Aerodynamic measurements on wind turbines

Received: 22 March 2017 Revised: 16 June 2018 Accepted: 18 June 2018
DOI: 10.1002/wene.320
ADVANCED REVIEW
Aerodynamic measurements on wind turbines

J. G. Schepers1 | S. J. Schreck2
1
Department of Wind Energy, Energy Research
Center of the Netherlands, ECN part of TNO, This article reviews the aerodynamic measurement programs on wind turbines that
Petten, The Netherlands have been performed in the last decades. It is largely based on results from four
2
National Wind Technology Center, National projects carried out under auspices of the International Energy Agency (IEA),
Renewable Energy Laboratory (NREL), Boulder,
which are denoted as IEA Tasks 14, 18, 20, and 29. The aim of these projects was
Colorado
to collect and analyze aerodynamic measurements on five field facilities (IEA
Correspondence
J. G. Schepers, Department of Wind Energy, Tasks 14 and 18), the National Renewable Energy Laboratory Phase VI turbine
Energy Research Center of the Netherlands, ECN placed in the large NASA Ames wind tunnel (IEA Task 20), and the Mexico tur-
part of TNO, Petten, P.O. Box 15, Petten 1755ZG, bine placed in the Large Low Speed Facility of the German Dutch Wind tunnel
The Netherlands.
Email: gerard.schepers@tno.nl
DNW (IEA Task 29). Other experimental programs with an important aerodynamic
content are touched upon as well. Research areas for which these measurements
have led to important progress are identified. The progress is illustrated with ana-
lyses on these experiments. It is shown that detailed aerodynamic measurements
are an absolute necessity for the validation and improvement of wind turbine
design codes, where it is also concluded that the amount of measurement data
which has been produced until now is still far too limited.
This article is categorized under:
Wind Power > Science and Materials
KEYWORDS
wind turbine aerodynamic measurements
1 | INTRODUCTION
The present article is devoted to detailed aerodynamic measurements on wind turbines. Such measurements are important for
a better understanding of the aerodynamic wind turbine behavior and the subsequent improvement of aerodynamic models in
wind turbine design codes. Thereto, it has to be realized that every aerodynamic process, in its basis, is described by means of
the so-called Navier Stokes equations which cannot be solved in an analytical way, where also a numerical solution of the
Navier Stokes equations is out of reach due to extreme calculational demands. The associated difficulty of aerodynamic
modeling is perhaps most convincingly illustrated by the fact that solving the Navier Stokes equations (as a matter of fact
“only” proving that a smooth solution exists) is one of the seven Millenium Prize Problems as formulated by the Clay Mathe-
matics institute in 2000.
As such every aerodynamic model inherently suffers from simplifications. For wind turbine aerodynamics, an additional
difficulty arises due to the fact that the computational effort for design calculations is more extreme than it is for most other
applications (e.g., fixed wing aerospace; see Schepers, 2012). This necessitates the use of very efficient, but also very simpli-
fied aerodynamic models (i.e., the Blade Element Momentum or BEM method). Obviously more advanced methods like com-
putational fluid dynamic (CFD) codes are applied in the wind energy society, but their use is, due to the calculational
demands, restricted to specific studies and load cases.
WIREs Energy Environ. 2018;e320. wires.wiley.com/energy © 2018 Wiley Periodicals, Inc. 1 of 25

https://doi.org/10.1002/wene.320
2 of 25 SCHEPERS AND SCHRECK
From a practical point of view these simplifications inevitably go together with a large uncertainty band, see for example,
Schepers et al. (2002) by which it is very difficult to design cost-effective and reliable wind turbines: Turbines behave unex-
pectedly experiencing instabilities or higher loads than expected leading to the risk of failure which can only be overcome
with high (costly) safety margins. Alternatively the loads may be lower than expected which implies an over dimensioned
(and costly) design. This was very well demonstrated in the EU project AVATAR (Schepers et al., 2018) where state of the
art aerodynamic models in design codes overestimated fatigue loads with 15% and, where dependent on the aerodynamic
model used, a stable or unstable blade state is predicted at high wind speeds at standstill, indicating the large uncertainties in
predicting the actual blade state.
From a more fundamental point of view, the simplifications in wind turbine aerodynamic models can obscure physical
phenomena crucial to wind turbine operation, which may remain hidden and unnoticed in the results obtained with these
models. Moreover, the load instabilities or extreme load excursions may occur due to the machine operating state, while others
are produced by inflow turbulence or extreme events. Unfortunately, it is generally not possible to discern whether the origin
lies with the machine itself or the inflow.
The availability of high-quality measurements is then the only way to unravel these hidden phenomena and to gain insight
into model uncertainties, leading to improved, more accurate, and reliable wind turbine aerodynamic models and a reduced
uncertainty band in design calculations. This observation is in line with the plea for better aerodynamic measurements in com-
bination with good analyses as given in the long-term research agenda from the European Academy of Wind Energy (van
Kuik et al., 2016).
It is this important area of high-quality rotor aerodynamic experiments that forms the subject of the present article. The
common denominator in these rotor aerodynamic experiments lies in the measurement of the pressure distributions at different
locations along the rotor blades. In these experiments, sometimes the underlying flow field around the wind turbine is mea-
sured too. Such dedicated aerodynamic experiments are carried out since the end of the 1980s and many of them were brought
into four subsequent IEA Tasks 14, 18, 20, and 29 (see Boorsma et al., 2018; Schepers et al., 1997, 2002, 2012, 2014;
Schreck, 2008). An IEA task is an international cooperation project organized under the auspices of the International Energy
Agency (IEA) Wind R&D Implementing agreement, in which participants share results from research which is usually carried
out on a national level. The IEA Tasks 14, 18, and 29 were coordinated by ECN and IEA Task 20 by NREL. It is noted that
IEA Task 29 is often denoted as Mexnext. The present review relies largely on results and analyses from these IEA tasks.
Thereto, Section 2 describes the content and the historical context of the IEA tasks.
Section 3 elaborates on measurement uncertainties, including those due to inflow spatial and temporal availability. It also
explains the difficulty in the determination of angle of attack and dynamic pressure in wind turbine measurements, which to
some extent can be seen as a measurement uncertainty too.
Then, Section 4 selects a few analyses on aerodynamic measurements and identifies research areas on which the present
measurements provided important information. It must be noted, however, that a huge amount of relevant results have been
generated over the last decades from which a small selection can be presented only. The chapter includes a long list of
citations.
It is emphasized that the present article limits itself to the application of aerodynamic measurements for validation and
model improvement purposes. Obviously, measurements could be used for the demonstration of aerodynamic innovations as
well (one can think of the demonstration of new airfoils or flow devices like vortex generators). The aerodynamic modeling of
these innovations does fall within the context of this article but the demonstration of them with, for example, practical recom-
mendations on how to apply these flow devices is excluded.
The article also limits itself to three-dimensional (3D) wind turbine rotor aerodynamic measurements, where the turbines
are considered to be relatively stiff, that is, the focus lies on “pure” aerodynamic effects, without aeroelastic interaction. The
3D measurements should be distinguished from another well-known category of wind turbine aerodynamic experiments that
is, the measurements on airfoils that are usually taken under two-dimensional (2D) steady conditions. Obviously, there is a
strong relationship between the 2D and 3D measurements. Among others, the 2D measurements are used to compare with the
3D measurements in order to extract the rotational effects. Moreover, the 2D airfoil measurements are input in so-called lifting
line codes. Hence, in the validation of lifting line codes with 3D measurements, these 2D measurements play an important role
in the interpretation of results.
The article also limits itself to public experiments. This among others implies that the very interesting DanAero experiment
in which aerodynamic measurements are carried out on a 2 MW turbine are only briefly touched upon. Although some results
of this experiment have been published in the works of Madsen et al. (2010) and Troldborg, Bak, Madsen, and Skrzypinski
(2013), the availability of measurements and machine data is generally restricted to the DanAero consortium.
The present article also excludes wind farm aerodynamic measurements as described in the studies by Barthelmie
et al. (2003, 2011), Schepers, Obdam, and Prospathopoulos (2012), and Vermeer, Sorensen, and Crespo (2003), to mention
SCHEPERS AND SCHRECK 3 of 25
only a few. Such measurements are generally taken in the far wake (several rotor diameters behind a wind turbine where one
might expect the next wind turbine in a wind farm). Still the distinction between wind farm and rotor aerodynamics is not
always very pronounced in view of the fact that the near wake (which is seen as part of rotor aerodynamics and considered in
this article) forms the starting point for the far wake. An example of an aerodynamic experiment relevant to both rotor aerody-
namics and wind farm aerodynamics is the wind tunnel measurement from Lignarolo (2016), where particle image velocime-
try (PIV) measurements are carried out in the near wake behind a model rotor and an “actuator disc” (represented through a
gauze). Hence, this experiment is important not only for understanding the aerodynamics of an actuator disc (i.e., a concept
often used in rotor aerodynamics), but also for understanding wind farm aerodynamics.
Acoustics is also considered to be outside the scope of this article, even though it is well known that wind turbine noise is
mainly aerodynamically driven (Schepers et al., 2007) and some of the experiments considered in this article include acoustic
measurements.
2 | W I ND T UR B I NE AE RO DY NA M I C M E ASU RE M E N T S : HI S T OR I CA L CO N T E X T
A N D R E L E V A N T I E A P RO J E C T S
The first detailed aerodynamic measurement programs, as described in this article, were performed at the end of the 1980s.
Until that time all experimental programs measured global loads only, that is, total (blade or rotor) loads. These loads consist
of an aerodynamic and a mass-induced component and they are integrated over the spanwise length. Various projects had been
performed in which wind turbine design codes were validated on basis of such global measurements, see for example, the
European “Benchmark exercise” as described in van Grol, Snel, and Schepers (1991). Differences between calculations and
measurements were found but the validations could not explain the physical causes of these discrepancies. This was only pos-
sible from more detailed measurements of aerodynamic properties along the blade as carried out in the wind tunnel or in the
field. As a result several institutes initiated programs to measure the pressure distributions at different radial positions. Thereto
Figure 1 shows how pressure taps around the airfoil, which are drilled into the surface, are connected with tubes to pressure
sensors. The sensors scan the pressure distribution almost instantaneously with rates up to 100 kHz. The pressure scanners
measure differential pressures relative to a reference pressure. This reference port was generally connected to the hub of the
rotor where the pressure is assumed to be close to the (known) atmospheric pressure, but Section 3.2 explains that this goes
together with an uncertainty.
Note that in the later Mexico experiment, individual absolute pressure sensors were mounted on the blade surface eliminat-
ing the uncertainty from the unknown reference pressure.
The resulting pressure distributions are then integrated to sectional forces, that is, the tangential force along the chord and
the force normal to it from which the lift and drag forces can be derived, although this requires the angle of attack, which as
explained in section 3.1, is difficult to determine.
The measurements of local aerodynamic loads are often complemented with probe measurements of inflow velocity and
angles ahead of the blade section, measurements of loads (e.g., thrust of blade root bending moments) and measurements of
the ambient (or tunnel) conditions. Some wind tunnel experiments also map the detailed flow field upstream, in and down-
stream of the rotor plane which is important information on induction aerodynamics which drives the rotor aerodynamics.
As mentioned in Section 1 many dedicated aerodynamic measurements are brought in four IEA Tasks. The first two tasks
(IEA Tasks 14 and 18), created a database of measurements taken under atmospheric conditions on various turbine where the
diameter of the turbines ranged from 10 to 27 m. The measurements were taken by the following institutes:
• Energy Research Center of the Netherlands, ECN, The Netherlands

• Delft University of Technology, TUDelft, The Netherlands
• Imperial College, IC and Rutherford Appleton Laboratory, RAL, United Kingdom
Lift Normal
Tangential Scanner
Drag
Probe
v
Reference pressure
FIGURE 1 Airfoil with pressure taps connected to pressure scanner (only a few pressure tubes between tap and scanner are indicated)
• National Renewable Energy Laboratory, NREL, USA

• Mie University, Japan
• RIS National Laboratory, Denmark
The various facilities from the participants and their instrumentation are described in detail in Schepers et al. (1997) and
Schepers, Brand, et al. (2002) and in Brand, Dekker, de Groot, and Spath (1996), Bruining (1997), Butterfield, Musial, and
Simms (1992), Madsen (1991a, 1991b, 1991c), Simms, Hand, Fingersh, and Jager (1999), and Spath (1993). It should be
known that NREL supplied measurements from three phases (Phase II to Phase IV) of the Unsteady Aerodynamics Experi-
ment (UAE). The difference between the phases lied in the blade design (untapered and untwisted for Phase II, untapered and
twisted for Phase II and IV, and in the instrumentation to measure the blade inflow angle (either a wind vane or a five hole
probe, see Figure 2).
The participants of IEA Task 14/18 agreed on a joint measurement program with rotating measurements at (roughly speaking)
yawed conditions (note that a field environment inevitably leads to yaw) and non-yawed conditions, with blade angles of attack
varying (roughly) between −5 and 40 . Angles of attack were set either by the wind speed, the pitch angle, or the rotor speed (rpm)
dependent on the experiment. Generally the time series had a length of 30–60 s often sampled with a frequency of 10 Hz or more.
Moreover, some parties supplied measurements for non-rotating (parked) conditions. Apart from time series, measurements of air-
foil coefficients (2D values as well as rotating values) as functions of angle of attack were added to the database.
One of the conclusions from IEA Task 14 and 18 was that the atmospheric free-stream environment in which the measure-
ments were taken led to a large uncertainty in the interpretation of many results (as is the case in all field measurements). This
was due to the unsteady inhomogeneous and uncontrolled wind conditions. The free stream conditions were usually measured
with a mast some distance away from the turbine (not necessarily upstream of the turbine) where this mast was instrumented
with only a few anemometers. In the best case, this yields a reasonable estimate for the mean wind conditions at the rotor
plane but it anyhow does not provide the instantaneous inflow conditions at every position in the rotor plane which is needed
to study high frequency, local, effects like dynamic stall. Although the inflow angles and velocities were sometimes measured
ahead of the airfoil with a five-hole pitot probe or a wind vane, these devices do not yield the exact instantaneous velocity with
a sufficient spatial resolution too. This problem was largely overcome in a next phase of NREL's Unsteady Aerodynamic
Experiment, that is, UAE Phase VI which was carried out in 2000. In this experiment a heavily instrumented rotor with a
diameter of 10 m was placed in the world's largest wind tunnel, that is, the NASA Ames (24.4 × 36.6 m2) facility with a
closed test section, see Hand et al. (2001) and Figures 3 and 4. As such measurements could be performed at stationary,
known and homogeneous conditions which facilitated the interpretation of measurement results and detection of subtle aero-
dynamic phenomena extending down to small, briefly lived structures and events.
The rotor was two-bladed where the blades have twist and taper and an S809 airfoil all along the blade. The turbine has a
synchronous generator with a rated rotor speed of 72 rpm. This results in a blade Reynolds number in the order of 1 million.
NREL performed measurements at very different conditions and very different configurations (unyawed/yawed conditions,
upwind/downwind rotor, teetered/fixed hub, with/without transition strips). The length of the measurement campaigns was
generally 30 s with a sampling frequency of 520.83 Hz.
NREL made the measurements from this experiment available to other institutes and they were analyzed within IEA Wind
Task 20, which started in June 2003 and ended in December 2007 (see Schreck, 2008). Nine institutes from eight different
countries participated.
FIGURE 2 NREL IEA Task 14/18 facility with probes

FIGURE 3 NASA Ames wind tunnel
Most of the analyses in IEA Task 20 were focused on measurements for the upwind rotor with fixed hub and without tran-
sition strips. The tunnel speed generally ranged between 5 and 25 m/s and pitch angles 0, 3, and 6 were considered. The rotor
speed of 72 rpm yields relatively low tip speed ratios and axial induction factors.
It is noted that within IEA Task 29 Mexnext additional Phase VI measurements were analyzed at higher rpm, in order to
investigate elevated rotor loading conditions.
The measurements from IEA Tasks 14/18 and 20 mainly focused on blade aerodynamics and they delivered little informa-
tion on induction aerodynamics. This problem was overcome in IEA Task 29 Mexnext. Mexnext is carried out in three phase
(Mexnext-I Schepers, Boorsma, et al., 2012, Mexnext-II Schepers et al., 2014, and Mexnext-III Boorsma et al., 2018).
The Mexnext consortium differed slightly from phase to phase but generally some 20 parties from 10 or 11 countries were
active in Mexnext.
One of the core activities in Mexnext is to analyze the measurement taken within the EU FP5 project “Mexico” (Model
Rotor Experiments in Controlled Conditions) and its successor project “New Mexico”). The Mexico experiment was carried
out in 2006 by 10 institutes from 6 countries on an instrumented, three-bladed wind turbine of 4.5 m diameter placed in the
largest European wind tunnel, the Large Low-speed Facility (LLF) of the German Dutch Wind Tunnel, DNW in the Nether-
lands with an open test section of size of 9.5 × 9.5m2, Figure 5.
FIGURE 4 NREL Phase VI turbine in NASA Ames wind tunnel

FIGURE 5 DNW-LLF wind tunnel
The rotor blades were twisted and tapered with three different aerodynamic profiles: The DU91-W2–250, the RISO-A2-
21, and the NACA 64-418. In Mexico the blades were tripped with zig-zag tape to avoid laminar separation phenomena and
to facilitate the validation of CFD codes in the sense that the uncertainty from modeling transition is avoided. The rotational
speed was either 424.5 or 324.5 rpm. At 424.5 rpm a chord-based Reynolds number of approximately 0.8 million was reached
at a tip speed of 100 m/s. This tip speed is relatively high to increase Reynolds number but sufficiently low to prevent notice-
able compressible conditions. Measurements were done at different tunnel speeds, tip speed ratios, and different yaw angles
and pitch angles. The setup of the Mexico experiment is given in Figure 6 and described in Schepers and Snel (2007) and
Boorsma and Schepers (2011).
In addition to pressure and load measurements, the 3D flow field was mapped using stereo PIV where two cameras are
mounted on a traversing tower which can move in the horizontal streamwise and radial direction (this tower can be seen in
Figure 6 at the right). They focus on a PIV sheet which is located horizontally in the symmetry plane of the rotor at the “9 o'
clock” position (see Figure 7).
FIGURE 6 Setup of the (new) Mexico turbine in the measurement section of the DNW LLF
FIGURE 7 PIV sheet at “9 o'clock position” in Mexico experiment

As such a comparison can be made between calculated and measured loads where the underlying flow field which drives
these loads can be assessed too. This comparison was intended to improve the understanding of the momentum balance
between loads and velocities but the results did not seem to obey this balance and several nonunderstood phenomena confused
the Mexnext-I consortium.
One of the recommendations from Mexnext-I was then the need to perform an additional experiment on the Mexico rotor,
which resulted in the New Mexico experiment as carried out in 2014 (Boorsma & Schepers, 2015).
The test matrix of New Mexico was designed to explain the nonunderstood phenomena from the Mexico experiment tak-
ing into account the lessons learned from the Mexico experiment. Prior to the New Mexico experiment in the DNW-LLF, as a
preparation to it, the instrumented Mexico blades were placed in the low-speed tunnel (LST) of the TU Delft at quasi 2D con-
ditions. In the LST the aerodynamic characteristics of the blades at standstill have been measured (including flow visualiza-
tion) where the blade instrumentation and data acquisition could also be tested and recalibrated. An important lesson from
these standstill measurements was the observation that the zig-zag tape as applied in the Mexico experiment triggered transi-
tion indeed but the associated thickening of the boundary layer and decambering of the flow is difficult to model. For this rea-
son the zig-zag tape in most New Mexico experiments was removed at the outer part of the blade, that is, the part with the
NACA 64-418 airfoil and the highest Reynolds number and the lowest risk of laminar separation.
The New Mexico experiment then partly repeated its predecessor experiment but an even wider variety of conditions in
terms of wind speed, yaw, etc, were considered where advantage was taken of the improved measurement techniques which
meanwhile became available (e.g., a wider PIV range could be covered). The test matrix also included measurements on flow
devices (Guerney flaps, root spoilers), measurements on “IEC aerodynamics” (measurements at standstill and faulty large
pitch misalignments) and acoustic measurements on blades with and without serrations (note the acoustic treatment of the tun-
nel in Figure 6). Moreover, several blade and wake flow visualization techniques were applied. The improved quality of the
New Mexico experiment compared to the Mexico experiment answered most of the nonunderstood questions from the first
experiment and the new results obeyed the momentum balance between loads and velocities.
Although the focus of Mexnext lies on the (New) Mexico experiment, the project is overarching, that is, it aims to consider
all public aerodynamic measurements. As such the measurements from IEA Tasks 14/18 and the NREL Phase VI experiments
are revisited, see for example, Boorsma et al. (2014). Moreover, some other experiments are considered of which the results
are not always fully explored yet. Among other the following experiments were found of relevance.
• Measurements in the wind tunnels from China Aerodynamics Research and Development Center, CARDC.
• Measurements carried out in the late 1980s under supervision of the Swedish FFA on a 5 m diameter rotor in a 12 × 16 m2
tunnel (Dahlberg et al., 1989; Ronsten, 1992).
• PIV data behind a one-eighth scaled model of the NREL Phase VI turbine which have been taken in a 3.8-m diameter
circular wind tunnel (Xiao, Wu, Chen, & Shu, 2011) These measurements are interesting because they complement the
original Phase VI measurements from NREL in which little flow field measurements were taken.
• Scaled down NREL Phase VI and Mexico experiments from the Korean Aerospace Institute (KARI; Cho & Kim, 2012,
2014) and Figure 8 and a scaled down from the Spanish National Institute for Aerospace Technology (INTA) section 10
of Schepers, Boorsma, et al. (2012). These experiments were less detailed than the original experiments but they still
formed interesting comparison.
• Measurements taken in the wind tunnel from Mie University on a rotor with diameter 2.4 m. The unique aspects from this
experiment lies in the Laser Doppler Anemometry data in the vicinity of the blades, see Maeda et al. (2012), Phengpom
FIGURE 8 KARI wind tunnel with scaled down model of Mexico rotor (compare with Figure 6)
et al. (2015), Phengpom, Kamada, Maeda, Nishimura, and Matsuno (2015), Phengpom, Kamada, Maeda, Matsuno, and
Sugimoto (2016).
• Measurements taken by TUDelft in their Open Jet Tunnel. An extensive database of hot-wire measurements in the near
wake of a 1.2-m rotor is reported in (Haans, 2011). In 2009 a new larger Open Jet Facility (OJF) opened in which several
experiments were taken on rotors with diameters up to 2 m, see for example, Micallef (2012).
• Field measurements from UAS-Kiel on a 30-m diameter E-30 turbine and TUDelft/ECN on a 25-m turbine with equip-
ment to detect transition (e.g., through microphones, hot films or high speed pressure sensors, see Schaffarczyk, Schwab,
Ingwersen, & Breuer, 2012, 2017; van Groenewoud, Boermans, & van Ingen, 1983). In this respect the measurements
from the Danaero experiment (Madsen et al., 2010; Troldborg et al., 2013) with microphones on the blade are very inter-
esting too but these measurements are not taken in the public area.
3 | UN CE RT AINT IE S
An important item in measurement programs is the identification of measurement uncertainties. These are usually assessed on
visual inspection, consistency checks, checks on reproducibility, checks of measurement sensor specifications, etc., leading to
at least some indication of the confidence level.
In general the uncertainties introduced by the measurement equipment itself (in particular the pressure sensors) were found
to be limited at least if frequent calibrations are carried out. An exception is formed by measurements at standstill or measure-
ments at the inner part of the blade where pressure levels were sometimes within the lower part of the measurement range.
One of the items investigated were the damping characteristics of the tubes between the pressure taps and scanners. Dependent
on the length and diameter of the tube acoustic damping was often found small up to frequencies of some 40 Hz.
Most of the results used in the IEA Tasks are averaged over a number of samples acquired during several consecutive rotor
revolutions. In general, the standard deviation for the wind tunnel measurements was limited (in the order of the plotting sym-
bol size!), although this is less true in stalled conditions (e.g., Schepers, Boorsma, et al., 2012).
However this does not imply that the results as supplied within the different IEA tasks should be considered as the “truth.”
In the wind tunnel experiments an uncertainty is introduced from tunnel boundary effects. In this respect a difference should
be made between closed and open tunnels. Tunnel effects in a closed tunnels are relatively easy to determine but they are also
known to be more severe than tunnel effects in an open tunnel (Barlow, Rae Jr., & Pope, 1999; Ewald, 1998) although the
huge size of the NASA-Ames wind tunnel makes that blockage effects on the UAE VI experiment remained limited to 1% or
less for most test conditions.
On the other hand, corrections for open tunnels are more difficult to determine due to the fact that the tunnel flow will
interact with the outer flow through a turbulent process. Such turbulent process is inherently subject to a large uncertainty
(e.g., Sorensen, Shen, & Mikkelsen, 2006; Voutsinas et al., 2003; Shen et al., 2010; Réthoré, Sørensen, Zahle, Bechmann, &
Madsen, 2011a, 2011b) from which it was concluded that tunnel effects in the (New) Mexico experiment are generally limited
apart from an acceleration in the far wake.
Other uncertainties may be introduced from the processing of data, for example, the integration of pressure distributions
into normal and tangential forces. Moreover, the use of processed data like angle of attack and dimensionless airfoil character-
istics require careful interpretation as will be explained in Section 3.1.
It is important to recall the difficult interpretation of field measurements due to the uncertainties from the limited spatial
resolution of inflow measurements.
Another uncertainty, which is not always realized when drawing conclusions from differences between calculations and
measurements, lies in the uncertainty from the aero-elastic model descriptions which are needed as input for the calculational
codes. Some data from these model descriptions may suffer from uncertainties, other data may be lacking or ignored (e.g., a
rigid construction is assumed in case structural input data are missing).
3.1 | Lifting line variables

The data from the experiments described in this article are often given as aerodynamic coefficients, that is, they are nondimen-
sionalized with the dynamic pressure and presented as function of angle of attack. Such data are usually referred to as lifting
line variables in view of the fact that aerodynamic coefficients, angles of attack, and the underlying-induced velocities are
important model parameters in lifting line codes.
Hence, the angle of attack is also a crucial quantity because it is needed to decompose the forces as determined in a chord-
related coordinate system into the wind-related lift and drag forces. It should be realized, however, that an angle of attack on a
rotating blade as an equivalent to the 2D wind tunnel angle of attack cannot be measured straightforwardly (nor modeled with
a flow field calculation from CFD).
This can be illustrated through Figure 9 which shows the angle of attack (αwindtunnel) in the 2D wind tunnel environment,
that is, the angle between the chord and the undisturbed wind vector which is aligned with the wind tunnel walls. Now sup-
pose that a device (i.e., a five-hole probe) which is placed ahead of the airfoil measures the angle of the incoming velocity.
The figure shows that αprobe is different from the angle of attack, due to the presence of the airfoil itself, that is, from the
upwash which is induced by the bound vorticity. In order to derive the angle of attack from this local inflow angle, this
upwash should be subtracted.
Some methods to derive the angle of attack are explained in section 5.2 of Schepers (2012), in Rahimi et al. (2018), Ship-
ley et al. (1995), and Vimalakanthan et al. (2018). The differences in methods are diverse. They amongst others use different
measurement data as starting point, for example, some methods require pressure distribution or integrated loads at blade sec-
tions where other methods require velocities around the blade sections and some other methods require both velocities and
pressures (loads). They are sometimes based on the measurement with a device where an experimentally determined 2D
upwash correction is applied (Schreck & Robinson, 2002) or the method from Shen, Hansen, and Sørensen (2009) which sub-
stracts the upwash from the bound vorticity to the measured velocities where the upwash is calculated from the pressure distri-
bution or integrated loads. Other methods act as inverse techniques (e.g., an inverse BEM method which calculates the
induced velocities backward from the measured loads or an inverse free wake method which, as intermediate step, calculates
the vorticity in the wake from the measured loads and from that the induced velocities; Sant, van Kuik, & van Bussel, 2006;
Micallef, Kloosterman, Ferreira, Sant, & van Bussel, 2010). Alternatively the measured pressure distribution or stagnation
method can be fitted to a known 2D pressure distribution or stagnation point at a particular angle of attack (e.g., van Rooij,
2003). The so-called Average Azimuthal Technique (Johansen & Sørensen, 2004) is based on annular average values of the
axial velocity by using the data at several upstream and downstream locations of the rotor. A modification to this method
which seems capable of determining angles of attack locally at the blade is the three-point technique as described in Rahimi,
Hartvelt, Peinke, and Schepers (2016). A method which determines the circulation around a blade section from the velocities
around the blade is described in Jost, Klein, Leipprand, Lutz, and Kramer (2017). Another method which can be applied for
axi-symmetric conditions and a three-bladed rotor is presented in Herráez, Daniele, and Schepers (2018). It makes use of the
fact that the upwash from the three blades at the angle bisector of two blades is equally canceled out by which the value of the
velocity at that position is, by definition, the induced velocity. Within the EU project AVATAR (Schepers et al., 2018) a
benchmark was organized in which different angles of attack methods were compared. The benchmark did not rely on mea-
sured results, instead CFD calculations were used as the starting point. These calculations were carried out by DTU with their
Ellipsys3D code on the 10-MW reference wind turbines which were designed in the AVATAR project and the EU project Inn-
wind.EU (Soerensen et al., 2017). The CFD results contained flow field data and pressure distributions. From these data,
induced velocities and angles of attack were derived by various participants with different methods for both axi-symmetric
and yawed conditions. In general, all methods showed a good mutual agreement at the mid span of the blade, say from 33 to
67% but near the outer part differences are present. Not all of these differences could be explained (e.g., Soerensen
et al., 2017).
α windtunnel
V
α
probe
FIGURE 9 Angle of attack in wind tunnel environment

3.2 | Dynamic pressure and nondimensionalization of aerodynamic coefficients in wind

One category of lifting line parameters is the aerodynamic coefficients, that is, the aerodynamic forces nondimensionalized by
the dynamic pressure. The establishment of the dynamic pressure from experiments which measure differential pressures rela-
tive to a reference pressure is often assumed to be the maximum pressure in the pressure distribution (local total pressure) or,
in case a pitot probe is added to the experiment, the pitot pressure which is also measured relative to a reference pressure.
These differential pressures correspond only to the dynamic pressures if the reference pressure is equal to the static pres-
sure at the blade section. However, this assumption is doubtful in view of the fact that the static pressure at the rotor plane is
unknown (e.g., in classical momentum theory a discontinuous pressure jump in the rotor plane occurs related to the axial force
on the rotor).
Alternatively some institutes determined the dynamic pressure by writing it as 0.5 ρ Veff2 where several procedures were
followed to estimate the value of Veff, that is, the incoming velocity at an airfoil. One possibility is to write Veff as the vectorial
sum of the rotational speed and the incoming wind speed with induced velocities added. However all of these methods are
subject to uncertainties (e.g., the induced velocities are not precisely known where the incoming wind speed in the case of
field experiments is not known with full spatial resolution).
An additional uncertainty, related to the unknown static pressure, appears in the determination of the pressure coefficients.
In many experiments the numerator of these pressure coefficients is based on the actually measured pressure difference at a
pressure tap, hence:

Cp = ptap – pref =q
Again this is only the correct definition if the reference pressure, pref, equals the freestream static pressure.
4 | AE RODY NA MIC M E ASU RE MEN TS: RE SUL TS AN D P ROGRE S S
In this section an inventory is given of the research on which progress has been possible through the use of detailed aerody-
namic measurements. It will largely rely on results from the IEA tasks 14, 18, 20, and 29 as described in Section 2. It should
then be realized that within these projects very many (more than 25) parties contributed over a very long (more than 25 years)
period, which leads to such breadth of results that it is virtually impossible to capture all research in a brief summary. The fact
that relevant research is also found outside the IEA framework complicates the task of summarizing all aerodynamic measure-
ment results even more.
4.1 | Literature on aerodynamic wind turbine measurements

A good starting point for a review on aerodynamic wind turbine measurements is formed by the final reports from the IEA Tasks
(Boorsma et al., 2018; Schepers et al., 1997, 2014; Schepers, Boorsma, et al., 2012; Schepers, Brand, et al., 2002; Schreck,
2008). These reports touch upon the relevant research carried out by the participants in the tasks and they include extensive litera-
ture lists. A large part of these lists are added to this article, where the references Shipley, Miller, Robinson, Luttges, and Simms
(1994), Hansen (1999), van Rooij, Timmer, and Bruining (2002), Acker and Hand (1999), Bermudez, Velasquez, and Matesanz
(2000), Bjorck (1995), Bruining (1993), Chaviaropoulos et al. (2001), Duque, van Dam, and Hughes (1999), Feigl (2003),
Graham & Brown, 2000, Leclerc and Masson (1999), Maeda and Schepers (2011), Miller et al. (1995), Schepers (2004),
Schepers, Feigl, Van Rooij, and Bruining (2004), Schepers and van Rooij (2005), Schreck, Robinson, Hand, and Simms
(2000), Simms, Robinson, Hand, and Fingersh (1996), Snel (1997), Strzelczyk (1998, 2000), van Rooij (2001, 2003),
van Rooij, Bruining, and Schepers (2003), and van Rooij and Schepers (2005) are related to IEA tasks 14/18.
The references Haans, Sant, van Kuik, and van Bussel (2006), Elgammi (2016), Guntur, Sørensen, and Schreck (2013),
Hand et al. (2001), Kuik, van Rooij, and Imamura (2004), Larwood and Chow (2016), Lindenburg (2003), Meng and van
Rooij (2007), Rooij and van Arens (2007), Sant (2007), Sant et al. (2006), Schepers (2007a, 2007b, 2007c), Schreck and Rob-
inson (2002), Schreck, Robinson, Hand, and Simms (2001), Schreck, Sant, and Micallef (2010), Schreck, Sørensen, and Rob-
inson (2007), Sørensen and Schreck (2012), van Rooij and Schepers (2005) are related to the NREL Phase VI experiment and
IEA Task 20 (although some have been produced after the end date of IEA Task 20).
The references Micallef et al. (2010, 2011), Breton, Sibuet, and Masson (2010a), Meister (2011), Hua et al. (2010, 2011),
Kuczaj (2009), Schepers, Pascal, and Snel (2013), Réthoré et al. (2011a, 2011b), Pereira, Schepers, and Pavel (2011), Gomez-
Iradi (2011), Shen et al. (2010), Guntur, Bak, and Sorensen (2012), Pascal (2009), Snel, Schepers, and Siccama (2009), Bech-
mann and Sørensen (2009), Bon (2011), Boorsma et al. (2014), Boorsma and Schepers (2014, 2016), Breton (2011), Breton,
Sibuet, and Masson (2010b), Hartveld (2016), Herraez, Stoevesandt, and Peinke (2014), Jeromin, Bentamy, and Schaffarczyk
(2014), Lutz, Meister, and Krämer (2011), Madsen et al. (2015), Mahmoodi and Schaffarczyk (2012), Micallef (2009a,
2009b), Micallef, Kloosterman, et al. (2010), Nilsson, Shen, Sorensen, and Ivanell (2011), Oggiano, Boorsma, Schepers, and
Kloosterman (2016); Parra, Boorsma, Schepers, and Snel (2016), Potentier (2013), Schepers (2013), Schepers and Boorsma
(2014), Schepers, Boorsma, Kim, and Cho (2011), Schepers, Boorsma, and Snel (2010), Schreck et al. (2010), Shen (2011),
Snel, Schepers, and Montgomerie (2007), Sorensen (2011), Sørensen, Zahle, Boorsma, and Schepers (2016), Stoevesandt
et al. (2010), and Szasz (2011) are related to IEA Task 29. Many of these references focus on specific research aspects
(e.g., 3D effects) or specific code validations. A study with a wider scope was the Dutch National Annexlyse project in which
all IEA Task 14/18 measurements were analyzed for validation and improvement of aerodynamic models (references from this
project are Feigl, 2003; Schepers, 2004; Schepers et al., 2004; Schepers & van Rooij, 2005; van Rooij, 2003; van Rooij et al.,
2003; another reference with a wider scope is Schepers (2012). This PhD thesis gives an extensive overview on the use of the
IEA Task 14/18/20/29 measurements in particular for the development and validation of engineering methods. Other PhD the-
ses which make use of aerodynamic experiments are Micallef (2012) (using TUDelft OJF measurements), Sant (2007) (using
UAE-VI and TUDelft OJF measurements), Elgammi (2016) (using UAE VI), and Micallef (2012) (using Mexico experiment
and TUDelft data).
It is also important to mention that literature on aerodynamic measurements has been generated in several cooperation pro-
jects outside the IEA framework. As an example IEA Task 14/18 measurements were brought into the EU project VISCEL
(see Hansen, 1999; Chaviaropoulos et al., 2001) in which the first generation of wind turbine CFD codes were validated.
Moreover IEA Task 14/18 measurements were used in the EU-Joule project Dynamic Stall and Three Dimensional Effect (see
Bjorck, 1995; Snel, 1997) to develop models on 3D and unsteady effects and the EU project ROTOW in which tower interfer-
ence models were validated (see Graham & Brown, 2000). The (New) Mexico data are used for validation of codes in INN-
WIND.EU (Madsen et al., 2015).
4.2 | Relevant research areas related to aerodynamic wind turbine measurements

Although the extensive list of references and projects from the previous section again illustrates the difficulty of capturing all
research in a brief summary, Section 4.3 will identify some main areas on which detailed aerodynamic measurements lead to
progress. As preparation to this section the measurements are first categorized through a distinction between field and wind
tunnel experiments (see Section 4.2.1). This is followed by a distinction on the application of data in Section 4.2.2: Measure-
ments can be used for validations but they can also be used for gaining fundamental physical insight. Finally Section 4.2.3.
distinguishes on the use of aerodynamic measurements for (stall) airfoil aerodynamics on one hand and induction aerodynam-
ics on the other hand.
4.2.1 | Aerodynamic measurements: Field versus wind tunnel measurements

From the previous sections it has become evident that detailed aerodynamic measurements are performed in the field and in
the wind tunnel.
At first sight measurements in the field would be preferred, since such measurements can be done on representative full-
size turbines, which avoids the inevitable scaling effects in a wind tunnel environment. Moreover the atmospheric conditions
are representative to the conditions which are experienced on a commercial wind turbine. On the other hand, this stochastic
atmosphere, though representative, makes interpretation of results extremely complicated due to the unknown instantaneous
inflow in the rotor plane.
However, wind tunnel measurements also suffer from drawbacks because the wind tunnel environment differs from the
full-scale environment in many aspects. In this respect wind tunnel blockage effects play an important role (see Section 3).
Also important are scaling effects, that is, the smaller size of the wind tunnel experiment makes the wind tunnel experiment
less representative. Some studies on scaling effects are reported in section 10 of Schepers, Boorsma, et al. (2012) and in Sche-
pers et al. (2011). These studies partly rely on the results from the scaled down Mexico experiments mentioned in Section 2.
Another important aspect of the wind tunnel environment lies in the smooth inflow which facilitates the interpretation of
results (As a matter of fact it forms the main motivation for taking measurements in the wind tunnel) but it may also lead to
different unsteady aerodynamic effects and for example, different boundary layer transition processes.
In general terms it should be concluded that there are inevitable drawbacks in both wind tunnel and field measurements
which definitely complicate the fulfillment of the research goals as mentioned in Section 1. The best remedy to overcome
these drawbacks is to combine the insights from full-scale field measurements with those from wind tunnel measurements.
In this respect it is interesting to mention the intermediate between field and wind tunnel measurements from Bellia
(1990), where a turbine was placed on a vehicle driving around on wind still days. Also interesting is the facility which was
recently being commissioned at Forwind Germany. This 3 × 3 m2 wind tunnel partly overcomes the problem of the nonrepre-
sentative turbulent inflow through a controllable turbulent inflow generated by an active grid in front of the measurement
section (a smaller wind tunnel with active grid has been used to measure the effect of turbulence on the airfoil characteristics,
see Heißelmann, Peinke, & Hölling, 2016). Another interesting possibility is offered by cryogenic tunnels, for example, the
European Transonic Wind Tunnel (ETW) in which high Reynolds numbers at low tunnel speeds (i.e., without compressibility
effects) are reached through the cooling and pressurizing of the tunnel.
4.2.2 | Validation versus physical understanding/model improvement

When categorizing the use of aerodynamic measurements a fundamental distinction can be made between the use of aerody-
namic measurements for validation on one hand and the use of measurements for physical understanding and consequent
model improvement/development on the other hand. Validation provides insight on differences between calculations and mea-
surements and eventually on the confidence level of the model. In principle the models remain untouched during validation.
Obviously the distinction between model validation and improvement is not very strict since insights from the validation phase
often form the basis for model improvement.
4.2.3 | Airfoil aerodynamics versus induction aerodynamics

Apart from the distinction mentioned in Section 4.2.2 between validation and understanding/model improvement, a distinction can
be made between the use of data for airfoil aerodynamics (in particular at stalled conditions) and induction aerodynamics. This dis-
tinction is partly historical since most of the research until, say 2000–2005 (i.e., the research from IEA Tasks 14, 18, and 20) was
devoted to stall aerodynamics (e.g., 3D and unsteady effects at stall) with a limited number of studies devoted to yaw induction, tip
effects, or dynamic inflow. The strong focus on stall aerodynamics in the 1990s can be understood by realizing that most of the tur-
bines at that time were controlled by means of stall, which was (and is) an extremely complicated physical phenomena.
This situation changed in the period after say 2005. Around that time most of the (new) turbines were not controlled by
stall anymore and interest in other subjects grew. In terms of airfoil aerodynamics more emphasis was put on boundary layer
transition, the aerodynamics of flow devices, and IEC aerodynamics, but in particular the subject of induction aerodynamics
got much more attention. Thereto the Mexico experiment as considered in IEA Task 29 provided valuable information since
this experiment mapped the flow field around the wind turbine together with measurements of loads along the blade. As such
the relation between loads and velocity in the rotor plane became one of the main focal points in Mexnext as well as induction
effects at, for example, dynamic conditions (like pitching steps), yawed flow, and the turbulent wake situation.
4.3 | Aerodynamic experiments: Specific progress

4.3.1 | Validation
Validation was one of the main reasons to initiate dedicated aerodynamic measurement programs and as soon as the first
results became available a wide range of rotor aerodynamic models, for example, engineering models, free wake methods, air-
foil design tools, and CFD methods have been compared with these measurements at both aligned and yawed flow conditions
(some more recent examples of validation can be found in Boorsma et al., 2014; Boorsma & Schepers, 2014; Herraez et al.,
2014; Jeromin et al., 2014; Madsen et al., 2015; Mahmoodi & Schaffarczyk, 2012; Sørensen et al., 2016; Boorsma & Sche-
pers, 2016; Hartveld, 2016).
It should be realized that a direct validation of lifting line codes inevitably remains limited to the comparisons of local
aerodynamic loads (normal and tangential forces) since the lifting line variables (e.g., angles of attack and underlying induced
velocities and aerodynamic coefficients) which are used as input in lifting line codes cannot be measured straightforwardly
from the present measurements (see Section 3.1).
CFD and panel methods do not rely on these lifting line variables and results from such codes can be compared directly
with measurements of pressure distributions or with flow field quantities, see the examples from Mexnext in Figures 10 and
11 (from Sørensen et al., 2016).
In many validation studies the comparison between calculated and measured aerodynamic properties (e.g., normal forces
along the blade or pressure distributions) showed a good to reasonable agreement below stall, but above stall the agreement
became poorer. Disagreements, in particular in engineering model results, were also found at yawed conditions.
An interesting observation is the added value from CFD compared to the more pragmatic models on several subjects. This
is even true at stalled conditions: although CFD predicts stalled conditions poorer than attached flow the agreement is rela-
tively good in comparison to the agreement from the engineering methods.
This is illustrated in Figure 12. It comes from Boorsma et al. (2018), and it shows large differences between calculated and
measured tangential force from the New Mexico experiment at a high tunnel speed (i.e., at stalled conditions) with the largest dif-
ferences for the lower fidelity engineering methods. The best agreement is found for the higher fidelity CFD methods.
The prediction of the flow field is often very well from CFD (and panel codes), see Figure 11 which shows that CFD pre-
dicts the velocity profiles in radial direction well in front of the rotor and it also predicts the lower velocity in the wake of the
r/R=0.35 r/R=0.60
2500 5000
Exp Exp
Ell, turb 4000 Ell, turb
2000 Ell, tran Ell, tran
Surface pressure (Nm–2)

3000
1500
2000
1000
1000
500
0
0
–1000
–500 –2000
–1000 –3000
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
Chord (m) Chord (m)
r/R=0.82 r/R=0.92
10000 12000
Exp Exp
8000 Ell, turb 10000 Ell, turb
Ell, tran Ell, tran
6000 8000

6000
4000
4000
2000
2000
0
0
–2000
–2000
–4000 –4000
–6000 –6000
–0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 –0.01 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Chord (m) Chord (m)
FIGURE 10 Pressure distributions at four radial positions measured in the New Mexico project and compared with Ellipsys3D calculations (fully turbulent
and transitional) at design conditions (15 m/s)
rotor with a sudden increase toward the free stream velocity at the edge of the wake (near r = 2.25 m) as induced by the tip
vortices. Later it will be shown that even at yawed conditions CFD predicts the velocity profiles reasonably well.
A very important outcome of the validation studies using detailed aerodynamic measurements was the conclusion that
such measurements are an absolute necessity when validating wind turbine design codes since a comparison with global mea-
surements of (rotor) blade loads does not give a decisive answer on the accuracy of such codes: This is due to the fact that a
good agreement between calculated and measured rotor (blade) loads is often caused by compensating errors (e.g., an overpre-
diction at the root can be compensated by an underprediction at the tip). In such cases a comparison with global load
Up-stream Down-stream
15 17
14.5 16
15
Axial velocity (ms–1)
14
Axial velocity (ms–1)
14
13.5
13
13
12
12.5
11
12 10
11.5 Exp 9 Exp
Ell, tran Ell, tran
11 8
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
Radius (m) Radius (m)
FIGURE 11 Radial profile of velocities measured in New Mexico and calculated with Ellipsys3D 0.3 m up and downstream of the rotor at design conditions
(tip of the blade at radius = 2.25 m)
Ft (N/m), U∞=24 m/s
NewMexico
120 BEM
FVW
AM
110 CFD_trans
CFD_turb
stdev(avg) band
100
Ft (N/m)
90
80
70
60
0.5 1.0 1.5 2.0

r (m)
FIGURE 12 New Mexico experiment: Tangential force as function of radial position, calculational results from different methodologies
measurement can yield a good agreement and the misleading conclusion that the code performance is very good since it
remains unnoticed that the underlying sectional loads are predicted poorly. This is illustrated in Figure 13, which shows the
ratios between ECN calculated and measured normal forces and tangential forces at five radial positions for the NREL UAE-
Phase IV experiment, as function of mean wind speed (from Schepers et al., 2004). Large disagreements can be seen at the
higher velocities but still Figure 14 shows a good agreement between calculated and measured rotorshaft torques due the large
disagreement in normal forces as compensated by the large disagreement in tangential force. These compensating errors could
be shown only by the availability of measured sectional forces.
4.3.2 | Aerodynamic understanding

The measurements considered in this article have not only provided validation material but they also have led to fundamental
model insights and aerodynamic understanding which clearly supported the improvement of aerodynamic wind turbine models
as implemented in design codes. This is illustrated by the statement from Schepers (2012) “nowadays there isn't a designer to
find who would dare to design a wind turbine with the very basic aerodynamic modeling as applied in the Wind Turbine
Benchmark Exercise on Mechanical Loads in the end 1980’s (van Grol et al., 1991)”.
1.5 10.0
Calculated/measured tangential force (–)
Calculated/measured normal force (–)
1.0 30% span

47% span
63% span
5.0 80% span
95% span
0.5
30% span
47% span
63% span
80% span
95% span
0.0 0.0
5 7 9 11 13 15 17 19 5 7 9 11 13 15 17 19
Wind speed (m/s) Wind speed (m/s)
FIGURE 13 Ratio between ECN calculated and NREL Phase IV measured normal force (left) and tangential force (right) as function of wind speed at five
radial positions
2.0
Calculated/measured blade loads (–)

M_flat
M_torque
1.0
0.0
5 7 9 11 13 15 17 19
Wind speed (m/s)
FIGURE 14 Ratio between ECN calculated and NREL Phase IV measured flatwise moment and rotorshaft torque as function of wind speed
Moreover Schepers (2012) mentions several modeling areas on which improvements were possible through the dedicated
aerodynamic measurements (e.g., 3D rotational effects, near wake and inflow aerodynamics, yawed flow, unsteady effects on
airfoil aerodynamics, and dynamic inflow). More recently subjects like boundary layer transition, “IEC Aerodynamics” and
devices have been put on the agenda of IEA Task 29. All these subjects are often considered independent but it must be real-
ized that several cross-links exists, for example, yaw aerodynamics and near wake aerodynamics are connected since it is the
near wake which determines the azimuthal variation of induced velocity where yaw and unsteady airfoil aerodynamics are also
connected in view of the angle of attack variation caused by yaw. In the sequel of this section the progress on several of these
research subjects is touched upon.
• 3D effects: One of the most important research areas for the present measurements was the 3D aerodynamic effects where
in particular the effect from rotation on the lift is studied. 3D geometrical effects (through taper, and twist) generally got
little attention although it is considered in van Rooij and Schepers (2005) and tip effects (e.g., the effect of the finite blade
geometry) also got attention in several studies, see below.
The effect of rotation is investigated by comparing measurements of rotating (mean) airfoil characteristics, with 2D airfoil
characteristics and where possible to standstill measurements on the 3D blade. Such studies require the angle of attack
which, as mentioned in Section 3.1 is difficult to retrieve from wind turbine measurements although in Schepers (2012) a
method is suggested, first introduced in Schepers et al. (2004) which tunes the airfoil characteristics as input to BEM to
the measurements of dimensional loads (e.g., normal forces and tangential forces). The tuned lift and drag coefficients at
different angles of attack can then be compared to the original 2D airfoil coefficients from which a quantitative correction
on the lift and drag coefficient can be derived. Still usually the angle of attack is derived, for example, through an experi-
mentally 2D upwash correction to inflow angle data as acquired from five-hole probes mounted ahead of the blade, for
example, in Schreck and Robinson (2002) UAE Phase VI measurements analyzed in this manner showed significant rota-
tional augmentation of aerodynamic forces compared to stationary 2D airfoil and parked 3D blade measurements. In
related work, the inverse free vortex method (Micallef, Kloosterman, et al., 2010; Sant et al., 2006) was used to derive
angle of attack for UAE Phase VI and MEXICO measurements. Both turbines exhibited substantial rotational augmenta-
tion relative to parked blade conditions, even though blade solidities and rotor control strategies differed greatly (Schreck
et al., 2010). Work that complemented UAE Phase VI measurements with Ellipsys3D computations clearly showed the
fluid dynamic structures and processes responsible for rotational augmentation (Shen, Mikkelsen, & Sørensen, 2005). The
insights from the measurements have led a wide variety of models to correct 2D airfoil data. Some of these modeling
activities are reported in Schepers (2012) (based on work from Schepers et al., 2004) and Bjorck (1995), Elgammi (2016);
Lindenburg (2003), Snel (1997), and Guntur et al. (2012) (to mention only a few).
Moreover, it was found that at the outer stations an over prediction of loads occurs when using 2D data. Models to
describe this geometrical tip effect on the airfoil coefficients (not to be confused with the Prandtl tip effects on induction)
are described in Shen et al. (2005), Schepers et al. (2004, 2014), Schreck et al. (2010). Little research has been carried out
on the drag at rotating conditions but in Schepers et al. (2004) an increased drag is found from rotation and a model for
that effect is derived. This model was a refinement to the model from Chaviaropoulos and Hansen (2000) in which an
engineering model for a drag increase at rotation was derived from CFD calculations. In Potentier (2013) the drag is stud-
ied using the experiment carried out by FFA in the large CARDC tunnel (Ronsten, 1992).
• Near wake and inflow aerodynamics: Several investigations, in particular within Mexnext, focused on the near wake
flow field, (e.g., Breton, 2011; Szasz, 2011) to mention a few. A clear dependency on the loading of the turbine was found
where a higher loading shows a much stronger wake expansion and a slower wake decay. The modeling of the high load-
ing cases (i.e., the turbulent wake state) with CFD initially led to convergence problems (Schepers, Boorsma, et al., 2012)
but after several code adjustments much better results were obtained. The path of the tip vortex in the wake was studied
in, for example, Nilsson et al. (2011) showing a wider expansion at higher loading where Xiao et al. (2011) find an initial
compression of the wake just downstream of the rotor, after which it expands. The initial wake compression is seen in
measurements from Micallef (2012) as well and explained in Van Kuik, Micallef, Herraez, Van Zuijlen, and Ragni (2014)
by the role of conservative forces which do not appear in approaches considering infinitely thin actuator discs or lines.
Also interesting to observe in the Mexico measurements is the vortex shedding in the wake from dynamic stall effects (see
the discussion on unsteady airfoil aerodynamics below). This vortex shedding is generally modeled well with CFD. More-
over the Mexico flow field measurements in the rotor plane have been used to analyze the nonuniformity of the flow
between the blades. It is this nonuniformity which forms the basis for the generally accepted tip loss factors as implemen-
ted in almost every design code. Some deficiencies and suggestions for improvement of the generally applied tip loss fac-
tor models from Prandtl are given in Bon (2011) and Ramdin (2017). However, an assessment of the tip loss factor from
the velocity measurements is far from straightforward due the upwash from the blade which should be substracted in order
to find the relevant induced velocities (see Section 3.1). Figure 15 shows the velocities as function of streamwise coordi-
nate (x = 0 m is in the rotor plane) for six-blade azimuth angles with a 20 interval. The yellow curve is the measured
result with one of the blades 10 below the PIV sheet at the 9 o'clock position where the blue curve has a blade 10 above
the PIV sheet (Figure 7). This result clearly shows a jump in velocity when the blade passes the PIV sheet which can be
attributed to the bound vortex.
• Relation between loads and rotor velocities: Much effort was spent on the relation between loads and velocities as mea-
sured in the Mexico experiment. Although a consistent dependency of the loading was found on the wake expansion, the
results did not seem to obey the momentum relation. In the later New Mexico experiment more reliable calibrations were
carried out after which the relation became much more consistent (Boorsma & Schepers, 2016) and Figure 16 shows the
axial force coefficients versus induction factor (determined with a variant of the AAT technique from Section 3.1 at differ-
ent radial positions and tip speed ratios). A detailed momentum analysis relating the loads and a full survey of the veloci-
ties from New Mexico is given in Parra et al. (2016) with very consistent results on the momentum balance in axial
direction and even in in-plane direction. On the other hand Figure 16 still shows some relatively large differences between
local thrust coefficient and local induction factor in particular at large thrust coefficient and induction factor.
• Yawed flow and unsteady airfoil aerodynamics: In the EU projects Dynamic Inflow as carried out in the 1990s
(Schepers & Snel, 1995; Snel & Schepers, 1994), much attention was paid to the azimuthal variation in induced velocity
u (m/s)
13
12
Axial velocity (m/s)
11
10
7 x (m)
–0.3 –0.2 –0.1 0.0 0.1 0.2 0.3

Axial location (m)
FIGURE 15 Velocities near the rotor plane as function of axial coordinate at 80% span and Vtun = 15 m/s for six blade azimuth angles, different lines
represent different blade positions
1.5
Thrust coefficient (–)

1.0
Momentum theory
Empirical approx.
λ=9.96, V=10 m/s
λ=6.67, V=15 m/s
0.5 λ=4.16, V=24 m/s
λ=6.67, V=15 m/s (rough)
r/R=0.25
r/R=0.35
r/R=0.60
r/R=0.82
0.0 r/R=0.92
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

Axial induction factor
FIGURE 16 Axial force coefficient versus axial induction factor from New Mexico experiment
at yaw. The developed models mainly relied on insights from the helicopter society from which a sinusoidal variation, as
induced by the tip vortices, was expected. In Schepers (1999) and Vermeer (1998) velocity measurements taken in the
OJF tunnel of TUDelft showed a deviation from this sinusoidal variation at the inner part of the blade due to the skewed
root vortex. These measurements were used to derive an engineering model for that effect which is described in detail in
Schepers (2012). The tip vortex path in yaw is studied and simulated in section 8.5.1 of Kuczaj (2009) and Schepers
(2012) and in a research led by University of Stuttgart, see section 9.4 of Schepers et al. (2014). These references show
that CFD can simulate the tip vortex path from the Mexico experiment reasonably well. Consistent to this observation is
the very good agreement between CFD calculated and measured flow field in the rotor plane at yawed conditions
(e.g., section 8.3 of Schepers, Boorsma, et al., 2012 and Figure 17).
• In a wind tunnel yaw leads to a well-defined periodic excitation by which it is often connected to the study of unsteady
airfoil aerodynamics. Thereto it should be known that an assessment of models for unsteady airfoil aerodynamics
through field measurements is difficult through the stochastic character of the atmosphere. This is illustrated in
Figures 18 and 19.
They show measurements on the RIS test facility (Madsen, 1991a) from IEA Task 14/18 of the normal force coeffi-
cients at 68% span as function of angle of attack (based on a five-hole pitot probe measurement) around a low and a
high angle of attack. A very disorderly pattern can be observed at high angle of attack by which these measurements
u (m/s), xm=+0.15 m, Az=240°, Yaw=30°, Vtun=15 m/s

20 MEXICO
DTU_MEK_AL_OAD
RISOE_3DTURB
TVD_PAN
TECHNION_CFD
Axial velocity (m/s)
15
10
–3 –2 –1 0 1 2 3
Radius (m)
FIGURE 17 Axial velocity as function of radial position for a yaw angle of 30 , 0.15 m downstream of the rotor. Measured in Mexico experiment and
calculated by Mexnext participants (measurements start at a radial positon of 1.2 m)
1.8
Normal force coefficient (-)

1.6
1.4
1.2
0.8
4 6 8 10 12 14 16 18 20 22
Angle of attack (deg)
FIGURE 18 Normal force as function of attack (low angle of attack) as measured by RIS in the field
cannot be used for a direct validation of dynamic stall models. To some extent the latter problem is overcome by select-
ing data according to a so-called baseline criterion as applied by NREL and TUDelft (Bruining, 1993; Miller et al.,
1995), which aims to select measurements which are taken at conditions which are as steady as possible.
Dynamic stall is caused by a vortex build up starting at the blade nose which is then convected downstream. The pres-
ence of this vortex and its convection is studied through the analysis of the surface pressure history at the different pres-
sure taps along the chord. Research along these lines initially was carried out using UAE Phase IV surface pressure
data, which were acquired in the atmosphere (Schreck et al., 2000) Though wind inflow variations introduced signifi-
cant anomalies, dynamic stall vortex kinematics were successfully isolated and characterized for yawed rotor condi-
tions, showing pronounced 3D deformations. These unique vortex kinematics were shown to be closely linked to
unsteady force amplifications that varied cyclically with blade rotation. Thereafter, similar analyses were applied to
UAE Phase VI (Schreck et al., 2001) and Mexico (section 12 of Schepers et al., 2014) wind tunnel yawed rotor surface
pressure data. Wind tunnel controlled conditions enabled more detailed characterizations of the vortex dominated flow
fields and associated unsteady aerodynamic forces, and results from these two studies were mutually corroborative, with
both showing 3D vortex deformations and force amplifications in response to rotor yaw and inflow speed. To gain
knowledge of flow field structures and interactions above the blade surface, detailed computational characterizations of
the UAE Phase VI blade oscillating in pitch with the rotor parked were validated with wind tunnel data (Sørensen &
Schreck, 2012). These revealed dynamic stall vortex reminiscent of those observed under operating rotor conditions.
More recently, Elgammi (2016) and Guntur et al. (2013) analyzed UAE Phase VI surface pressure data, exploiting con-
trolled conditions to separate the effects of dynamic stall and rotational augmentation on rotor blade vortex flow fields
and unsteady force production.
Modeling efforts based on Mexico measurements at yawed conditions took place in Pereira et al. (2011) where the Bed-
does Leisman dynamic stall model was validated and improved.
1.8
Normal force coefficient (-)
1.6
1.4
1.2
0.8
4 6 8 10 12 14 16 18 20 22
Angle of attack (deg)
FIGURE 19 Normal force as function of attack (high angle of attack) as measured by RIS in the field
• Tower effects: Tower effects were found to be overpredicted in comparison to IEA Task 14/18 measurements when a
standard potential dipole model common in BEM tools is used (see Graham & Brown, 2000). Recent research
(Larwood & Chow, 2016) has compared the aerodynamics of the UAE Phase VI rotor for upwind and downwind configu-
rations, focusing on aerodynamic load production during passage of the blade through the tower wake.
• Dynamic inflow: Dynamic Inflow (unsteady induction) effects occur at fast thrust changes due to, for example, pitching
or rotor speed variations. In such cases the wake behind the rotor, and subsequently, the induction responds with a certain
delay resulting into an overshoot in the load response. Dynamic inflow effects were studied in the 1990s in Schepers and
Snel (1995) and Snel and Schepers (1994) based on measurements of power and blade root bending moments These mea-
surements were useful for the development of first generation dynamic inflow based on a time constant which depends on
the radial position. However, this radial dependency could not be assessed from the above-mentioned measurements from
power and blade root bending moments. Among others the NASA Ames experiment and the New Mexico experiment
measured the response at pitching steps of the aerodynamic loads at different radial positions leading to information of the
radial dependency of the time constant (see section 7.3 of Schepers, 2012) (based on Schepers, 2007c), and in Boorsma
and Schepers (2016), Hartveld (2016). In Yu, Hong, Ferreira, and van Kuik (2016)) the flow behind an actuator disc
undergoing a sudden change in porosity (leading to a sudden change in thrust) is measured in the TUDelft Open Jet
Facility.
• Boundary layer transition: Precise knowledge of the transition location on an airfoil of a rotating wind turbine blade is
extremely important since it determines the airfoil characteristics (in particular the drag) and so the power and loads to a
large extent. The precise prediction of the transition point turns out to be a difficult issue. In Section 2 it is mentioned that
in most New Mexico experiments the zig-zag tape which was present in the Mexico experiment was removed at the outer
part of the blade. In Figure 10 results are shown from fully turbulent and transitional calculations which show a better
agreement with the New Mexico experiment for the transitional cases.
Moreover airfoil data used in BEM generally originate from wind tunnel measurements at very low turbulence levels.
Within the wind energy society the usefulness of “smooth” wind tunnel data for the calculation of the power has often
been questioned because it is believed that the turbulent environment in which a wind turbine operates causes a significant
forward shift of the transition point on the suction side, “by-pass transition.” This would yield a higher drag and conse-
quently a considerably lower power. In Section 2 specific experiments are mentioned which investigate transition. More-
over it was possible to determine the state of the boundary layer with the fast Kulite sensors applied in the New Mexico
experiment (Brendon, 2018). They showed a change in the spectrum of the pressure signals between laminar and turbulent
flow at those frequencies where instable Tollmien–Schlichting disturbances appear. These instable Tollmien–Schlichting
disturbances are known to be the primary cause for boundary layer transition. The relevant frequencies were calculated
from the shape factor with the airfoil design code RFOIL (Snel et al., 1993) and the database of stability diagrams supplied
together with the eN method (van Ingen, 2008).
• Devices: Nowadays many turbines are equipped with flow devices (e.g., vortex generators, flaps, root spoilers). The flow
physics of these devices (in particular at rotating conditions) is still an unexplored area for which little measured informa-
tion is available. Within New Mexico measurements have been performed on root spoilers and guerney flaps. The latter
were applied in two different configurations, namely up to 60%R and 46%R, and are assessed using the pressure sensors
and generator torque. The effects of guerney flap were visible in the pressure distributions at the inner part. The shorter
Guerney flap increases power for all measured tip speed ratios where the longer flaps decrease the power at higher tip
speed ratios due to the drag penalty which becomes apparent at those tip speed ratios (see Boorsma & Schepers, 2016).
It turned out that the (New) Mexico set-up with a maximum blade thickness of only 25% is less optimal for spoilers (root
spoilers are mainly meant to improve performance of very thick transition airfoils). A follow-up experiment on root
spoilers was carried out within the framework of AVATAR in the TUDelft OJF (see Gonzalez et al., 2016).
• “IEC Aerodynamics”: Parked conditions are among the design driving load conditions (storm loads) where results from
the AVATAR project shows these conditions very challenging to model for large blades (Heinz et al., 2016). In the vari-
ous IEA Tasks several standstill measurements have been carried out. Calculations at the parked UAE-PhaseVI at pitch
motion are presented in Sant (2007). The measurements are often compared with those on an airfoil section at 2D condi-
tions so that the effect of 3D blade geometry can be assessed, where the velocities induced by the trailed vortices from the
3D load distributions should be substracted (e.g., section 6 of Schepers et al., 2014).
Another challenge lies in the modeling of some unconventional IEC load cases. One can think of a pitch failure leading to
a pitch unbalance between the blades. Such cases are obviously difficult to measure (no wind turbine owner will allow a
deliberate pitch unbalance on his/her turbine) but within New Mexico a pitch unbalance of 20 could be applied to one of
the blade. It was then found that this unbalance effects the pressure distribution of the unaffected blades too (see Oggiano
et al., 2016). This reference shows simulation results from models of different complexity on this challenging load case.
All the models are only partially able to reproduce the complex flow generated by the off-pitched blade and the loads on
the immediately following blade but the results from CFD show a relatively good match due to a better prediction of the
near wake and flow conditions in the rotor plane.
5 | RE COMME ND ATIONS
Although the detailed aerodynamic measurements from this paper have led to a significant improvement in aerodynamic
models several comparisons were shown between calculations and measurements where differences are large. So further
improvement of aerodynamic models is still urgently needed. It should be realized that these insights on large differences
between calculations and measurements are based on local aerodynamic measurements and that differences in terms of inte-
grated loads often turned out to be relatively limited.
The most important recommendation from this article then follows from the observation that a comparison of results from
wind turbine codes with measurements of integrated blade and rotor loads does not give a decisive answer on the accuracy of
those codes and that this answer and the subsequent improvement of codes can only be based on detailed aerodynamic mea-
surements. Moreover a sound scientific approach requires data on a large variety of wind turbine configurations in order to
assess the general validity of aerodynamic models. This holds for the assessment and calibration of engineering models but
also for CFD models in which closure relations need to be calibrated. The amount of data presented in this article, even though
it is seemingly large, is still far from sufficient to meet these requirements. As such the quantity of data should be increased
but the quality of data is essential too. This is illustrated by the fact that, in retrospective, part of the efforts which the aerody-
namic research society spent on the first Mexico measurements turned out to be useless since some Mexico measurements suf-
fered from insufficient quality leading to wrong conclusion. This had to be overcome by another (costly) experiment, that is,
the New Mexico experiment.
Apart from wind tunnel measurements, field measurements need to be carried out on a modern state of the art wind tur-
bine. Obviously the aerodynamics on such large turbine is inextricable related to aero-elastic effects. It is important (though
difficult) that such experiment is carried out in a public framework, that is, the measurements data but also the underlying
aerodynamic design data should be shared within the wind energy community. In line with these recommendations, the sub-
group aerodynamics of the European Energy Research Alliance has appointed such experiment to be one of its main focal
points (Schepers, Madsen, Soerensen, & Munduate, 2012) where the description of the experiment has been detailed in the
AVATAR project (Schepers et al., 2018) and the recently started fourth phase of IEA Task 29. In these projects it will be sug-
gested to instrument a blade which is designed by the research community in order to overcome the problem of confidentiality
of blade data.
Moreover wind tunnel measurements have to be performed in the largest wind tunnel possible in order to achieve Reyn-
olds numbers on the blade which are at least to some extent representative for Reynolds numbers on modern wind turbines.
Alternatively the potential of cryogenic tunnels should be explored.
The test programs should consider the most recent and advanced measurement techniques and they should record more
data than those considered in this article. The most essential addition should lie in the measurement of boundary layer parame-
ters (transition, shear stresses) and (viscous) drag and detailed inflow measurements, for example, from LIDAR. These are
very important parameters but they are, apart from a few rare boundary layer transition experiments largely concealed in the
present experiments. An extensive uncertainty analysis should be carried out on every experiment where a description of the
input data (with knowledge of its accuracy) is needed too. Detailed information on the aerodynamics of flow devices is also
seen as very important. Moreover flow visualization of both the blade flow and the flow around the rotor is needed.
ACKNOWLEDGMENTS
The support from the IEA Wind Executive Committee which made the cooperation of measurement programs possible is
greatly appreciated. The authors are also grateful to the contribution of participants to the IEA Tasks described in this article.
A special word of thanks should go to Koen Boorsma from ECN for his unyielding support in the analysis of the (New)
Mexico data.
CONFLICT OF INTEREST
The authors have declared no conflicts of interest for this article.
RELATED WIREs ARTICLE

An overview of the state of the art technologies for multi-MW scale offshore wind turbines and beyond
REFERENC ES
Acker, T., & Hand, M. (1999, January). Aerodynamic performance of the NREL unsteady experiment (Phase IV) twisted rotor. Paper presented at Proceedings of the
1999 ASME Wind Energy Symposium, Reno, NV, USA.
Barlow, J. B., Rae, W. H., Jr., & Pope, A. (1999). Low-speed wind tunnel testing (3rd ed.). New York: John Wiley & Sons.
Barthelmie, R. J., Larsen, G., Pryor, S., Ejsing Jørgensen, H., Bergström, H., Magnusson, M., … Folkerts, L. (2003). Efficient development of off-shore wind farms
(Final Report of the ENDOW Project No. RISO-R-1407(EN)). RISO Danish National Laboratory.
Barthelmie, R. J., Frandsen, S. T., Rathmann, O. S., Hansen, K.S., Politis, E., Prospathopoulos, J., … Heath, M. (2011). Flow and wakes in large off-shore wind farms
(Final report for Upwind No. WP8 RISO-R-1765(EN)). RISO Danish National Laboratory.
Bechmann, A. & Sørensen, N. (2009). CFD simulation of the Mexico rotor wake. Paper presented at European Wind Energy Conference. Marseille, France.
Bellia, J. M. (1990, July). Aerodynamic measurements on a small HAWT rotor in axial and yawed flow (PhD thesis). Cranfield Institute of Engineering, School of
Mechanical Engineering.
Bermudez, L., Velasquez, A., & Matesanz, A. (2000). Numerical simulation of unsteady aerodynamic effects in horizontal Axis wind turbines. Solar Energy, 68, 9–21.
Bjorck, A. (1995). Dynamic stall and three dimensional effect (Report No. FFA-TN-1995-31). FFA.
Bon, A. (2011, March). Non-uniformities of the flow between blades of wind turbine Analysis of Mexico measurements and simulation (Report No. ECN Wind
Memo-11-026). Energy Research Center of the Netherlands.
Boorsma, K., & Schepers, J. G. (2011). Description of experimental set-up, Mexico measurements (Report No. ECN-X-11-120). Energy Research Center of the
Netherlands
Boorsma, K. & Schepers, J. G. (2014, September). New MEXICO experiment. Preliminary overview with initial validation (Report No. ECN-E-14-048). Energy
Research Center of the Netherlands.
Boorsma, K., & Schepers, J. G.. (2015). Description of experimental set-up, New-Mexico experiment (Report No. ECN-X-15-093). Energy Research Center of the
Netherlands
Boorsma, K. & Schepers, J. G. (2016, October). Rotor experiments in controlled conditions continued: New Mexico. Paper presented at Science of Making Torque Con-
ference, Munich, Germany.
Boorsma, K., Schepers, J. G., Gomez-Iradi, S., Aagaard Madsen, H., Sørensen, N., Shen, W. Z., …, & Schreck, S. (2014, March). Mexnext-II: The latest results on
experimental wind turbine aerodynamics. Paper presented at EWEA 2014, Barcelona, Spain.
Boorsma, K., Schepers, J. G., Schaffarczyk, P., Rahimi, H., Oggiano, L.; Gomez-Iradi, S., …, Herraez, I. (2018, May). Final report of IEA wind task 29 Mexnext
(Phase 3) (Report No. ECN-E-18-003), Energy Research Center of the Netherlands.
Brand, A. J., Dekker, J. W. M., de Groot, C. M., & Spath, M. (1996). Overview of aerodynamic measurements on an Aerpac 25 WPX wind turbine blade at the HAT
25 experimental wind turbine (Report No. ECN-DE-Memo-96-014). Energy Research Center of the Netherlands.
Brendon, A. I. (2018, June). Investigation into boundary layer transition on the New Mexico Blade. Science of Making Torque Conference, Milano, Italy.
Breton, S. (2011, June). Numerical analysis of the vorticity structure of the MEXICO rotor in the near wake. Paper presented at Session on Mexico Experiments at
Wake Conference, Visby, Sweden: Gotland University.
Breton, S., Sibuet, C., & Masson, C. (2010a). Using the actuator surface method to model the three-bladed MEXICO wind turbine. Paper presented at 48th AIAA Aero-
space Sciences, Orlando, FL.
Breton, S., Sibuet, C., & Masson, C. (2010b). Analysis of the inflow conditions of the MEXICO rotor: Comparison between measurements and numerical simulations.
The Science of Making Torque from the Wind, Heraklion, Greece.
Bruining, A. (1993). Pressure distributions from a rotating wind turbine blade measured on the Open Air Research Facility of Delft University of Technology. Paper
presented at Proceedings of AWEA Conference, San Fransisco, CA.
Bruining, A. (1997, September). Aerodynamic characteristics of a 10m diameter rotating wind turbine blade (TUDelft Report No. IW95-084R), Delft, the Netherlands.
Butterfield, C. P., Musial, W. P., & Simms, D. A. (1992, October). Combined experiment phase 1 (Final Report No.NREL/TP-257-4655). National Renewable Energy
Laboratory.
Chaviaropoulos, P. K., & Hansen, M. O. L. (2000). Investigating 3D and rotational effects on wind turbine blades by means of a quasi-3D Navier stokes solver. Journal
of Fluids Engineering, 122(2), 330–336.
Chaviaropoulos, P. K. et al. (2001, July). Viscous and aeroelastic effects on wind turbine blades. The viscel project. Paper presented at Proceedings of European Wind
Energy Conference (pp. 347–351). Copenhagen, Denmark.
Cho, T., & Kim, C. (2012). Wind tunnel test results for a 2/4.5 scale Mexico rotor. Renewable Energy, 42, 152–156.
Cho, T., & Kim, C. (2014). Wind tunnel test for the NREL phase VI rotor with 2m diameter. Renewable Energy, 65, 265–274.
Dahlberg, J. A. et al. (1989, July 10–13). Wind tunnel measurements of load variations for a yawed turbine with different hub configurations. Paper presented at Pro-
ceedings of EWEC Conference (pp. 95–100). Glasgow, Scotland.
Duque, E. P. N., van Dam, C., & Hughes, S. (1999, January). Navier-Stokes simulations of the NREL combined experiment Phase II rotor. Paper presented at Proceed-
ings of the 1999 ASME Wind Energy Symposium, Reno, NV.
Elgammi, M. (2016, December). Modelling the self-induced cycle to cycle variations in the aerodynamic blade loads of a yawed wind turbine under controlled operat-
ing conditions (PhD thesis). University of Malta.
Ewald, B. F. R. (1998, October). Wind tunnel wall correction. AGARDograph 336.
Feigl, L. (2003, September). Analysis of aerodynamic field measurements on instrumented wind turbine rotors (Master Project No. ECN 2003). EUREC-Agency.
Gomez-Iradi, S. (2011). A CFD investigation of the influence of trip-tape on the MEXICO wind turbine blade sections. Paper presented at Science of Making Torque
from the Wind, Heraklion, Greece.
Gonzalez, A., Baldacchino, D., Caboni, M., Kidambi, A., Manolesos, M., & Troldborg, N. (2016, December). Aerodynamic flow control. Deliverable 3.5 AVATAR
project.
Graham, J. M. R., & Brown, C. J.. (2000, September). ROTOW-investigation of the aerodynamic interaction between wind turbine rotor blades and the tower and its
impact on wind turbine design (Final Publishable Report to the EU, Contract Nr: JOR3-CT98-0237). London, England: Department of Aeronautics, Imperial Col-
lege of Science and Medicine.
Guntur, S., Sørensen, N., & Schreck, S. (2013). Dynamic stall on rotating airfoils: A look at the n-sequence data from the NREL phase VI experiment. Key Engineering
Materials, 569-570, 611–619. https://doi.org/10.4028/www.scientific.net/KEM.569-570.611
Guntur, S. K., Bak, C., & Sorensen, N. N. (2012). Analysis of 3D stall models for wind turbine blades using data from the Mexico experiment. Paper presented at 13th
International Conference on Wind Engineering, ICWE, Amsterdam, Holland.
Haans, W. (2011). Wind turbine aerodynamics in yaw, unravelling the measured rotorwake (PhD thesis). Technical University of Delft.
Haans, W., Sant, T., van Kuik, G. A. M., & van Bussel, G. J. W. (2006). Stall in yawed flow conditions: A correlation of blade element momentum predictions with
experiments. Journal of Solar Energy Engineering, 128, 472–408.
Hand, M. M., Hand, M. M., Simms, D. A., Fingersh, L. J., Jager, D. W., Cotrell, J. R., . . ., Larwood, S. M. (2001). Unsteady aerodynamics experiment phase vi wind
tunnel test configurations and available data campaigns (Report No. NREL/TP-500-29955). National Renewable Energy Laboratory.
Hansen, M. O. L. (1999). Viscous and aeroelastic effects on wind turbine blades. Viscel, Task 1 Report, ET-Viscel-TR1.
Hartveld, M. (2016). Comparison of the Aeroelastic Free Vortex Wake Code AWSM with Conventional BEM Based Codes on 10MW+ Wind Turbines. (ECN/TUDelft
Master thesis project).
Heinz, J., Sørensen, N. N., Riziotis, V., Schwarz, M., Gomez Iradi, S., & Stettner, M. (2016, August). Aerodynamics of large rotor, Deliverable 4.5 of the AVATAR
project.
Heißelmann, H., Peinke, J. & Hölling, M. (2016, October). Experimental airfoil characterization under tailored turbulent conditions. Paper presented at Science of
Making Torque, Munich, Germany.
Herráez, I., Daniele, E., & Schepers, J. G. (2018). Extraction of the wake induction and angle of attack on rotating wind turbine blades from PIV and CFD results. Wind
Energy Science, 3, 1–9. https://doi.org/10.5194/wes-3-1-2018
Herraez, I., Stoevesandt, B., & Peinke, J. (2014). Insight into rotational effects on a wind turbine blade using Navier–stokes computations. Energies, 7(10), 6798–6822.
Hua, Y., et al. (2011). Extraction of airfoil data using PIV and pressure measurements. Journal of Wind Energy, 14, 539–556.
Jeromin, A., Bentamy, A., & Schaffarczyk, A. P. (2014). Actuator disk modeling of the Mexico rotor with OpenFOAM. Paper presented at ITM Web of Conferences
2, 06001.
Johansen, J., & Sørensen, N. (2004). Aerofoil characteristics from 3D CFD rotor computations. Journal of Wind Energy, 7, 283–294.
Jost, E., Klein, L., Leipprand, L., Lutz, T., & Kramer, E. (2017). Extracting the angle of attack on rotor blades from CFD simulations. Wind Energy, 2018, 1–16. https://
doi.org/10.1002/we.2196
Kuczaj, A. K. (2009). Virtual blade simulations of the Mexico experiment (Report No. NRG-21810/09.97106). NRG
Kuik, G.A.M., van Rooij, R., Imamura M (2004, November). Analysis of the UAE Phase VI wind turbine tunnel results in the non-yawed flow. Paper presented at the
meeting of the Conference Proceedings of EWEA, London, England.
Larwood, S., & Chow, R. (2016). Comparison of upwind and downwind operation of the NREL Phase VI Experiment, torque 2016. Journal of Physics: Conference
Series, 753, 022041. https://doi.org/10.1088/1742-6596/753/2/022041
Leclerc, C., & Masson, C. (1999, January). Predictions of aerodynamic performance and loads of HAWTS operating in unsteady conditions. Paper presented at Proceed-
ings of the 1999 ASME Wind Energy Symposium, Reno, NV.
Lignarolo, L. (2016, April). On the turbulent mixing of horizontal axis wind turbine wakes (PhD thesis). TUDelft.
Lindenburg, C. (2003, June). Investigation into rotor blade aerodynamics. Analysis of the stationary measurements of the UEA Phase VI rotor in the NASA Ames wind
tunnel (Report No. ECN-C-03-025). Energy Research Center of the Netherlands
Lutz, T., Meister, K., & Krämer, E. (2011). Near wake studies of the Mexico rotor. Paper presented at EWEA Annual Event, Brussels, Belgium.
Madsen, H. A. (1991a, September). Aerodynamics of a horizontal-axis wind turbine in natural conditions (Report No. RISO M-2903). RISO National Laboratory.
Madsen, H. A.. (1991b, September). Aerodynamics of a horizontal-axis wind turbine in natural conditions—Raw data overview (Report No. RISO M-2901). RISO
National Laboratory.
Madsen, H. A. (1991c, February). Structural dynamics of a 100 kW HAWT (Report No. RISO M-2887). RISO National Laboratory.
Madsen, H. A. et al. (2010, September). The DAN-AERO MW experiments final report (Report No. Risø-R-1726(EN)). RIS National Laboratory
Madsen, H. A. et al. (2015, June). Validation of models by the New Mexico data. Deliverable II.13 of the INNWIND.EU Project, Part 2.
Maeda, T., Kamada, Y., Murata, J., Suzuki, D., Kaga, N., & Kagisaki, Y. (2012, October). LDV measurement of boundary layer on rotating blade surface in wind tun-
nel. Paper presented at the Science of Making Torque Conference, Oldenburg, Germany.
Maeda, T., & Schepers, G. (2011). Wind turbine performance assessment and knowledge management for aerodynamic behaviour modelling and design: IEA experi-
ence chapter 12 of ‘wind energy systems. In Optimising design and construction for safe and reliable operation (pp. 350–365). Cambridge, UK: Woodhead Publish-
ing Ltd. ISBN 978-1-84569-527-9.
Mahmoodi, E. & Schaffarczyk, A. P. (2012, February 22–24). Actuator disc modeling of the Mexico rotor (Euromech Colloquium No. 528). Oldenburg, Germany:
Wind Energy and the Impact of Turbulence on the Energy Conversion Process.
Meister, K. (2011). Grid dependency studies on tip vortex preservation in wind turbine CFD simulations. Paper presented at Wake Conference, Visby, Sweden: Gotland
University.
Meng, F. & van Rooij, R. P. J. O. M. (2007). CFD investigations with respect to model sensitivity for the non-rotating flow around the NREL Phase VI blade. Paper pre-
sented at Proceedings of the 2007 European Wind Energy Conference and Exhibition, 2007 European Wind Energy Conference and Exhibition in Milan, May
7–10, 2007, EWEA, Brussels, Belgium.
Micallef, D. (2009a). MEXICO data analysis. Stage I—MEXICO data validation and reliability tests. Delft: TUDelft.
Micallef, D. (2009b). MEXICO data analysis. Stage V—investigation of the limitations of inverse free wake vortex codes on the basis of the MEXICO experiment.
Delft: TUDelft.
Micallef, D. (2012). 3D flows near a HAWT rotor: A dissection of blade and wake contributions (PhD thesis). Delft, the Netherlands: Technical University of Delft.
Micallef, D., Kloosterman, M., Ferreira, C., Sant, T., & van Bussel G.J.W. (2010, January). Validating BEM, direct and inverse free wake models with the MEXICO
experiment (Report No. AIAA-2010-0462). AIAA.
Micallef, D. et al. (2010). Validating BEM, direct and inverse free wake models with the Mexico experiment. Paper presented at 48th AIAA Aerospace Sciences,
Orlando, FL.
Micallef, D., et al. (2011). The relevance of spanwise flows for yawed horizontal-axis wind turbines. Paper presented at 13th International Conference on Wind Engi-
neering, ICWE, Amsterdam, Holland.
Miller, M. S., Shipley, D. E., Young, T. S., Robinson, M. C., Luttges, M.W., & Simms, D. A. (1995). The baseline data sets for phase II of the combined experiment
(Report No. NREL/TP/-442-6915). National Renewable Energy Laboratory.
Nilsson, K., Shen, W. Z., Sorensen, J. N., & Ivanell, S. (2011, June). Determination of the tip vortex trajectory behind the Mexico rotor. Paper presented at Session on
Mexico experiments at Wake Conference, Visby, Sweden: Gotland University.
Oggiano, L., Boorsma, K, Schepers, G. & Kloosterman, M. (2016, October). Comparison of simulations on the New Mexico rotor operating in pitch fault conditions.
Paper presented at the meeting of the Science of Making Torque Conference, Munich, Germany.
Parra, A., Boorsma, K., Schepers, J. G., & Snel, H. (2016, October). Momentum considerations on the New MEXICO experiment. Paper presented at Science of Making
Torque Conference, Munich, Germany.
Pascal, L. (2009). Analysis of Mexico measurements (Report No. ECN-Wind Memo-09-010). Energy Research Center of the Netherlands.
Pereira, R., Schepers, J. G., & Pavel, K. M. (2011). Validation of the Beddoes Leishman dynamic stall model for horizontal axis wind turbines using Mexico data. Paper
presented at 49th AIAA Aerospace Sciences Meeting. Orlando, FL.
Phengpom, T., Kamada, Y., Maeda, T., Matsuno, T., & Sugimoto, N. (2016). Analysis of wind turbine pressure distribution and 3D flows visualization on rotating con-
dition. IOSR Journal of Engineering, 06(02), 18–30.
Phengpom, T., Kamada, Y., Maeda, T., Murata, J., Nishimura, S., & Matsuno, T. (2015). Study on blade surface flow around wind turbine by using LDV measure-
ments. Journal of Thermal Science, 24(2), 131–113.
Phengpom, T., Kamada, Y., Maeda, T., Nishimura, J. M. S., & Matsuno, T. (2015). Experimental investigation of the three-dimensional flow field in the vicinity of a
rotating blade. Journal of Fluid Science and Technology, 10(2). https://doi.org/10.1299/jfst.2015jfst0013
Potentier, T. (2013, December). Improved aerodynamic modeling focused on drag investigation by analysis of advanced measurements. (ECN/Eurec Master thesis
report).
Rahimi, H., Hartvelt, M., Peinke, J., & Schepers, J. G. (2016, October) Investigation of the current yaw engineering models or simulation of wind turbines in BEM and
comparison with CFD and experiment. Paper presented at the Science of Making Torque Conference, Munich, Germany.
Rahimi, H., Schepers, J. G., Shen, W. Z., Ramos García, N., Schneider, M. S., Micallef, D., … Herráez, I. (2018). Evaluation of different methods for determining the
angle of attack on wind turbine blades with CFD results under axial inflow conditions. Renewable Energy, 125, 866–876.
Ramdin, S. F. (2017). Prandtl tip loss factor assessed, comparison using other methods (ECN/TUDelft thesis report).
Réthoré, P.-E., Sørensen, N. N., Zahle, F., Bechmann, A., & Madsen, H. A. (2011a). CFD model of the MEXICO wind tunnel. Paper presented at EWEA Annual Event,
Brussels, Belgium.
Réthoré, P.-E., Sørensen, N. N., Zahle, F., Bechmann, A., & Madsen, H. A. (2011b). MEXICO wind tunnel and wind turbine modelled in CFD. Paper presented at
AIAA Conference, Honolulu, HI.
Ronsten, G. (1992). Static pressure measurements on a rotating and a non-rotating 2.375 meter wind turbine blade, comparison with 2D calculations. Journal of Wind
Engineering and Industrial Aerodynamics, 39(1-3), 105–118.
Rooij R. P. J. O. M. & van Arens, E. A., (2007, August 28–31). Analysis of the experimental and computational Flow Characteristics with respect to the augmented lift
phenomenon caused by blade rotation. Paper presented at the Journal of Physics, Conference Series, The Science of Making Torque from Wind, Lyngby, Denmark.
Retrieved from http://www.iop.org/EJ/toc/1742-6596/75/1
Sant, T. (2007). Improving BEM based aerodynamic model in wind turbine design codes (PhD thesis). Technical University of Delft.
Sant, T., van Kuik, G., & van Bussel, G. J. W. (2006). Estimating the angle of attack from blade pressure measurements on the NREL phase VI rotor using a free wake
vortex model: Axial conditions. Journal of Wind Energy, 9, 549–577.
Schaffarczyk, A. P., Schwab, D., Ingwersen, S., & Breuer, M. (2012, October). Pressure and hot film measurements on a wind turbine blade operating in the atmo-
sphere. Paper presented at Science of Making Torque Conference, Oldenburg, Germany.
Schaffarczyk, A. P., Schwab, D., Ingwersen, S., & Breuer, M. (2017). Experimental detection of laminar-turbulent transition on a rotating wind turbine blade in the
free atmosphere. Journal of Wind Energy, 20, 211, 220. https://doi.org/10.1002/we.2001
Schepers, G., Boorsma, K., Gomez-Iradi, S., Schaffarczyk, P., Madsen, H. A., Sørensen, N. N., . . . , Schreck, S. (2014, December). Final report of IEA Wind Task 29:
Mexnext (Phase 2) (Report no. ECN-E-14-060). Energy Research Center of the Netherlands
Schepers, J. G. (2004, October). Annexlyse: Validation of yaw models on basis of detailed aerodynamic measurements on wind turbine blades (Report
No. ECN-C-04-097). Energy Research Center of the Netherlands.
Schepers, J. G. (2007a, December). IEA Wind Task XX: Comparison between calculations and measurements on a wind turbine in the NASA Ames wind tunnel (Report
No. ECN-E-07-066), Energy Research Center of the Netherlands.
Schepers, J. G. (2007b, December). IEA Wind Task XX: Comparison between calculations and measurements on a wind turbine in yaw in the NASA Ames wind tunnel
(Report No. ECN-E-07-072). Energy Research Center of the Netherlands.
Schepers, J. G. (2007c, December). IEA Wind Task XX: Dynamic inflow effects at fast pitching steps on a wind turbine placed in the NASA Ames wind tunnel (Report
No. ECN-E-07-085). Energy Research Center of the Netherlands.
Schepers, J. G. (1999, January). An engineering model for yawed conditions developed on basis of wind tunnel measurements. Paper presented at proceedings of ASME
Wind Energy Symposium, Reno, NV.
Schepers, J. G. (2012 November 27). Engineering models in aerodynamics. (PhD thesis). Technical University of Delft, Netherlands.
Schepers, J. G. (2013, November), Aerodynamic and acoustic international cooperation projects: How they (should) come together. Paper presented at IQPC Confer-
ence Noise and Vibration, Hamburg, Germany
Schepers, J. G. & Boorsma, K. (2014, September 17). The important role of aerodynamic measurements for the reliable and cost effective design of wind turbines.
Invited presentation at the 10th European Fluid Mechanics Conference.
Schepers, J. G., Boorsma, K., Cho, T., Gomez-Iradi, S., Schaffarczyk, P., Jeromin, A., …, Sorensen, N. (2012, February). Final report of IEA Task 29, Mexnext (Phase
1). Analysis of Mexico wind tunnel measurements (Report No. ECN-E-12-004). Energy Research Center of the Netherlands.
Schepers, J. G., Boorsma, K., Kim, C., & Cho, T. (2011). Results from Mexnext: Analysis of detailed aerodynamic measurements on a 4.5 m diameter rotor placed in
the large German Dutch Wind Tunnel DNW. Paper presented at EWEA Annual Event, Brussels, Belgium.
Schepers, J. G., Boorsma, K., & Snel, H. (2010). IEA task 29 Mexnext: Analysis of wind tunnel measurements from the EU project Mexico. The Science of Making Tor-
que from the Wind, Heraklion, Greece.
Schepers, J. G., Brand, A. J., Bruining, A., Graham, J. M. R., Hand, M. M., Infield, D. G., . . ., Simms, D. A. (1997, June). Final report of IEA Annex XIV field rotor
aerodynamics (Report No. ECN-C-97-027), Energy Research Center of the Netherlands.
Schepers, J. G., Brand, A. J., Bruining, A., Graham, J. M. R., Hand, M. M., Infield, D. G., . . ., Stefanatos, N. (2002, February). Final report of IEA Annex XVIII,
‘enhanced field rotor aerodynamics database (Report No. ECN-C-02-016). Energy Research Center of the Netherlands.
Schepers, J. G., Curvers, A. Oerlemans, S. Braun, K. Lutz, T. Herrig, A., . . ., T. Maeder (2007, September 20–21). SIROCCO: Silent rotors by acoustic optimisation.
Paper presented at the Second International Meeting on Wind Turbine Noise, Lyon, France.
Schepers, J. G., Feigl, L., Van Rooij, R., & Bruining, A. (2004). Analysis of detailed aerodynamic field measurements using results from an aeroelastic code. Journal of
Wind Energy, 7(4), 357–371.
Schepers, J. G., Heijdra, J. J., Foussekis, D., ye, S., Smith, R., Belessis, M., . . ., Drost, H. L. (2002). Verification of European wind turbine design codes, VEWTDC
(Final Report No. ECN-C-01-055). Energy Research Center of the Netherlands.
Schepers, J.G., Madsen, H.A., Soerensen, N., & Munduate, X. (2012, April).EERA aerodynamic test facility. Discussion document for EERA SP Meeting Aerodynam-
ics, DTU, Copenhagen, April 2012.
Schepers, J. G., Obdam, T., & Prospathopoulos, J. (2012). Analysis of wake measurements from the ECN wind turbine test site Wieringermeer, EWTW. Wind Energy,
15(4), 575–591.
Schepers, J. G., Pascal, L., & Snel, H. (2013). First results from Mexnext: Analysis of detailed aerodynamic measurements on a 4.5 m diameter rotor placed in the large
German Dutch Wind Tunnel DNW. Paper presented at the meeting of the European Wind Energy Conference, EWEC, Warsaw, Poland.
Schepers, J. G. & Snel, H. (1995). JOULE2: dynamic inflow: Yawed conditions and partial span pitch (Report No. ECN-C-95-056). Energy Research Centre of the
Netherlands.
Schepers, J. G. & Snel, H. (2007). Model experiments in controlled conditions, final report (ECN Report No. ECN-E-07-042). Energy Research Center of the Nether-
lands. Retrieved from http://www.ecn.nl/publicaties/default.aspx?nr=ECN-E–07-042
Schepers, J. G. & van Rooij, R. (2005). Final report of the Annexlyse project: Analysis of aerodynamic field measurements on wind turbines (Report
No. ECN-C-05-064). Energy Research Center of the Netherlands, Energy Research Center of the Netherlands.
Schepers, J. G. et al. (2018, February). Final report of the EU project AVATAR: Aerodynamic modelling of 10 MW turbines (Final report to EU, Contract No
FP7-ENERGY-2013-1/no. 608396). Retrieved from http://www.eera-avatar.eu/publications-results-and-links/
Schreck, S. (2008). IEA wind Annex XX: HAWT aerodynamics and models from wind tunnel measurements (Report No. TP-500-43508). National Renewable Energy
Laboratory, Golden, CO. Retrieved from http://www.nrel.gov/docs/gen/fy09/43508.pdf
Schreck, S., & Robinson, M. (2002). Rotational augmentation of horizontal axis wind turbine blade aerodynamic response. Wind Energy, 5(2/3), 133–150.
Schreck, S., Robinson, M., Hand, M., & Simms, D. (2000). HAWT dynamic stall response asymmetries under yawed flow conditions. Wind Energy, 3(4), 215–232.
Schreck, S., Robinson, M., Hand, M., & Simms, D. (2001). Blade dynamic stall vortex kinematics for a horizontal Axis wind turbine in yawed conditions. Journal of
Solar Energy Engineering, 123, 272–281.
Schreck, S., Sant, T., & Micallef, D., (2010, June). Rotational augmentation disparities in the Mexico and UAE Phase VI experiments. Paper presented at the Proceed-
ings of the 2010 Torque from Wind Conference, Heraklion, Greece.
Schreck, S., Sørensen, N., & Robinson, M. (2007). Aerodynamic structures and processes in rotationally augmented flow fields. Wind Energy, 10(2), 159–178.
Shen, W. Z. (2011, June). Actuator Line/Navier Stokes computations for flows past the yawed MEXICO rotor. Paper presented at Session on Mexico Experiments at
Shen, W. Z., Hansen, M. O. L., & Sørensen, J. N. (2009). Determination of the angle of attack on rotor blades. Journal of Wind Energy, 12, 91–98.
Shen, W. Z., Mikkelsen, R., & Sørensen, J. N. (2005). Tip loss corrections for wind turbine computations. Journal of Wind Energy, 8, 457–475.
Shen, W. Z., et al. (2010). Validation of the actuator line/Navier stokes technique using Mexico measurements. Paper presented at Science of Making Torque from the
Wind, Heraklion, Greece.
Shipley, D. E., Miller, M. S., Robinson, M. C., Luttges, M. W., & Simms, D. A. (1994). Evidence that aerodynamic effects, including dynamic stall, dictate HAWT
structural loads and power generation in highly transient time frames.
Shipley, D. E. et al. (1995, May). Techniques for the determination of local dynamic pressure and angle of attack for a HAWT (Report No. NREL-TP-44-7393).
National Renewable Energy Laboratory.
Simms, D. A., Hand, M. M., Fingersh, L. J., & Jager, D. W. (1999). Aerodynamics experiment phases II-IV test configurations and available data campaigns (Report
No. NREL/TP-500-25950). National Renewable Energy Laboratory.
Simms, D. A., Robinson, M. C., Hand, M.M., & Fingersh, L. J. (1996, January). Characterisation and comparison of baseline aerodynamic performance of optimally
twisted versus non-twisted HAWT blades. Paper presented at Proceedings of the 1996 ASME Wind Energy Symposium. Houston, TX.
Snel, H. (1997, June). The project ‘Dynamic stall and 3D effects’, a summary (in Dutch) (Report No. ECN-C-97-034). Energy Research Center of the Netherlands.
Snel, H. & Schepers, J. G. (1994). JOULE1: Joint investigation of dynamic inflow effects and implementation of an engineering method (Report No. ECN-C-94-107).
Energy Research Centre of the Netherlands.
Snel, H., Schepers, J. G., & Montgomerie, B. (2007). The MEXICO project (model experiments in controlled conditions): The database and first results of data proces-
sing and interpretation. Paper presented at the Science of Making Torque from the Wind. Lyngby, Denmark: Technical University of Denmark.
Snel, H., Schepers, J. G., & Siccama, A. (2009). Mexico, the database and results of data processing and analysis. Paper presented at the 47th AIAA Aerospace Sci-
ences, Orlando, FL.
Snel, H., et al. (1993, May). Sectional prediction of lift coefficients on rotating wind turbine blades in stall (Report No. ECN-C-93-052). Energy Research Centre of the
Netherlands.
Soerensen, N. N., et al. (2017, November) Engineering models for complex inflow situations, AVATAR Deliverable D2.8. Retrieved from http://www.eera-avatar.eu/
publications-results-and-links/
Sorensen, J. N., Shen, W. Z., & Mikkelsen, R. (2006). Wall correction model for wind tunnels with open test section. AIAA Journal, 44(8), 1890–1894.
Sorensen, N. (2011, June). Near wake predictions behind the MEXICO rotor in axial and yawed flow conditions. Paper presented at Session on Mexico Experiments at
Sørensen, N., & Schreck, S. (2012). Computation of the NREL phase-VI rotor in pitch motion during standstill. Wind Energy, 15(3), 425–442.
Sørensen, N. N., Zahle, F., Boorsma, K., & Schepers, G. (2016, October). CFD computations of the second round of MEXICO rotor measurements. Paper presented at
the meeting of the Science of Making Torque Conference, Munich, Germany.
Spath, M. (1993). Implementation of a pressure scanner system for a field rotor aerodynamics experiments (Report No. ECN-R-93-017). Energy Research Center of the
Netherlands.
Stoevesandt, B., et al. (2010). OpenFOAM: RANS-simulation of a wind turbine and verification. The Science of Making Torque from the Wind, Heraklion, Greece.
Strzelczyk, P. (1998). On simple vortex theory of horizontal axis wind turbines. Rzeszow, Poland: Faculty of Mechanical Engineering and Aviation, Rzeszow University
of Technology.
Strzelczyk, P. (2000). Determination of the horizontal axis wind turbine performance by use of simplified vortex theory (Transactions of the Institute of Aviation
No. 2/2000 (161)), pp. 24–33.
Szasz, R. (2011, June). LES of the near wake of the MEXICO wind turbine. Paper presented at Session on Mexico experiments at Wake Conference, Visby, Sweden:
Gotland University.
Troldborg, N., Bak, C., Madsen, H. A., & Skrzypinski, W. (2013, April). DANAERO MW II: Final report (DTU Wind Energy Report No. E-0027).
van Groenewoud, G. J. H., Boermans, L. M. M., & van Ingen, J. L. (1983). Investigation of laminar-turbulent transition of the boundary layer on the 25 m HAT wind
turbine. (in Dutch) (TUDelft Report No. LR-390). Faculty of Aerospace Engineering, Technical University of Delft.
van Grol, H. J., Snel, H., & Schepers, J. G. (1991). Wind turbine benchmark exercise on mechanical loads (A State of the Art Report. Volume 1, part A & B'.
ECN-C-91-031). Energy Research Center of the Netherlands.
van Ingen, J. L. (2008). The eN method for transition prediction. Paper presented at Historical Review of Work at TU Delft 38th Fluid Dynamics Conference and
Exhibit, Seattle, Washington, USA. AIAA 2008-3830.
Van Kuik, G., Micallef, D., Herraez, I., Van Zuijlen, A., & Ragni, D. (2014). The role of conservative forces in rotor aerodynamics. Journal of Fluid Mechanics, 750,
284–315. https://doi.org/10.1017/jfm.2014.256
van Kuik, G. A. M., Peinke, J., Nijssen, R., Lekou, D., Mann, J., Sørensen, J. N., … Skytte, K. (2016). Long-term research challenges in wind energy – A research
agenda by the European Academy of Wind Energy. Wind Energy Science, 1, 1–39. https://doi.org/10.5194/wes-1-1-2016
van Rooij, R. (2003). Determination of the angle-of-attack for the test field experiments with RFOIL. Paper presented at IEA Joint Action, Aerodynamics of Wind Tur-
bines 16th Symposium, Boulder, CO.
van Rooij, R., Bruining, A., & Schepers, J. G. (2003, June). Validation of some rotor stall models by analysis of the IEA Annex XVIII field data. Paper presented at Pro-
ceedings of EWEC European Wind Energy Conference. Madrid, Spain.
van Rooij, R. & Schepers, J. G. (2005, January). The effect of blade geometry on the normal force distribution of a rotating blade. Paper presented at the meeting of the
AIAA Conference, Reno NV.
van Rooij, R. P. J. O. M. (2001). Experiments on wind turbine airfoils and wind turbine rotors. Paper presented at International Wind Tunnel Symposium 2001. Tsu,
Japan.
van Rooij, R. P. J. O. M., Timmer, W. A., & Bruining, A. (2002). Determination of the local inflow angle on Rotating blades. Paper presented at the meeting of the 1st
World Wind Energy Conference and Exhibition, July, 2-6, Berlin, Germany.
Vermeer, L. J. (1998, June). Windtunnel experiments on a rotor model under yawed conditions (in Dutch) (Report No. TUD-IW-98136R). Institute for Wind Energy.
Vermeer, N. J., Sorensen, J. N., & Crespo, A. (2003). Wind turbine wake aerodynamics. Progress in Aerospace Sciences, 39, 467–510.
Vimalakanthan, K., Schepers, J. G., Shen, W. Z., Micallef, D., Rahimi, H., Jost, E., & Simao Ferreira, C. J. (2018, June). Evaluation of different methods of determining
the angle of attack on wind turbine blades under yawed inflow conditions. Abstract accepted for the Science of Making Torque Conference, June 2018, Milano,
Italy.
Voutsinas, S. et al. (2003, November). Mexico WP2: Navier Stokes and Euler simulations to determine the maximum rotor size (Report
No. MEXICO-DOC-WP2-01-01). National Technical University of Athens.
Xiao, J.-p., Wu, J., Chen, L., & Shu, Z.-y. (2011). Particle image velocimetry (PIV) measurements of tip vortex wake structure of wind turbine. Applied Mathematics
and Mechanics, 32(6), 729–738. https://doi.org/10.1007/s10483-011-1452-x
Yang, H., Shen, W. Z., Sørensen, J. N., & Zhu, W. J. (2010). Determination of the angle of attack on the mexico rotor using experimental data. The Science of Making
Torque from Wind, Heraklion, Greece.
Yu, W., Hong, V. W., Ferreira, C., & van Kuik, G. A. M. (2016). Validation of engineering dynamic inflow models by experimental and numerical approaches. Journal
of Physics: Conference Series, 753, 022024.
How to cite this article: Schepers JG, Schreck SJ. Aerodynamic measurements on wind turbines. WIREs Energy Envi-
ron. 2018;e320. https://doi.org/10.1002/wene.320

Aerodynamic measurements on wind turbines

Uploaded by

Copyright:

Available Formats

Aerodynamic measurements on wind turbines

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Aerodynamic measurements on wind turbines

Uploaded by

Copyright:

Available Formats

Received: 22 March 2017 Revised: 16 June 2018 Accepted: 18 June 2018

Aerodynamic measurements on wind turbines

wind turbine aerodynamic measurements

WIREs Energy Environ. 2018;e320. wires.wiley.com/energy © 2018 Wiley Periodicals, Inc. 1 of 25

• Energy Research Center of the Netherlands, ECN, The Netherlands

• National Renewable Energy Laboratory, NREL, USA

FIGURE 2 NREL IEA Task 14/18 facility with probes

FIGURE 3 NASA Ames wind tunnel

FIGURE 4 NREL Phase VI turbine in NASA Ames wind tunnel

FIGURE 5 DNW-LLF wind tunnel

FIGURE 7 PIV sheet at “9 o'clock position” in Mexico experiment

3.1 | Lifting line variables

FIGURE 9 Angle of attack in wind tunnel environment

3.2 | Dynamic pressure and nondimensionalization of aerodynamic coefficients in wind

4 | AE RODY NA MIC M E ASU RE MEN TS: RE SUL TS AN D P ROGRE S S

4.1 | Literature on aerodynamic wind turbine measurements

4.2 | Relevant research areas related to aerodynamic wind turbine measurements

4.2.1 | Aerodynamic measurements: Field versus wind tunnel measurements

4.2.2 | Validation versus physical understanding/model improvement

4.2.3 | Airfoil aerodynamics versus induction aerodynamics

4.3 | Aerodynamic experiments: Specific progress

Surface pressure (Nm–2)

Surface pressure (Nm–2)

Ft (N/m), U∞=24 m/s

0.5 1.0 1.5 2.0

4.3.2 | Aerodynamic understanding

1.0 30% span

Calculated/measured blade loads (–)

–0.3 –0.2 –0.1 0.0 0.1 0.2 0.3

Thrust coefficient (–)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

u (m/s), xm=+0.15 m, Az=240°, Yaw=30°, Vtun=15 m/s

Normal force coefficient (-)

RELATED WIREs ARTICLE

You might also like