Journal of Fluids and Structures 57 (2015) 206218

Contents lists available at ScienceDirect

Journal of Fluids and Structures

journal homepage: www.elsevier.com/locate/jfs

An experimental investigation on the characteristics

of fluidstructure interactions of a wind turbine model sited
in microburst-like winds
Yan Zhang, Partha P. Sarkar, Hui Hu n
Department of Aerospace Engineering, Iowa State University, Ames, IA 50011, United States

a r t i c l e in f o abstract

Article history: An experimental investigation is performed to assess the characteristics of the fluid
Received 6 December 2014 structure interactions and microburst-induced wind loads acting on a wind turbine model
Accepted 23 June 2015 sited in microburst-liked winds. The experiment study was conducted with a scaled wind
Available online 18 July 2015
turbine model placed in microburst-like winds generated by using an impinging-jet-typed
Keywords: microburst simulator. In addition to quantifying complex flow features of microburst-like
Wind turbine aeromechanics winds, the resultant wind loads acting on the turbine model were measured by using a
Microburst-like wind high-sensitive forcemoment sensor as the turbine model was mounted at different radial
Fluidstructure interaction locations and with different orientation angles with respect to the oncoming microburst-
like winds. The measurement results reveal clearly that, the microburst-induced wind
loads acting on the turbine model were distinctly different from those in a conventional
atmospheric boundary layer (ABL) wind. With the scales of the wind turbine model and
the microburst-like wind used in the present study, the dynamic wind loadings acting on
the turbine model were found to be significantly higher (i.e., up to 4 times higher for the
mean loads, and up to 10 times higher for the fluctuation amplitudes) than those with the
same turbine model sited in ABL winds. Both the mean values and fluctuation amplitudes
of the microburst-induced wind loads were found to vary significantly with the changes of
the mounted site of the turbine model, the operating status (i.e., with the turbine blades
stationary or freely rotating), and the orientation angle of the turbine model with respect
to the oncoming microburst-like wind. The dynamic wind load measurements were
correlated to the flow characteristics of the microburst-like winds to elucidate underlying
physics. The findings of the present study are helpful to gain further insight into the
potential damage caused by the violent microbursts to wind turbines to ensure safer and
more efficient operation of the wind turbines in thunderstorm-prone areas.
& 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Wind power market is growing rapidly in recent years in many countries around the world. With an average growth rate
about 30% during the past 10 years, the total installed wind energy capacity has reached 321 GW globally by the end of 2013
(Wiser and Bolinger, 2014). As both the total number and the size of wind turbines increase, the structural integrity and

n
Corresponding author.
E-mail address: huhui@iastate.edu (H. Hu).

http://dx.doi.org/10.1016/j.jfluidstructs.2015.06.016
0889-9746/& 2015 Elsevier Ltd. All rights reserved.
 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218 207

Fig. 1. Schematic of a microburst.

operational safety of wind turbines are receiving more and more attentions. According to the information provided by 2014
Caithness Windfarm Information Forum (2014) available at the website of http://www.caithnesswindfarms.co.uk/accidents.
pdf, there were about 1500 wind turbine accidents and incidents in the UK alone in the past 5 years. Among all the turbine
safety issues, structural failure, typically caused by extreme winds in thunderstorms, has contributed about one fifth of the
total accidents, and resulted in much more property losses than any other types of accidents.
Downburst, as a particular example of the extreme winds, could be a serious wind hazard to the structural safety of wind
turbines. Downbursts are quite common in many areas of the world. According to the 2011 Extreme Weather Sourcebook of
National Center for Atmospheric Research (NCAR), approximately 5% of thunderstorms would produce downbursts that is
one of the primary factors responsible for the estimated $1.4B of insured property loss each year in USA alone (data taken
from 1950 to 1997). Fujita (1985) first classified the downbursts into microbursts and macrobursts, based on the horizontal
extension of the divergent outburst flows. A microburst, as defined by Fujita (1985), is a strong downburst which produces
an intense outburst of damaging wind with the radial extent being less than 4.0 km, or else is defined as a macroburst.
Although a microburst has a smaller size than its counterpart, macroburst, it could produce a much stronger outflow
with the maximum wind speed up to 270 km/h.
The flow characteristics of a microburst are dramatically different from those of conventional atmospheric boundary
layer (ABL) winds and other wind hazards of wide concerns, e.g., tornadoes. As shown schematically in Fig. 1, a microburst
can produce an impinging-jet-like outflow profile diverging from its center with the maximum velocity occurring at an
altitude of less than 50 m above ground (Hjelmfelt, 1988). Such extreme high wind speed and wind shear (i.e., velocity
gradient) near the ground could produce a significantly greater damaging potential to built structures. Furthermore, a
microburst would also produce strong vertical velocity component in both the core region and the leading edge of the
outburst, which is very different from conventional ABL winds. Therefore, microburst-induced wind loading pattern is quite
different from what is usually expected with conventional ABL winds. Due to the extreme damaging potential of microbursts
to built structures, a number of experimental and numerical simulation studies have been conducted in recent years to
quantify the flow characteristics of the microburst-like winds and to assess the effects of the microburst-induced wind
loading on various built structures on the ground, such as transmission towers, grain bins, low-rise residential houses and
high-rise buildings (Savory et al., 2001; Chay and Letchford, 2002; Sarkar et al., 2006; Sengupta and Sarkar, 2008; Zhang
et al., 2013a,b, 2014a,b).
Generally, wind turbines are designed to operate in conventional ABL winds. The wake characteristics and resultant wind
loads acting on horizontal-axis wind turbines sited in conventional ABL winds have been studied extensively over the past
years (Vermeer et al., 2003; Cal et al., 2010; Chamorro and Porte-Agel, 2010, 2011; Lebron et al., 2012; Yang et al., 2012; Hu
et al., 2012; Zhang et al., 2012; Tian et al., 2014; Jeong et al., 2014). A number of numerical simulations have also been
conducted by coupling stochastic or CFD turbulence models with aeroelastic models (e.g., FAST Jonkman and Buhl, 2005) to
investigate the wind turbine loads subject to turbulent atmospheric boundary layer winds (Moriarty et al., 2004; Lee et al.,
2011). Although extreme situations, such as Extreme Coherent Gust with Direction Change (ECD) and Extreme Direction
Change (EDC), have already been considered in the IEC standards for wind turbine design (International Standard, 2005),
such standards are not applicable for non-conventional ABL wind conditions. While microburst-like winds have been shown
to generate significantly different fluidstructure interaction characteristics and wind loading effects on both low-rise and
high-rise structures (Savory et al., 2001; Chay and Letchford, 2002; Sengupta and Sarkar, 2008; Zhang et al., 2014a,b), only
few analytical studies can be found in literature to investigate microburst-induced wind loads acting on wind turbines. For
example, Nguyen et al. (2011) studied the wind loads acting on a wind turbine sited in a simulated translational microburst
wind by using an analytical model as suggested by Chay et al. (2006). They found that the simulated microburst would
impose 86% higher out-of-plane bending moment on turbine blades than a typical EDC load case defined in IEC standards
(International Standard, 2005), and 20% higher bending moment than that of an ECD load case when there was no yaw
control applied. Kwon et al. (2012) introduced a concept of gust loading factors into the analysis of the wind loads acting on
208 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218

wind turbines sited in gust-front winds, and suggested that the gust-front winds would induce approximately twice higher
of the static loads on the wind turbine tower system than those in conventional ABL winds defined by ASCE-7 standards.
However, these analytical studies were based on idealized mathematical models without taking the complicated flow
features and turbulence nature of microburst winds into account. As revealed in the Zhang et al. (2013a,b), microbursts are
actually very complex vortex flows with intense downdrafts and violent outburst winds. While the intense downdraft
dominates the flow feature in the core region of a microburst, the characteristics of the divergent outburst flow resembles
those of a wall jet well with the maximum velocity occurring very close to the ground, i.e., lower than the hub-height of a
large-sized wind turbine. Furthermore, the much higher turbulence intensity levels in violent microburst winds may result
in much more significant peak loads acting on the wind turbine, in comparison with those in conventional ABL winds, which
are extremely dangerous to the structural integrity of the wind turbine. Therefore, it is necessary and highly desirable to
improve our understanding about the characteristics of the fluidstructure interaction between wind turbines and violent
microbursts and the resultant wind loads acting on the wind turbines in order to improve our understanding about the
potential damage caused by violent microbursts to wind turbines.
In the present study, a comprehensive experimental investigation was conducted to assess the characteristics of the
fluidstructure interaction and the microburst-induced wind loads acting on wind turbines. The experimental study was
performed with a scaled horizontal axis wind turbine (HAWT) model placed in a microburst-like wind generated by using
an impinging-jet-typed microburst simulator available at Iowa State University. With the turbine model being mounted at
different radial locations away from the core center of the microburst-like wind and at different orientation angles with
respect to the oncoming outburst wind, the microburst-induced wind loads (i.e., both the aerodynamic forces and
corresponding moments) acting on the wind turbine model were measured quantitatively by using a high-sensitive load cell
mounted under the tower base of the turbine model. During the experiments, the rotor blades of the turbine model was set
in either free rotating or stationary to simulate the scenario of the wind turbine in either normal operation or shutdown
status during a microburst event. The present study simulates a wide range of the possible situations when a microburst in
nature occurs nearby a wind turbine. To the best knowledge of the authors, this is the first effort of its nature. It is hoped that
the findings derived from the present study would be helpful to improve our understanding about the potential damages
caused by violent microbursts to ensure safer and more robust operation of the wind turbines sited in thunderstorm-
prone areas.

2. Experimental setup and wind turbine model

While microbursts in nature are transient phenomena with a life time of about 10 min, a steady impinging-jet flow was
found to resemble the major features of a microburst at its maximum strength reasonably well (Hjelmfelt, 1988). Such
similarity has been proven by many researchers who successfully used impinging-jet-models to produce outburst flow
profiles to simulate microburst-like winds (Holmes and Oliver, 2000; Choi, 2004; Chay et al., 2006; Mason et al., 2005).
Moreover, the flow characteristics of a microburst wind represented by a steady impinging jet flow would also be the most
critical scenario to study the microburst-induced wind loading effects on built structures. It should be noted that, dynamic
similarity is one of the greatest challenges to conduct laboratory experiments to simulate meteorological phenomena such
as microbursts. It will be very difficult, if not impossible, to match the Reynolds numbers of the microbursts in nature with
those of the impinging jet flows generated in the laboratories due to the significant scale difference of the two compared
cases. However, it has been found that, although the dynamic similarity is difficult to match, the measurement results
obtained from laboratory experiments are still useful to reveal the flow characteristics of microburst-like winds and to
predict the winds loads acting on test models induced by microburst-like wind as long as the Reynolds number of the
laboratory experiments is high enough. Therefore, impinging-jet-typed microburst simulators have been widely adopted to
generate microburst-like winds in laboratories to produce outburst flow velocity profiles resembling those of microburst
winds (Wood et al., 2001; Chay et al., 2006; Das et al., 2010; Zhang et al., 2013a,b).
In the present study, an impinging-jet-based microburst simulator hosted in the Department of Aerospace Engineering of
Iowa State University (ISU) is used to generate microburst-like winds, as shown in Fig. 2. A schematic of this simulator with
detailed dimensions can be found in Zhang et al. (2013a,b). A downdraft flow is generated through an axial fan driven by a
step motor. The exhaust nozzle diameter (D) of the ISU microburst simulator is 610 mm (i.e., D610 mm). The distance
between the nozzle exit and the ground plane (H) is adjustable up to 2.3 times the nozzle diameter. Honeycomb and screen
structures are placed upstream of the nozzle exit in order to produce a uniform jet flow exhausted from the ISU microburst
simulator. During the experiments, a three-component cobra anemometer probe (Turbulent Flow Instrumentation Pvt.,
Ltd.), which is capable of simultaneously measuring all three components of the wind velocity vector, was used to quantify
the flow characteristics of the jet flow at the points of interest. It was found that the jet flow exhausted from ISU microburst
simulator was quite uniform across the nozzle exit, and the turbulence level of the core jet flow was found to be within 2.0%.
For the measurement results given in the present study, the ground floor was fixed at 2D below the ISU microburst
simulator (i.e., H/D2.0). The flow velocity at the nozzle exit of the ISU microburst simulator was set to 13.0 m/s (i.e.,
Ujet 13 m/s), which corresponds to a Reynolds number of 5.2  105 based on the nozzle diameter, D, of the ISU microburst
simulator. Further information about the design, construction, and performance of the ISU microburst simulator as well as
quantitative comparisons of the microburst-like winds generated by using the ISU microburst simulator with the
microbursts occurring in nature can be found in Zhang et al. (2013a,b).
 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218 209

Fig. 3 shows the schematic of the triple-blade horizontal axis wind turbine (HAWT) model used in the present study,
which has a hub height of 184 mm and rotor diameter of 254 mm. With the geometric scale of approximately 1:500, the
turbine model would represent a 2 MW utility scale wind turbine with a rotor diameter of about 130 m and hub-height of

Fig. 2. ISU microburst simulator used in the present study.

Fig. 3. Schematic of the wind turbine model and the coordinate system used in the present study. (a) Schematic of the wind turbine model (b) The
coordinate system and the orientation angle of the turbine model.
210 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218

Table 1
The primary design parameters of the wind turbine model used in the present study.

Parameter Rrotor Hhub drod dnaccele a a1 a2

Dimension (mm) 127 225 18 28 61 100 15 70

90 m in the field. As a result, the present study would simulate the scenario of a 2 MW utility scale wind turbine interact
with a microburst of 300 m in diameter in the field. The rotor blades of the turbine model are MA0530TE blades (Windsor
Propeller Inc.), which are twisted blades with the pitch angle ranging from 201 at the root to 101 at the tip of the blades. The
blades have a chord length of 12 mm at tip, 19 mm in the middle, and 16 mm at root. The airfoil cross-section of blades has a
concave pressure surface and is well adapted for low Reynolds number applications. Since the blades were originally
designed for propeller applications, they were mounted reversely with the pressure side of the blades facing the oncoming
airflow during the experiments to improve their aerodynamic performance when used as wind turbine rotor blades. The
primary design parameters of the wind turbine model are listed in Table 1. It should be noted that, the same turbine model
was used by Hu et al. (2012) to study the characteristics of the dynamic wind loads acting on wind turbine sited in a
conventional ABL wind.
In the present study, the microburst-induced wind loads acting on the turbine model were measured by using a high-
sensitive forcemoment sensor (JR3, model 30E12A-I40). The JR3 load cell is capable of measuring forces in three directions
and the moment (torque) about each axis. The measurement uncertainty of the JR3 load cell is 70.25% of the full range
(40 N). The wind loading measurements were performed with the turbine model mounted at different radial locations (r/
D 0.0 to 2.0, with a spacing of 0.5, where r is radial distance from the imping jet center) and with different orientation
angles (OA) with respect to the oncoming outburst flow direction (i.e., OA0 to 180 deg, with a spacing of 22.5 deg), as
shown in Fig. 3. For each test run, the wind load data were taken for 30 s with a sampling frequency of 1000 Hz. Both the
mean values and the standard deviations of the instantaneous wind loads were analyzed in the present study to quantify
the characteristics of the microburst-induced wind loads acting on the wind turbine model. Since the radial location of r/
D 1.0 is a critical location where maximum wind speed typically occurs in the microburst-like wind as reported in Zhang
et al. (2013a,b), the radial component of the flow velocity at the turbine hub height, Uhub, at the radial location of r/D 1.0
was used as the reference velocity (i.e., Uhub E 3.5 m/s) in the present study to normalize the measured force and moment
data for all the tested cases. It should also be noted that, when the turbine model was mounted at the radial location of r/
D 1.0 in the microburst-like wind, the tip-speed-ratio of the rotor model turbine was found to be Rrotor =U hub 4:1,
where is the angular speed of rotation measured using a laser tachometer. It is in the working range of a typical large-scale
wind turbine on a modern wind farm i.e., E4.08.0, as described in Burton et al. (2001).
As shown schematically in Fig. 3, CFr is defined as the coefficient of the radial component of the aerodynamic force acting
on the turbine model (i.e., it is also usually referred as the thrust coefficient for the case with a wind turbine sited in a
conventional ABL wind); CFt is the coefficient of the tangential component of the aerodynamic force with respect to the
direction of the outburst flow; CFz refers the coefficient of the vertical component of the aerodynamic force. CMr, CMt, CMz
are the corresponding moment coefficients about each axis with direction obeying the right-hand rule. The coefficients of
the Force and moment were calculated by using following equations:
F
CF i 1 2 i 2 1
2U hub Rrotor

Mi
CM i 1 2 2
2
2U hub Rrotor H hub

where F i and M i refer to the components of the aerodynamic force and moment in radial, tangential and vertical directions;
is density of the air; Rrotor is the radius of the turbine blades; Hhub is the hub height of the wind turbine model (i.e.,
Hhub 184 mm for the present study).
It should be noted that, while the flow velocity at the turbine hub height would change greatly as the turbine model was
mounted at different radial locations in the microburst-like wind, the force and moment coefficients were defined by using
the same reference velocity of Uhub (i.e., the velocity at turbine hub height with the turbine model mounted at the radial
location of r/D 1.0) in the present study. Therefore, the force and moment coefficients reported in the present study can be
compared directly for the measurement data obtained at different test conditions.
It is well known that wind turbines will cease power generation and even shut down at high wind speeds. The wind
speed at which shut down occurs is called the cut-out speed, which is usually at 25 m/s (i.e., Ucut-out 25 m/s) for modern
utility scale wind turbines. Having a cut-out speed is a safety feature which protects the wind turbines from damage. In
practice, wind turbine shut down may occur in one of several ways. While an automatic brake may be activated by a wind
speed sensor for some wind turbines, twist or pitch the rotor blades are also used to spill the wind for the shutdown of
wind turbines. Still others use spoilers, drag flaps mounted on the blades or the hubs, which are automatically activated by
high rotor rpms, or mechanically activated by spring loaded devices which turn the wind turbines sideways to the
oncoming wind stream.
 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218 211

Fig. 4. PIV measurement results of the microburst-like wind and the scale of the wind turbine model used in the present study.

It should be noted that the wind speed detected by the anemometers located at the top of the turbine nacelle is widely
used for wind turbine control to determine whether the turbine should be shut down or not. Since a severe microburst in
nature could induce a strong outflow with the maximum wind speed up to 270 km/h, i.e., 170 mph (Fujita, 1985), wind
turbines would be shut down during a severe microburst event (i.e., the turbine blades would be kept stationary during a
severe microburst event) if the wind speed detected by the anemometer probes at the turbine hub height is greater than the
cut-out speed of the wind turbines. Since a microburst would produce an impinging-jet-like outflow profile with the
maximum velocity occurring at an altitude of less than 50 m above ground (Hjelmfelt, 1988), it is also possible that the wind
speed detected by the anemometer probes located at the turbine hub height (e.g., 90 m for modern utility-scale wind
turbines) may still be smaller than the cut-out speed when the wind turbine is sited in the outburst region of the microburst
(e.g., at downstream location of r/D41.0 as shown in Fig. 4). As a result, the wind turbine would be in free rotation in the
microburst-like wind. With this in mind, two different cases, i.e., the rotor blades of the turbine model were set in either free
rotation or stationary, were considered in the present study, and the measurement data for the two cases were compared
quantitatively. For the case with the turbine blades freely rotating, the rotor blades of the wind turbine model would rotate
freely driven by the oncoming microburst-like wind. For the case with the turbine blades stationary, the rotor blades were
pitched to minimize the aerodynamic forces acting on the rotor blades and tapped firmly to ensure that the rotor blades
would not rotate during the experiments. For the case with the blades stationary, the three rotor blades of the wind turbine
model were fixed in such a way that one rotor blade is in horizontal, and the other two blades having phase angles of 601
and  601 related to the horizontal plane, respectively.

3. Measurement results and discussions

3.1. Flow characteristics of the microburst-like wind

As aforementioned, microbursts are very complex flows with intense downdrafts and violent outburst winds, which have
distinctly different flow features in comparison with conventional ABL winds. Fig. 4 gives the PIV measurement result
reported in Zhang et al. (2013b) to visualize the unique flow features of the microburst-like wind generated by using the ISU
microburst simulator. The cartoon of a wind turbine was also added into the figure in order to provide a visual comparison
of the scale of the turbine model with that of the microburst-like wind. As visualized clearly in Fig. 4, the intense downdraft
would cause airflow to stagnate at the center of the microburst-like wind as it impinges onto the ground, forming a dome of
high static pressure near the stagnation point as reported in Zhang et al. (2013b). After impingement, the streamlines of the
diverging airflow were found to be curved, and the airflow would transition from downdraft to outburst flow in the region
of 0.5 rr/Dr1.0. At the further downstream locations of r/DZ1.0, the radial component (Vr) of the flow velocity was found
to become dominant, and the vertical velocity component (Vz) becomes almost negligible. The profile of the radial flow
velocity in the outburst region resembles that of a wall jet flow, with the maximum velocity occurring at an elevation very
close to the ground (i.e., z/Dr0.15), i.e., at a height lower than the hub-height of the wind turbine model as shown clearly in
Fig. 4. Further detailed discussions about the flow characteristics of the microburst-liked wind based on PIV measurements
can be found in Zhang et al. (2013a,b).
The distribution of turbulence intensity was also found to vary significantly in the microburst-like wind, which has great
impacts on the dynamic wind loadings on the high-rise built structures as suggested by Zhang et al. (2014b). More
specifically, in the core region of the microburst-like wind, the turbulence intensity is very small (  2.0%). The turbulence
intensity was found to increase rapidly as the radial distance from the center increases and become much greater in the
transition region and outburst region in comparison with than those in the core center. It should also be noted that, as
reported in Hu et al. (2012), the turbulence intensity in a conventional ABL wind would decrease gradually with the
increasing elevation level away from the ground. However, the region with higher turbulence intensity in the microburst-
like wind was found to concentrate at a higher elevation (i.e., near the turbine hub height), which is quite far away from the
ground plane.
In summary, the measurement results described above reveal clearly that the microburst-like wind is a complex
turbulent flow with significantly different flow characteristics (i.e., different magnitude and distinct distribution pattern in
212 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218

Fig. 5. Force and moment coefficients vs. the radial location of the wind turbine model. (a) Force coefficients (b) Moment coefficients.

both the mean velocity and turbulence intensity) from those of a conventional ABL wind. As a result, the characteristics of
the fluidstructure interaction between wind turbines and microburst-like winds are expected to become much more
complicated, in comparison with those in conventional ABL winds.

3.2. Characteristics of the microburst-induced wind loads acting on the wind turbine model

3.2.1. Mean values of the microburst-induced wind loads acting on the turbine model
As described above, the dynamic wind loads acting on the turbine model were measured by using a JR3 load cell with a
sampling frequency of 1000 Hz for 30 s for each tested cases. Both the mean values and standard deviations of the
instantaneous wind loads were analyzed in the present study to reveal the characteristics of the microburst-induced wind
loads. Fig. 5 shows the mean (i.e., time-averaged) values of measured aerodynamic forces and moments acting on the wind
turbine model as the turbine model was mounted at the different radial locations in the microburst-like wind. For the
measurement data given in Fig. 5 the turbine blades were set either stationary or freely rotating with the orientation angle
of the turbine model being 0 deg (i.e., OA 0 deg) with respect to the oncoming microburst-like wind, i.e., the turbine rotor
disk is normal to the radial component of the microburst-like wind as shown in Fig. 3. It can be seen that, corresponding to
the complexity of the flow features of the microburst-like wind, the microburst-induced wind loadings acting on the wind
turbine model were found to vary significantly as a function of the mounted location of the turbine model. In general, the
radial-component of the aerodynamic force Fr and the bending moment Mt were found to be the most prominent among all
the force and moment components for all the tested cases, except for the case with the turbine model sited at the center of
the microburst-like wind. When turbine model sited at the center of the microburst-like wind (i.e., r/D0.0), corresponding
 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218 213

to the intense downdraft in the core region of the microburst-like wind, while the vertical component of the aerodynamic
force was found to be the most significant component as expected, all the other force components are negligibly small. As
the turbine model is moved away from the core center of the microburst-like wind (i.e., r/D increases), the radial-
component of the aerodynamic force (Fr) and the bending moment (Mt) were found to increase rapidly. Both Fr and Mt were
found to reach their maximum values as the turbine model was mounted at the radial location of r/D0.5, while the
maximum radial velocity was found to be reached at the radial location of r/D 1.0. This is because the depth of the outburst
flow at the radial location of r/D0.5 is much greater than those at the radial location of r/D 1.0 and other further
downstream locations. As shown in Fig. 4, the submerged area of the turbine model in the high-speed outburst flow
would be much larger when mounted at the radial location of r/D 0.5, thereby, the turbine model would experience much
larger resultant wind loads, in comparison with those at r/D1.0 and the further downstream locations. Due to the
symmetric nature of the turbine model at OA0 deg in respect to the oncoming microburst-like wind, the mean lateral
force, Ft, and corresponding radial moment, Mr, were found to be always very small, which are negligible.
It should also be noted the CFr values of the wind turbine model sited in the microburst-like wind (CFr E0.451.2 for the
radial location of r/DE0.52.0) was found to be much higher than the value reported in Hu et al. (2012) by using the same
turbine model sited in a conventional ABL wind (i.e., CFr E0.3 in a conventional ABL wind). It suggests that, with the same
oncoming flow velocity at the turbine hub height, the absolute values of the mean wind loads would become much greater
(i.e., up to 4 times greater) when the turbine is sited in a microburst-like wind, in comparison with those in a conventional
ABL wind. Furthermore, the turbine model was also found to experience a considerably large downward force Fz as it is
mounted near the core region of the microburst-like wind (i.e., at r/DE00.5). Such substantial downward force induced by
the intense downdraft of the microburst-like winds could be very hazardous to the turbine structure integrity since wind
turbines are designed by assuming only to operate in conventional ABL winds without considering the downward force
induced by the microburst-like wind.
As shown in Fig. 5, the microburst-induced wind loads experienced by the turbine model were also found to be quite
different when the turbine blades were set stationary, in comparison with the freely rotating case. As reported in Tian et al.
(2014), when sited in a conventional ABL wind, a wind turbine would experience much larger thrust and bending moment
for the case with its blades in rotation than those with the rotor blades stationary. However, interestingly, a very different
fashion was observed when the same turbine model was mounted in the microburst-like wind. As shown clearly in Fig. 5,
when the turbine model was mounted at r/D0.5 (i.e., near the boundary of the core region of the microburst-like wind),
the values of CFr and CMt for the case with the turbine blades stationary were found to be considerably larger than those
with the turbine blades freely rotating. When the turbine model was moved further downstream into the outburst region
(i.e., r/D E1.0), the values of CFr and CMt for the stationary case were still found to be slightly larger than those of the
rotating case. The differences between the two compared cases were found to decrease gradually as the distance away from
the center of the microburst-like wind increases.
The distinct differences in the characteristics of the microburst-induced wind loads acting on the turbine model are
believed to be closely related to the unique flow features of the microburst-like wind. When a wind turbine is sited in a
conventional ABL wind, the resultant wind loads acting on the wind turbine are expected to be much higher at the upper
half (i.e., the region above the turbine hub height) of the wind turbine due to the much higher wind speed at the higher
elevations away from the ground. However, since the microburst-like wind would have a wall-jet-liked velocity profile in
the outburst flow with the maximum wind velocity located at a height much lower than the hub-height of the wind turbine
model, only the lower half of the wind turbine would be submerged into the high-speed diverging airflow. As a result, the
wind loads acting on the wind turbine would mainly come from the lower half of the turbine rotation disk (i.e., the region
below the turbine hub height). As illustrated schematically in Fig. 6(a), for the case with the rotor blades freely rotating,
when the turbine blades rotate into the lower half of the rotation disk, they will be submerged into the high speed outburst
wind near the ground. A substantially large aerodynamic force would be generated to drive the blades to rotate
continuously, which results in a positive thrust force acting on the turbine blades. As the turbine blades rotate into the
upper half of the rotation disk, the turbine blades would rotate with the same rotational speed as they were at the lower half
of the rotation disk due to the inertial effects. However, since the velocity of the oncoming airflow would decrease
dramatically in vertical direction as shown in Fig. 4, the turbine blades would actually work as propeller blades to add
energy to the oncoming low-speed airflow, instead of extracting kinetic energy from the oncoming airflow, at the higher
elevations. As revealed clearly in the schematic of the velocity vector triangles given in Fig. 6(b), the resultant wind speed
may cause a negative thrust force acting on the turbine blades, which would counterbalance the positive thrust and bending
moment acting on the turbine blades as they were at the lower half of the rotation disk.

3.2.2. The variations of the microburst-induced wind loads as a function of the orientation angle
Since the direction of the oncoming airflow in a real microburst event in nature could be arbitrary in respect to the rotor
disks of the wind turbines, a parametric study was conducted in the present study to assess the variations of the microburst-
induced wind loads as a function of the orientation angle (OA) with respect to the oncoming microburst-like wind, and
some of the measurement results are given in Fig. 7. For the measurement data given in Fig. 7, the turbine model was
mounted at the radial location of r/D0.5, where the turbine model was found to experience the maximum radial
aerodynamic force and bending moment. It can be seen that, for the case with the turbine blades stationary, both values of
CFr and CMt would decrease as the orientation angle increases, reaching their minimum values at OA90 deg, and then
214 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218

Fig. 6. Schematic of the flow velocity vectors relative to the cross-section of the turbine blade.

increasing again as the orientation angle increases from OA 90 deg to OA180 deg. Such trend is found to be consistent
with the variation of the projected area of the turbine rotation disk along with the direction of the oncoming outburst flow.
It can also be seen that, while CFr and CMt were found to reach their maximum values at OA0 deg and OA180 deg, i.e.,
with the turbine rotation disk normal to the oncoming outburst wind, the tangential component of the aerodynamic force Ft
and its corresponding moment, Mr, were found to be greatest when the turbine model is mounted with an approximate
orientation angle of 112.5 deg with respect to the oncoming outburst wind (i.e., OAE112.5 deg). Since the turbine model
was exposed to the same downwash flow when mounted at the radial location of r/D0.5, the vertical component of the
aerodynamic force, Fz, and its corresponding moment, Mt, were found to be almost independent on the orientation angle.
For the case with the turbine blades freely rotating, since the rotor blades of the turbine model would not able to rotate
as the turbine model was set at a relatively high orientation angle (i.e., for the cases OA445 deg) with respect to the
oncoming outburst wind, the measurement data can be obtained only when the turbine model has a relatively small
orientation angle. As shown clearly in Fig. 7, the values of Fr and Mt for the case with the turbine blades freely rotating were
found to be consistently smaller than those with the turbine blades stationary at all the three tested orientation angles, i.e.,
OA 0.0, 22.5, and 45 deg. The differences between the two compared cases were found to become smaller and smaller as
the orientation angle increases. It should be noted that, a series of experiments were also conducted to investigate the
effects of the turbine orientation angle on the microburst-induced wind loads with the turbine model mounted at further
downstream locations (i.e., r/D Z1.0). Since the measurement results with the turbine model located at further downstream
locations reveal a very similar trend as those given in Figs. 6 and 7, the measurement data are not shown here for
conciseness.

3.2.3. The standard deviations of the dynamic wind loads acting on the wind turbine model
As revealed from the measurement results given above, microburst-like winds usually have much higher turbulence
intensity levels and more complicated distribution pattern, in comparison to conventional ABL winds. As a result, the
resultant wind loads acting on wind turbines are expected to fluctuate much more significantly when sited in microburst-
like winds. In the present study, the standard deviations of the instantaneous wind load acting the wind turbine model was
also analyzed in order to assess the fluctuation extent of the dynamic wind loads induced by the violent microburst-like
winds. Table 2 summarizes the standard deviations (i.e., root-mean-square values) of the dominant components of the
dynamic wind loads (i.e., CF r and CMt ) as the turbine model was mounted at different radial locations in the microburst-
like wind. It should be noted that, with the same wind turbine model sited in a conventional ABL wind, as reported by Hu
et al. (2012), the standard deviations of the coefficients of the thrust forces and the corresponding bending moment was
found to be about 0.11 and 0.15, respectively. However, as shown in Table 2, due to the more complicated distribution
pattern and much higher turbulence levels in the microburst-liked winds, the standard deviations of the dynamic wind
loads acting on the turbine model (i.e., CF r and CMt ) were found to be significantly higher (i.e.,  up to 10 times higher)
than those reported in Hu et al. (2012) with the same turbine model sited in a conventional ABL wind.
 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218 215

Fig. 7. Force and moment coefficients vs. the orientation angle of the turbine model at r/D 0.5.

It can also be seen that, both CF r and CMt values were found to vary substantially as the mounted location of the turbine
model changes in the microburst-like wind. For the case with the turbine blades stationary, both CF r and CMt were found to
reach their peak values when the model turbine was mounted at the radial location of r/D E1.0, where the maximum wind
speed would occur as shown in Fig. 4. However, for the case with the turbine blades freely rotating, the maximum values for
both CF r and CMt were found to be reached at the radial location of r/D 0.5, which is believed to be corresponding to the
highest rotation speed of the turbine blades at the radial location of at r/D0.5. Interestingly, while the turbine model was
found to experience considerably larger mean wind loads for the case with the turbine blades stationary as shown in Fig. 5,
the standard deviations of the dynamic wind loads acting on the turbine model were found to be always greater for the case
with the turbine blades freely rotating at all the tested radial locations. The larger fluctuation amplitudes of the dynamic
wind loads acting on the wind turbine for the case with the turbine blades freely rotating are believed to be closely related
to the shedding of the unsteady wake vortex structures from the rotating rotor blades, similar as those visualized
quantitatively in Yang et al. (2012). It should be noted that, the extreme high peak loads acting on wind turbines induced by
microburst-like winds are very hazardous to the structure integrity of the wind turbines, which can cause permanent
structure failures of wind turbines when they are exposed in violent microburst-like winds.
Based on Fast Fourier transform (FFT) analysis of the instantaneous wind load measurement data obtained at a sampling
rate of 1000 Hz, the power spectra of the dynamic wind loads acting on the turbine model can be determined. Fig. 8 gives an
example of the power spectra of the measurement data as the turbine model was mounted at the radial location of r/
D 0.50 with the turbine blades either freely rotating or stationary. Since very similar features were also seen in the power
216 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218

Table 2
The standard deviations of the dynamic wind loads acting on the wind turbine model sited in the microburst-like wind.

The location of the turbine model in the microburst- Turbine blades stationary Turbine blades freely rotating
like wind
Radial-component force, Bending moment Radial-component force, Bending moment
CF r CMt CF r CMt

r/D E0.0 0.76 0.58 0.80 0.67

r/D E0.5 0.61 0.58 1.51 1.57
r/D E1.0 0.78 0.64 1.36 1.29
r/D E1.5 0.69 0.47 0.84 0.77
r/D E2.0 0.63 0.50 0.98 0.95

Fig. 8. Power spectrum of the bending moment with the turbine model mounted at the radial location of r/D 0.5 and 0 deg orientation angle.
(a) Stationary, (b) free rotation.

spectra of the other components of the force and moment measurement data, only the power spectra of the bending
moment (i.e., Mt) were presented here for conciseness.
As shown in clearly in Fig. 8(a), a well-defined dominant peak (fvortex) can be identified in the power spectrum for the
case with the turbine blades stationary, which is believed to be closely related to the periodical shedding of the large-scale
primary vortices in the microburst-like wind. As described in Zhang et al. (2013a), the large-scale primary vortices are
generated due to KelvinHelmholtz instabilities at the strong shear layer between the high-speed outburst flow of the
microburst-like wind and the low-speed ambient flow. As shown in Fig. 8(b), in addition to the well-defined dominant peak
(fvortex) corresponding to the periodic shedding of the large-scale primary vortices in the microburst-like wind, a series of
obvious peaks can also be identified in the power spectrum for the case with the turbine blades freely rotating. The
corresponding frequencies of the series of the obvious peaks identified in the power spectrum were found to be correlated
very well with the rotation frequency of the turbine blades f0 and its harmonic frequencies, i.e., nn f0. It indicates that the
instantaneous wind loads acting on the turbine model would be influenced not only by the shedding of the large-scale
primary vortices in the microburst-like wind, but also by other factors such as wind shear, yaw, rotation-induced
unbalances, and the periodically-shedding of the tip and root vortices as visualized in Hu et al. (2012). Due to the
superposition of the effects of the large-scale primary vortices and the unsteady wake vortex structures shedding
periodically from the rotating turbine blades, the fluctuation amplitudes of the microburst-induced wind loads acting on
the turbine model would become much greater for the case with the turbine blades freely rotating, which is confirmed
quantitatively by the measurement data given in Table 2.
It should be noted that, while the test conditions and controlling parameters of the present study with a microburst
simulator are intrinsically different from those of the experimental study described in Hu et al. (2012) with a conventional
ABL wind tunnel, the values of the measured mean and fluctuating wind loads described above were obtained with the facts
that the same turbine model was used in the two studies and the coefficients of the wind loads are defined based on the
same reference velocity, i.e., the wind speed at the turbine hub height. However, since the microburst-induced wind loads
experienced by a wind turbine may vary significantly as a function of the scale ratio between the wind turbine and the
microburst, the mounted site of the turbine in relation to the microburst center, the operating status (i.e., with the turbine
blades stationary or freely rotating), the orientation angle of the turbine with respect to the oncoming microburst-like wind,
 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218 217

extra caution is needed to directly compare the wind loads acting on wind turbines in microburst winds with those in ABL
winds. While it is almost impossible to simulate all the possible scenarios experienced by wind turbines operating in
thunderstorm-prone areas, the findings of the present study are believed to very valuable to gain further insight into the
potential damage caused by the violent microbursts to wind turbines to ensure safer and more efficient operation of the
wind turbines.

4. Conclusions

An experimental study was conducted to assess the characteristics of the fluidstructure interactions and resultant
dynamic wind loads acting on a wind turbine model sited in a microburst-like wind. In addition to quantifying the unique
flow features of the microburst-like wind, the microburst-induced wind loads acting on the wind turbine model was
measured quantitatively as the turbine model was mounted at different radial locations and orientation angles with respect
to the oncoming microburst-like wind. The measurement results reveal clearly that, in comparison with those with the
same turbine model sited in a conventional atmospheric boundary layer (ABL) wind, the characteristics of the fluid
structure interactions between the wind turbine and the microburst-like winds become much more complicated, and the
resultant wind loads acting on the turbine model become significantly higher for the case with the turbine model sited in
the microburst-liked wind. Some major findings of the present study are summarized as follows:

1) Due to the violent and turbulent nature of the microburst-like winds, the microburst-induced wind loads acting on the
turbine model were found to be up to 4 times higher for the mean values of the wind loads, and up to 10 times greater in
the term of the fluctuation amplitudes (i.e., with the scales of the wind turbine model and the microburst-like winds
used in the present study), compared with those with the same turbine model sited in a conventional ABL wind.
2) The radial component of the aerodynamic force (Fr) and bending moment (Mt) were found to be the most dominant
components among all the microburst-induced wind loads acting on the wind turbine model. Both the radial force and
bending moment were found to reach their peak values as the turbine model was mounted at the radial location of r/
D0.5, where the microburst-like wind would transition from intense downdraft to violent outburst flow.
3) The intense downdraft near the core region of the microburst-like wind would result in significant downward wind loads
acting on the wind turbine model when sited near the core region of the microburst-like wind, which may pose a serious
hazard to the turbine structure integrity.
4) Unlike those in conventional ABL winds, the wind turbine model sited in microburst-liked wind was found to experience
larger mean radial aerodynamic force (Fr) and bending moment (Mt) for the case with the turbine blades stationary, in
comparison with the case with the turbine blades freely rotating. The difference between the two compared cases was
found to decrease as the orientation angle increases.
5) Interestingly, while the mean values of the microburst-induced wind loads acting on the wind turbine model were found
to be higher for the case with the turbine blades stationary, the fluctuation amplitudes of the dynamic wind loads acting
on the wind turbine model were found to be much higher for the case with the turbine blades freely rotating due to the
superposition of the effects of the large-scale primary vortices in the microburst-like wind and the unsteady wake vortex
structures shedding periodically from the rotating turbine blades.
6) The microburst-induced wind loads was found to be a function of the orientation angle with respect to the oncoming
microburst-like wind. The radial-component of the aerodynamic force (Fr) and bending moment (Mt) were found to
reach their peak values at OA 0 deg and OA180 deg, corresponding to the largest blockage area of the turbine rotation
disk along with the direction of the diverging microburst-like wind.

The findings derived from the present study are believed to be very helpful to gain further insight into the potential
damage caused by the violent microbursts to wind turbines to ensure safer and more efficient operation of the wind
turbines in thunderstorm-prone areas.

Acknowledgments

The project is funded by National Science Foundation (NSF) under award numbers CMMI-1000198 and CBET-1133751.

References

Burton, T., Jenkins, N., Sharpe, D., Bossanyi, E., 2001. Wind Energy Handbook. John Wiley & Sons Ltd., England.
Caithness Windfarm Information Forum, 2014. Summary of Wind Turbine Accident data to 30 June 2014. Available online at http://www.
caithnesswindfarms.co.uk/accidents.pdf.
Cal, R.B., Lebrn, J., Castillo, L., Kang, H.S., Meneveau, C., 2010. Experimental study of the horizontally averaged flow structure in a model wind-turbine array
boundary layer. J. Renewable Sustainable Energy 2, 013106.
Chamorro, L.P., Porte-Agel, F., 2010. Thermal stability and boundary-layer effects on wind turbine wakes: a wind tunnel study. Boundary Layer Meteorol.
136, 515533.
Chamorro, L.P., Porte-Agel, F., 2011. Turbulent flow inside and above a wind farm: a wind-tunnel study. Energies 4, 19161936.
218 Y. Zhang et al. / Journal of Fluids and Structures 57 (2015) 206218

Choi, E.C.C., 2004. Field measurement and experimental study of wind speed profile during thunderstorms. J. Wind Eng. 92, 275290.
Chay, M., Letchford, C., 2002. Pressure distributions on a cube in a simulated thunderstorm downburstPart A: Stationary downburst observations. J. Wind
Eng. Ind. Aerodyn. 90 (7), 711732.
Chay, M.T., Albermani, F., Wilson, R., 2006. Numerical and analytical simulation of downburst wind loads. Eng. Struct. 28, 240254.
Das, K.K., Ghosh, A.K., Sinhamahapatra, K.P., 2010. Investigation of the axisymmetric microburst flow field. J. Wind Eng. Ind. Aerodyn. 7, 115.
Extreme Weather Sourcebook of National Center for Atmospheric Research (NCAR) , URL: www.sip.ucar.edu/sourcebook/contact.jsp.
Fujita, T.T., 1985. The Downburst. SMRP Research Paper, University of Chicago.
Hjelmfelt, M.R., 1988. Structure and life cycle of microburst outflows observed in Colorado. J. Appl. Meteorol. 27 (8), 900927.
Holmes, J.D., Oliver, S.E., 2000. An empirical model of a downburst. Eng. Struct. 22, 11671172.
Hu, H., Yang, Z., Sarkar, P., 2012. Dynamic wind loads and wake characteristics of a wind turbine model in an atmospheric boundary layer wind. Exp. Fluids
52, 12771294.
International Standard, 2005. Wind TurbinesPart 1: Design Requirements (IEC 61400-1). International Electrotechnical Commission, 2005. www.iec.ch.
Jeong, M.S., Lee, S.W., Yoo, S.J., 2014. Wake impacts on aerodynamic and aeroelastic behaviors of a horizontal axis wind turbine blade for sheared and
turbulent flow conditions. J. Fluids Struct. http://dx.doi.org/10.1016/j.jfluidstructs.2014.06.016.
Jonkman, J.M., Buhl Jr, M.L., 2005. FAST Users Guide. National Renewable Energy Laboratory, Golden, CO.
Kwon, D.K., Kareem, A., Butler, K., 2012. Gust-front loading effects on wind turbine tower systems. J. Wind Eng. Ind. Aerodyn. 104106, 109115.
Lebron, J., Castillo, L., Meneveau, C., 2012. Experimental study of the kinetic energy budget in a wind turbine streamtube. J. Turbul. 13, N43.
Lee S., Churchfield M., Moriarty P., Jonkman J., Michalakes J. Atmospheric and Wake Turbulence Impacts on Wind Turbine Fatigue Loading. National
Renewable Energy Laboratory, Technical Report No. NREL/CP. , -500053567, 2011.
Mason, M.S., Letchford, C.W., James, D.L., 2005. Pulsed wall jet simulation of a stationary thunderstorm downburst, Part A: Physical structure and flow field
characterization. J. Wind Eng. Ind. Aerodyn. 93, 557580.
Moriarty P.J., Holley W.E., Butterfield S.P. Extrapolation of Extreme and Fatigue Loads Using Probabilistic Methods. National Renewable Energy Laboratory,
Technical Report No. NREL/TP. , -50034421, 2004.
Nguyen, H.H., Manuel, L., Veers, P.S., 2011. Wind turbine loads during simulated thunderstorm microbursts. J. Renewable Sustainable Energy 3 (5), 053104.
Sarkar, P.P., Haan, F.L., Balaramudu, V., Sengupta, A., 2006. Laboratory simulation of tornado and microburst to assess wind loads on buildings. Struct. Congr.
2006, 110.
Savory, E., Parke, G.A., Zeinoddini, M., Toy, N., Disney, P., 2001. Modelling of tornado and microburst-induced wind loading and failure of a lattice
transmission tower. Eng. Struct. 23, 365375.
Sengupta, A., Sarkar, P.P., 2008. Experimental measurement and numerical simulation of an impinging jet with application to thunderstorm microburst
winds. J. Wind Eng. Ind. Aerodyn. 96 (3), 345365.
Tian W., Ozbay A., Hu H., An Experimental Investigation on Dynamic Wind Loads Acting on a Wind Turbine Model in Atmospheric Boundary Layer Winds,
AIAA-20141221, 2014 AIAA Science and Technology Forum and Exposition (SciTech2014), , , National Harbor, Maryland, 1317 January, 2014.
Vermeer, L.J., Srensen, J.N., Crespo, A., 2003. Wind turbine wake aerodynamics. Prog. Aerosp. Sci. 39, 467510.
Wiser, R.H., Bolinger M., 2014. 2013 Wind Technologies Market Report. Report on Energy Efficiency and Renewable Energy, Department of Energy. .
Wood, G.S., Kwok, C.S., Motteram, N.A., Fletcher, D.F., 2001. Physical and numerical modelling of thunderstorm downbursts. J. Wind Eng. Ind. Aerodyn. 89,
535552.
Yang, Z., Sarkar, P., Hu, H., 2012. Visualization of the tip vortices in a wind turbine wake. J. Visual. 15, 3944.
Zhang, W., Markfort, C.D., Porte-Agel, F., 2012. Near-wake flow structure downwind of a wind turbine in a turbulent boundary layer. Exp. Fluids 52,
12191235.
Zhang, Y., Hu, H., Sarkar, P.P., 2013a. Modeling of microburst outflows using impinging jet and cooling source approaches and their comparison. Eng. Struct.
56, 779793.
Zhang, Y., Hu, H., Sarkar, P.P., 2014a. Comparison of microburst-wind loads on low-rise structures of various geometric shapes. J. Wind Eng. Ind. Aerodyn..
Zhang, Y., Sarkar, P.P., Hu, H., 2013b. An experimental study of flow fields and wind loads on gable-roof building models in microburst-like wind. Exp. Fluids
54 (5), 1511.
Zhang, Y., Sarkar, P.P., Hu, H., 2014b. An experimental study on wind loads acting on a high-rise building model induced by microburst-like winds. J. Fluids
Struct. http://dx.doi.org/10.1016/j.jfluidstructs.2014.07.010.