Characterisation of single wind turbine wakes with static and scanning WINTWEX-W LiDAR data

Abstract

With further development of LiDAR technology wake measurements by use of LiDAR became of common interest in the wind energy community. To study new measurement strategies of scanning and nacelle LiDARs, in combination with already existing measurement principles of static LiDARs, Norcowe conducted in collaboration with the Energy research Centre of the Netherlands (ECN) the Wind Turbine Wake Experiment Wieringermeer (WINTWEX-W). In this study we use data from the static Windcubes V1 to illustrate a proof of concept of wake effects at 1.75 and 3.25 rotor diameter downstream distance. After validating Windcube data against sonic anemometers from the met mast, we compare downstream velocity deficits and turbulence intensities between measurements of static and scanning WindCubes. To further characterize single wind turbine wakes and their frequencies of occurrence we analysed the results in terms of atmospheric stability. Wake measurements are of great importance to further developing tools for optimising wind farm layouts and operations.

Full text

1876-6102 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of SINTEF Energi AS
doi: 10.1016/j.egypro.2015.11.428
Energy Procedia 80 ( 2015 ) 245 – 254
ScienceDirect
12th Deep Sea Offshore Wind R&D Conference, EERA DeepWind’2015
Characterisation of single wind turbine wakes with static and
scanning WINTWEX-W LiDAR data
Valerie-M. Kumera,b, Joachim Reudera,b, Benny Svardalc, Camilla Sætrec, Peter Eecend
aGeophysical Institute, University of Bergen, Allegaten 70, 5007 Bergen, Norway
bBjerknes Center for Climate Research, Allegaten 70, 5007 Bergen, Norway
cChristian Michelsen Research AS, Fantoftvegen 38, Bergen, Norway
dEnergy research Centre of the Netherlands, Westerduinweg 3, Petten, The Netherlands
Abstract
With further development of LiDAR technology wake measurements by use of LiDAR became of common interest in the wind
energy community. To study new measurement strategies of scanning and nacelle LiDARs, in combination with already existing
measurement principles of static LiDARs, Norcowe conducted in collaboration with the Energy research Centre of the Netherlands
(ECN) the Wind Turbine Wake Experiment Wieringermeer (WINTWEX-W). In this study we use data from the static Windcubes
V1 to illustrate a proof of concept of wake effects at 1.75 and 3.25 rotor diameter downstream distance. After validating Windcube
data against sonic anemometers from the met mast, we compare downstream velocity deficits and turbulence intensities between
measurements of static and scanning WindCubes. To further characterize single wind turbine wakes and their frequencies of
occurrence we analysed the results in terms of atmospheric stability. Wake measurements are of great importance to further
developing tools for optimising wind farm layouts and operations.
c
2015 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of SINTEF Energi AS.
Keywords: LiDAR, wind turbine wakes, planetary boundary layer, atmospheric stability
1. Introduction
The wake region of a wind turbine is in general characterized by a reduction in average wind speed and an increase
in turbulence intensity. Both effects have negative implications for the operation of a wind farm by reducing the power
output and increasing loads and fatigue for downstream turbines. The accurate characterization of the structure and
dynamics of single turbine wakes is therefore the key for understanding the wind field inside a wind farm and of
uttermost importance for wind farm designers and operators.
∗Valerie-M. Kumer. Tel.: +47-55-58-2672.
E-mail address: valerie.kumer@gfi.uib.no
Available online at www.sciencedirect.com
© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of SINTEF Energi AS

246 Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254
Wakes are a complex interaction between the turbulent atmospheric boundary layer (ABL), the static structure
of the turbine tower and the aerodynamics of the rotating turbine blades. In the past, wake effects have mainly
been investigated by scaled wind tunnel studies (e.g. [1–4]) and model simulations of different complexity (e.g.
[5], [6], [7], [8]). These approaches have contributed to a reasonable understanding of wind turbine wakes under
idealized conditions, i.e. the assumption of neutral atmospheric stratification and corresponding use of a logarithmic
wind profile. It is however evident that atmospheric stability plays an important role for the development, structure,
dynamics and decay of wind turbine wakes ([9–13]) and that its’ appropriate description requires corresponding long-
term observations.
The recent development in the field of sodar(e.g. [14]), radar (e.g. [15]) and in particular lidar based remote sensing
capabilities (e.g. [16] has opened new perspectives for such full scale wake observations under different atmospheric
stability conditions. Static wind lidar systems, available for about one decade now, have been used to determine the
average wake deficit in the wind profile behind a turbine (e.g. [17]) or to determine single volume turbulence statistics
by operating the instruments in staring mode ([18,19]). With scanning wind lidars entering the market during the
last years, the measurement capabilities related to structure and dynamics of turbine wakes have distinctly increased.
However the corresponding campaigns have usually been limited to shorter measurement periods (e.g. [20–23]).
The presented study describes a 7 month multi-lidar deployment at the test site Wieringermeer operated by the
Energy research Centre of the Netherlands (ECN). It combined 4 static wind lidar systems upstream and downstream
of a row of test turbines with one scanning unit located around 12 rotor diameters downstream in the main wind
direction. In addition 2 nacelle mounted horizontally looking lidars, operated part of the campaign one looking in the
inflow and the other in the wake, and part of the campaign with both measuring the wake simultaneously. The main
purpose of this campaign was to test a novel approach for the characterization of the structure and dynamics of single
turbine wakes by the extensive use and combination of static and scanning wind lidar systems. The primary focus of
this manuscript is the description of the campaign and the presentation of first results on the effects of atmospheric
stability on the strength and extension of the wake downstream of single wind turbines.
2. Campaign Setup
Data used in this study origins from the WInd Turbine Wake EXperiment Wieringermeer (WINTWEX-W) con-
ducted by the University of Bergen (UiB), Christian Michelsen Reserach (CMR) and the Energy research Centre
of the Netherlands (ECN), all partners of the NORwegian Center of Offshore Wind Energy (NORCOWE). It took
place from November 2013 until mid of May 2014 in ECN’s Wind turbine test facilities in Wieringermeer, 1 km of
the Wieringermeer coastline at -5 m below sea level, to test wake measurement of three different Doppler LiDAR
instruments.
Fig. 1. a) Map of the Netherlands with a black box indicating the location of ECN’s wind turbine test site Wieringermeer. b) Map of the wind
turbine test site with a box indicating the row of research turbines. c) Map of the WINTWEX-W measurement setup around the research turbine
number 6.

Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254 247
The test facility consists of a row of prototype turbines and a row of five Nordex 2.5 MW research wind turbines
with a hub height and rotor diameter of 80 m. Around 3 rotor diameter upstream of the research turbine number 6 a met
mast of 108 m measures upstream wind and temperature conditions at three and two different altitudes respectively
(figure 1). In this study we use wind information of the 3D Gill Sonic at 80 and 108 m height and the temperature
difference between 37 and 10 m. Additional to the standard met-mast instrumentation, we deployed one Windcube V1
upstream of the research turbine number 6 and two Windcube V1 and a Windcube 100s downstream. The downstream
devices were located at 2, 5 and 12 rotor diameter distance aligned for winds coming from 210◦degrees south-west.
After relocation on November 29th 2013, the locations changed to 1.75, 3 and a 12 rotor diameter downstream along
the wake line aligned for winds coming from 218◦degrees south-west (figure 1). The Windcube V1s measured
standard profiles at 40, 52, 60, 100, 108, 120, 140, 160, 200 m above ground level with a sampling frequency of 1 Hz.
The scanning configuration of the Windcube 100s consisted of three Plan Position Indicator (PPI) and three Range
Hight Indicator (RHI) scans repeating every minute (table 2).
scan type azimuth [ ◦] elevation [ ◦] scan speed [ ◦/sec ] scan duration [ sec ]
PPI 198 - 258 2.4 6 10
PPI 198 - 258 4.7 6 10
PPI 198 - 258 7.1 6 10
RHI 228 60-0 6 10
RHI 228 0-60 6 10
RHI 228 60-0 6 10
Table 1. Scan configuration of the Windcube 100S, consisting of three Plan Position Indicator (PPI) and three Range Hight Indicator (RHI) scans.
On the nacelle of the research turbine number 6 an upstream looking Wind Iris and a downstream looking Zephir
(ZPH328) nacelle LiDAR also measured hub height up and downstream flow condition, but are not part of this study.
3. Data and Methods
The sonic and the upstream Windcube WLS67 data were available throughout the campaign duration. Due to the
relocation and farm work the downstream Windcube WLS37 and WLS65 have some data gaps in end of November
2013 and from March until May 2014. For some days in December the daily average data availability at 108 m was
below 50% (figure 2). Before relocation the downstream Windcube WLS37 had a directional offset form geographic
north of 45◦, which we corrected for the further analysis.
As a data quality control of the Windcube measurements we compared their ten minute mean wind speed and
direction measurements to ten minute mean sonic measurements at 108 m height, which is located at the top of the
upstream met-mast number 3. Following the data calibration recommend in the IEA Task 32 ([24]) we correlated
the data according to wind speeds greater than 4ms−1, dry weather conditions, data availability greater than 90%
and turbulence intensities smaller than 1%. Additionally, we highlighted correlations for undisturbed wind direction.
Following correlation coefficients will refer to this undisturbed data. The first downstream Windcube (WLS37) was
influenced by the wake of almost all research turbines. Therefore, only winds in the sectors between 150 to 210◦and
250 to 260◦were used for validation. For the second downstream Windcube we picked sectors between 330 to 20◦
and 100 to 310◦. The upstream Windcube (WLS67) was validated against winds between 330 to 20◦and 50 to 180◦.
Following Courtney ([25]), the Windcube 100s was validated not against the meteorological wind speeds measured
by the sonic anemometer but against its converted radial wind speed component (equation 1) for all wind directions
vrsonic =u∗cos(α)+v∗sin(α),(1)
where u and v are the zonal and meridional wind components and αis the angle between geographic north and the
line of sight measurement of the Windcube 100S that we choose for comparison. The collected data by the Windcubes
V1 show a generally good quality with correlation coefficients varying between R2=0.946 and R2=0.973 for the
horizontal wind speed and between R2=0.944 and R2=0.998 for the wind direction (figure 3). The ten minute mean
radial wind speeds of the Windcube 100s, compare with a R2value of 0.645 not as good to the ten minute mean radial
wind speeds of the sonic anemometer, which might be related to data processing issues. A comparison of horizontal

248 Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254
Nov Dec Jan Feb Mar Apr May Jun
sonic
wls67
wls37
wls65
wls100s
Fig. 2. Daily mean data availability greater than 50% at 108 m height of the sonic anemometer (grey), the Windcubes WLS67 (blue), WLS37 (red),
WLS65 (green) and WLS100S (black).
wind speeds calculated from Windcube 100s measurements to Windcube V1 data by Hu show a higher correlation of
R2=0.95 ([26]).
The campaign was designed to study single wind turbine wakes. Therefore the downstream wind turbines were
aligned in the main wind direction of south westerly winds (210 degrees) and only the sector around the main wind
direction will be analysed to study wind turbine wake characteristics. Figure 4 shows the wind speed distribution of
the analysed period from beginning of November 2013 until the mid of March 2014 at 108 m height, while highlight-
ing the analysed wind sector. Narrowing the data to a 45 degree wind sector leaves us with 22 % of the analysed
data set. 45 degrees were chosen over a smaller sector in order to increase the amount of data to allow for a more
general conclusion. However, this leaves us with a more complex interpretation of the data inside the wake, since the
Windcube profiles face wake position fluctuations.
The position of the Windcubes inside the wake are crucial for further interpretations of the data since, wake char-
acteristics and measurement accuracies are different in the wake central line or at the edges ([27]). In order to print
the average wake location in which the downstream Windcubes measured their profiles, a total average of each 10
second PPI scan was calculated. Additional to the wake characteristics at different positions the measurement princi-
ple of the profiling Windcubes is sensitive to complex flow. Wind vectors are retrieved by the Doppler Beam Swift
(DBS) method using the last four radial wind speed measurements, assuming uniform flow. In non-uniform flow this
assumption is broken and the measurement uncertainty of 0.5 m/s is expected to increase. However, measurement
inter-comparisons by Rhodes et al. ([28]) and LES simulations of LiDAR measurements inside wake by Lundquist
et al. ([27]) have shown that measurement errors will be ≤1ms
−1. In order to analyse the profiles dependent on at-
mospheric stability, we put the data into three categories representing stable, unstable and neutral atmospheric states.
Apositive, non existent or a temperature gradient smaller than the dry adiabatic lapse rate was used to define stable,
neutral and unstable conditions respectively.
4. Results
An analysis of the data shows a general proof of concept for both LiDAR measurement strategies. With the above
described scan pattern we were able to catch single wind turbine wakes and some of their meandering characteristics
(figure 5). An instantaneous PPI scan cutting though hub height shows a wave pattern with an increasing wave length
downstream characterized by a band of radial velocity deficits of around 7 ms−1, similar in magnitude to vertical
scans by Iungo et al. ([20]). It is also interesting to see the signature of the upstream prototype turbine wakes with a
much longer wave length and lower but still significant velocity deficits of around 4 ms−1. On a 10 minutes average

Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254 249
0 5 10 15 20 25 30
−5
0
5
10
15
20
25
30
Met.mast: Radial mean wind speed [m/s]
Lidar: Radial mean wind speed [m/s]
y=0.978x−0.456
R2=0.645
Bin. data (N=3114)
1:1
OLS
Oct 27 Nov 03 Nov 10 Nov 17 Nov 24 Dec 01 Dec 08 Dec 15
−30
−25
−20
−15
−10
−5
0
5
10
15
20
Radial mean wind speed [m/s]
METMAST3 @ 108m
WLS100S @ 144m
0 5 10 15 20 25 30
0
5
10
15
20
25
30
Met.mast: Mean wind speed [m/s]
Lidar: Mean wind speed [m/s]
y1=0.911x+1.51
R2
1=0.923
y2=0.948x+0.941
R2
2=0.946
Bin. data (N=4096)
OLS
Bin. data sector (N=3655)
OLS sector
1:1
0 5 10 15 20 25 30
0
5
10
15
20
25
30
Met.mast: Mean wind speed [m/s]
Lidar: Mean wind speed [m/s]
y1=0.859x+1.77
R2
1=0.889
y2=0.939x+0.897
R2
2=0.951
Bin. data (N=3189)
OLS
Bin. data sector (N=1725)
OLS sector
1:1
050 100 150 200 250 300 350
0
50
100
150
200
250
300
350
Met.mast: Mean wind direction [ o ]
Lidar: Mean wind direction [ o ]
y1=0.702x+54.3
R2
1=0.421
y2=0.875x+22.6
R2
2=0.944
Bin. data (N=3041)
OLS
Bin. data sector (N=1746)
OLS sector
1:1
0 5 10 15 20 25 30
0
5
10
15
20
25
30
Met.mast: Mean wind speed [m/s]
Lidar: Mean wind speed [m/s]
y1=0.981x+0.324
R2
1=0.943
y2=0.974x+0.148
R2
2=0.973
Bin. data (N=6027)
OLS
Bin. data sector (N=1428)
OLS sector
1:1
050 100 150 200 250 300 350
0
50
100
150
200
250
300
350
Met.mast: Mean wind direction [ o ]
Lidar: Mean wind direction [ o ]
y1=1x−4.71
R2
1=0.994
y2=1x−5.64
R2
2=0.998
Bin. data (N=5876)
OLS
Bin. data sector (N=1401)
OLS sector
1:1
050 100 150 200 250 300 350
0
50
100
150
200
250
300
350
Met.mast: Mean wind direction [ o ]
Lidar: Mean wind direction [ o ]
y1=1.01x−3.15
R2
1=0.999
y2=1.01x−3.12
R2
2=0.998
Bin. data (N=3949)
OLS
Bin. data sector (N=3525)
OLS sector
1:1
a) b)
c) d)
e) f)
g) h)
Fig. 3. a) to f) Scatter plots of 10 minutes mean wind direction (left) and horizontal wind speed (right) measured by a sonic anemometer and
Windcubes V1 at 108 m height. Blue, green and red colors indicate the different Windcubes WLS67, WLS65 and WLS37 respectively. Disturbed
wind directions are more transparent and solid lines indicate the 1:1 diagonal and the ordinary least square fit of the binned data points.
g) and h) Time series and scatter plot of 10 minutes mean radial wind speeds measured by a sonic anemometer at 108 m (gray) and a Windcube
100s range gate at 144 m height (black).

250 Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254
Fig. 4. Windrose of Sonic anemometer at 108 m height from November 2013 until March 10th 2014. The highlighted sector represents winds
coming from the south-west (202.5◦- 247.5◦).
meandering effects vanish and the wake becomes a straight band of deficits still around 7 ms−1. On a daily average
the wake gets of a more conical shape with smooth deficits of 4 ms−1in the near wake area and to 2 ms−1in the far
wake area (>5D).
A high enough sampling rate allowed us to calculate the turbulence intensity of the radial velocities during a period
of ten minutes. During ten minutes, turbulence intensities with 30% are highest at the flanks of the wake (figure 6).
Signatures of the prototype turbine wakes are also visible in lines of TI with only a 10% smaller magnitude than the
wakes of the research turbines. When averaging over a day, TI decays from 40% at 2D to 20% at 5D and forms
a more uniform area. After 5D downstream distance of the turbine TI seams to decrease by another 10%. At this
distance the height of the scanned line of sight is at 46 m and TI induced by the wake is of the same magnitude than
the ambient air turbulence. Averaging over the whole analysed period for south-westerly winds makes the wake areas
more symmetric compared to the daily mean picture.
Further analysis of all south-westerly winds measured by the three Windcubes V1 at three fixed positions show
that wind profiles follow nicely the theoretical approach of an upstream logarithmic profile and downstream velocity
deficits with increased turbulence intensities below blade tip height (figure 7). Velocity deficits range from 4 ms−1to
1ms
−1dependent on downstream distance and weather condition. The same applies to differences in TI, which vary
between 10 and 22%. The wind speeds standard deviations are generally higher for the downstream devices which
can be related to higher measurement uncertainties and the varying position inside the wake.
Filtering the data according to their stability class revealed highest TI values for stable weather conditions going
along with highest velocity deficits for both downstream devices. Under these conditions the power output of the
research turbine number 6 is lower than the average of the analysed period. During unstable cases the velocity deficit
is smaller than in stable cases and the wake seems to almost recover at 5D downstream distance. Enhanced vertical
mixing also expands the vertical extent of the wake and decreases the vertical wind shear. Under neutral conditions
wake characteristics are weaker as TI and velocity deficits are smaller. Similar to the unstable conditions the velocity
deficit is almost gone at 5D but TI is still 3% higher compared to upstream conditions. Weaker characteristics can
be explained by stronger winds that introduce vertical mixing through vertical wind shear. The daily mean radial TI
measured during unstable conditions by the Windcube 100s fits with 20 % at the location of the WLS7-37 and 10%
at the location of the WLS7-65 well to their measured TI ranges.
5. Conclusion and Outlook
In this study we analysed static and scanning Windcube data collected during the WINd Turbine Wake EXperiment
Wieringermeer (WINTWEX-W), in terms of wake characteristics. The data was collected from November 2013 until
May 2014 in the collaboration with University of Bergen, CMR and the ECN.

Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254 251
1
000
500
0
1
200
1000
800
600
4
00
2
00
0
|
vr
|
1
5
1
0
5
0
1
000
5
00
0
1200
1
000
800
600
400
2
00
0
| vr |
15
1
0
5
0
1
000
500
0
12
00
1000
800
600
400
200
0
|vr|
1
5
1
0
5
0
1
000
5
00
0
1
200
1
000
800
600
400
2
00
0
|
vr
|
15
1
0
5
0
1
000
500
0
1
2
00
1000
800
600
400
2
00
0
|vr|
15
1
0
5
0
1
000
5
00
0
12
00
1
000
800
6
00
4
00
200
0
|vr|
1
5
1
0
5
0
distance towards west
[
m
]
di
s
t
ance
t
owar
d
s sou
th
[
m
]
di t t d th [ ]
a
)
c)
e
)
f)
d)
b)
Fi
g
.
5
. Contour plots of 6
0
◦
PPI scans at
4
.
7
◦
(l
e
f
t
)
an
d7
.
1
◦
(r
igh
t) e
l
evat
i
on an
gl
eo
f
a) &
b
)
i
nstantaneous, c) &
d
)10m
i
nutes mean an
d
e) &
f)
d
ail
y
mean radial wind speeds on November 1st, 2013. Circles, squares and the cross indication the locations of turbines, Windcubes and the me
t
m
ast, respect
i
ve
ly.
W
i
n
d
cu
b
e pro
fil
e measurements s
h
ow at neutra
l
con
di
t
i
ons a c
l
ass
i
ca
l
wa
k
ep
i
cture w
i
t
h
at
h
ree mont
h
avera
ge
fl
ow deceleration and Turbulence Intensit
y
of 3 m
s
−
1
and 14% at t
w
o rotor diameter do
w
nstream distance. These fou
r
month avera
g
es are lower compared to results of the scannin
g
Windcube 100S with deficits of
5
m
s
−
1
and TI of 20%.
The Windcube 100S results are consistent with wind tunnel and model studies
(
[29]
)
.D
i
ff
erences to the static devices
ff
ff
c
an
b
eexp
l
a
i
ne
dby
t
h
evar
yi
n
gl
ocat
i
on o
f
t
h
e measure
d
pro
fil
e
i
ns
id
et
h
ewa
k
e
.

252 Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254
1
000
5
00
0
12
00
1
000
800
600
400
2
0
0
0
TI
1
0
.
8
0
.
6
0.
4
0.
2
1
000
500
0
12
00
1
000
800
600
400
200
0
TI
1
0
.
8
0
.
6
0.
4
0
.
2
1000 500
0
1
200
1
000
800
600
4
00
2
00
0
TI
1
0
.
8
0
.
6
0
.
4
0
.2
1000
5
00
0
12
00
1000
800
600
4
00
2
00
0
TI
1
0
.
8
0
.6
0
.
4
0
.
2
a
)
b)
c)
d)
1000
5
0
0
0
12
00
1
000
800
600
400
200
0
T
I
1
0
.
8
0
.
6
0
.
4
0
.
2
1
000
5
00
0
12
0
0
1
00
0
800
600
4
0
0
200
0
distance towards west
[
m
]
distance towards south
[
m
]
T
I
1
0
.
8
0.
6
0
.
4
0
.
2
e
)
f)
Fi
g
.
6
. Contour plots of
60
◦
P
PI
sca
n
sat4
.
7
◦
(l
e
f
t
)
an
d7
.
1
◦
(r
igh
t) e
l
evat
i
on an
gl
eo
f
tur
b
u
l
ence
i
ntens
i
t
y
o
f
a) &
b
)10m
i
nutes, c) &
d
) one
d
a
y
on Novem
b
er 1st, 2013 an
d
e) &
f
)t
h
e ana
ly
se
d
per
i
o
df
or sout
h
-wester
ly
w
i
n
d
s. C
i
rc
l
es, squares an
d
t
h
e cross
i
n
di
cate t
h
e
l
ocat
i
ons o
f
tur
bi
nes
,
Windcubes and the met mast, respectively
.
However, the Windcube v1 measurements have proven to be of valuable information althou
g
h measurement error
s
o
f
≤
1
m
s
−
1
and a variation of the profile location are expected. A look at d
i
ff
erent atmospheric stability conditions
ff
ff
reveal that turbulence intensities and velocit
y
deficits are stron
g
est durin
g
stable conditions, as ne
g
ative temperatur
e
g
radients prohibit mixin
g
with the ambient flow. As the turbines are influence b
y
the upstream protot
y
pe turbines the
y
g
enerate less ener
gy
in stable conditions. Durin
g
unstable conditions turbulence intensities var
y
in
g
between 10 an
d

Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254 253
0
40
80
120
160
2
00
h
e
igh
t
[
m
]
0
0
.
5
1
1
.
5
p
ower
[
m
W
]
0
40
80
12
0
160
2
00
h
e
igh
t
[
m
]
0
0
.
5
1
1
.
5
p
ower
[
m
W
]
0
0
.
1
0
.
2
0
.
3
0
.
4
0
40
80
12
0
1
60
2
00
TI
h
e
igh
t
[
m
]
0
5
10
15
20
wi
n
d
s
p
ee
d
[
m
/
s
]
0
5
10
15
20
0
0
.
5
1
1
.
5
p
ower
[
m
W
]
w
i
n
d
s
p
ee
d
[
m
/
s
]
u
n
stable
n
eut
r
al
stable
F
i
g
. 7. Solid lines show Windcube V1 mean turbulence intensit
y
and wind speed upstream (blue), near wake (red) and fare wake (
g
reen) profile
s
of the anal
y
sed period from November 2013 until mid March 2014 in the left and central column. Dots indicate the standard deviation. The ri
g
h
t
co
l
umn s
h
ows 10 m
i
nutes mean an
d0
.
5
m
s
−
1
binned power output of turbine number
6
in dots and solid lines. Data belon
g
in
g
into the horizontall
y
separated stabilit
y
classes are hi
g
hli
g
hted in blue. Onl
y
data durin
g
south-westerl
y
winds are shown
.
19% an
d
compare we
ll
to w
i
n
d
tunne
l
stu
di
es, w
hil
eve
l
oc
i
t
yd
e
fi
c
i
ts are not as stron
g
as
i
nt
h
ew
i
n
d
tunne
l
set u
p
([30]). Enhanced vertical mixin
g
mi
g
ht be underestimated in the wind tunnel and is leadin
g
in our measurements t
o
a
g
reater vertical expansion of the wake. In the fare wake re
g
ion the flow deficit has recovered almost to upstrea
m
conditions, but still has 3% hi
g
her turbulence intensities. Hi
g
h enou
g
h wind speed durin
g
neutral condition induc
e
s
i
m
il
ar m
i
x
i
n
g
e
ff
ects as during unstable conditions.
ff
ff
In the future the data will be anal
y
sed in terms of spectral ener
gy
densit
y
to learn more about the turbulence inpu
t
b
y
the wind turbine dependent on d
i
ff
erent stability conditions. This information and the experienced gained by
ff
ff
W
I
N
T
W
EX-
W
can
b
eta
k
en
o
ff
s
h
ore as wa
k
eo
f
upstream tur
bi
nes cou
ld i
n
fl
uence t
h
e natura
lf
requenc
i
es o
ffl
oat
i
n
g
wi
n
d
tur
bi
nes
.
R
efe
r
e
n
ces
[
1
]
Chamorro, L.P., Por
t
e-Agel, F.. Turbulent flow inside and above awind farm: A wind-tunnel study. Energies 2011;4:1916–1936.
´
doi:10.3390
/
0
en4111
9
1
6
.
[
2] C
h
amorro, L., Arn
d
t, R.E.A., Sot
i
ropou
l
os, F.. Tur
b
u
l
ent F
l
ow Propert
i
es Aroun
d
a Staggere
d
W
i
n
d
Farm. Boun
d
ary-Layer Meteoro
l
og
y
2011;141(3):349–3
6
7. URL: htt
p
:
//
dx.doi.org
/
g
g
1
0
.1
00
7
/
s10
5
46-011-9649-6. doi:10.1007
/
s10
5
46-011-9649-6.
[
3] Hancoc
k
, P.E., Pasc
h
e
k
e, F.. W
i
n
d
-Tunne
l
S
i
mu
l
at
i
on o
f
t
h
eWa
k
eo
f
a Lar
g
eW
i
n
d
Tur
bi
ne
i
n a Sta
bl
e Boun
d
ar
y
La
y
er
:
Part 2, the Wake Flow. Boundar
y
-La
y
er Meteorolo
gy
2014;1
5
1(1):23–37. URL: http
:
//
dx.doi.org
/
g
g
1
0
.1
00
7
/
s10
5
46-013-
9
887-x
h
tt
p
:
//
link.springer.com
/
m
10
.
100
7
/
s10
5
46-013-
9
887-x. doi:10.1007
/
s10
5
46-013-
9
887-x.
[
4] Krogstad, P.A.g., Sæ tran, L., Adaramola, M.S.. Blind Test 3 calculations of the performance and wake development behind two in-lin
e
a
n
do
ff
set mo
d
e
l
w
i
n
d
tur
bi
nes. Journa
l
o
f
F
l
u
id
san
d
Structures 2014;URL:
h
tt
p:
//
linkinghub.elsevier.com
/
m
retrieve
/
/
/
e
pii
/
/
i
S088
99
7461400217
5
.
doi:10.1016
/
6
j.jfluidstructs.2014.10.002.
/
/
[
5
] Mehta, D., van Zui
j
len, A., Koren, B., Holierhoek, J., Bi
j
l, H.. Lar
g
e Edd
y
Simulation of wind farm aerod
y
namics: A review. Journa
l
of Wind Engineering and Industrial Aerodynamics 2014;133:1–17. URL: http
:
//
linkinghub.elsevier.com
/
m
retrieve
/
/
/
e
pii
/
/
i
S0167610
5
140013
9
1.
doi:10.1016
/
6
j.jweia.2014.07.002.
/
/
[
6
] Zhan
g
, W., Markfort, C.D., Por
t
e-Agel, F.. Near-wake flow structure downwind of a wind turbine in a turbulent boundary layer. Experiments
´
in Fluids 2011;
5
2(
5
):1219–123
5
. URL: htt
p:
//
link.springer.com
/
m
10
.
100
7
/
s
00348-011-12
5
0-8. doi:10.1007
/
s
00348-011-12
5
0-8.

254 Valerie-M. Kumer et al. / Energy Procedia 80 ( 2015 ) 245 – 254
[7] Wu, Y.T., Port´
e-Agel, F.. Simulation of Turbulent Flow Inside and Above Wind Farms: Model Validation and Layout Effects. Boundary-Layer
Meteorology 2013;146(2):181–205. URL: http://dx.doi.org/10.1007/s10546-012-9757-y http://link.springer.com/10.1007/s10546-012-9757-y.
doi:10.1007/s10546-012-9757-y.
[8] Sanderse, B., Pijl, S., Koren, B.. Review of computational fluid dynamics for wind turbine wake aerodynamics. Wind Energy 2011;14(7):799–
819. URL: http://doi.wiley.com/10.1002/we.458. doi:10.1002/we.458.
[9] Pe˜
na, A., Rathmann, O.. Atmospheric stability-dependent infinite wind-farm models and the wake-decay coefficient. Wind Energy
2014;17(8):1269–1285. URL: http://dx.doi.org/10.1002/we.1632 http://doi.wiley.com/10.1002/we.1632. doi:10.1002/we.1632.
[10] Abkar, M., Port´
e-Agel, F.. The effect of free-atmosphere stratification on boundary-layer flow and power output from very large wind farms.
Energies 2013;6:2338–2361. doi:10.3390/en6052338.
[11] Sathe, A., Mann, J., Barlas, T., Bierbooms, W., van Bussel, G.. Influence of atmospheric stability on wind turbine loads. Wind Energy
2013;16(7):1013–1032. URL: http://dx.doi.org/10.1002/we.1528 http://doi.wiley.com/10.1002/we.1528. doi:10.1002/we.1528.
[12] Friedrich, K., Lundquist, J.K., Aitken, M., Kalina, E.A., Marshall, R.F.. Stability and turbulence in the atmospheric bound-
ary layer: A comparison of remote sensing and tower observations. Geophysical Research Letters 2012;39(3):n/a–n/a. URL:
http://dx.doi.org/10.1029/2011GL050413 http://doi.wiley.com/10.1029/2011GL050413. doi:10.1029/2011GL050413.
[13] Wharton, S., Lundquist, J.K.. Atmospheric stability affects wind turbine power collection. Environmen-
tal Research Letters 2012;7(1):14005. URL: http://stacks.iop.org/1748-9326/7/i=1/a=014005 http://stacks.iop.org/1748-
9326/7/i=1/a=014005?key=crossref.df26a50fc7064a0a7c4f98db89353d29. doi:10.1088/1748-9326/7/1/014005.
[14] Barthelmie, R.J., Larsen, G.C., Frandsen, S.T., Folkerts, L., Rados, K., Pryor, S.C., et al. Comparison of Wake Model Simulations
with Offshore Wind Turbine Wake Profiles Measured by Sodar. Journal of Atmospheric and Oceanic Technology 2006;23(7):888–901. URL:
http://dx.doi.org/10.1175/JTECH1886.1. doi:10.1175/jtech1886.1.
[15] Trombe, P.J., Pinson, P., Vincent, C., Bø vith, T., Cutululis, N.A., Draxl, C., et al. Weather radars - the new eyes for offshore wind farms?
Wind Energy 2013;:n/a–n/aURL: http://dx.doi.org/10.1002/we.1659 http://doi.wiley.com/10.1002/we.1659. doi:10.1002/we.1659.
[16] Gottschall, J., Courtney, M.S., Wagner, R., Jø rgensen, H.E., Antoniou, I.. Lidar profilers in the context of wind energy-a verification
procedure for traceable measurements. Wind Energy 2012;15(1):147–159. URL: http://doi.wiley.com/10.1002/we.518. doi:10.1002/we.518.
[17] Kumer, V.M.. Analysis of Lidar Wind Measurements at the Bruck an der Leitha Wind Park. Master thesis; Universit{¨
a}t Wien; Institut f{¨
u}r
Meteorologie und Geophysik UZA II, Althanstra{ß}e 14, 1090 Wien; 2012.
[18] Mann, J., Cariou, J.P., Courtney, M.S., Parmentier, R., Mikkelsen, T., Wagner, R., et al. Comparison of 3D
turbulence measurements using three staring wind lidars and a sonic anemometer. Meteorologische Zeitschrift 2009;18(2):135–
140. URL: http://openurl.ingenta.com/content/xref?genre=article&issn=0941-2948&volume=18&issue=2&spage=135. doi:10.1127/0941-
2948/2009/0370.
[19] Sathe, A., Mann, J.. A review of turbulence measurements using ground-based wind lidars. Atmospheric Measurement Techniques
2013;6(11):3147–3167. URL: http://www.atmos-meas-tech.net/6/3147/2013/. doi:10.5194/amt-6-3147-2013.
[20] Iungo, G.V., Port´
e-Agel, F.. Measurement procedures for characterization of wind turbine wakes with scanning Doppler wind LiDARs.
Advances in Science and Research 2013;10:71–75. URL: http://www.adv-sci-res.net/10/71/2013/. doi:10.5194/asr-10-71-2013.
[21] Iungo, G.V., Port´
e-Agel, F.. Volumetric LiDAR scanning of wind turbine wakes under convective and neutral atmospheric stability regimes.
Journal of Atmospheric and Oceanic Technology 2014;:140822100122000URL: http://journals.ametsoc.org/doi/abs/10.1175/JTECH-D-13-
00252.1. doi:10.1175/JTECH-D-13-00252.1.
[22] K¨
asler, Y., Rahm, S., Simmet, R., K¨
uhn, M.. Wake Measurements of a Multi-MW Wind Turbine with Coher-
ent Long-Range Pulsed Doppler Wind Lidar. Journal of Atmospheric and Oceanic Technology 2010;27(9):1529–1532. URL:
http://journals.ametsoc.org/doi/abs/10.1175/2010JTECHA1483.1. doi:10.1175/2010JTECHA1483.1.
[23] Trujillo, J.J., Bing ¨
ol, F., Larsen, G.C., Mann, J., K ¨
uhn, M.. Light detection and ranging measurements of wake dynam-
ics. Part II: two-dimensional scanning. Wind Energy 2011;14(1):61–75. URL: http://onlinelibrary.wiley.com/doi/10.1002/we.402/full
http://doi.wiley.com/10.1002/we.402. doi:10.1002/we.402.
[24] Clifton, A., Elliott, D., Courtney, M.. RECOMMENDED PRACTICES FOR REMOTE SENSING FOR WIND ENERGY APPLICATIONS.
IEA wind 2012;.
[25] M. Courtney, R. Wagner, T. Frils Pedersen, M. Bardon, S.D.. Calibrating Nacelle Lidars. January; 2012. ISBN 9788792896292.
[26] Hu, B.. Wind speed and wake studies using scanning LiDAR measurements. Tech. Rep. February; 2015.
[27] Lundquist, J.K., Churchfield, M.J., Lee, S., Clifton, a.. Quantifying error of lidar and sodar Doppler beam swinging measurements of wind
turbine wakes using computational fluid dynamics. Atmospheric Measurement Techniques 2015;8:907–920. URL: http://www.atmos-meas-
tech.net/8/907/2015/. doi:10.5194/amt-8-907-2015.
[28] Rhodes, M.E., Lundquist, J.K.. The Effect of Wind-Turbine Wakes on Summertime US Midwest Atmospheric Wind Profiles as Observed
with Ground-Based Doppler Lidar. Boundary-Layer Meteorology 2013;149:85–103. doi:10.1007/s10546-013-9834-x.
[29] Troldborg, N., Sø rensen, J.. A simple atmospheric boundary layer model applied to large eddy simulations of wind turbine wakes. Wind
Energy 2014;17(August 2010):657–669. URL: http://onlinelibrary.wiley.com/doi/10.1002/we.1608/full. doi:10.1002/we.
[30] Zhang, W., Markfort, C.D., Port ´
e-Agel, F.. Wind-Turbine Wakes in a Convective Boundary Layer: A Wind-Tunnel Study. Boundary-Layer
Meteorology 2013;146(2):161–179. URL: http://link.springer.com/10.1007/s10546-012-9751-4. doi:10.1007/s10546-012-9751-4.