Journal of Fluids and Structures 50 (2014) 547564

Contents lists available at ScienceDirect

Journal of Fluids and Structures

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

An experimental study on wind loads acting on a high-rise

building model induced by microburst-like winds
Yan Zhang, Partha Sarkar, Hui Hu n
Department of Aerospace Engineering, Iowa State University, Ames, IA 50011-2271, USA

a r t i c l e i n f o abstract

Article history: In the present study, an experimental investigation is conducted to quantify the char-
Received 4 October 2013 acteristics of the microburst-induced wind loads (i.e., both static and dynamic wind loads)
Accepted 12 July 2014 acting on a high-rise building model, compared to those with the test model placed in
Available online 20 August 2014
conventional atmospheric boundary layer (ABL) winds. The experimental study is
Keywords: performed by using an impinging-jet-based microburst simulator available at Iowa State
Microburst University. In additional to conducting flow field measurements to quantify the flow
High-rise building characteristics of the microburst-like wind, both mean and dynamic wind loads acting on
Fluidstructure interactions the test model induced by the microburst-like wind are assessed in detail based on the
Microburst-induced wind loads
quantitative measurements of the surface pressure distributions around the test model
Laboratory simulation of microburst-like
and the resultant aerodynamic forces. It is found that the microburst-induced wind loads
wind
Impinging-jet model acting on high-rise buildings would be significantly different from their counterparts in
conventional ABL winds. Both the static and dynamic wind loads acting on the high-rise
building model were found to change significantly depending on the radial locations and
the orientation angles of the test model in respect to the oncoming microburst-like wind.
The dynamic wind loads acting on the test model were found to be mainly influenced by
the periodical shedding of the primary vortices and the high turbulence levels in the
microburst-like wind. The findings derived from the present study are believed to be
useful to gain further insight into the underlying physics of the flowstructure interac-
tions of high-rise buildings in violent microburst winds for a better understanding of the
damage potential of microburst winds to high-rise buildings.
& 2014 Elsevier Ltd. All rights reserved.

1. Introduction

High-rise buildings are commonly designed to resist extreme wind conditions with long lifespans. Although catastrophic
structural failures induced by strong winds are almost unlikely to occur, researches about wind loading effects on high-rise
buildings are still more than necessary from serviceability and economic points of view. Local wind damages, such as broken
glass and local component failures for high-rise buildings, could be induced by extreme winds due to the effects of either the
external pressure fluctuations or wind-borne debris. To better understand the mechanisms of wind-induced static and
dynamic loads and reduce the risk of damage, wind tunnel testing has been carried out by many researchers and proven to

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

http://dx.doi.org/10.1016/j.jfluidstructs.2014.07.010
0889-9746/& 2014 Elsevier Ltd. All rights reserved.
548 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

be an effective tool to investigate wind loads acting on high-rise buildings. For example, Melbourne (1980) compared the
measurement results of pressure distributions and force responses on the CAARC (i.e., Commonwealth Advisory
Aeronautical Research Council Coordinators) standard rectangular tall building model (Wardlaw and Moss, 1970) from six
establishments and obtained a good consistency among different research groups. Tanaka and Lawen (1986) studied the
same building model with a different scale (1:1000) and concluded that almost no deficit could be found due to the
exaggerated small length scale of the test model. Lin et al. (2005) conducted extensive experiments to study the local wind
loads on nine high-rise building models with different rectangular cross-sections and revealed different parametric effects
on wind loadings. Fruitful research accomplishments have also been achieved over the past years covering a wide variety of
interesting topics related to wind loadings effects on high-rise buildings, such as across-wind responses of tall buildings
(Kwok, 1982; Kawai, 1992; Marukawa et al., 1992; Kareem, 1992), mitigation of cross-wind response by aerodynamic
modifications (Kwok, 1988; Kim et al., 2008; Kim and Kanda, 2010), and interference effects on wind loads among multiple
high-rise buildings (Sykes, 1983; Kareem, 1987; Taniike, 1992; Lam et al., 2011). While most of those previous studies were
performed with atmospheric boundary layer (ABL) winds, a great amount of wind hazards, however, could be contributed
by other non-conventional ABL winds. Chen and Letchford (2004) compared the maximum dynamic magnification factor
(MDMF) of the CAARC building model induced by standard wind profile and conceptual generic downburst wind profiles.
Sengupta et al. (2008) performed a laboratory test to study the transient loads of a cubic building in a translating tornado
and microburst winds. Both tornado and microburst loads were found to exceed the design standard of ASCE 7-05.
Yang et al. (2011) studied the flowstructure interactions and the resultant wind loads acting on a high-rise building model
in a tornado-like wind. Nevertheless, the investigations on the wind loading effects on high-rise buildings induced by
extreme non-conventional ABL winds are still very scarce and the impact of these non-standard wind profiles on tall
buildings needs further research as suggested by Irwin (2009).
Extreme winds can be produced either by tropical cyclonic systems, such as typhoons or hurricanes, or by localized
severe weather conditions, such as thunderstorms. Compared with the tropical storms, local wind storms are usually more
devastating to the affected area and much more difficult to predict due to the small length scale and short lifespan.
Downburst is one kind of such local storms usually hidden within a thunderstorm, whose flow regime is analogous to
a reversed tornado. As a tornado causes a low-pressure core and sucks air inwards and upwards (Yang et al., 2011),
a downburst, originated from an intense downdraft of air, usually produces radial outburst winds due to the high pressure in
the core (Chay and Letchford, 2002; Zhang et al., 2013a, 2013b). More specifically, Fujita (1985) defined that a microburst is a
MISO-scale downburst which extends less than 4 km radially. It could cause a damaging outburst wind speed up to 270 km/
h (168 mph) based on the information available at NOAA website of http://www.erh.noaa.gov/cae/svrwx/downburst.htm.
Fig. 1 shows schematically the flow features of a microburst along with the relative scales of a low-rise building and a high-
rise building compared with the outburst profile of the microburst. Basically, the flow field of a microburst can be divided
into four regions according to different flow characteristics, i.e., downdraft region, stagnation region (r/D r0.5), transition
region (r/D E0.51.0), and outburst region (r/DZ1.0), where r denotes the radial distance from the microburst center and D
denotes the diameter of the downburst jet flow as shown in Fig. 1. A microburst wind usually has the following
characteristics: (1) high static pressure in the stagnation region. This is often referred to as the pressure nose, which is
opposite to the pressure drop in a tornado core. (2) Impinging-jet-like flow characteristics in the outburst region.
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 the ground (Hjelmfelt, 1988). The velocity profile of a microburst wind no
longer follows the log-law function of conventional ABL winds. (3) High turbulence levels and strong wind shear near the
ground in the outburst region. Due to the strong shear at the jetambient interface, turbulence levels in the outburst region
of a microburst could be much higher than those in ABL winds. As shown in Fig. 1, since the height of high-rise buildings is

Fig. 1. Schematic of a microburst.

 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564 549

normally greater than the depth of the outburst in a microburst wind, the microburst-induced wind loadings acting on high-
rise buildings would be affected greatly by all the unique wind characteristics described above. Therefore, while microburst
winds are often found to be more catastrophic for low-rise buildings due to the extreme near-ground speed and wind shear
(Hjelmfelt, 1988; Zhang et al., 2013a, 2013b), microburst-induced wind loads acting on high-rise buildings could also
become much more complicated, compared with those in conventional ABL winds. It should be noted that, based on the
report of 2001 Extreme Weather Sourcebook of National Center for Atmospheric Research, approximately 5% of thunder-
storms would produce downbursts. Walter (2007) showed that average potential microburst days per month for July and
August during the period of 19972003 are around 1520 for the Great Plains and the 50-year return period microburst-
wind gusts were generally over 137 km/h (85 mph) in the vast Midwest region of the United States. While the probability for
a microburst occurring near a specific high-rise building is difficult to estimate due to the lack of climatological data, the
potential threat of microburst winds to high-rise buildings could not be ignored. A better understanding about the
characteristics of microburst-like winds and the microburst-induced wind loading acting on high-rise buildings is highly
desirable.
In the present study, an experiment investigation is carried out to quantify the microburst-induced wind loads (both the
static and dynamic loads) acting on a high-rise building model based on the quantitative measurements of the surface
pressure distributions and resultant aerodynamic forces acting on the high-rise building model. The microburst-induced
wind loads are compared with those published in the literature with conventional ABL winds in order to reveal the
uniqueness characteristics of the microburst-induced wind loads which have not been covered by the existing building
design standards. The objective of the present study is to gain further insight into the underlying physics of the flow
structure interactions of high-rise buildings in violent microburst winds for a better understanding of the damage potential
of the microburst winds to high-rise buildings.

2. Experimental setup and test model

While microbursts in nature are transient phenomena with a lifetime 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). This
similarity has been proven by many researchers who successfully used impinging-jet-model to produce outburst flow
profiles to simulate microburst 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 buildings. Therefore, impinging-jet-model has
been widely adopted to simulate microburst-like wind in laboratory experiments 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, 2013b).
In the present study, an impinging-jet-based microburst simulator located in the Department of Aerospace Engineering at
Iowa State University (ISU) is used to generate microburst-like wind. Fig. 2 shows the schematic and a photo depicting the
flow circuit and dimensions of an ISU microburst simulator. A downdraft flow is generated through an axial fan driven by a
step motor. The exhaust nozzle diameter of the ISU microburst simulator is 610 mm (i.e., D 610 mm). A geometric scale of
approximately 1:650 was reached if comparing this diameter to a small-size microburst in the nature with 400 m diameter
(Hjelmfelt, 1988). The distance between the nozzle exit and the ground plane (Hjet) was set to be 2 times of the jet diameter
(Hjet/D 2.0), which falls into a reasonable range of microburst in nature (Hjet/DE0.757.5). Honeycomb and screen
structures were placed upstream of the nozzle exit in order to produce a uniform jet flow exhausted from the microburst
simulator.
During the experiments, a multi-hole cobra-probe (Turbulent Flow Instrumentation Pvt. Ltd.s), which is capable of
simultaneously measuring all three components of the flow velocity vectors, was used to quantify the flow characteristics of
the impinging 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 about 2.0%. The flow
velocity at the nozzle exit of the ISU microburst simulator was set to 13.0 m/s (i.e., Ujet 13.0 m/s), which corresponds to a
Reynolds number of 5.2  105 based on the nozzle diameter, D, of the ISU microburst simulator. 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 microbursts
in nature with those of the impinging jet flows generated in the laboratories due to the significant scale difference of the
two cases. It has been proved that, although the Reynolds numbers of the laboratory experiments may not be able to match
those of microbursts in nature, the measurement results obtained from laboratory experiments are still very useful to reveal
the flow characteristics of microburst-like winds and to predict the wind loads acting on test models induced by microburst-
like winds as long as the Reynolds number of the laboratory experiments is high enough (i.e., Re 4105).
Fig. 3 shows the quantitative comparisons of the measured outflow velocity profile of the microburst-like wind
generated by ISU microburst simulator versus the NIMROD (Fujita, 1981) and JAWS (Hjelmfelt, 1988) field measurement data
of real microbursts occurring in nature along with the published data of previous studies (Wood et al., 2001; Kim and
Hangan, 2007; Sengupta and Sarkar, 2008). The outflow velocity profiles given in Fig. 3 were taken in the vicinity of the
radial location where the maximum wind speeds in the microburst-like winds occur. As shown in Fig. 3, while the radial
velocity of the microburst-like wind was normalized by the maximum radial velocity Vr,max (i.e., Vr/Vr,max), the height where
half of the maximum radial velocity occurred (i.e., bE 170 mm for the present study) was used to normalize the vertical
550 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

Fig. 2. ISU microburst simulator and the high-rise building model used in the present study: (a) schematic of ISU microburst simulator and (b) photo of ISU
microburst simulator and a high-rise building model.

Fig. 3. The measured outflow velocity profile of the microburst-like wind generated by ISU microburst simulator compared to the field measurement data
of microbursts occurring in nature and the published results of previous studies.

height in the microburst-like wind (i.e., z/b). As suggested in previous studies, while the detailed flow features of each
microburst may vary from case to case, all the microbursts were found to have a similar trend in terms of normalized
outflow velocity profiles. It can also be seen clearly that, even though the simulated microburst-like winds by using ISU
microburst simulator and the real microbursts occurring in nature are significantly different in their size (e.g., the one
generated by using ISU microburst simulator with a diameter of 0.6 m vs. approximately 4004000 m for the microbursts in
 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564 551

Fig. 4. The high-rise building model in the microburst-like wind.

nature), the unique features of the outburst flows in microburst-like winds are captured reasonably well using the
impinging-jet-based ISU microburst simulator. Further information about the design, construction, and performance of ISU
microburst simulator as well as the quantitative comparisons of the microburst-like wind generated using ISU microburst
simulator with the microbursts occurring in nature can be found at Zhang (2013).
As shown in Fig. 2(b), the high-rise building model used in the present study has a 45 mm  45 mm (B  B) square cross-
section and a total height of 180 mm (H 4B). Based on the 1:650 scale ratio of the simulated microburst-like wind for the
present study, the test model would represent a high-rise building with a 29 m  29 m square cross section and a 117 m
total height. Fig. 4 shows the relative scale of the high-rise building model used in the present study mounted in the
microburst-like wind generated using the ISU microburst simulator. As shown in Fig. 4, the height of the high-rise building
model fits well within the extent of the outburst flow of the microburst-like wind generated using the ISU microburst
simulator.
To measure the surface pressure distributions around the high-rise building model, 30 pressure taps were distributed in
a 3  10 grid-pattern on each of the four sidewalls of the test model, and 9 pressure taps were distributed in a 3  3 grid-
pattern on the roof of the test model. The pressure taps were connected to four units of pressure scanners (DSA3217, Digital
Sensor Array, Scanivalve Corp.s, 16 channel per unit) for surface pressure measurements. The four units of pressure
scanners were connected in parallel for the measurements. The pressure scanners incorporate temperature-compensated
piezoresistive pressure sensors with a pneumatic calibration valve, RAM, 16 bit A/D converter, and a microprocessor in a
compact self-contained module. The precision of the pressure acquisition system is 70.2% of the full scale (710 in. H2O).
During the experiments, the instantaneous surface pressure measurement data were acquired for 100 s with the data
acquisition rate of 100 Hz for each test case. It should be noted that since the Tygon tubing used in the present study to
connect the pressure taps and pressure scanners is quite short (i.e., 1.5 mm in diameter and 0.3 m in length), the amplitude
attenuation and phase lag of the surface pressure signals caused by the short tubing were expected to be negligible at the
data acquisition rate of 100 Hz (Irwin et al., 1979).
In the present study, in addition to integrating the surface pressure distributions over the test model to determine the
resultant aerodynamic forces induced by the microburst-like wind, the resultant wind loads acting on the test model were
also measured directly using a high-sensitivity force-moment sensor (i.e., JR3 load cell, model 30E12A-I40). The JR3 load cell
is capable of measuring forces in three directions and the moment (torque) about each axis. For each test run, 30,000 data
points were taken with a sampling frequency of 1000 Hz. The measurement uncertainty of the JR3 load cell is 70.25% of the
full range (40 N). During the experiments, the wind loads acting on the test model were measured while the high-rise
building model was mounted in the microburst-like wind at five typical radial locations (i.e., r/DE0.0, 0.5, 1.0, 1.5 and 2.0)
with three different orientation angles (i.e., OAE0.01, 22.51 and 451) in respect to the oncoming microburst-like wind. In the
present study, the orientation angle of OA E0.01 refers to the position with the wall of the high-rise building model being
normal to the radial component of the microburst-like wind.

3. Results and discussions

3.1. The flow characteristics of the microburst-like wind

The flow characteristics of the microburst-like winds generated by the ISU microburst simulator were quantified using a
high-resolution PIV system and a cobra-probe at the points of interest before the high-rise building model was mounted on
the ground plane. While detailed PIV measurements and associated discussions to reveal the flow features of the
microburst-like winds were well documented in Zhang et al. (2013a, 2013b), Fig. 4 shows an example of the PIV
measurement results along with the high-rise building model placed in the microburst-like wind. The color contours in the
q
plot represent the magnitude of the time-averaged velocity flow velocity of the microburst-like wind (i:e:; jV j V 2r V 2z ,
where V r is the radial velocity component, and V z is the axial/vertical velocity component) normalized by the jet flow
velocity V jet at the exit of the ISU microburst simulator. As shown clearly in Fig. 4, the wind speed in the microburst-like
wind was found to vary greatly, depending on the radial location in respect to the center of the microburst. Corresponding to
the impingement of the downward jet flow, the wind speed was found to be quite small in the core region around the
stagnation point, as expected. Outside the impinging core, the airflow was found to be accelerated while diverging from the
552 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

center of the microburst-like wind, and reach its maximum speed in the region around r/D E1.0. Then, the diverging flow
was found to slow down gradually further downstream in the outburst region of the microburst-like wind. The wind speed
in the outburst region of the microburst-like wind was also found to vary significantly along the vertical direction. As shown
clearly in Fig. 4, the wind speed was found to be significant in the region very near the ground plate with the maximum
wind speed at the elevation height of z/DE0.12 (i.e., 50 m above the ground with the scale ratio of 1:650 of the present
study). As the elevation height keeps on increasing, the wind speed in the outburst flow was found to decrease gradually
(at r/DZ1.0), instead of increasing continuously as those in conventional ABL winds (Lin et al., 2005; Kim and Kanda, 2010).
As a result, the wind speed at the roof height of the high-rise building model (i.e., the white dashed line in the plot at
z/DE0.3) was found to become very small. It suggests that, unlike low-rise buildings with their heights typically much
lower than the elevation height where the maximum wind speed would reach in the outburst flows of microbursts, high-
rise buildings would cover a much wider range vertically in microburst winds; hence, they would experience much more
complicated wind loading effects induced by the unique flow features of the microburst winds.
Fig. 5 shows the measured velocity and turbulence intensity profiles in the microburst-like wind at five typical radial
locations (i.e., r/D E0.0, 0.5, 1.0, 1.5, and 2.0). The measured data given in the plots were obtained using the multi-hole
cobra-probe to measure all three components of the velocity vectors at the points of interest. It can be seen clearly that, near
the core center of the microburst-like wind (i.e., r/D E0.0), the flow velocity of the impinging jet flow, V z , was found to
decrease gradually from V z =V jet 1:0 around z/H E2.0 (zE0.6Hjet), i.e., 2.0 times the height of the test model, as it
approaches the ground plane. It should be noted that such flow deceleration towards the ground plane would result in the
formation of a high-pressure dome around the stagnation point as reported in Zhang et al. (2013a, 2013b). As a result, the
high-rise building model would be completely immersed in the high-pressure dome when it was mounted near the core
center of the microburst-like wind. The turbulence level at the center was found to be around 2%, which is much less than
that in the outburst region of the microburst-like wind. At the radial location of r/D E0.5 (i.e., in the transition region of the
microburst-like wind), the radial velocity, Vr, was found to increase gradually as the elevation height away from the ground
plate increases. It should also be noted that the downdraft velocity component (i.e., the negative value of Vz) is still
significant at this point since the airflow was in transition from the downdraft to the radial outburst flow in the transition
region. The turbulence intensity at this location was found to be greater than that at the core center of the microburst-like
wind, but still much smaller compared with that at further downstream locations in the outburst flow. At further
downstream locations of r/D E1.0, 1.5, and 2.0 (i.e., in the outburst region of the microburst wind), the profiles of the radial
velocity, Vr, were found to be very similar to those of a wall jet flow with significant radial velocity concentrated in the
region very near the ground plate. As the distance away from the center of the microburst-like wind increases, the
magnitude of the maximum wind speed in the outburst region was found to decrease gradually. Meanwhile, the vertical
velocity component, Vz, becomes almost negligible compared with that at the microburst center (i.e., r/D E0.0) and in the
transition region (i.e., at the radial location of r/D E0.5). Turbulence levels in the outburst region were found to become
much higher (410%) compared with those in the core center (i.e., r/D E0.0) and the transition region (r/D E0.5) of the
microburst-like wind. Corresponding to the moving paths of the primary vortex structures of the impinging jet flow as
shown in Zhang et al. (2013), the turbulence intensity in the microburst-like wind was found to increase rapidly as the
vertical distance increases from z/H 0.5 to the roof height of the high-rise building model. Such substantial variations of the
turbulence levels over the height of the high-rise building model would add more complexity to the dynamic wind loads
acting on the test model induced by the microburst-like wind.

Fig. 5. Mean velocity and turbulence intensity profiles in the microburst-like wind at five typical radial locations.
 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564 553

Fig. 6. Mean static pressure coefficients on the ground surface induced by microburst-like wind.

It should also be noted that a microburst can also induce significant changes for the surface pressure on the ground. Fig. 6
shows the measured static pressure distribution on the ground induced by the microburst-like wind. As shown clearly in
Fig. 6, a high static pressure region was formed due to the direct impingement of the jet flow near the center of the
microburst-like wind. The size of the positive pressure region was found to be much greater than the diameter of the
impinging core jet flow (i.e., r/D o0.5), which almost reached the radial location of r/D E1.0. The Gaussian-like distribution
pattern of the surface pressure in the microburst-like wind was found to be similar to those reported in the previous studies
of Tu and Wood (1996) and Sengupta and Sarkar (2008). Due to the significant changes of the surface pressure on the
ground induced by the microburst-like wind, the radial location of a building in relation to the center of the microburst-like
wind would be critical in determining the microburst-induced wind loadings acting on the building.

3.2. Time-averaged surface pressure distributions around the high-rise building model in the microburst-like wind

As aforementioned, since the flow features of a microburst-like wind at its maximum strength are resembled well by an
impinging jet flow, the quantification of the microburst-induced wind loads acting on the high-rise building model using
ISU microburst simulator would be akin to evaluating the worst-case-scenario for high-rise buildings in microburst winds.
Fig. 7 shows the time-averaged surface pressure distribution around the test model in terms of surface pressure coefficient,
C p P P atm =0:5V 2jet , with the test model mounted in the microburst-like wind at five typical radial locations.
Apparently, the surface pressure distributions around the high-rise building model were found to change remarkably as
the location of the test model in respect to the center of the microburst-like wind changes. As shown in Fig. 7(a), when the
test model was located at the center of the microburst-like wind, the strong downdraft jet flow would impinge directly onto
the roof of the test model with the surface pressure coefficient on the roof of the test model reaching approximately 1.0.
Corresponding to the high static pressure in the stagnation region of the microburst-like wind, the surface pressure
coefficients on the four sidewalls of the test model were found to be quite high, i.e., Cp 40.75. It indicates that, when the
building is located at the center region of a microburst, since the entire building would be wrapped by high surface pressure,
the pressure differences between the inside and outside glass walls may pose some potential safety problems for a well-
sealed high-rise building.
As shown in Fig. 7(b), when the test model was moved into the transition region at r/DE0.5, a region with very high
positive pressure was found to exist on the windward wall of the test model, which could be attributed to both the local
high static pressure and the dynamic head of the accelerating diverging airflow. While flow separations would result in
lower surface pressure regions at sidewall corners and the upper half of the leeward wall, most of the surfaces on the
leeward wall and sidewalls were still found to have quite high surface pressure. It indicates that, when a high-rise building is
located in the transition region of a microburst, both the direct impingement of the diverging airflow and the high static
pressure in the core region would contribute to the extreme pressure distributions on the surfaces of the high-rise building.
This is exactly the opposite situation of that in a tornado-like wind, in which the significant pressure drop in the core region
of a tornado is the major concern.
When the test model was moved further away from the microburst center into the outburst region, i.e. at r/DE1.0, 1.5,
and 2.0, the effects of the local static pressure would become small, and the surface pressure distributions on the test model
would be mainly determined by the characteristics of the local wind speed. Since the high speed flow in the outburst region
of the microburst-like wind was found to concentrate near the ground plane, the regions with relatively high surface
pressure values were found to be in the lower half of the windward wall of the test model, i.e. z/Ho0.5, as shown clearly
in Fig. 7(c)(e) The surface pressure on the windward wall of the test model was found to decrease gradually towards the
roof-height, which has an upside-down pattern compared with that observed in a conventional ABL winds (e.g.
Melbourne, 1980; Kim and Kanda, 2010). Corresponding to the decay of the radial flow velocity in the outburst region of
554 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

Fig. 7. The time-averaged surface pressure distributions around the test model in the microburst-like wind at five typical radical locations: (a) r/D E0.0,
(b) r/D E0.50, (c) r/DE 1.0, (d) r/D E1.5 and (e) r/D E2.0.

the microburst-like wind, the peak value of the surface pressure on the test model was found to decrease gradually as the
model was moved further away from the microburst center, i.e., moved from the radial location of r/D E1.0 to 2.0.
It can also be seen that the regions with significant negative pressure values were found on both sidewalls and leeward
wall of the test model due to the flow separations over the wall surfaces. As shown in Fig. 7(c), when the test model was
mounted in the outburst region at r/DE1.0, the peak value of the negative surface pressure on the sidewalls of the test model
was found to reach at much higher elevation levels (i.e., z/HE0.70.8) compared with that of the maximum positive surface
pressure on the windward wall (i.e., z/Ho0.3). The peak values of the negative surface pressure were found to exist near the
leading edge of the two sidewalls, and recover quickly at the upper part of the sidewalls. It indicates that the separated flow
would reattach at the rear portions of the sidewalls to form separation bubble at higher elevation levels of the test model. On
the contrary, the negative surface pressure at the lower elevation levels of the sidewalls was found to have a much smaller
absolute value but more uniformly distributed, suggesting that the separated flows would not reattach to the sidewalls at the
lower half of the test model. Such pattern in the surface pressure distributions on the sidewalls is believed to be closely related
to the flow characteristics of the microburst-like wind at the upper and lower elevation levels of the high-rise building model,
as shown in Fig. 4. While the relatively low wind speed and high turbulence level in the outburst region would be helpful for
the separated flows to reattach onto the sidewalls at the higher elevation levels of the test model, the high wind speed and low
turbulence at lower elevation levels would make the airflow separate from the sidewalls completely. Since the wind speed and
the turbulence intensity level become more uniform vertically at the further downstream locations of r/DE1.5 and 2.0, the
elevation heights of the region with peak negative surface pressure values on the sidewalls were found to decrease gradually
to approach the elevation level where the maximum positive surface pressure was found on the windward wall of the
test model.

3.3. The mean wind loads acting on the high-rise building model induced by the microburst-like wind

Following the work of Lin et al. (2005) and Kim and Kanda (2010), the coefficients of the time-averaged local
aerodynamic forces (i.e., both local along-wind and cross-wind force coefficients) of the test model induced by the
microburst-like wind were calculated in the present study based on the pressure differences between the windward and
 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564 555

leeward walls and the two sidewalls of the test model at different elevation levels. The local aerodynamic force coefficients
of the test model in the microburst-like wind at five typical radial locations (i.e., r/DE 0.0, 0.5, 1.0, 1.5, and 2.0) are compared
with those reported in Lin et al. (2005) and Kim and Kanda (2010) with conventional ABL winds. It should be noted that,
since different reference velocities were used to define the non-dimensional parameters (i.e., the pressure coefficient, CP,
and force coefficients, CF) in microburst-like winds (i.e., the impinging jet velocity, V jet , for the present study) and in the
conventional ABL winds (i.e., the longitudinal velocity at building roof height, U H , for conventional ABL winds as used by
Lin et al. (2005) and Kim and Kanda (2010)), the pressure and force coefficients given in the present study would only be
used qualitatively to reveal the significant differences in the distribution patterns of the microburst-induced wind loads in
comparison with those in conventional ABL winds.
Fig. 8 shows the profiles of the local along-wind force coefficients (i.e., drag coefficients, CFr) of the test building model
when the test model was mounted at different radial locations with 0 orientation angle in respect to the oncoming
microburst-like wind (i.e., the wall of the test model is normal to the radial velocity component of the microburst-like
wind). As shown clearly in Fig. 8(a), when the model was mounted at the core center of the microburst-like wind (i.e., at
r/D E0.0), the local along-wind force coefficients, CFr, were found to be very small, which are almost negligible. When the
test model was mounted in the transition region of the microburst-like wind at r/DE0.5, the local drag coefficients of the
test model were found to increase greatly. While the variations of the local drag coefficients, CFr, were found to be quite
small over the entire height of the high-rise building model in general, the local along-wind coefficients at the upper half of
the test model were found to be slightly greater than those of the lower half. However, when the test model was moved into
the outburst region, i.e. at r/D E1.0, 1.5, and 2.0, corresponding to the much higher surface pressure on the lower half of the
windward wall as shown clearly in Fig. 7, the local along-wind coefficients at the lower half of the test model (z/Ho0.5)
were found to become much larger than those at the upper half of the test model. It can also be seen that the local drag
coefficients, CFr, would reach their peak values when the test model was mounted at the radial location of r/D E1.0.
It indicates that the test model would experience the maximum along-wind (i.e., drag) force when it is mounted in the
outburst region at r/D E1.0, which is correlated well with the appearance of the maximum wind speed in the outburst
region of the microburst-like flow, as shown in Fig. 4.
Fig. 8(b) shows the comparison of the microburst-induced local along-wind force coefficients of the test model when it is
mounted in the outburst region of the microburst-like wind at r/D E1.0 and those reported in Lin et al. (2005) and Kim and
Kanda (2010) with conventional ABL winds. It should be noted that the geometry shape of the high-rise building model used
in the present study is the same as those used by Lin et al. (2005) and Kim and Kanda (2010), i.e., the same square cross-
section and the same H/B ratio of 4 for the test model. While the impinging jet velocity, Vjet (which is also the maximum
wind speed in the outburst region of the microburst-like wind as shown in Fig. 4), is used in the present study as the
reference velocity to define the local aerodynamic force coefficients, the wind speed at the roof height U H of the test model
was used as the reference velocity by Lin et al. (2005) and Kim and Kanda (2010) in order to define the coefficients of local
aerodynamic forces in conventional ABL winds. It can be seen clearly that when the high-rise building model was placed in
conventional ABL winds, the local along-wind force coefficients were found to increase monotonically with the increasing
elevation height before reaching their peak values at the stagnation region around z/HE0.85. Apparently, the distribution of
the local along-wind force coefficients in the microburst-like wind showed a quite different trend in comparison with those
in conventional ABL winds. Since the height of the high-rise building model exceeds the thickness of the outburst flow, the
regions with higher local along-wind force coefficients were found to concentrate at the lower half of the high-rise building

Fig. 8. Comparison of the local along-wind force coefficients in the microburst-like wind versus those in conventional ABL winds.
556 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

model as the test model was mounted in the outburst region of r/DZ1.0. While such microburst-induced wind load
distributions could pose much higher local wind loads acting on the lower half of the high-rise building model, it would
result in less bending moment acting on the high-rise buildings compared with those produced by an equivalently strong
ABL wind. More specifically, when the high-rise building model was mounted in the microburst-like wind at the location of
r/D E1.0, the bending moment acting on the test model would be approximately 57% and 62% of those produced by
equivalently strong conventional ABL winds based on the results given in Lin et al. (2005) and Kim and Kanda (2010),
respectively. However, it should be noted that, due to the extreme high wind speed in microburst winds, overall mean wind
loads acting on high-rise buildings could be much greater compared with those in conventional ABL winds.
As aforementioned, in addition to mapping the surface pressure distributions around the test model, the magnitude of
the microburst-induced wind loads acting on the test model was also measured directly using a high-sensitive force-
moment sensor unit (i.e., JR3 load cell). Fig. 9 gives the measured microburst-induced wind loads (i.e., both time-averaged
along-wind and cross-wind forces) acting on the test model when the model was mounted in the microburst-like wind at
five typical radial locations (i.e., r/D E0.0, 0.5, 1.0, 1.5 and 2.0) and with three different orientation angles (i.e., OA 0.01,
22.51 and 451) in respect to the oncoming microburst-like wind. In the present study, the along-wind force coefficient (i.e.,
drag coefficient) of the test model is defined as CF r F r =0:5V 2jet A, and the cross-wind force coefficient is defined as
CF t F t =0:5V 2jet A, where F r and F t are the aerodynamic forces along the radial and tangential directions of the microburst-
like wind, and A is the reference area which is defined as AB  H. In the plots, while the results obtained by integrating the
surface pressure distributions around the test model surfaces are shown as hollow dots, the data derived from direct force
measurements with force-moment sensor unit (i.e., JR3 load cell) are shown as solid dots. It can be seen clearly that the
results derived from the integration of the surface pressure distributions around the test model agree well with those
measured directly with the JR3 load cell.
As shown in Fig. 9(a), when the test model was moved away from the microburst center, the time-averaged along-wind
force was found to increase rapidly at first, reach its peak value in the transition region at r/D E0.51.0, and then decrease
gradually in the outburst region of the microburst-like wind (i.e., r/D4 1.0). This indicates that a high-rise building would
experience the maximum static wind loads when it is in the transition region of the microburst-like wind (i.e. at r/D E0.5
1.0), which was found to be closely related to the unique flow characteristics of the microburst-like wind discussed above.
It is interesting to note that, while the along-wind force acting on the test model was found to vary greatly with the
orientation angle of the test model in respect to the oncoming microburst-like wind when the model was mounted at the
radial location of r/D E0.5, the effects of the orientation angle were found to become much smaller in the outburst region of
the microburst-like wind, i.e. r/D Z1.0. This can be explained based on the facts revealed from the changes in the surface
pressure distributions on the test model at different orientation angles, given in Fig. 10. In general, the overall drag force
acting on the test model will be determined by the pressure difference on the windward and leeward walls and the
projected area of the test model along the flow direction. As shown clearly in Fig. 10(a), when the test model was mounted
in the transition region of the microburst-like wind, i.e., r/D E0.5, while the surface pressure on the leeward walls stays
almost the same as the orientation angle of the test model was changed from 0.01 to 451, the regions with relatively high
surface pressure on the windward walls of the test model were found to increase significantly. As a result, corresponding to
the greater projected area of the test model at a higher orientation angle, the along-wind force acting on the test model
was found to increase monotonically as the orientation angle increases from 0.01 to 451. However, when the test model
was moved into the outburst region of the microburst-like wind, such as at r/D E1.0 as shown in Fig. 10(b), the effects of
the increasing projected area of the test model would be canceled out by the decreasing size of the high-pressure regions
on the windward walls, corresponding to the inclined surface toward the oncoming microburst-like wind for the cases
with non-zero orientation angles. Therefore, time-averaged along-wind force acting on the test model was found to be

Fig. 9. The time-averaged wind loads acting on the test model as a function of the radial location and the orientation angle of the test model.
 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564 557

Fig. 10. Comparison of the time-averaged surface pressure distributions around the test model at different orientation angles in respect to the oncoming
microburst-like wind: (a) at the radial location of r/DE0.5 and (b) at the radial location of r/DE 1.0.

almost independent of the orientation angle of the test model when it was mounted in the outburst region of the
microburst-like wind.
As shown in Fig. 9(b), since the flow field of the microburst-like wind is essentially axisymmetric, the time-averaged
across-wind force acting on the test model was found to be quite small, which is almost negligible, when the building model
was mounted in a symmetric way in respect to the oncoming microburst-like wind, i.e., for the test cases of OA E0.01 and
451. As shown clearly in Fig. 10, the surface pressure distributions on the test model would become asymmetric for the test
cases with the orientation angle OA E22.51. While the values of the across-wind force coefficients were found to be always
quite small in general, its peak value was found to be about  0.2 when the test model was mounted in the transition region
of the microburst-like wind (i.e., r/D E0.5). Since the magnitude of the across-wind force acting on the test model is found
to be always much smaller compared with the along-wind force (i.e., drag force), the across-wind force would not be a
serious concern for the safety of high-rise buildings in microburst-like winds.

3.4. The characteristics of dynamic wind loads induced by microburst-like wind

While the time-averaged surface pressure distributions around the test model given above are very helpful to reveal the
global features of the microburst-induced mean wind loads acting on high-rise buildings, it would be very insightful and
essential to take the turbulent nature of the microburst-like wind into account in order to assess its damage potential more
accurately. In the present study, the fluctuations of the surface pressure distributions around the high-rise building model
were also investigated for a better understanding of the unsteady nature of the wind loads induced by the microburst-like
wind. The maps of the standard deviations of the surface pressure distribution around the test model, which is defined as
C p;rms P stdev =0:5V 2jet ; where Pstdev denotes the standard deviation of the instantaneous surface pressure data, are shown
in Fig. 11. It can be seen clearly that when the test model was mounted near the core center of the microburst-like wind (i.e.,
r/D E0.0), while the time-averaged values of the surface pressures on the test model are quite high as shown in Fig. 7(a), the
fluctuations of the surface pressures around the test model were found to be rather small, in general. When the test model
was moved into the transition region of the microburst-like wind (i.e., r/D E0.5), the regions with larger surface pressure
fluctuation amplitudes were found to appear near the upper edge of the sidewalls and the mid-height of the leeward wall of
558 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

Fig. 11. The distributions of the RMS values of the surface pressures around the test model in the microburst-like wind at five typical radial locations:
(a) r/D E0.0, (b) r/DE0.50, (c) r/D E1.0, (d) r/D E1.5, and (e) r/D E2.0.

the test model. Such a distribution pattern is believed to be closely related to the unique flow features of the microburst-like
wind described above.
When the test model was moved further away into the outburst region of the microburst-like wind (i.e., r/D 41.0), the
fluctuation amplitudes of the surface pressures on the test model were found to become much greater, corresponding to the
much higher turbulence levels in the outburst region of the microburst-like wind (Zhang et al., 2013a). Due to the flow
separations on the sidewalls of the test model, the pressure fluctuations over the sidewalls were found to be much greater
than those on the windward and leeward walls. Similar findings were also reported by Melbourne (1980) and Tanaka and
Lawen (1986) with the high-rise building placed in conventional ABL winds. It is interesting to note that the regions with
larger surface pressure fluctuation amplitudes were found to be the regions where the peak negative surface pressure was
observed in the time-averaged surface pressure distributions as given in Fig. 7. While the maximum flow velocity in the
microburst-like wind was found to occur at the radial location r/D E1.0 as shown in Fig. 4, the fluctuations of the surface
pressures on the sidewalls of the test model were found to become more significant when the test model was mounted at
further downstream locations, i.e., r/D E1.5 and 2.0. Meanwhile, the fluctuations of the surface pressure on the windward
wall of the test model, which was considered to be directly related to the turbulence levels of the microburst-like wind,
were also found to increase greatly as the radial distance increases to r/D E1.5 and 2.0.
In the present study, the root-mean-square (RMS) values of the local aerodynamic forces acting on the test model in the
microburst-like wind were also calculated in order to better reveal the unsteady nature of the microburst-induced wind
loads acting on the high-rise building model. Figs. 12(a) and 13(a) present the profiles of the RMS values of the local along-
wind and across-wind force coefficients as a function of the elevation height when the test model was mounted in the
microburst-like wind at five typical radial locations. The measurement results reported by Lin et al. (2005) with a high-rise
building model placed in a conventional ABL wind are presented in Figs. 12(b) and 13(b) for comparison. In general, the
fluctuation amplitudes of both local along-wind and across-wind forces acting on the test model were found to be closely
related to the local turbulence levels of the microburst-like wind at different radial locations as shown in Fig. 5.
Corresponding to the much higher turbulence levels in the outburst region of the microburst-like wind, the RMS values
of both local along-wind and across-wind forces acting on the test model were found to become apparently greater when
the test model was mounted in the outburst region (i.e., r/DZ1.0) compared to those with the test model mounted within
 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564 559

Fig. 12. Comparison of the RMS values of the local along-wind force coefficients in the microburst-like wind versus those in a conventional ABL wind.
(a) Microburst (Normalized by Vjet) and (b) ABL wind (Normailzed by UH).

Fig. 13. Comparison of the RMS values of the local cross-wind force in the microburst-like wind versus those in a conventional ABL wind. (a) Microburst
(Normalized by Vjet) and (b) ABL wind (Normailzed by UH).

the stagnation region (i.e., r/D E0.0) or the transition region (i.e., r/DE0.5) of the microburst-like wind. The RMS values of
both local along-wind and cross-wind forces were found to reach their peak values when the test model was mounted in the
outburst region at the location of r/DE1.5, where the local turbulence intensity levels of the microburst-like wind were
found to reach the maximum values. Corresponding to the higher turbulence intensity levels associated with the strong
shear layer at the higher elevation level of z/H 40.5 as shown in Fig. 5, the RMS values of the local along-wind and across-
wind forces at the higher elevation levels were found to be generally quite high, particularly for the cases with the test
model mounted at the radial locations of r/D E0.5 and 1.0. It was also found that the RMS values of the local across-wind
force would be greater than those of the along-wind force. This is believed to be closely related to the larger surface pressure
fluctuations on the sidewalls of the test model due to the flow separations on the surfaces, as revealed clearly from the
surface pressure distributions given in Fig. 7.
In comparison with those in the microburst-like wind, the profiles of the RMS values of both the local along-wind and
across-wind forces showed much simpler distributions when the high-rise building model was placed in conventional ABL
winds. As shown clearly in Figs. 12(b) and 13(b), while the RMS values of the local along-wind forces were found to increase
monotonically as the elevation height increases, the RMS values of the across-wind forces would increase slightly before it
decreases towards the roof height. It indicates that, for the same high-rise buildings, much more complicated dynamic wind
loads would be induced by microburst winds compared with those in conventional ABL winds. It should also be noted that
the magnitude of the RMS values of the local force coefficients induced by the microburst-like wind should not be compared
directly with those reported by Lin et al. (2005) with conventional ABL winds since the two studies used two different
reference flow velocities for the normalization. However, since the reference flow velocity in a microburst wind would
usually be much greater (i.e., up to 270 km/h) than that of a conventional ABL wind, the fluctuations of the microburst-
induced wind loads acting on high-rise buildings would be much more significant than those in a conventional ABL wind.
560 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

Based on the time-sequences of the dynamic wind load measurements with the JR3 load cell at the data acquisition rate
of 1000 Hz for 30 s, the standard deviations of the total along-wind and across-wind forces induced by the microburst-like
wind were determined in the present study, which can be used to reveal the unsteady nature of the microburst-induced
wind loads more clearly and quantitatively. Table 1 lists the standard deviations of the dynamic wind loads (i.e., both along-
wind and across-wind forces CFr and CFt) acting on the test model as it was mounted in the microburst-like wind at five
typical radial locations (i.e., r/D E0.0, 0.5, 1.0, 1.5, and 2.0) with three orientation angles (i.e., OA 0.01, 22.51 and 451) in
respect to the oncoming microburst-like wind. It can be clearly seen that the fluctuation amplitudes of the dynamic wind
loads (i.e., both the along-wind and cross-wind forces) acting on the test model would change greatly depending on both
the radial location and the orientation angle of the test model. Even though the time-averaged values of the cross-wind
forces, CFt, were found to be always quite small as shown in Fig. 9(b), the fluctuation amplitudes of the dynamic cross-wind
forces, i.e., CFt, were found to be much greater than those of the dynamic along-wind forces, i.e., CFr, except for the cases
with the test model mounted at the core center of the microburst core (i.e., r/D E0.0). When it was mounted at the same
radial location in the microburst-like wind, the test model was found to suffer the largest aerodynamic force fluctuations
when it was mounted at 01 orientation angle in respect to the oncoming microburst-like wind (i.e., the position with the
wall of the test model being normal to the radial component of the oncoming microburst-like wind). Among all the test
cases, the dynamic wind loads acting on the high-rise building model were found to have the maximum fluctuation
amplitudes when it was mounted in the outburst region of the microburst-like wind at the radial location of r/D E1.5.
It should be noted that this location does not coincide with the location where the test model would experience the
maximum time-averaged wind loads (i.e., the test model would experience the maximum mean wind loads at the radial
location of r/DE0.51.0, as shown clearly in Fig. 9). It indicates that, due to the complex flow features of the microburst-like
wind, the greatest dynamic wind loads and the maximum mean wind loads acting on high-rise buildings would not occur
simultaneously during a microburst event. The finding suggests that special attention should be paid to both the maximum
mean wind loads and the greatest dynamic wind loads for the safety of high-rise buildings in violent microburst winds.

3.5. Power spectrum density (PSD) analysis of the turbulent flow and microburst-induced wind loads acting on the test model

Based on the time-sequences of the measured instantaneous flow velocity in microburst-like winds with the Cobra Probe
at the data acquisition rate of f1000 Hz, the power spectrum density (PSD) analysis was conducted to reveal the flow
characteristics of the turbulent microburst-like wind. Fig. 14 presents the PSD analysis results of the measured turbulent
flow velocities with the measurement point located in the core region (i.e., at the location of r/D E0 and z/H 0.5 for Fig. 14
(a)) and the outburst region of the microburst-like wind (i.e., at the location of r/D E1.0 and z/H 0.10, 0.33 and 0.67 for
Fig. 14(b)(d), respectively). In the present study, the reduced frequency, N f f B=V jet , was normalized using the jet velocity
V jet and the side length of the square across-section of the test model, B.
As shown in Fig. 14(a), due to the continuous energy input in the jet flow exhausted from the ISU microburst simulator, the PSD
values at all frequencies were found to remain almost constant, and no decay was observed at the relatively high frequencies. No
dominant frequency could be found from the PSD spectrum when the measurement point was located at the center of the
impinging jet flow, as shown in Fig. 14(a). However, when the measurement point was moved into the outburst region of the
microburst-like wind at r/DE1.0, a dominant frequency peak, around Nf 0.060.07, can be identified in the power spectrum plots
at all the three studied elevation levels, as shown in Fig. 14(b)(d). Similar findings can also be found at other radial locations in the
outburst region of the microburst-like wind, which are not presented here. The peak PSD value in the energy-containing range
indicates that a large-scale flowstructure would be dominant in the outburst region of the microburst-like wind. At relatively high
frequencies, the normalized PSD values of the turbulent flow velocity were found to decrease with a slope of  23 obeying
 2=3
Kolmogorovs 5/3 law in the inertial sub-range, i.e., nSz; n=u2n 0:26f , as described in Simiu and Scanlan (1996), particularly
at z/H 13 and 23 , where the flow was less affected by wall boundary than that at z/H1=10.
It should be noted that the dominant frequency in the power spectrum plots could be corresponding to the periodic
shedding of the primary vortices of the impinging jet flow, as reported in Zhang et al. (2013a). The periodic shedding of

Table 1
The standard deviations of the along-wind and cross-wind force coefficients (i.e., CFr and CFt) of the test model induced by the microburst-like wind.

The location of the test model in Orientation angle of the test model, Orientation angle of the test model, Orientation angle of the test model,
the microburst-like wind OA 0.01 OA 22.51 OA 45.01

Along-wind Cross-wind Along-wind Cross-wind Along-wind Cross-wind

component, CFr component, CFt component, CFr component, CFt component, CFr component, CFt

r/D E0.0 0.17 0.13 0.14 0.10 0.13 0.10

r/D E0.5 0.12 0.15 0.12 0.13 0.10 0.12
r/D E1.0 0.21 0.28 0.15 0.16 0.15 0.15
r/D E1.5 0.21 0.32 0.15 0.19 0.18 0.19
r/D E2.0 0.21 0.31 0.14 0.19 0.14 0.15
 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564 561

Fig. 14. Power spectrum density analysis results of the turbulent flow velocities in the microburst-like wind.

Fig. 15. The power spectrum density analysis of the dynamic wind loads acting on the test model in the microburst-like wind at the radial location of
r/D E1.0.

primary vortices in impinging jet flows has been observed in many previous studies such as Ho and Nosseir (1981) and
Didden and Ho (1985). Meanwhile, many previous studies on impinging-jet heat transfer have also reported that the
periodic shedding of the primary vortices in impinging jet flows would result in oscillating behavior of the instantaneous
Nusselt number distributions on the impinging plate (Chung and Luo, 2002). In order to reveal more clearly how the
periodic shedding of the primary vortices may affect the microburst-induced dynamic wind loads acting on the high-rise
building model, PSD analysis was also conducted for the measured dynamic forces acting on the test model. Fig. 15 shows
the PSD results of the microburst-induced dynamic wind loads when the test model was mounted in the outburst region at
r/D E1.0.
As described in Lin et al. (2005), while the PSD analysis results of the along-wind force acting on a high-rise building in
conventional ABL winds would often show a wide-band distribution as it is mainly affected by the turbulence levels in the
oncoming airflow, the PSD analysis results of the across-wind force usually has a narrow-band peak, which is induced by the flow
separation and vortex shedding on the sidewalls of the test models. However, as shown clearly in Fig. 15, the PSD analysis results of
the microburst-induced along-wind and across-wind forces were found to be correlated much better in the microburst-like wind
than those in conventional ABL winds, particularly for the region with relatively high frequencies, which is believed to be closely
related to the much higher turbulence levels in the outburst region of the microburst-like wind. It can also be seen that a dominant
frequency can be identified at Nf f B=V jet 0:065 in the PSD plot of the along-wind force given in Fig. 15. The dominant
frequency was found to be very close to the peak frequency of the turbulent flow velocity at this radial location as given in Fig. 14.
It suggests that the low-frequency component of the microburst-induced wind loads would be dominated by the periodic shedding
562 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

of the primary vortices when the test model was mounted in the microburst-like wind. It should also be noted that the peak
frequency in the PSD plot of the across-wind force was found to be Nf f B=V jet 0:08, which is slightly higher than the dominant
frequency of the dynamic along-wind force. This peak frequency is believed to be closely related to the flow separations on the
sidewalls of the test model, which could also be expected in conventional ABL winds as described in Lin et al. (2005). The
corresponding reduced frequency of the peak frequency at N f f B=V jet 0:08 in the PSD plot of the microburst-induced dynamic
cross-wind force would become approximately 0.12 when it is re-normalized using the mean velocity at the building height (i.e.,
 8.6 m/s for the present study), which was found to agree with the vortex-shedding frequency from a square cylinder in a
turbulence flow as given in Vickery (1966).
In summary, the PSD analysis given above reveals clearly that the characteristics of the microburst-induced wind loads
acting on the high-rise building model are much more complex than those in conventional ABL winds. All the factors
discussed above, including the periodic shedding of the primary vortices, the massive flow separations on the sidewalls of
the test model, and the high turbulence levels in the microburst-like wind, would contribute to the characteristics of the
dynamic wind loads acting on the building model. Such unique features of microburst-induced dynamic wind loads should
be taken into account to gain a better understanding of the underlying physics and to provide a better prediction of the
damage potential of high-rise buildings subject to violent microburst winds.

4. Conclusions

An experimental study was conducted to quantify the flow characteristics of microburst-like wind and to assess both the static
and dynamic wind loads acting on high-rise buildings induced by violent microburst-like winds in comparison with those in
conventional atmospheric boundary layer (ABL) winds. The experimental study was conducted using an impinging-jet-based
microburst simulator available in the Department of Aerospace Engineering of Iowa State University. A high-rise building model,
which has a 45 mm  45 mm cross-section and a total height of 180 mm (i.e., an aspect ratio of 4), was used in the present study.
During the experiments, the test model was mounted in the microburst-like wind at five typical radial locations (i.e., r/DE0.0, 0.5,
1.0, 1.5, and 2.0) with different orientation angles (01, 22.51, and 451) in respect to the oncoming flow streams. In addition to
conducting flow field measurements to quantify the unique flow characteristics of microburst-like winds, both the mean and
dynamic wind loads acting on the test model induced by the microburst-like wind are assessed in detail based on the quantitative
measurements of the aerodynamic forces and the surface pressure distributions around the high-rise building model. While the
microburst-induced wind loadings acting on high-rise building model were found to be significantly different from their
counterparts in conventional ABL winds in general, the major findings of the present study are listed as follows:

(1) The time-averaged surface pressure distributions around the high-rise building model were found to vary significantly
depending on the radial location of the test model in respect to the core center of the microburst-like wind. When the
test model was mounted within the core center (i.e., r/D E0.0) or in the transition region (i.e., r/D E0.5) of the
microburst-like wind, the test model was found to be wrapped around by high positive pressure, which may pose
potential safety problems for well-sealed high-rise buildings due to the significant pressure differences between the
inside and outside glass walls. When the test model was moved farther away into the outburst region of the microburst-
like wind (i.e., r/D Z1.0), while the regions with significant negative pressure values would cover most of the sidewalls,
a region with quite high positive pressure values was found to concentrate on the lower half of the windward wall.
Compared with those in conventional ABL winds, a upside-down pattern was observed for the surface pressure
distribution on the windward wall of the test model due to the unique flow characteristics of the microburst-like wind.
(2) Both the radial location and the orientation angle of the test model were found to affect the mean wind loads acting on
the high-rise building model greatly. With an orientation angle of OA E0.01 (i.e., with the wall of the test model being
normal to the radial component of the oncoming microburst-like wind), the test model was found to experience the
maximum mean wind loads at the radial location of r/DE1.0. When the test model was mounted with an orientation
angle of OA E22.51 or 451 in respect to the oncoming microburst-like wind, the test model would experience the
maximum mean wind loads in the transition region of the microburst-like wind at r/DE 0.5.
(3) The fluctuation amplitudes of the surface pressures on the high-rise building model were found to be much greater
when the test model was located in the outburst region of the microburst-like wind (r/D Z1.0), compared with those
with the test model mounted at the core center (i.e., r/DE 0.0) or in the transition region of the microburst wind
(r/D E0.5). The large fluctuation amplitudes of the surface pressures on the test models would imply significant peak
wind loads acting on the high-rise building models, which would greatly increase the damage potential of high-rise
buildings subject to microburst-like wind. While the time-averaged values of the across-wind force acting on the test
model in the microburst-like wind were found to be very small in general, which are almost negligible, in comparison
with those of the along-wind force, the fluctuation amplitudes of the dynamic across-wind forces were found to be
much greater than those of the along-wind forces acting on the test model in the microburst-like wind.
(4) While the test model was found to experience the maximum mean wind loads in the microburst-like wind at the radial
location of r/D E0.51.0, the dynamic wind loads acting on the high-rise building model were found to have the largest
fluctuation amplitudes when it was mounted in the outburst region at r/D E1.5 due to the highest turbulence intensity
levels at the radial location in the microburst-like wind. It suggests that, due to the complex flow features of the
 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564 563

microburst-like wind, the greatest dynamic wind loads and the maximum mean wind loads acting on high-rise
buildings may not occur simultaneously during a microburst event. Special attentions should be paid to both the
maximum mean wind loads and the greatest dynamic wind loads for the safety of high-rise buildings during violent
microburst winds.
(5) The power spectrum density (PSD) analysis of the turbulent flow velocity and the dynamic wind loads reveal clearly that
the characteristics of the dynamic wind loads acting on the high-rise building model in the microburst-like wind are
much more complex than those in conventional ABL winds. The periodic shedding of the primary vortices, the massive
flow separations on the sidewalls of the test model, and the high turbulence levels in the outburst region of the
microburst-like wind would contribute to the characteristics of the dynamic wind loads acting on the high-rise building
model in the microburst-like wind. Such unique features of the microburst-induced dynamic wind loads should be
taken into account in order to gain a better understanding of the underlying physics to provide a better prediction of the
damage potential of high-rise buildings subject to microburst-like wind.

In summary, the findings of the present study reveal clearly that the microburst-induced wind loadings acting on high-
rise buildings would be significantly different from their counterparts in conventional ABL winds. It suggests that the
current design standard for high-rise buildings, which are generally based on ABL winds, may not be sufficient to cover the
damage potentials induced by microburst-like winds due to the flow complexity. It should also be noted that, while only a
single high-rise building model sited over a flat surface was considered in the present study, an actual situation may become
much more complicated due to the existence of surrounding buildings and effects of different types of surface terrains.
More comprehensive studies and systematic laboratory testing are needed to take the complexities into account in order to
improve our understanding about the flowstructure interactions between the high-rise buildings and microburst-like
winds for a more accurate prediction of the microburst-induced wind damage potentials to high-rise buildings.

Acknowledgments

This research work was funded by the US National Science Foundation (NSF) under Award no. CMMI-1000198. The
authors would like to thank NSF for the financial support and acknowledge the help of Bill Rickard, department technician,
and Nick Krauel, undergraduate student, at Iowa State University in building test models and experimental setups.

References

Chay, M.T., Albermani, F., Wilson, R., 2006. Numerical and analytical simulation of downburst wind loads. Engineering Structures 28 (2), 240254.
Chay, M.T., Letchford, C.W., 2002. Pressure distributions on a cube in a simulated thunderstorm downburstPart A: stationary downburst observations.
Journal of Wind Engineering and Industrial Aerodynamics 90 (7), 711732.
Chen, L., Letchford, C.W., 2004. Parametric study on the along-wind response of the CAARC building to downbursts in the time domain. Journal of Wind
Engineering and Industrial Aerodynamics 92 (9), 703724.
Choi, E.C., 2004. Field measurement and experimental study of wind speed profile during thunderstorms. Journal of Wind Engineering and Industrial
Aerodynamics 92 (3), 275290.
Chung, Y.M., Luo, K.H., 2002. Unsteady heat transfer analysis of an impinging jet. Journal of heat transfer 124 (6), 10391048.
Das, K.K., Ghosh, A.K., Sinhamahapatra, K.P., 2010. Investigation of the axisymmetric microburst flow field. Journal of Wind Engineering 7, 115.
Didden, N., Ho, C.M., 1985. Unsteady separation in a boundary layer produced by an impinging jet. Journal of Fluid Mechanics 160, 235256.
Fujita, T.T., 1985. The Downburst. SMRP Research Paper. University of Chicago Press, Chicago.
Fujita, T.T., 1981. Tornadoes and downbursts in the context of generalized planetary scales. Journal of the Atmospheric Sciences 38 (8), 15111534.
Hjelmfelt, M.R., 1988. Structure and life cycle of microburst outflows observed in Colorado. Journal of Applied Meteorology 27 (8), 900927.
Ho, C.M., Nosseir, N.S., 1981. Dynamics of an impinging jet. Part 1. The feedback phenomenon. Journal of Fluid Mechanics 105, 119142.
Holmes, J.D., Oliver, S.E., 2000. An empirical model of a downburst. Engineering Structures 22 (9), 11671172.
Irwin, H.P.A.H., Cooper, K.R., Girard, R., 1979. Correction of distortion effects caused by tubing systems in measurements of fluctuating pressures. Journal of
Wind Engineering and Industrial Aerodynamics 5 (1), 93107.
Irwin, P.A., 2009. Wind engineering challenges of the new generation of super-tall buildings. Journal of Wind Engineering and Industrial Aerodynamics 97
(7), 328334.
Kareem, A., 1987. The effect of aerodynamic interference on the dynamic response of prismatic structures. Journal of Wind Engineering and Industrial
Aerodynamics 25 (3), 365372.
Kareem, A., 1992. Dynamic response of high-rise buildings to stochastic wind loads. Journal of Wind Engineering and Industrial Aerodynamics 42 (1),
11011112.
Kawai, H., 1992. Vortex induced vibration of tall buildings. Journal of Wind Engineering and Industrial Aerodynamics 41 (1), 117128.
Kim, J., Hangan, H., 2007. Numerical simulations of impinging jets with application to downbursts. Journal of Wind Engineering and Industrial
Aerodynamics 95 (4), 279298.
Kim, Y.M., You, K.P., Ko, N.H., 2008. Across-wind responses of an aeroelastic tapered tall building. Journal of Wind Engineering and Industrial Aerodynamics
96 (8), 13071319.
Kim, Y., Kanda, J., 2010. Characteristics of aerodynamic forces and pressures on square plan buildings with height variations. Journal of Wind Engineering
and Industrial Aerodynamics 98 (8), 449465.
Kwok, K.C.S., 1982. Cross-wind response of tall buildings. Engineering Structures 4 (4), 256262.
Kwok, K.C.S., 1988. Effect of building shape on wind-induced response of tall building. Journal of Wind Engineering and Industrial Aerodynamics 28 (1),
381390.
Lam, K.M., Zhao, J.G., Leung, M.Y.H., 2011. Wind-induced loading and dynamic responses of a row of tall buildings under strong interference. Journal of
Wind Engineering and Industrial Aerodynamics 99 (5), 573583.
Lin, N., Letchford, C., Tamura, Y., Liang, B., Nakamura, O., 2005. Characteristics of wind forces acting on tall buildings. Journal of Wind Engineering and
Industrial Aerodynamics 93 (3), 217242.
564 Y. Zhang et al. / Journal of Fluids and Structures 50 (2014) 547564

Marukawa, H., Ohkuma, T., Momomura, Y., 1992. Across-wind and torsional acceleration of prismatic high rise buildings. Journal of Wind Engineering and
Industrial Aerodynamics 42 (1), 11391150.
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. Journal of Wind Engineering and Industrial Aerodynamics 93 (7), 557580.
Melbourne, W.H, 1980. Comparison of measurements on the CAARC standard tall building model in simulated model wind flows. Journal of Wind
Engineering and Industrial Aerodynamics 6 (1), 7388.
Sengupta, A., Haan, F.L., Sarkar, P.P., Balaramudu, V., 2008. Transient loads on buildings in microburst and tornado winds. Journal of Wind Engineering and
Industrial Aerodynamics 96 (10), 21732187.
Sengupta, A., Sarkar, P.P., 2008. Experimental measurement and numerical simulation of an impinging jet with application to thunderstorm microburst
winds. Journal of Wind Engineering and Industrial Aerodynamics 96 (3), 345365.
Simiu, E., Scanlan, R.H., 1996. Wind Effects on Structures, An Introduction to Wind Engineering, second edition John Wiley & Sons, New York5354.
Sykes, D.M., 1983. Interference effects on the response of a tall building model. Journal of Wind Engineering and Industrial Aerodynamics 11 (1), 365380.
Tanaka, H., Lawen, N., 1986. Test on the CAARC standard tall building model with a length scale of 1: 1000. Journal of Wind Engineering and Industrial
Aerodynamics 25 (1), 1529.
Taniike, Y., 1992. Interference mechanism for enhanced wind forces on neighboring tall buildings. Journal of Wind Engineering and Industrial
Aerodynamics 42 (1), 10731083.
Tu, C.V., Wood, D.H., 1996. Wall pressure and shear stress measurements beneath an impinging jet. Experimental Thermal and Fluid Science 13 (4),
364373.
Vickery, B.J., 1966. Fluctuating lift and drag on a long cylinder of square cross-section in a smooth and in a turbulent stream. Journal of Fluid Mechanics 25
(03), 481494.
Walter, J.C., 2007. A new climatology of 225-year, 50-year, and 100 year microburst winds. In: Proceedings of the 22nd Conference on Weather Analysis and
Forecasting/18th Conference on Numerical Weather Prediction.
Wardlaw, R.L., Moss, G.F., 1970. A standard tall building model for the comparison of simulated natural winds in wind tunnels. Report CC-662 Tech, 25.
Wood, G.S., Kwok, K., Motteram, N.A., Fletcher, D.F., 2001. Physical and numerical modelling of thunderstorm downbursts. Journal of Wind Engineering and
Industrial Aerodynamics 89 (6), 535552.
Yang, Z.F., Sarkar, P.P., Hu, H., 2011. An experimental study of flow field around a high-rise building model in tornado-like winds. Journal of Fluids and
Structures 27 (4), 471486.
Zhang, Y., Hu, H., Sarkar, P.P., 2013a. Modeling of microburst outflows using impinging jet and cooling source approaches and their comparison. Engineering
Structures 56, 779793.
Zhang, Y., Sarkar, PP, Hu, H, 2013b. An experimental study of flow fields and wind loads on gable-roof building models in microburst-like wind.
Experiments in Fluids 54 (5), 1511.
Zhang, Y., 2013. A Study of Microburst-like Wind and its Loading Effects on Structures Using Impinging-jet and Cooling-source Approaches (Ph.D.
Dissertation). Iowa State University, Iowa, USA.