 previously existing airfoils.

1 Introduction Section 2 describes the airfoil design objectives and methodol-

ogy. A brief description of the wind-tunnel facility, measure-
Over the last 12 years, a considerable numbe~ of ~oils ~ave ment techniques, and data reduction method is provided in Sec-
been specifically developed for horizon~ aXIS. w~nd turbi?es tion 3. Section 4 provides a discussion of the wind-tunnel test
(HAWfs). For example, 25 specially tailored airfmls .for wmd results, including tests with fixed transition to simulate leading-
turbines originated from the joint work between the NatJ.onal Re- edge roughness caused by blade erosion, accumulation of insect
newable Energy Laboratory (NREL) and Airfoils, Inc. (Tangier debris, etc. Finally, conclusions are given in Section 5.
and Somers, 1995). For wind-turbine applications, the advanced
NREL airfoils have been shown to provide aerodynamic and struc-
tural advantages as compared with airfoils originally designed f?r 2 Airfoil Design Objectives and Methodology
airplane applications. In addition, BjOrk ( 1988, 1989, 1990), Hill In this design work, the focus has been on designing airfoils
and Garrard ( 1989), Madsen and Rasmussen ( 1992), and Timmer that can be used along the entire blade span of small variable-
and Rooy (1992) have also designed airf~ils for ~v.:s .. There speed HA Wfs. Three airfoils, referred to as the primary airfoils,
are, however, only a limited number of wmd-turbme rurfo!ls that were designed to provide optimum performance over a broad
have been developed exclusively for small blades. range of operating conditions. Considering the low operating
Typically, the entire span of the blades ?f small ~wrs Reynolds numbers and beneficial centrifugal stiffening effects
operate at low Reynolds numbers where lanunar .sep~tton ef- of small HA Wfs, the airfoil thickness for the primary airfoils
fects can severely degrade the performance of airfmls not de- was fixed at ten percent. A 16 percent root airfoil was also
signed for this flow regime. In additi~n, the ty~ically f~t r~ta­ designed to accommodate possible large root bending moment
tional speeds of small HAwrs provide centrifugal stJ.ffemng and large blade-stiffness :equirements. With the root airfoil over
that reduces the blade bending loads and thus, allows for the the inboard 30 percent of the blade, the primary airfoils are
use of thinner airfoils as compared with those designed for recommended for use over the last 25 percent of the blade (from
larger wind turbines. Reductions in laminar separation effec~ the 75 percent station to the tip), with blending between 30
and in airfoil thickness provide an increase m aerodynarmc percent and 75 percent of the blade span. The root airfoil
performance, which yields better energy .capture. The~efore, ~w­ (SG6040) and the three primary airfoils (SG6041, SG6042,
ing to low Reynolds number aerodynanu.cs and centnfu?al stiff- and SG6043) are depicted in Fig. l.
ening effects, wind-turbine airfoils designed for medmm and During normal operation, variable-speed HA Wfs operate below
large blades are not particularly suitable for smal~er ?lades .. stall and over a relatively limited lift range as compared with
A number of existing low Reynolds number rurfods, which constant-speed wind turbines. For these operating conditions, min-
were designed for small unmanned aerial vehicles, have been imizing leading-edge roughness effects, such as no loss in the
shown to be applicable to small wind turbines (Giguere and maximum lift coefficient, is not particularly critical (Tangier and
Selig, 1996, 1997a). Specially tailored airfoils for small Somers, 1995; Tangier, 1997). For optimum aerodynamic perfor-
HA Wfs, however, should provide enhanced performance. From mance during variable-speed operation, the low-drag lift range
the literature, it appears that the S822 and S823 airfoils are the (drag bucket) can be reduced in favor of greater lift-to-drag ratios.
only two airfoils that have been specially designed for small Accordingly, the primary airfoils were designed to achieve as high
blades. These two airfoils are principally intended for stall- a lift-to-drag ratio as possible at the design lift coefficient, ranging
regulated HA wrs and form one of the nine advanced NREL from 0.6 to 1.2. The primary airfoil having the lowest design lift
airfoil families (Tangier and Somers, 1995), Consequently, coefficient was expected to yield large blade chords and conse-
there is a need for new wind-turbine airfoils for small HAWfs, quently, the root airfoil was designed to be mainly used with the
especially for those operating in variable-speed ~ode. two other primary airfoils. To account for possible variations in
In a continuing effort to support new blade designs for .sm~ll the tip-speed ratio (TSR) caused by atmospheric turbulence, the
HAWfs and improve their performance, a total of four rurfotls best lift-to-drag conditions were designed to occur over a range
were designed and wind-tunnel tested. These airfoils are in- of lift coefficients centered about the design lift ~fficient.
tended primarily for 1-5 kW rated power variable-speed wind The operational Reynolds number range for HAWTs having
turbines for conventional (tapered/twisted) or pultruded blades. a rated power between 1-5 kW is typically below 1,000,000.
At these Reynolds numbers and particularly below 500,000,
minimizing laminar seperation effects is an important design
Contributed by the Solar Energy Division of THE AMERICAN SOCIETY OF ME- driver. Accordingly, low design Reynolds numbers were fa-
CHANICAL ENGINEERS for publication in the ASME JOURNAL OF SoLAR. ENERGY
ENGINEERING. Manuscript received by the ASME Solar Energy DIVISIOn, July vored to control laminar seperation effects. The design Reynolds
1997; final revision. Dec. 1997. Associate TechnicB.l Editor: P. S. Veers. numbers for the four airfoils were selected according to the

108 I Vol. 120, MAY 1998 Copyright 0 1998 by ASME Transactions of the ASME
 root airfoil, a design Reynolds number of 200,000 was used. The
-~- objectives for the four new airfoils are shown in Table 1.
Thickness: 16% : Cambe~ 2.5% To achieve the design objectives, the new airfoils were sys-
tematically designed using PRO FOIL (Selig and Maughmer,
--C~- 1992), which is an inverse airfoil design method. In addition,
Thidcness: 10%: Cambe~ 2.0%

_c-== SG6042 ==-=--------

Thidcness: 10% : Camber: 3.8%
the Eppler code (Eppler, 1990) and XFOIL (Drela, 1989) were
used to obtain preliminary airfoil performance characteristics
to guide the design. With PRO FOIL, the desired velocity distri-
bution is prescribed, from which the corresponding airfoil shape
-L---sG6043 ~- is determined. In tailoring the velocity distribution for each
Thidcness: 10% : Camber: 5.5%
airfoil, a gradual transition ramp was introduced on the upper
Fig. 1 Contours of the root and primary airfoils surface to minimize laminar separation effects such as bubble
drag. The need for a shallow pressure gradient combined with
a high-lift requirement lead to aft-loading and, consequently, a
Table 1 Airfoil design objectives
relatively high pitching moment coefficient ( -0.08 < Cm.cl• <
Airfoil tic Design C1 Design Re -0.14). Note that the reduction in the size of the drag bucket
SG6040 16% 1.1 200,000 caused by the use of aft-loading is not a drawback for variable-
SG6041 10% 0.6 500,000 speed HA WTs because, as mentioned previously, they operate
SG6042 10% 0.9 333,333
over a relatively limited lift range under normal operation. The
SG6043 10% 1.2 250,000
primary airfoils have similar velocity distributions that differ
mainly by their respective amount of aft-loading. Therefore, the
aerodynamic characteristics of the three primary airfoils are
10 related and these airfoils can be considered as an airfoil series.
· SG6040 Avg. accuracy: 0.23 mm (0.0069 in)
E The addition of the root airfoil to the primary airfoils provides
5 0.5
a unique airfoil family for variable-speed HA WTs.
~ 0.0 _______ ......
~ -0.5 / --\;;.:.,:. -:;.,;;.,- ---- 3 Testing Apparatus and Methods
i:S The experiments were performed in the University of Illinois
at Urbana-Champaign low-turbulence subsonic wind tunnel
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
xlc having a 0.857 m (2.81 ft) high and 1.219 m (4ft) wide test
section. For the airfoil tests, two 1.829 m ( 6 ft) long Plexiglas
Fig. 2 Contour accuracy plot for the SG6040 airfoil splitter plates were inserted into the test section and reduced
the test section width to 0.854 m (2.802 ft). With an empty
test section, the turbulence intensity of the tunnel is less than
tradeoff between operating lift coefficient C1 and Reynolds num- 0.1 percent for the test Reynolds numbers (Guglielmo, 1996;
ber Re. It has been shown that for an optimum variable-speed Selig et al., 1995). The 0.305 m (1 ft) chord airfoil models
rotor, the product of the lift coefficient and Reynolds number were made of foam with a fiberglass finish and were inserted
is nearly constant for a particular blade station (Giguere and horizontally between the splitter plates with nominal gaps of
Selig, 1997b), i.e., 1-2 mm (0.040-0.080 in.). Using a coordinate measuring ma-
(l) chine, each airfoil model was digitized at midspan. As an exam-
ple of the digitized results, Fig. 2 presents the accuracy plot for
Here, R is the reduced Reynolds number which is constant so long the model of the SG6040 airfoil, which was the least accurate
as the TSR and the local velocity, which includes the axial induc- of the four models used in these tests. The average accuracy of
tion factors, of a given blade station remain constant. Therefore, the SG6040 airfoil model is, nevertheless, 0.23 mm (0.0089
the potential increase in the maximum lift-to-drag ratio from op- in.) over the 305 mm ( 12 in.) chord. Similar plots for the pri-
erating at a high lift coefficient is reduced by a lower operating mary airfoils can be found in Lyon eta!. ( 1997a). The airfoil
Reynolds number according to Eq. ( 1) and vice versa The design coordinates are tabulated in the Appendix.
Reynolds number for each primary airfoil was determined using The lift was directly measured using a strain-gage load cell, and
Eq. (I ) and a reduced Reynolds number R of 300,000. For the the drag was determined from the average of four different span-

SG6040 v Re=100,000 o Re:300,000

Free transition o Re=150,000 "' Rt!=<IOO,OOO

' .
1.5
: : : . :
~--·--··-~---------+---------!---------~----------~-------- 1.5

Ct 1.0
-----~~-~:~. ' : . :
--:---------r-----··r··---- 1.0

0.5 0.5

0.0
----T·------r··---- 0.0

-0.5 w...................l-.4...L-J:...::O...U::.'<>'--iU.............J...o...............L~w...J
0.00 0.01 O.o2 0.03 0.04 0.05 0.06 0 10 20
c. . a (deg)

Fig. 3 Drag polars and corresponding lift curves for the SG6040 root airfoil (free
transition)

Journal of Solar Energy Engineering MAY 1998, VoL 120 I 109

 SG6041 "' Re-1 00,000 o Re-300,000
Free transition o Re-150,000 "' Re-400,000
o Ae=200,000 t> Re-500,000
2,0 f-.r-.-,--,C"T"'-.-."""T""T"T'T-rT-,-,-.,-,-:-r,-......-,-.,....,..,...,~

1.5
1.5 -------r·------r-------
1.0
c, c/1.0 -----·-·r··--
0.5 0.5 -~------~ ______ ..... ______ _

0.0 •_________ . . . _______ _

0.0

-0.5 -0.5 L,c_.................__._._............._._._._,

o.oo om o.o2 o.o3 o.o4 o.os o.os -1 o 0 10 20
c. . a(deg)
SG6042 "' Re-1 00,000 o Re-300,000
Free transition o Ae-150,000 A Re-400,000

1.5 -------T·-------~--------- -------- 1.5 --------r-------~

1.0
Ce 1.0 -------r-- T------
0.5 . ------r-------
0.0
0.0 -- r-------r·----
·0.5 -0.5 w,...,_....................................................w
0.00 0.01 0.02 0.03 0.04 0.05 0.06 -10 0 10 20
c. . a(deg)
SG6043 v Re-100,000 o Ae-300,000
Free transition o Ae-150,000 A Re-400,000
2.0 j-r-,..,.....,...,..........,""T"'!..,.......,.....,_,...,.;.r""T.,....,..",..Rer-·rSOO,..,..:.;,OOO..T-1

1.5
. ------ r·----- 1.5

1.0 1.0

~~T=-
----r·----
0.5

: '
. -------r-----
--------j----------:----------t---------t------
: 1 :
0.0

1 ! ! !
0.0
·r-------r··---
-0.5 ..() .5 L.;....._............L..L...........r.....L..........'-L-.1
0.00 0.01 0.02 0.03 0.04 0.05 0.06 -10 0 10 20
c... a(deg)

Fig. 4 Drag polars and corresponding lift curves for the SG6041, SG6042, SG6043
primary airfoils (free transition)

wise wake surveys spaced 76.2 mm (3 in.) apart (Selig et a!., roughness elements has the advantage of being a more repeatable
!995; Lyon eta!., 1997a). The overall uncertainty in both the lift method, but the results obtained with the standard trips should bC
and drag measurements was estimated to be 1.5 percent ( Gug- considered as a worst case scenario.
lielmo, 1996; Selig eta!., 1995; Lyon eta!., I997a). All measure-
ments were corrected for wind .. tunnel interference effects ac- 4 Wind Tunnel Test Results
cording to a method that has been validated with data from the This section provides an overview of the airfoil data obtained
NASA Langley Low Turbulence Pressure Tunnel (Giguere and from the wind-tunnel tests conducted with the four new airfoils.
Selig, 1997c; Lyon eta!., 1997b). Finally, to simulate leading- The complete data set for a Reynolds number range of 100,000-
edge roughness effects, a 0.58 mm (0.023 in.) zigzag trip was 500,000 can be found in Lyon et al. ( 1997a). Eiren though this
fixed to the upper and lower surface of the airfoils (Lyon et a!., Reynolds number range might not covers the entire operational
I997a). In this paper, this trip size is referred to as the "standard" range of smalll-5 kW HA WTs, data above a Reynolds number
trip height. The trips were positioned in the vicinity of the suction of 500,000 can generally be obtained quite accurately from
peaks at two percent chord on the upper surface and at five percent logarithmic and linear extrapolation for the drag and lift, respec-
chord on the lower surface. Note that transition does not occur tively. The drag variations are largest below a Reynolds number
immediately behind the trips but rather takes place over a finite of 300,000 and thus, data in that range is more critical to docu-
distance (Lyon et a!., I997a). The use of trips instead of grit ment for small wind turbines.

110 I Vol. 120, MAY 1998 Transactions of the ASME

 Re =300,000
• SG604x Airfoil Family
100 o Previously existing airfoils
FX 63-137 (Vc=13.6%)'-
0

SG6042 (Vc=10%)'- , /
.g 90
('! ,//•'
Ol SD7032 (Vc=10%)
~ /' E387
''6 _-'' (Vc=9%) \
/
0
/GO 417a (Vc=2.9%)

~ 80 / o-A18 (Vc=7.3%) /SG6040 (Vc=16%)

E S6062 ,.. ... "'

/ Q"Clari<-Y (Vc=11.7%)
0
•
::I '-SD7037 (Vc=9.2%)
E (Vc=B%)'\ /
&'
\
-~ ~SD6060
(Vc=10.4%) o'S822 (Vc=16%)
:E 70 SG6041/
0 0
(Vc=10%)/ '-BW-3 (Vc=S%)
57012 0
(Vc=8.8%) '-NACA2414
RG15 (Vc=14%) 0
(Vc=8.9%)
'-se23 (Vc=21%)
60

0.6 0.7 0.8 0.9 1.0 1.1 1.2 u 1.4 1.5

Lift coefficient for maximum lift-to-drag ratio

Fig. 5 Maximum lift-to-drag ratio versus the corresponding lift coefficient of various airfoils for small HAWTs
(Re = 300,000)

Tests with free transition (no trips applied to the airfoil) were numbers, and the thickness of each airfoil is also shown. Note that
conducted at six Reynolds numbers from 100,000 to 500,000. all the data shown in Fig. 5 is based on wind tunnel experiments
Figures 3 and 4 present the drag polars and corresponding lift conducted with the testing apparatus and methods described in
curves for the root and primary airfoils, respectively. These figures Section 3. As indicated in Fig. 5, the new airfoils provide lift-to-
show that the design objectives have been met. For example, drag ratios that are equivalent or exceeding those of previously
the maximum lift-to-drag ratio of the SG6042 occurs near a lift existing low Reynolds airfoils over a wide range of design lift
coefficient of 0.9 at its design Reynolds number and also over coefficients. Therefore, small variable-speed HAWTs are likely to
most of the Reynolds number range considered. Furthermore, the benefit from enhanced energy capture from the use of the new
objective of providing a range of lift coefficients for which best lift- airfoils. Note, however, that comparing data for constant Reynolds
to-drag ratio performance occurs has also been satisfied. Therefore, number can be misleading based on the previously mentioned
performance penalties owing to off--design conditions should be tradeoff between operating lift coefficient and Reynolds number.
relatively small except for a Reynolds number of 100,000 where A better figure of merit to use in comparing the airfoils would be
laminar separation effects caused a large increase in drag. It should the maximum lift-to-drag ratio for a given reduced Reynolds num-
be emphasized, however, that such performance at a Reynolds ber (Giguere and Selig, 1997a). Nonetheless, the results shown
number of 100,000 is typical of most low Reynolds airfoils (Gi- in Fig. 5 are indicative of the potential of the new airfoils for small
guere and Selig, 1996; Selig et al., 1995). variable-speed wind turbines.
Figure 5 indicates the maximum lift-to-drag ratio and corre- For the tests with ''fixed'' transition, three Reynolds numbers
sponding lift coefficient under clean conditions of the SG604x were considered for each airfoil: 150,000, 300,000, and
airfoil family and other low Reynolds number airfoils applicable 500,000. An overview of these results is shown in Fig. 6 with
to small HAWTs. The results are shown for a Reynolds number results for a Reynolds number of 300,000. To facilitate compari-
of 300,000, which is representative of the data for other Reynolds son, the drag polar of each airfoil is shown for both free and

Re=300,000 o SG6040 v SG6041 Re=300.ooo o SG6040 v SG6041

Free transition "- SG6042 Fixed transition A SG6042
zigzag trip@ 2%us, 5%1s
2 _0 ,....._,,...,...,.,..,...,....,.--r-r~~T"'""..-.-..,...,...,...,~~..,.;;.sGS-T-T-04..:.:-t3

1.5

1.0
c,
0.5

0.0

-0.5 l...L.J'-'--'-'-'..............L-D..L..L..L..L..W....J-L.L...L..............'--'-'..........- ' - ' - '

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.01 0.02 0.03 0.04 0.05 0.06
c,.. c..
Fig. 6 Drag polars for the root and the primary airfoils at a Reynolds number of 300,000 (free and fixed
transition)

Journal of Solar Energy Engineering MAY 1998, Vol. 120 I 111

 _ SG6040 Re=300 000 SG6041 Re=300 000 SG6042 Re=300000 SG6043 Re=300 000
20 2.0 2.0 2.0

1.5

1.0
1.5

1.0
1.5

1.0
r~'{ 1.5

1.0

Ct c, c, c,
0.5 0.5 0.5 0.5

0.0 0.0 0.0 0.0

·0.5 L*-"'-'-"W.....-'-'...LJ....................., ..().5 ...................J..L........."-'............'-L.) ·0.5 =-................................................ . . . .

·10 0 10 20 ·10 0 10 20 ·10 0 10 20 0 10 20
o: (deg) o: (deg) o: (deg) o: (deg)

Fig. 7 Lift curves for the root and the primary airfoils at their design Reynolds numbers (free and fixed transition):
dark symbols, increasing a and open symbols, decreasing a

fixed transition. In these tests, the standard trip as defined in the airfoils with fixed transition are for the most part indepen-
Section 3 was used. As expected, the higher the design lift dent of the Reynolds number. Consequently, blades using the
coefficient and relative thickness of the airfoil, the higher is the SG6043 airfoil are likely to yield the best energy capture in
loss in performance caused by forcing transition. Furthermore, the presence of leading-edge roughness elements. Such blades,
the lift coefficient corresponding to the maximum lift-to-drag however, are also likely to be most affected by Reynolds num-
ratio is also influenced by the simulated roughness at the leading ber effects owing to the high design lift coefficient of the
edge. The change in lift coefficient for maximum lift-to-drag SG6043 airfoil.
ratio can be explained by the results presented in Fig. 7. Even The data shown so far with fixed transition has been for a
though the lift coefficients for maximum lift-to-drag ratio vary fixed trip height, namely the standard trip. The effect of reducing
between the free and fixed transition cases (Fig. 6), it can be the height of the trip was also investigated, and the results for
seen from Fig. 7 that the angle of attack for which maximum the SG6042 airfoil at a Reynolds number of 300,000 are shown
lift-to-drag ratio occurs remains, however, relatively constant. in Fig. 8. A total of five different zigzag trip heights were used
For example, Fig. 6 indicates that the lift coefficient for maxi- ranging from 0.13 mrn (0.005 in.) or 0.04 percent chord to
mum lift-to-drag ratio of the SG6043 airfoil with free and fixed 0.56 mm (0.023 in.-standard trip) or 0.19. percent chord. The
transition is 1.16 and 0.93, respectively. The lift curves for the results indicate that trips thicker than the smallest one caused
SG6043 airfoil shown in Fig. 7 indicates that these lift coeffi- a rapid degradation in airfoil performance.
cients correspond to an angle of attack of 4 deg. Therefore, Further tests were performed with plain trips to investigate
a blade designed with the new airfoils should yield optimum whether or not the new airfoils were optimized for low drag at
performance even with leading-edge roughness. Based on the their respective design points. The tradeoff to consider involves
loss in maximum lift coefficient shown in Fig. 7, the use of the reducing bubble drag at the cost of increased device drag-
new airfoils should not be extended to small stall-regulated drag caused by the trip itself. For the SG6042 airfoil, a 0.15
HA WTs. Airfoils such as the NREL S822 and S823 are better mm (0.006 in.) thick plain (tape type) trip was positioned at
candidates in that case. The new airfoils, due largely to lower different chord wise positions on the upper surface at the respec-
thickness, provide much better lift-to-drag ratio performance as tive design lift coefficient. The results shown in Fig. 9 clearly
compared with the S822 and S823 airfoils. indicate that the addition of the plain trip did not lead to any
Table 2 summarizes the maximum lift coefficient, maximum drag reduction beyond the experimental uncertainty of the mea-
lift-to-drag ratio and the corresponding lift coefficient for both surements. In addition, three additional plain trip heights were
the free and fixed transition cases of the SG604x airfoil family. also tested at 35 percent chord and the same Reynolds number
As expected, the results presented in Table 2 indicate that the of 300,000. As indicated in Fig. I 0, trips thicker than the base-
sensitivity to roughness of the airfoil performance increases line (0.15 mrn) were not beneficial in reducing the overall drag.
with the design lift coefficient (Tangier, 1997). Note, however, Similar results were also obtained for the SG6043 airfoil at
that for a given Reynolds number, the maximum lift-to-drag Reynolds numbers of 200,000 and 300,000. Therefore, the air-
ratios with fixed transition of the airfoil having the highest foils do have low bubble drag that likely cannot be reduced
design lift coefficient ( SG6043) remain the largest of all the through the use of a trip or turbulator. Finally, the four airfoils
new airfoils. In addition, the lift-to-drag ratio characteristics of considered as a group provide excellent lift-to-drag ratios over
a broad range of lift conditions and are well suited for their
intended application.
Table 2 Performance summary for the SG604x airfoil family

SG6040 root airfoil SG6041 (primary airfoil 5 Conclusions

FIH tnmaltlon Fixed lnmaiUon Free trlnsttlon Fixed transition
Re c...., IC>:{Udl- {Ud)- c...., ;(Ud),. (UdJ- c..., :(UdJ- {Ud),.. C...tt C,;{Udl- (Ud),.
The airfoils designed during this work form a unique airfoil
100,000 1.21l 1.16 46.0 1.08 1.15 0.66 51.5 1.08 family for small variable-speed HA WTs. Wind-tunnel tests of

1r
1.21 57.0 1.09 0.70 33.1 1.16 0.67 57.5 1.09 0.80 33.3
1.35 1.17 66.3 1.09 1.22 0.70 64.1 1.14
the root airfoil ( SG6040) and three primary airfoils ( SG6041,
1.39 1.11 78.5 1.11 0.84 34.7 1.29 0.65 72.2 1.16 0.87 36.7 SG6042, and SG6043) provided an extensive airfoil data set
1.42 1.13 83.5 1.13 1.34 0.60 60.0 1.19 - that can be used in the design of small blade~; The two primary
1.42 1.13 66.6 1.14 0.76 36.7 1.36 0.61 84.4 1.20 0.77 39.4
SG6042.JJ>rlm!ry airfoil SG6043(prtmary airfoil airfoils having the highest design lift coefficients (SG6042 and
Free trancltlon Fixedtnnaltlon Free transition Fixed transition SG6043) yielded enhanced lift-to-drag performance over many
c...., c...., C,;(Vd\- {Ud)-

er~
;{Ud)_ (1/d)_ ;(Ud). (Ud\- C...tt ;(1/d\- ((Ud),.
1.10 55.6 1.26 1.52 1.37 59.4 1.36
other low Reynolds number airfoils. Consequently, small vari-
1.29 0.89 59.7 1.27 0.83 36.7 1.56 1.31 74.2 1.36 1.01 42.0 able .. speed wind turbines are likely to benefit from the use of
1.41 1.01 77.8 1.26 1.59 1.33 66.6 1.40
300.00011 1.47 0.92 90.3 1.32 0.66 41.5 1.65 1.16 105.3 1.42 0.93 45.2
these two airfoils. The SG6040 root airfoil and low-lift SG604l
400,000111.50 0.93 101.0 1.33 1.68 1.17 118.0 1.44 airfoil are also likely to enhance the energy capture of small
500,000 1.52 0.84 105.9 1.34 0.90 45.7 1.70 1.10-1.2 125.1 1.43 0.96 46.4
variable-speed HA WTs owing to structural requirements and

112 I Vol. 120, MAY 1998 Transactions of the ASME

 SG6042 v No trip o hlc=0.12%
Re=300,000 o hlc=0.04% t> hlc=0.15%
zigzag trip 2%us, 5%1s <> hlc=O.OS% I> hlc=0.19%
2.0 2.0 r'""'~.,....,.~...-~r"T'"]

1.5 --- ·---r--------r--------:---------,·---------:-·------

' ' ' ' ' 1.5

. . .
Ct
1.0
. ----r--------r --- --;---- --- 1.0

0.5 0.5
---r-------r-------T---- ---:--------
0.0 -----~---------:-----·---'f·----~-- 0.0

0.02 0.03 0.04 0.05 0.06 0 10 20

C.t a (deg)

Fig. 8 Drag polars and corresponding lift curves for the SG6042 airfoil with varying
zigzag trip heights

the tradeoff between operating lift coefficient and Reynolds Acknowledgments

numbers. Overall, the new airfoils provide excellent lift-to-drag
The support from the National Renewable Energy Laboratory
ratios over a broad range of lift conditions and are well suited
under subcontract XAF-4-14076-03 is gratefully acknowledged.
for small variable-speed wind turbines.
Also, the authors would like to thank James L. Tangier of NREL
for many useful discussions throughout this work. Finally, the au-
thors are also thankful to Mark Allen of Allen Developments for
0.014
building the four wind-tunnel models used in this study.

0.012 References
E- Bjork, A .• 1988, "Airfoil Design for Horizontal Axis Wind Turbines," Pro-
0.010
~
-.- - ~ - ceedings of the Second lEA Symposium on Joint Aclion on Aerodynamics of Wind

~ Baseline drag coefficient (0.0103)

Turbines, Lyngby, Nov., pp. 204-230.
BjOrk, A., 1989, "Airfoil Design foc Variable RPM Horizontal Axis Wind Tur-
0.008 with experimental uncertainty of 1.5% bines," Proceedings EWEC 89, European Wind Energy Conference, Glasgow, Scot-
(J land.
SG6042 Bjork, A., 1990, "Coordinates and Calculation for the FFA-Wl-xxx, FFA-W2-
0.006 Plain trip h=0.05%c (0.006") x.u, FFA-W3-xxx Series of Airfoils for Horizontal Axis Wind Turbines," The
=
Re 300,000 Aeronautical Research Institute of Sweden, Technical Report, FFA TN 1990-15.
a =4.2 deg Drela, M., 1989, "XFOIL: An Analysis and Design System for Low Reynolds
0.004 '----· Number Airfoils,'' Lecture Notes in Engineering: Low Reynolds Number Aerody-
namics, T. J. Mueller, ed., Vol. 54, Springer-Verlag, New York, June.
0.002 Eppler, R., 1990, Airfoil Design and Data, Springer-Verlag, New York.
Giguere, P., and Selig, M. S., 1996, "Aerodynamic Performance of Small
Wind Turbines Operating at Low Reynolds Numbers," American Wind Energy
0.000 Association WIND POWER '96 Conference, June 23-27, Denver, CO.
0 10 20 30 40 50 60 Giguere, P., and Selig, M. S., 1997a, "Low Reynolds Number Airfoils for Small
Trip location on upper surface (% chord) Horizontal Axis Wmd Turbines," Wind Engineering, 1997, Vol. 21, No.6, 1997.
Giguere, P., and Selig, M. S., 1997b, "Desirable Airfoil Characteristics for
Fig. 9 Drag coefficient as a function of the chordwise position of a hI Large Variable-Speed Horizontal Axis Wind Turbines," ASME JoURNAL OF SO-
c = 0.05 percent plain trip on the upper surface LAR ENERGY ENGINEERING, Vol. 119, August, pp. 253-260.
Giguere P., and Selig, M.S., l997c, "Velocity Correction for Two Dimensional
Tests with Splitter Plates," AJAA J., Vol. 35, No.7, pp. 1195-1200.
Guglielmo, J., 1996, "Spanwise Drag Variations at Low Reynolds Numbers,"
Department of Aeronautical and Astronautical Engineering Department, Univer-
sity of Illinois at Urbana-Champaign, Urbana, IL, May.
Hill, G. C., and Garrard, A. D., 1989, "Design of Aerofoils for Wind Turbine
0.012
~
Use," Proceedings of the European Wind Energy Conference, Glasgow, Scotland.
Lyon, C. A., Broeren, A. P., Giguere, P., Gopalarathnam, A., and Selig, M. S.,
0.010 t=~;;:;.::r::;p~~~i:i'iSs~ts-EE'~-s-.-;:J:-:;]3.~~ 1997a, Summary of Low-Speed Airfoil Data-Volume 3, SoarTech Publications,

~Baseline drag coefficient (0.0103)

1504 N, Horseshoe Circle, Virginia Beach, VA.
Lyon, C. A., Selig, M. S., and Broeren, A. P .• l997b, "Boundary-Layer Trips
0.008 with experimental uncertainty of 1.5% on Airfoils at Low Reynolds Numbers," AIAA Paper 97-051l.
Madsen, H. A., and Rasmussen, F., 1992, "Experiences in Airfoil Analysis and
SG6042 Design," Proceedings of the Sixth lEA Symposium on the Aerodynamics of Wind
0.006 Plain trip at 35%c upper surface Turbines, Petten, Netherlands.
Re =300,000 Selig, M.S., and Maughmer, M.D., 1992, "Generalized Multipoint Inverse
a =4.2 deg Airfoil Design," AIAA J., Vol. 30, No. II, pp. 2618-2625.
0.004
Selig, M.S., Guglielmo, J. J., Broeren, A. P., and Giguere, P., 1995, Summary
of Low·Speed Airfoil Data- Volume 1, SoarTech Publications,. 1504 N. Horseshoe
0.002 Circle, Virginia Beach, VA.
Tangier, J. L., and Somers, D. M., 1995, "NREL Airfoil Families for
HA WTs," American Wind Energy Association WIND POWER '95 Conference,
May 9-12, Washington, DC.
Q05 Q10 Q15 0.20 Tangier, J. L., 1997, "Influence of Pitch, Twist and Taper on a Blade's Perfor-
Trip height(% chord) mance Loss Due to Roughness," ASME JOURNAL OF SoLAR ENERGY ENGI-
NEERtNG, Vol. 119, pp. 248-252.
Fig. 1 0 Drag coefficient as a function of the normalized height of a plain Timmer, W. A., and Rooy, R. P. J. 0. M. van, 1992, "Thick Airfoils for
trip positioned at 35 percent chord on the upper surface HA WTs," J. of Wind Engineering and Industrial Aerodynamics. Vol. 39, May.

Journal of Solar Energy Engineering MAY 1998, Vol. 120 I 113

 APPENDIX: Airfoil Coordinates
[SG6040 Jl:ili~U4 ISG6042 [SG6043
XJC ylc xlc yc x/C Yfc x/C y/c
1.00000 o.ouuuu I.OUUUU .. 0.00000 1.00000 0.00000 1.llOi:lOO u.ooooo
0.99788 0.00048 0.99801 0.00029 0.99806 0.00048 0.99811 0.00066
0.99183 0.00207 0.99227 0.00132 0.99249 0.00203 0.99274 0.00271
0.98233 0.00477 0.98311 0.00310 0.98372 0.00483 0.98439 0.00607
0.96977 0.00845 0.97084 0.00560 0.97210 0.00811 0.97343 0.01047
0.95442 0.01283 0.95568 0.00867 0.95784 0.01223 0.96007 0.01552
0.93637 0.01766 0.93778 0.01217 0.94102 0.01672 0.94429 0.02092
0.91559 0.02285 0.91722 0.01600 0.92161 0.02148 0.92597 0.02655
0.89221 0.02845 0.89412 0.02018 0.89968 0.02657 0.90516 0.03248
0.86647 0.03446 0.86872 0.02469 0.87546 0.03195 0.88207 0.03868
0.83661 0.04080 0.84127 0.02946 0.84915 0.03756 0.85688 0.04510
0.80887 0.04739 0.81200 0.03438 0.82098 0.04333 0.82979 0.05165
0.77752 0.05415 0.78119 0.03933 0.79117 0.04914 0.80101 0.05824
0.74482 0.06094 0.74901 0.04408 0.75995 0.05469 0.77074 0.06478
0.71104 0.06759 0.71552 0.04849 0.72757 0.06038 0.73922 0.07114
0.67632 0.07394 0.68087 0.05261 0.69410 0.06544 0.70666 0.07717
0.64081 0.07992 0.64526 0.05635 0.65962 0.07006 0.67320 0.08268
0.60477 0.08541 0.60884 0.05967 0.62435 0.07417 0.63889 0.08761
0.56837 0.09032 0.57186 0.06258 0.58844 0.07772 0.60396 0.09190
0.53180 0.09461 0.53448 0.06502 0.55208 0.08071 0.56854 0.09551
0.49526 0.09815 0.49692 0.06699 0.51546 0.08310 0.53276 0.09842
0.45891 0.10089 0.45939 0.06847 0.47877 0.08486 0.49685 0.10059
0.42297 0.10279 0.42211 0.06944 0.44221 0.08600 0.46096 0.10201
0.38758 0.10378 0.38529 0.06991 0.40597 0.08648 0.42528 0.10269
0.35286 0.10386 0.34915 0.06987 0.37024 0.08634 0.39000 0.10260
0.31903 0.10299 0.31389 0.06931 0.33522 0.08556 0.35527 0.10175
0.26619 0.10116 0.27971 0.06624 0.30110 0.08415 0.32131 0.10017
0.25449 0.09843 0.24683 0.06666 0.26808 0.08215 0.28827 0.09788
0.22405 0.09479 0.21540 0.06458 0.23633 0.07956 0.25633 0.09491
0.19496 0.09028 0.18563 0.06201 0.20603 0.07639 0.22567 0.09127
0.16734 0.08506 0.15767 0.05896 0.17735 0.07269 0.19645 0.08702
0.14137 0.07921 0.13169 0.05546 0.15044 0.06847 0.16882 0.08220
0.11719 0.07283 0.10783 0.05149 0.12544 0.06379 0.14293 0.07684
0.09496 0.06598 0.08618 0.04710 0.10249 0.05866 0.11890 0.07100
0.07481 0.05875 0.06690 0.04231 0.08169 0.05313 0.09687 0.06473
0.05686 0.05123 0.05002 0.03711 0.06316 0.04726 0.07694 0.05809
0.04123 0.04347 0.03555 0.03159 0.04898 0.04105 0.05920 0.05117
0.02798 0.03559 0.02361 0.02578 0.03315 0.03458 0.04376 0.04398
0.01725 0.02771 0.01411 0.01971 0.02179 0.02795 0.03062 0.03659
0.00913 0.01987 0.00702 0.01366 0.01284 0.02119 0.01983 0.02916
0.00362 0.01212 0.00241 0.00767 0.00619 0.01446 0.01142 0.02171
0.00072 0.00449 0.00016 0.00194 0.00197 0.00797 0.00528 0.01436
0.00027 -0.00268 0.00029 -0.00270 0.00011 0.00182 0.00150 0.00741
0.00289 -0.00884 0.00378 -0.00645 0.00035 -0.00341 0.00002 0.00094
0.00906 -0.01450 0.01080 -0.01010 0.00371 -0.00730 0.00059 -0.00488
0.01829 -0.02023 0.02073 -0.01346 0.01068 -0.01042 0.00393 -0.00860
0.03027 -0.02580 0.03361 -0.01837 0.02068 -0.01311 O.Q1077 -0.01120
0.04483 -0.03103 0.04948 -0.01888 0.03369 -0.01516 0.02087 -0.01316
0.06202 -0.03566 0.06829 -0.02104 0.04982 -0.01659 0.03409 -0.01430
0.08193 -0.03971 0.08990 -0.02294 0.06901 -0.01751 0.05050 -0.01460
0.10448 -0.04336 0.11413 -0.02461 0.09117 -0.01805 0.07015 -0.01422
0.12939 -0.04657 0.14081 -0.02605 0.11611 -0.01826 0.09292 -0.01329
0.15655 -0.04937 0.16976 -0.02728 0.14365 -0.01821 0.11864 -0.01197
0.18574 -0.05172 0.20077 -0.02828 0.17358 -0.01792 0.14712 -0.01034
0.21677 -0.05361 0.23363 -0.02907 0.20570 -0.01744 0.17815 -0.00848
0.24943 -0.05503 0.26812 -0.02964 0.23979 -0.01679 0.21150 -0.00645
0.29349 -0.05594 0.30402 -0.02998 0.27559 -0.01599 0.24693 -0.00430
0.31873 -0.05832 0.34108 -0.03010 0.31286 -0.01506 0.28417 -0.00210
0.35490 -0.05613 0.37907 -0.02997 0.35135 -0.01401 0.32296 0.00011
0.39177 -0.05528 0.41772 -0.02959 0.39078 -0.01286 0.36301 0.00232
0.42916 -0.05371 0.45679 -0.02894 0.430.99 -0.01159 0.40401 0.00450
0.46691 -0.05132 0.49604 -0.02797 0.47140 -0.01016 0.44571 0.00669
0.50479 -0.04801 0.53524 -0.02663 0.51208 -0.00853 0.48785 0.00893
0.54298 -0.04368 0.57423 -0.02493 0.55277 -0.00671 0.53029 0.01117
0.58164 -0.03663 0.61284 -0.02285 0.59330 -0.00475 0.57286 0.01327
0.62049 -0.03323 0.65090 -0.02037 0.63345 -0.00273 0.61525 0.01504
0.65925 -0.02772 0.88832 -0.01783 0.67298 -0.00077 0.65704 0.01635
0.69762 -0.02231 0.72492 -0.01482 0.71161 0.00102 0.69785 0.01721
0.73526 -0.01718 0.76041 -0.01210 0.74904 0.00255 0.73735 0.01760
0.77181 -0.01250 0.79448 -0.00956 0.78494 0.00377 0.77520 0.01751
0.80690 -0.00840 0.82683 -0.00728 0.81900 0.00462 0.81109 0.01695
0.84016 -0.00497 0.85718 -0.00532 0.85093 0.00510 0.84488 0.01592
0.87120 -0.00228 0.88524 -0.00370 0.88041 0.00521 0.87567 0.01447
0.89963 -0.00033 0.91073 -0.00243 0.90718 0.00496 0.90377 0.01265
0.92511 0.00090 0.93339 -0.00149 0.93097 0.00442 0.92871 0.01056
0.94728 0.00148 0.95300 -0.00083 0.95153 0.00364 0.95024 0.00828
0.96587 0.00154 0.96938 -0.00036 0.96868 0.00273 0.96814 0.00591
0.98063 0.00122 0.98244 -0.00005 0.98224 0.00178 0.98215 0.00361
0.99132 0.00070 0.99204 0.00008 0.99204 0.00090 0.99209 ~ ,.0.00170
0.99782 0.00022 0.99798 0.00006 0.99800 0.00026 0.99803 0.00045
1.00000 0.00000 1.00000 0.00000 1.00000 0.00000 1.00000 0.00000

114 I Vol. 120, MAY 1998 Transactions of the ASME