Energy 46 (2012) 62e71

Contents lists available at SciVerse ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Airfoil optimization for noise emission problem and aerodynamic performance

criterion on small scale wind turbines
Tuhfe Gmen a, *, Bars zerdem a, b
a
b

_
Department of Mechanical Engineering, Izmir
Institute of Technology, Glbahe, 35430 Urla-Izmir, Turkey
Department of Energy Systems Engineering, Bahcesehir University, 34353 Besiktas-Istanbul, Turkey

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 13 September 2011
Received in revised form
17 May 2012
Accepted 18 May 2012
Available online 18 June 2012

Noise emission is one of the major concerns in wind turbine industry and especially for small scale wind
turbines, which are mostly erected to the urban areas; the concern is turning into a problem. This paper
focuses on the optimization of six airfoils which are widely used on small scale wind turbines in terms of
the noise emission and performance criteria and the numerical computations are performed for a typical
10 kW wind turbine. The main purpose of this optimization process is to decrease the noise emission
levels while increasing the aerodynamic performance of a small scale wind turbine by adjusting the
shape of the airfoil. The sources of the broadband noise emission are dened and their dominancy is
investigated with respect to the operating conditions. While redesigning, together with the principals of
reducing the airfoil self-noise, the aerodynamic prospects of increasing the performance have been taken
into account. The codes which are based on aero-acoustic empirical models and a collection of wellknown aerodynamic functions are used in this study. The results obtained from the numerical analysis
of the optimization process have shown that, the considered commercial airfoils for small scale wind
turbines are improved in terms of aero-acoustics and aerodynamics. The pressure sides of the baseline
airfoils have been manipulated together with the trailing edge and redesigned airfoils have lower levels
of noise emission and higher lift to drag ratios.
2012 Elsevier Ltd. All rights reserved.

Keywords:
Small scale wind turbine
Noise emission
Aerodynamic performance
Aero-acoustics
Airfoil optimization

1. Introduction
The technological research on sustainable energy has been
propagating and accelerating all around the world as a result of the
energy politics gained today. In this context, the wind power is
playing an important role contributing to this development. The
small scale wind turbines seem to contribute a lot to this development, especially in Turkey, because of its potential and the recent
legal arrangement done on this subject. However, the wind turbine
noise is one of the major obstacles on the widespread use of wind
power. The main sources of wind turbine noise may be divided into
two: mechanical noise which includes the noise from the fans,
generator, gear box etc., and the aerodynamic noise which is originated from the interaction between the rotor and the wind.
Nowadays the mechanical noise is mostly handled; therefore the
main engineering effort is being made on the subject of aerodynamic noise research [1]. There are ve self-noise mechanisms
on an airfoil, aerodynamically: Turbulent boundary layer trailing

* Corresponding author. Tel.: 90 232 7506789; fax: 90 232 7506701.

E-mail address: tuhfegocmen@iyte.edu.tr (T. Gmen).
0360-5442/$ e see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.energy.2012.05.036

edge noise, separation e stall noise, laminar boundary layer vortex

shedding noise, trailing edge bluntness e vortex shedding noise
and turbulent inow noise [2]. Some early researchers were
considering the tip vortex formation as the fth noise mechanism
[3].
Turbulent boundary layer trailing edge noise; is the noise that is
occurred due to the formation of turbulence and wakes around the
trailing edge of the airfoil. Separation e stall noise exists where the
ow is separated around the airfoil. Laminar boundary layer vortex
shedding noise occurs when a laminar boundary layer exists on any
side of an airfoil. Trailing edge bluntness e vortex shedding noise is
the noise that is formed due to the vortex shedding caused by the
bluntness of the trailing edge. Turbulent inow noise is due to the
characteristics of the incoming ow.
The optimization process, which is based on the geometrical
changes done on the airfoils, has been applied to proles of FX 63137, S822, S834, SD2030, SG6043 and SH3055. They are chosen to
be optimized since they may be considered to be in widely use in
small scale wind turbines [4].
The optimization process starts with the prediction of noise
generation and the results of such analysis can be found in the
studies of Kim et al. [5] in which turbulent inow and turbulent

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

63

Fig. 1. The baseline airfoils.

boundary layer trailing edge noises are analyzed and Son et al. [6] in
which the calculations of self-noise are adjusted to the calculations
of large range noise propagation with the terrain effects. Also, Jiang
et al. [7] experimentally investigated the sound generation due to
vortex shedding. In order to achieve airfoil geometry that emits
lower levels of noise and yet aerodynamically sufcient enough, the
algorithmic optimization methods have been used in the literature.
Bertagnolio et al. [8] has considered the Trailing Edge Noise (TNO)
model by Parchen et al. [9] and the model is put into the optimization program called AIRFOILOPT designed at Ris Laboratories.
Marsden et al. [10] has applied the derivative-free optimization
technique which minimizes the airfoil noise output due to vortex
shedding. Another optimization model has been developed by
Jouhaud et al. [11] and 2D Naca subsonic airfoil has been modied.

Those methods are shown to be quite successful but require

detailed study on the optimization techniques and computationally
affordable for only two airfoil self-noise mechanisms at most.
Therefore, more practical methods needed to be found. Bai et al.
[12] placed a ow deector on leading and trailing edges of S809
and FX60-100 airfoils and the ow separation is reduced, therefore

Fig. 2. Angles used in directivity functions.

Table 1
Turbine geometry factors and operating conditions.
Average wind speed (m/s)
Hub height (m)
Blade radius (m)
Rated power (kW)
Number of blades

6
25
7
10
3

Fig. 4. Lift to drag ratios for FX 63-137 and modied.

Fig. 3. Airfoil geometry comparison (FX 63-137 and modied).

64

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

Fig. 5. SPL spectra for FX 63-137 and modied.

noise is decreased where lift to drag ratio is increased. A more

practical airfoil shape optimization has been done by decreasing
the maximum thickness of the baseline airfoil basically [13]. This
study is based on the geometrical approaches around the pressure
side and the trailing edge of the prole. Those geometrical
approaches include thinner trailing edge and engraved surface in
the lower part of the airfoil and all ve mechanisms of airfoil selfnoise have been taken into account. With this study, it was planned
to reveal the basic geometrical specications that an airfoil should
have in order to reduce small scale wind turbine noise emission
levels and six airfoils have been studied as examples. It is concluded
that considered commercially available airfoils can be improved to
emit lower self-noise and have better aerodynamic performances.

and is said to have capability to perform separation and stall

calculations accurately.
While comparing the aerodynamic performances, according to
the study edited by Spera et al. [16] the ratio of lift coefcient to
drag coefcient is considered as the most important criterion.
The NAFNoise code [20] is used for the prediction of the airfoil
self-noise and the numerical algorithm is based on the semiempirical models that are achieved by Brooks et al. [3], for the
rst four noise sources. For the last noise source that is listed as the
turbulent inow noise, the semi-empirical formula obtained by
Amiet et al. [17] is taken into account. The validation of NAFNoise is
studied by Moriarty et al. [18] for Mach number of 0.2 which is the
case of this study.

2. Material and method

2.1. Semi-empirical noise models

The XFOIL code is used to predict the aerodynamic performances of the airfoils [14] which are shown in Fig. 1. The linear
vorticity stream function panel method is used in this numerical
approach and 2D ow eld is constructed by three components:
free stream ow, airfoil surface vortex sheet, and source sheet. The
domain which is composed of airfoil contour and the wake is discretized into panels and each panel is assumed to have a linear
vorticity prole and constant source strength. According to the
study done by Hoogedoorn et al. [15], XFOIL has been validated
especially for high Reynolds numbers which is the case of this study

2.1.1. Turbulent boundary layer trailing edge noise

This is the rst and perhaps the most common airfoil self - noise
source especially for high Reynolds numbers. Based on the studies
done by Brooks et al. [3] the turbulent boundary layer trailing edge
noise along the pressure side of the airfoil is modeled as;

SPLp 10log

d*p M5 LDh
re2


A

And for the suction side;

Fig. 6. Airfoil geometry comparison (S822 and modied).

Stp
St1

K1  3 DK1

(1)

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

65

where d* d*(a, Rec) is the boundary layer displacement thickness,

based on a, the angle of attack, and Rec, the Reynolds number based
on chord. The subscript p refers to the pressure side and s to the
suction side of the airfoil. Other parameters in Eqs. (1) and (2) are L,
the span of the airfoil section; re, the effective observer distance;
and A, an empirical spectral shape based on the Strouhal number,
St (fd*/U), where f is the frequency, and U is the local mean
velocity. Three other empirical relations are also used,
St1 0.02M0.6, K1 K1 (Rec), and DK1 DK1 (a, Red*). Also, Dh, the
directivity function is dened in the equations since it is a correction factor considering the distance between source and the
observer and therefore the Doppler effect and it may be described
as [1];

Dh Qe Fe z



2sin2 1=2Qe sin2 Fe
1 Mcos Qe 1 M  Mc cos Qe

(3)

where M is the Mach number for airfoil section and Mc is the

convective Mach number and Qe and Fe are the directivity angles
which are determined by Fig. 2.
2.1.2. Separation e stall noise
For an airfoil, as the angle of attack, a, increases within its boundaries, the separation around the suction side of the prole increases
and that situation becomes more dominant on producing the noise.
The semi e empirical formula of separation e stall noise, SPLa is very
similar to Eq. (2) except that it has a remarkable dependency on a.
The resulting scaling model for turbulent boundary layer trailing
edge noise with noise due to separation and stall characteristics is
then [1];



SPL
SPL
SPL
SPLTBLTE 10log 10 p=10 10 s=10 10 a=10

(4)

Fig. 7. Lift to drag ratios for S822 and modied.

SPLs 10log

d*s M5 LDh
re2

 
Sts
K1  3
A
St1

(2)

2.1.3. Laminar boundary layer vortex shedding noise

Since the characteristic of that type of noise is feedback amplication, it is tonal and therefore is not disturbing for larger scale
wind turbines with higher Reynolds number. However, it is still
a major concern for small scale ones which are the subject of this

Fig. 8. SPL spectra for S822 and modied.

66

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

Fig. 9. Airfoil geometry comparison (S834 and modied).

study. The laminar boundary layer vortex shedding noise is

described by Brooks et al. [3] as;

d*p M5 LDh

SPLLBLVS 10log

!
G1

re2

St 0
0
Stpeak


G2

Rec
Rec 0

G3 a

K pfc/U, is the local wave number, where f is the frequency of

interest, c is the local airfoil chord length, and U is the local velocity
over the airfoil section, DL is a low-frequency directivity function,
LFC is a low e frequency correction factor, S is the compressible
Sears function; and b2 1M2.
3. Results and discussion

(5)
where most of the variables are already dened in Eq. (1) Other
than those, which are G1, G2, and G3 are empirical functions, St0 is
0
0
Stpeak
(Rec) and it is the
the Strouhal number based on dp, Stpeak
peak Strouhal number and (Rec)o is a reference Reynolds number
depending on the angle of attack.
2.1.4. Trailing edge bluntness e vortex shedding noise
Since this source of noise emission from wind turbines is
directly related with the airfoil geometry, the model includes some
geometry-dened parameters. Trailing edge bluntness e vortex
shedding noise is described as;

d*p M5 LDh

SPLTEBVS 10log

re2

0
G5 @

d*avg

1
h
@
G4 * ; JA
1

; J;

During the optimization process, main idea was to change the

shape of the pressure side of the airfoil since it is aerodynamically
critical. Also the trailing edge of each airfoil has been well studied
since most of the self-noise is originated from that location. Note
that the calculations are done for a typical 10 kW wind turbine with
a radius of 7 m. The ow characteristics around the airfoils have
been identied using the studies of nveren et al. [19] and the wind
speed is averagely taken to be 6 m/s for a hub height of 25 m. The
Reynolds number of that ow is calculated around 3.0  105
and Mach number calculation gives a value of 0.02 for an

davg

St 00 A
00
Stpeak

(6)
*

where h is the trailing edge thickness, davg is the average

displacement thickness for both sides of the airfoil, J is the solid
angle between both airfoil surfaces just upstream of the trailing
*
00
00
Stpeak
(h/davg ) is
edge, St00 is the Strouhal number based on h, Stpeak
the peak Strouhal number; and G4 and G5 are empirical functions of
these parameters.
2.1.5. Turbulent inow noise
The incoming ow characteristics play an important role in this
type of noise emission from the wind turbines. Different from the
previous four models, this semi e empirical formula is studied by
Amiet et al. [17].


LFC
1 LFC

SPLInflow SPLH
Inflow 10log
SPLH
Inflow

10log

r20 c20 lL
2re2

3 2 2

M u I 
7 DL
1 K2 3

2

LFC 10S2 MK 2 b

2pK

b2

1 2:4

K3

(7)
!
58:4

(8)

(9)
K

b2

!1 !1
(10)

where ro is the air density, co is the speed of sound, l is a turbulence

length scale, u is the mean wind speed, I is the turbulence intensity,

Fig. 10. Lift to drag ratios for S834 and modied.

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

Fig. 11. SPL spectra for S834 and modied.

Fig. 12. Airfoil geometry comparison (SD2030 and modied).

incompressible ow. The results are satised with the desired

aerodynamic performances (Table 1).

3.1. Optimization of FX 63-137

FX 63-137 has maximum thickness of 13.71% at 30.9% of the
chord and it has maximum camber of 5.97% at 53.3% of the chord
and it is optimized to have 9.05% thick at 26.7% of the chord with
a maximum camber of 7.75% at 43.6% at chord which is graphically
shown in Fig. 3.
XFOIL code has been used to analyze and compare the most
critical aerodynamic performance parameter which is the ratio of
lift coefcient, cL, and drag coefcient, cD, with respect to the angle
of attack, alpha and the outcome is shown graphically in Fig. 4.
As can be seen from Fig. 4, the lift to drag ratio of the modied
version of FX 63-137 is more than 110 where for FX 63-137 that
value is slightly above 100.
The noise emission is compared by using the sound pressure
level (SPL) spectra which is the outcome of the noise analysis tool,
NAFNoise. For FX 63-137 airfoil optimization, the comparison
between the SPL in dB is shown in Fig. 5.
The SPL spectra of FX 63-137 and FX 63-137modied show that
the optimized version has a little lower levels of noise emission.
Since the baseline prole is originally thin, the difference between
the sound pressure levels is not that signicant.

3.2. Optimization of S822

The maximum thickness of that airfoil is 16.01% at 39.2% of the
chord and it has maximum camber of 1.92% at 59.5% of the chord.
The optimized geometry is 14.10% thick at 24.3% of the chord with

Fig. 13. Lift to drag ratios for SD2030 and modied.

67

68

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

Fig. 14. SPL spectra for SD2030 and modied.

Fig. 15. Airfoil geometry comparison (SG6043 and modied).

a maximum camber of 5.15% at 61.0% of the chord. The geometry

comparison between S822 and S822modied is shown in Fig. 6.
As can be seen from the Fig. 7, the pressure side of the baseline
airfoil has been changed by increasing its curvature towards the
trailing edge which is more effective in aerodynamics point of view.
The comparison of performance obtained from the XFOIL tool is
shown in Fig. 7.
Fig. 7 shows that, the lift to drag ratio has been increased from
70 to almost 95 with the help of the optimized geometry.
The noise emission levels are shown for S822 and its modied
prole in Fig. 8.
The difference in the sound pressure levels of S822 and
S822modied can be seen clearly in Fig. 8. Since the baseline prole
is relatively thick, the geometry is easier to modify and the
improvements in aerodynamics and noise emission level is
signicant.
3.3. Optimization of S834
S834 has 15.0% thickness at 39.5% at the chord and its maximum
camber is 1.63% at 60.0% of the chord. The modied shape is 12.93%
thick at 23.3% of the chord with a maximum camber of 4.69% at
63.1% of the chord. The comparison of those two geometries is
shown in Fig. 9.
Using the same approach with the modication of S822, the
pressure side of S834 is carved smoothly towards the trailing edge.
The performance outcome of that geometric change may be seen in
Fig. 10.
The maximum lift to drag ratio has been shifted from 70 to 90
which means a valuable progress in the performance of the airfoil
in terms of wind turbine applications.

Fig. 16. Lift to drag ratios for SG6043 and modied.

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

Fig. 17. SPL spectra for SG6043 and modied.

Fig. 18. Airfoil geometry comparison (SH3055 and modied).

The noise levels are compared in Fig. 11 which has been obtained by the tool of NAFNoise.
The sound pressure level spectra for S834 optimization shows
that the noise emission level of modied shape is clearly reduced. It
can also be interpreted from the Fig. 11 that the decrease is more
signicant in the frequency range of 50e2000 Hz.

3.4. Optimization of SD2030

SD2030 is the baseline airfoil for this optimization and it has
geometrical properties as; maximum thickness of 8.56% at 35.2% of
the chord and maximum camber of 2.25% at 45.7% of the chord. The
resulting optimized airfoil has; 8.52% thickness at 30.2% at the
chord with a maximum camber of 2.94% at 54.0% of the chord. The
geometries are shown in Fig. 12.
Since the airfoil is relatively thin, the effective way of modifying
may be to change the camber characteristics and Fig. 12 shows the
modication.
The aerodynamic performance comparison of those airfoils is
given in Fig. 13.
Even though geometrically the difference is not that signicant,
since the pressure side of the baseline airfoil has been studied, the
performance increased in certain extent. In Fig. 13, it can easily be
seen that with the modication, not only the lift to drag ratio has
been increased but also the distribution of that ratio becomes more
stable with respect to angle of attack.
Since the baseline airfoil is thin, as mentioned, the difference not
being signicant was expected. However, the SPL spectra show in
Fig. 14 that the decrease in the noise level is achieved especially
around 4000e7000 Hz ranges.

Fig. 19. Lift to drag ratios for SH3055 and modied.

69

70

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

Fig. 20. SPL spectra for SH3055 and modied.

3.5. Optimization of SG6043

SG6043 is another relatively thin airfoil used in the small scale
wind turbines with a thickness of 10.0% at 32.3% of the chord and
maximum camber of 5.45% at 48.8% of the chord. SG6043modied
which is the optimized airfoil has a maximum thickness of 7.13% at
25.9% of the chord and its camber is distributed with a maximum
value of 6.98% at 46.6% of the chord. They are seen schematically in
Fig. 15.
The optimization process of SG6043 includes the changes made
in thickness and camber distribution, as can be seen in Fig. 15. The
outcome of this optimized pressure side aerodynamically is shown
in Fig. 16.
SG6043modied has higher maximum lift to drag ratio around
its design angle of attack, adesign. Fig. 16 also shows that the values
of cL/cD of the optimized shape are below the ones that belong to
the baseline airfoil. However, since corresponding angle of attacks
are already in the region where stall occurs, the performance of the
wind turbine will be at its minimum regardless to its lift to drag
ratio. Thus, the optimization may still be considered satisfactory in
terms of aerodynamic performance.
The airfoil self-noise characteristics of those two proles are
given in Fig. 17.
The noise emission levels differ tremendously especially for
relatively lower frequencies because of the fact that the modied
shape of SG6043 prole has smoother trailing edge and the overall
thickness is lower.
3.6. Optimization of SH3055
SH3055 is one of the widely used airfoil in small scale wind
turbine business and it is really successful in terms of consumer
satisfaction. However, this optimization process shows that even
this prole may be modied to have lower noise emission levels
while its performance is increased.
The maximum thickness of SH3055 is 15.11% at 25.2% of the
chord and the maximum camber of this prole is 5.40% at 62.9% of
the chord. The modied version, SH3055 modied, is 13.14% thick
at 26.5% of the chord where the maximum camber is 6.57% at 57.4%

of the chord. The geometrical comparison of those two airfoils is

included in Fig. 18.
As can be seen from Fig. 18, especially towards the leading edge,
the overall thickness of the baseline airfoil is reduced. The reason
for that is the trailing edge performance of SH3055 is satisfactory
and no need for further improvements. However, by sharpening the
nose section of the prole has improved the ow uniformity
around the airfoil and therefore reduced the additional boundary
layer formation which directly reduces noise and increases the
performance.
The performance criteria of the airfoils mentioned above is
compared in Fig. 19.
The cL/cD distribution of SH3055 modied is not only higher but
also more stable meaning that there is no signicant peak points
which is important for the turbine to perform at its best in various
angles of attack.
As can be seen from Fig. 20, the noise emission level of SH3055 is
also decreased quite signicantly.
4. Conclusion
Airfoil low noise and high performance design optimization has
been carried out for a typical three bladed, 10 kW wind turbine
which has a radius of 7m. In this study, FX 63-137, S822, S834,
SD2030, SG6043 and SH3055 which are the most widely used
airfoils on the small scale wind turbine have been studied and each
of them have been developed to have lower noise emission and
higher aerodynamic performance, separately. In order to achieve
that goal, the ow analysis tool XFOIL and airfoil self-noise analysis
tool NAFNoise have been used. While changing the geometry of the
airfoils, especially the pressure side and the trailing edge characteristics have been modied since they are the most critical locations particularly for aerodynamic and aero-acoustic concerns,
respectively. In order to compare those two outcomes of the optimization, lift to drag criterion and SPL spectra has been drawn. The
difference between the baseline and optimized versions of relatively thinner airfoils are not seen signicantly because the
geometrical changes made on those thin proles are not allowed
to be substantial. However, even for the thin proles, the

T. Gmen, B. zerdem / Energy 46 (2012) 62e71

improvement in noise and performance characteristics is clearly

shown. It can be seen that, with the geometrical optimization
applied, the lift to drag ratios are increased whereas the noise
emission levels are decreased up to 5 dB. Each of the optimized
airfoils has been shown to have better aerodynamic performance
and lower noise emission than their baseline proles.
Nomenclature

d*
boundary layer displacement thickness, m
a, alpha the angle of attack, deg
Reynolds number based on chord
Rec
L
span of the airfoil section, m
effective observer distance, m
re,
A
empirical spectral shape based on the
St
Strouhal number
f
frequency, Hz
U
local mean velocity, m/s
directivity function
Dh
M
Mach number
convective Mach number
Mc
Qe , Fe directivity angles
separation e stall noise, dB
SPLa
SPLTBL-TE turbulent boundary layer trailing edge noise, dB
trailing edge noise along the pressure side of the airfoil
SPLp
trailing edge noise along the suction side of the airfoil
SPLs
SPLLBL-VS laminar boundary layer vortex shedding noise, dB
G1, G2, G3 empirical functions
Strouhal number based on dp*
St0
0
peak Strouhal number
Stpeak
reference Reynolds number depending on the angle of
(Rec)o
attack
SPLTEB-VS trailing edge bluntness e vortex shedding noise, dB
h
trailing edge thickness, m
d*avg
average displacement thickness for both sides of the
airfoil, m
J
solid angle between both airfoil surfaces just upstream of
the trailing edge, deg
Strouhal number based on h
St00
00
peak Strouhal number based on h
Stpeak
empirical functions
G4 , G 5
SPLInow turbulent inow noise, dB
ro
air density, kg/m3
co
speed of sound, m/s
l
turbulence length scale, m
u
mean wind speed, m/s
I
turbulence intensity, %
K
local wave number

c
DL
LFC
S

71

local airfoil chord length, m

low-frequency directivity function
low-frequency correction factor
compressible Sears function

References
[1] Wagner S, Bareiss R, Guidati G. Wind turbine noise. Germany: Springer Verlag;
1996.
[2] Leloudas G, JZhu W, Srensen JN, Shen WZ, Hjort S. Prediction and reduction
of noise for a 2.3 MW wind Turbine, the Science of making torque from wind.
Journal of Physics: Conference Series 2007;75:1e9.
[3] Brooks TF, Pope DS, Marcolini MA. Airfoil self-noise and prediction. NASA
Ref. Pub; 1989. p. 1218.
[4] Migliore P, Oerlemans S. Wind tunnel aeroacoustic tests of six airfoils for use
on small wind turbines. National Renewable Energy Laboratory; 2003.
[5] Kim H, Lee S, Son E, Lee S, Lee S. Aerodynamic noise analysis of large horizontal axis wind turbines considering uid-structure interaction. Renewable
Energy 2012;42:46e53.
[6] Son E, Kim H, Kim H, Choi W, Lee S. Integrated numerical method for the
prediction of wind turbine noise and the long range propagation. Current
Applied Physics 2010;10:S316e9.
[7] Jiang M, Li XD, Zhou JJ. Experimental and numerical investigation on sound
generation from airfoil-ow interaction. Applied Mathematics and
Mechanics-English Edition 2011;32(6):765e76.
[8] Bertagnolio F, Madsen HA, Bak C. Trailing edge noise model validation and
application to airfoil optimization. Journal of Solar Energy-Transactions of
ASME 2010;132(3).
[9] Parchen R. Progress report DRAW: a prediction scheme for trailing edge noise
based on detailed boundary-layer characteristics. The Netherlands: TNO
Institute of Applied Physics. Report No. HAG-RPT-980023, http://www.tno.nl/
index.cfm; 1998.
[10] Marsden AL, Wang M, Dennis JE, Moin P. Suppression of vortex-shedding
noise via derivative-free shape optimization. Physics of Fluids 2004;16(10):
L83e6.
[11] Jouhaud JC, Sagaut P, Montagnac M, Laurenceau J. A surrogate-model based
multidisciplinary shape optimization method with application to a 2D
subsonic airfoil. Computers and Fluids 2007;36(3):520e9.
[12] Bai YL, Ma XY, Ming X. Lift enhancement of airfoil and tip ow control for
wind turbine. Applied Mathematics and Mechanics-English Edition 2011;
32(7):825e36.
[13] Kim T, Lee S, Kim H, Lee S. Design of low noise airfoil with high aerodynamic performance for use on small wind turbines. Science China 2010;
53:75e9.
[14] Drela M. XFOIL: an analysis and design system for low Reynolds number
airfoils. In: Mueller TJ, editor. Low Reynolds numbers aerodynamics, lecture
notes in engineering. New York: Springer; 1989. p. 54e61.
[15] Hoogendoorn E, Jacobs GB, Beyene A. Aero-elastic behavior of a exible blade
for wind turbine application: a 2D computational study. Energy 2010;35:
778e85.
[16] Spera DA. Wind turbine technology. 2nd ed. New York: ASME Press; 2009.
[17] Amiet RK. Acoustic radiation from an airfoil in a turbulent stream. Journal of
Sound Vibration 1975;41(4):407e20.
[18] Moriarty P. NAFNoise users guide; 2005. Golden Colorado.
[19] nveren C. Wind Resource Assessment on the Campus Area of Izmir Institute
of Technology: Use of Multi Point Data Source MSc thesis, Izmir Institute of
Technology, 2010.
[20] Moriarty P, Migliore P. Semi-empirical aeroacoustic noise prediction code for
wind turbines. National Renewable Energy Laboratory; 2003.