AIAA 2010-4682
28th AIAA Applied Aerodynamics Conference
28 June - 1 July 2010, Chicago, Illinois
Unsteady Flow Structure Interactions of a Gable-roofed
Building Model in Tornado-like Winds
Zifeng Yang1, Partha Sarkar2, Fred Haan3 and Hui Hu4 ()
Department of Aerospace Engineering, Iowa State University, Ames, IA, 50010
An experimental study was conducted to quantify the characteristics of wake vortex and
flow structures around a gable-Roof building model as well as the resultant pressure
distribution and wind loads (both forces and moments) acting on the test model in tornadolike winds. In addition to measuring the pressure acting on the tested high-rise building
model using pressure transducers and wind loads using a high-sensitivity load cell, a digital
Particle Image Velocimetry (PIV) system was used to conduct detailed flow field
measurements to quantify the evolution of the unsteady vortex and turbulent flow structures
around the test model in tornado-like winds. The measurement results revealed clearly that
the evolution of the wake vortex and turbulent flow structures around the test model as well
as the resultant wind loads induced by tornado-like winds were significantly different from
those in conventional straight-line winds. The detailed flow field measurements were
correlated with the pressure and wind load measurement data to elucidate the underlying
physics to gain further insight into the flow-structure interactions between the tested gableroof building model and tornado-like winds.
Nomenclature
A1
A2
L
Projected area of the test building model along roof direction
Projected area of the test building model in height direction
Width of the building model along roof direction
1
2
( P P0 ) /( UVO )
2
1
2
CFX FX /( UVO A1 )
2
1
2
CFY FY /( UVO A1 )
2
1
2
CFZ FZ /( UVO A2 )
2
1
2
CM X M X /( UVO A1 L )
2
1
2
CM Y M Y /( UVO A1 L)
2
1
2
CM Z M Z /( UVO A1 L)
2
R0
Pressure coefficient
CP
Force coefficient in the x direction
Force coefficient in the y direction
Force coefficient in the z direction
Moment coefficient in the x direction
Moment coefficient in the y direction
Moment coefficient in the z direction
Core Radius of the tornado-like vortex
1
Graduate Student, Department of Aerospace Engineering.
Professor, Department of Aerospace Engineering.
3
Associate Professor, Department of Aerospace Engineering.
4
Associate Professor, Department of Aerospace Engineering, AIAA Senior Member, Email: huhui@iastate.edu
2
Copyright © 2010 by Zifeng Yang , Partha Sarkar, Fred Haan and Hui Hu . Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.
R
Fx , Fy , Fz
U, V, W
X, Y, Z
V0
P
P0
The distance from the center of the tornado-like vortex
X, Y, Z component of the aerodynamic force
X,Y,Z component of the flow velocity vectors
Cartesian axes as defined
Maximum tangential velocity of the tornado like vortex
Air density
Pressure on the building model
Atmosphere pressure in the test room
I. Introduction
T
Ornadoes are powerful columns of winds spiraling around centers of low atmospheric pressure, which are often
(but not always) visible as funnel clouds. Tornadoes are considered as nature’s most violent storms, which can
produce the highest nature wind speeds on earth, i.e., with the maximum wind speed up to 400 mph. Large-scale
damage and heavy loss of life caused by intense tornado outbreaks in populated areas have been reported frequently
in the past years (Grazulis 1993; Brooks and Doswell 2001; Speheger et al. 2002; Doswell et al. 2006; Forbes 2006).
Some of the well-known tornadoes include the Tri-State tornado of March 18, 1925 killed 695 people and the
Natchez, Mississippi tornado of May 7, 1840 claimed 317 lives. The 1965 Palm Sunday tornado outbreak was the
second deadliest of the 20th century, killing 258 people and injuring 3,148. The 1999 Oklahoma tornado outbreak
that lasted from May 3 to May 6, 1999 stands as the costliest tornado in United States history having destroyed
nearly 11,000 buildings and causing over $1.9 billion in damage.
According to Wind Hazard Reduction Coalition statistics, an average of 800 ~ 1000 tornados would occur each
year in the United States alone. Statistics also show that almost 90% of all recorded tornados are rated F2 or less on
the Fujita Scale, i.e., they involve wind speeds less than 160 mph (Bluestein & Golden, 1993). It has been suggested
that it is economically feasible to design built structures such as commercial buildings, bridges, hospitals, power
plants and airports to resist F2 or even stronger tornados. Any such design work, however, requires a keen
understanding of the nature of tornadoes and accurate estimations about the tornado-induced wind loads and wind
field information around built civil structures due to the presence of tornados. With the consideration of buildings as
surface-mounted obstacles, extensive experimental and numerical studies have been carried out in recently years to
investigate the flow-structure interactions between the building models and the incoming turbulent winds as well as
the resultant wind loads acting on the building models (Cermak 1975; Castro and Robins 1977; Hunt et al. 1978;
Martinuzzi and Tropea 1993; Shah and Ferziger 199; Yakhot et al. 2006). Besides the studies using prismatic
obstacles to represent cube-shaped buildings, several studies have also been conducted to consider more realistic
low-rise residential building models with various roof shapes (Kanda and Maruta, 1993; Peterka et al., 1998;
Stathopoulos et al., 2001, and Sousa, 2002; Sousa and Pereira 2004) to quantify the effects of the roof shape on the
wake flow behaviors as well as the resultant wind loads acting on the low-rise building models.
While a great deal about the characteristics of the flow-structures interactions between the building models and
incoming turbulent winds as well as the resultant wind loads have been uncovered by those studies, all the previous
studies mentioned above were conducted with straight-line boundary layer winds. Tornadoes are strong vortices
with significant tangential component, radial inflow/outflow, and vertical updraft/downdraft. Tornado-induced wind
loads and the characteristics of the flow-structure interaction between the buildings with the spiraling tornado winds
would be quite different from those in straight-line boundary-layer flows (Chang, 1971; Ward, 1972; Markowski et
al., 2002).
Surprisingly, very few studies can be found in the literature that specifically addresses the wind loads and flowstructure interactions between building models and swirling, tornado-like winds. Jischke & Light (1983) and
Bienkiewicz & Dudhia (1993) conducted comparative studies to measure the wind loads and surface pressure on
small building models in swirling, tornado-like winds and straight-line winds. They found that the wind loads acting
on the tested models are significantly higher (3 ~ 5 times) in swirling, tornado-like winds. The surface pressure
distributions on the tested building models in tornado-like winds are also quite different compared to those in
straight-line winds. The findings highlights that it is incorrect, at least incomplete, to use a conventional straight-line
boundary layer wind tunnel running with maximum tornado wind speed to estimate tornado-induced wind loads on
built structures. It should also be noted that almost all the previous work on building models in tornado-like winds
were conducted by measuring wind loads and/or surface pressure distributions only (Nolan & Farrell, 1999;
Sengupta et al., 2008; Mishra et al., 2008a,b, Haan et al. 2010), no study has ever been conducted so far to provide
detailed flow field measurements to quantify the characteristics of tornado-like wind and to reveal the dynamical
flow-structure interactions between building models with spiraling, turbulent tornado-like winds.
While tornado-induced damages have been reported for both low-rise buildings and high-rise buildings, the
patterns of the tornado-induced damages to high-rise buildings and low-rise buildings are quite different in nature.
While tornado-induced damages to low-rise residential buildings are usually associated with structural or large
component failures such as complete collapses or roofs being torn-off, tornado damage to high-rise buildings are
generally associated with building envelope failures with no or little structural damage. There has been no report of
structural collapse of tall buildings due to tornadoes. As a result, tornado-induced damages to low-rise buildings are
more fatal and more commonly seen in the aftermaths of intense tornadoes.
In the present study, a comprehensive experimental study was conducted to quantify the characteristics of
swirling, tornado-like winds and to reveal the flow-structure interactions between a low-rise, gable-roof building
model with swirling, turbulent tornado-like winds. The experimental work was conducted by using a large-scale
tornado simulator located in the Aerospace Engineering Department of Iowa State University (ISU). The effects of
important parameters, such as the distance between the centers of the tornado-like vortex and the test model and the
orientation angles of the test model related to the tornado-like vortex, on the wake vortex and flow structures around
the test model as well as the wind loads induced by the tornado-like vortex were assessed quantitatively. While
pressure taps and a high-sensitivity load cell were used to map the pressure distributions around the building model
and to measure the wind loads (forces and moments) acting on the test model induced by tornado-like winds, a
digital Particle Image Velocimetry (PIV) system was used to conduct detailed flow field measurements to quantify
the evolution of the unsteady vortex and turbulent flow structures around the tested building model in tornado-like
winds. To our best knowledge, this is the first effort of its kind. The detailed flow field measurements were
correlated with the wind load measurements to elucidate the underlying physics to gain further insight into flowstructure interactions between the low-rise building model and tornado-like winds in order to provide more accurate
prediction of wind damage potential to built environment with the ultimate goal of reducing life loss, injury and
economic loss that results from violent tornado-like winds.
II. Tornado Simulator and Experimental Setup
ISU Tornado Simulator
The experiments were conducted by using a large-scale tornado simulator located in the Aerospace Engineering
Department of Iowa State University (ISU). Figure 1 shows the schematic and picture depicting the flow circuit and
dimensions of the ISU tornado simulator. A circular duct of 5.49m in diameter and 3.35m in height is suspended
from a heavy duty overhead crane. A 1.83m diameter fan (maximum flow rate is 59.0 m3/s, 125,000 cfm) is
mounted concentrically inside the circular duct to generate a strong updraft. The flow from the fan is redirected
downward in a 0.30m wide annular duct to simulate the rear flank downdraft (RFD) encirclement found in natural
tornados (Chang, 1971). Swirling is imparted to the airflow in the duct by adjusting the angle of the vanes at the top
of the tornado simulator. The downdraft air diverges upon hitting the ground with most of the flow moving inward
toward the fan. The fan updraft stretches the low-level vorticity into a tornado-like vortex. A unique feature of the
ISU tornado simulator is that the tornado-like vortex can travel along the ground plane as the entire fan/downdraftproducing mechanism translates. This translation, along with the fact that there is a good clearance between the
translating duct and the ground plane, allows a wide range of building models to be placed in the path of the
tornado-like vortex for testing. The ISU tornado simulator can generate a tornado-like vortex with a maximum
diameter of 1.2 m and maximum tangential velocity of 14.5 m/s. The maximum swirl ratio (Church et al., 1979)
achieved is 1.14, and the translation speed of the tornado-like vortex can reach up to 0.8m/s. The vortex height can
vary from 1.2m to 2.4m by adjusting the ground plane upward or downward. Further information about the design,
construction and performance of the ISU tornado simulator as well as the quantitative comparisons of the tornadolike vortex generated by using the ISU tornado simulator with the tornados found in nature can be found at Haan et
al. (2008).
For the measurement results given in the present study, the ground floor was fixed at 0.457m below the exit of
the outer duct, and the fan speed was fixed at 20Hz (1/3rd of the full speed). The radius of the tornado-like vortex
core, RO, measured at the 30mm above the ground plate (i.e., the horizontal plane passing the eave of the gable-Roof
building model) was found to be about 0.16m (i.e., RO | 0.16m), at where the maximum tangential speed
( VO | 10 .0 m/s) was observed. The corresponding (radial) Reynolds number of the tornado-like vortex is about
100,000 for the present study. In the present study, the swirl ratio of the tornado-like vortex, which was calculated
according to the expression of S
S VO RO 2 / Q
with Q being the flow rate through the fan, was about 0.1. The
aspect ratio, which is defined as the ratio between the height and the radius of the tornado-like vortex, was about 4.0
for the present study. According to Haan et al. (2008), with the parameter setting described above, the vortex
generated by ISU tornado simulator would be a single cell typed tornado-like vortex, which is also confirmed from
the PIV measurement results to be discussed later.
a.
Schematic diagram
b. Picture of the ISU tornado simulator
Fig. 1: Schematic and picture of the ISU tornado simulator
(a). The gable-roof building model
(b). experimental setup
Fig. 2: The gable-roof building model and experimental setup.
Fig. 2 shows schematic of the low-rise, gable-roof building model used in the present study. The test model has
a base of 90 mm u 90 mm in cross section and an eave height of 30 mm. The gable roof angle is about 35° and the
maximum height of the model is 60 mm. It has been found that tornadoes can occur in a wide range of sizes with
the core radius varying from a few meters to kilometers. By setting a length scale of 1:200, the work produced here
will be corresponding to the study of the flow-structure interactions of a tornado with the core diameter about 65
meters, i.e., approximately the size of the Dallas tornado of April 2, 1957 (Haan et al. 2010), with a low-rise gableroof building of 18 m u 18 m in cross section and an eave height of 6.0m. Figure 2(b) shows the relative position of
the tornado-like vortex center with respect to the center of the gable-roof building model and the orientation angle
(OA) of the building model, where R is the distance between the centers of the tornado-like vortex and the test
building model.
In the present study, three identical gable-roof building models were made for the present study. The first one
was made of thin plastic plates with 89 pressure tabs on the surface to measure the surface pressure distributions
around the house model: the roof contains a total of 43 pressure taps with 24 taps on the side first to be struck with
tangential winds and 19 pressure taps on the opposite side of the roof. The building walls contain nine pressure taps
on each side, and the two triangular gable end sections contain five taps each. The pressure taps were conducted to
two high-speed electronic pressure scanners (Scanivalve-ZOC33/64Px) with the pressure data sampled at the rate of
430 Hz for 30 second for each tested cases.
The second model was constructed as a single unit using rapid prototyping technique with an aluminum rod
through the center as the main structural support. This rod was connected through a hole on the ground plane to a
high-sensitivity force-moment sensor (JR3, model 30E12A-I40) to measure the resultant wind load (both force and
moment) induced by the tornado-like winds. The JR3 load cell is composed of foil strain gage bridges, which are
capable of measuring the forces on three orthogonal axes, and the moment (torque) about each axis. The precision of
the force-moment sensor cell for force measurements is ±0.25% of the full range (40N). In the present study, the
aerodynamic force and moment coefficients were calculated by using the equations of CFX
CFY
FY
1
UVO 2 A
2
;
CFZ
FZ
1
UVO 2 A
2
;
CM X
MX
1
UVO 2 AH
2
;
CM Y
FX
1
UVO 2 A
2
MY
;
and
1
UVO 2 AH
2
MZ
,where U is the air density, A is the projected area of the gable-roof building model along the
1
UVO 2 AH
2
Y-direction as shown in Fig. 2 (a), H is the peak building height. It should be noted that, although the projected
areas of the gable-roof building model along X, Y, and Z directions are different, the same area, i.e., the projected
area along the Y-direction (the direction normal to the tangential velocity in the tornado-like wind), was used in the
present study to normalize the wind loads acting on the gable-roof building model induced by tornado-like winds in
all the directions in order to make direct comparisons of the magnitude of the aerodynamic forces and moments
along X, Y, and Z directions.
CM Z
The third test model one was made of transparent Plexiglas, which was used for PIV measurements of flow
fields around the tested building model. A digital Particle Image Velocimetry (PIV) system was used to conduct
detailed flow field measurements to quantify the evolution of the unsteady vortex and turbulent flow structures
around the gable-roof building model. For the PIV measurements, the flow was seeded with ~ 1 m oil droplets by
using a droplet generator. Illumination was provided by a double-pulsed Nd:YAG laser (NewWave Gemini 200)
adjusted on the second harmonic and emitting two pulses of 200 mJ at the wavelength of 532 nm with a repetition
rate of 10 Hz. The laser beam was shaped to a sheet by a set of mirrors with spherical and cylindrical lenses. The
thickness of the laser sheet in the measurement region was about 1.0mm. A high resolution 12-bit CCD camera
(Pixelfly, CookeCorp) was used for PIV image acquisition with the axis of the camera perpendicular to the laser
sheet. The CCD camera and the double-pulsed Nd:YAG lasers were connected to a workstation (host computer) via
a Digital Delay Generator (Berkeley Nucleonics, Model 565), which controlled the timing of the laser illumination
and the image acquisition. Instantaneous PIV velocity vectors were obtained by a frame to frame cross-correlation
technique involving successive frames of patterns of particle images in an interrogation window of 32u32 pixels.
An effective overlap of 50% of the interrogation windows was employed in PIV image processing. The timeaveraged velocity ( U , V ) distributions were obtained from a cinema sequence of 500 frames of measured
instantaneous velocity fields. The measurement uncertainty level for the velocity vectors was estimated to be within
2.0%.
III. Results and Discussion
3.1 Characteristics of the tornado-like vortex
In the present study, PIV measurements were conducted to quantify the characteristics of the tornado-like vortex
generated by the ISU tornado simulator before the low-rise, gable-roof building models were mounted on the ground
plane. Fig. 3 shows the typical instantaneous and time-averaged PIV measurement results (i.e., velocity distributions
and the corresponding streamlines) at different elevation planes above the test ground plate. It can be seen clearly
that, with the experimental controlling parameters described above for the present study, the vortex generated by the
ISU tornado simulator would be a single-cell-typed tornado-like vortex, which confirmed the conjectures of Haan et
al. (2008) quantitatively. The instantaneous PIV measurement results also reveal clearly that the tornado-like vortex
was highly turbulent with its size, shape and the position of the vortex center varying significantly from one frame to
another. The center of the tornado-like vortex was tracked based on the time-sequence of the instantaneous PIV
measurements to investigate the wandering behavior of the tornado-like vortex. Fig. 4 shows the moving trajectory
of the center of the tornado-like vortex based on 500 frames of instantaneous PIV measurement results in the Z/RO |
0.10 horizontal plane. It can be seen clearly that, while the instantaneous center of the tornado-like vortex was found
to be scattering around the time-averaged center of the vortex randomly, the wandering of the tornado-like vortex
were found to be almost symmetrical to its time-averaged center with the maximum wandering amplitude about 0.6
times of the vortex radius. Further data reduction reveals that the probability density function (pdf) of the
instantaneous location of the tornado-like vortex center during the wandering movement can be represented well by
using a circular Gaussian function with the standard deviation Vx | Vy | 0.3RO .
While the instantaneous PIV measurement results reveal clearly that the tornado-like vortex was highly turbulent
with its center wandering around randomly, flow pattern in the form of a well-defined single counter-clockwise
vortex structure can be identified clearly from the time-averaged PIV measurement results in the horizontal planes at
different elevations. As visualized clearly in the plots, the streamlines of the time-averaged flow field in the core
region of the tornado-like vortex were found to be a group of concentric circles. While the magnitude of the flow
velocity was found to increase almost linearly with the increasing radial distance away from the vortex center, the
radial components of the flow velocity vectors in the core region of the tornado-like vortex were found to be very
small, which is almost negligible compared with the tangential components. Such measurement results indicate that
the flow in the core region of the tornado-like vortex would be very much like that of a potential vortex with the
flow velocity vectors being mainly tangential and the radial components of the flow velocity almost negligible. For
the comparison of the time-averaged flow field at different elevation planes, it can be seen clearly that the size of the
regions with concentric circular streamlines at the center of the vortex was found to become bigger and bigger as the
PIV measurement planes were further away from the ground plate. It indicates that the core region of the tornadolike vortex would become bigger and bigger as the height of the elevation increases, which agrees with the
observations of the funnel-shaped cores of the tornadoes found in nature.
Figure 5 shows the time-averaged PIV measurement results in the Z/RO | 0.10 plane as the center of the
tornado-like vortex moved away from the center of the PIV measurement window. While the streamlines of the
flow field in the core region of the tornado-like vortex were found to be concentric circles, the streamlines in the
outer region of the tornado-like vortex revealed an interesting flow feature of the tornado-like vortex, i.e., a spiral
motion. As shown clearly in the plots, the flow velocity vectors in the region further away from the center of the
tornado-like vortex were found to have significant radial components (VR) in addition to tangential components (VT).
Such PIV measurements suggest a strong spiral flow would be generated in the outer region of the tornado-like
vortex, which made the surrounding air streams flow towards the core of the tornado-like vortex.
The PIV measurements described above revealed clearly that the tornado-like vortex in each horizontal plane
can be divided into two regions: an inner core region and an outer region. In the inner core region, air streams rotate
concentrically with the wind speed increasing linearly with the increasing radial distance away from the rotation
center. The flow inside the tornado-like vortex core rotates like a rigid column. The flow velocity vectors in the
vortex core region were found to be almost tangential with the radial components being almost negligible (i.e., VR |
0). After reaching its peak value at the outer boundary of the tornado-like vortex core, wind speed was found to
decrease gradually with the increasing distance away from the center of the tornado-like vortex in the outer region.
The flow velocity vectors in the outer region of the tornado-like vortex have significant radial components flowing
towards the center of the tornado-like vortex. As a result, a strong spiral motion is generated, which makes
surrounding air streams flowing towards the core of the tornado-like vortex. It should be noted, as revealed clearly
from the PIV measurement results in the vertical plane shown in Fig. 6, there also exists a strong upward flow in the
outer region and downdraft jet flow in the vortex core, which makes the tornado-like vortex being a very complex,
turbulent three-dimensional vortex flow.
In-plane Wind
Speed (m/s)
In-plane Wind
Speed (m/s)
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
50
50
Y (mm)
100
Y (mm)
100
0
0
-50
-50
-100
-100
-150
-100
-50
0
50
100
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
150
-150
-100
-50
X (mm)
0
50
100
150
X (mm)
Z/RO | 0.10
In-plane Wind
Speed (m/s)
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
100
100
50
50
Y (mm)
Y (mm)
In-plane Wind
Speed (m/s)
0
0
-50
-50
-100
-100
-150
-100
-50
0
50
100
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
150
-150
-100
-50
X (mm)
0
50
100
150
X (mm)
Z/RO | 0.40
In-plane Wind
Speed (m/s)
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
100
100
50
50
Y (mm)
Y (mm)
In-plane Wind
Speed (m/s)
0
0
-50
-50
-100
-100
-150
-100
-50
0
50
100
150
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
-150
X (mm)
-100
-50
0
50
X (mm)
C. Z/RO | 0.70
Figure 3: Typical instantaneous and time-averaged flow fields of the tornado-like vortex
100
150
Fig 4: Wandering of the tornado-like vortex center in the Z/RO | 0.10 horizontal plane
In-plane Wind
Speed (m/s)
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
100
100
50
50
Y (mm)
Y (mm)
In-plane Wind
Speed (m/s)
0
0
-50
-50
-100
-100
-150
-100
-50
0
50
100
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
150
-150
-100
-50
X (mm)
a). R/RO | 0.50
In-plane Wind
Speed (m/s)
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
100
100
50
50
0
-50
-100
-100
-100
-50
0
X (mm)
100
150
50
100
150
1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
0
-50
-150
50
b). R/RO | 1.0
Y (mm)
Y (mm)
In-plane Wind
Speed (m/s)
0
X (mm)
-150
-100
-50
0
50
100
X (mm)
c). R/RO | 2.0
d). R/RO | 4.5
Fig. 5: The flow characteristics in the outer region of the tornado-like vortex ( Z/RO | 0.10)
150
Fig. 6: The flow characteristics of the tornado-like vortex
Based on the PIV measurements in different elevation planes, the 3-D flow structure of the tornado-like vortex
was reconstructed, which is given in Fig. 6. A funnel shape profile was plotted in the figure to highlight the size
increase of the core region of the tornado-like vortex as the elevation level increases. The PIV measurement results
in a vertical plane passing the passing the time-averaged center of the tornado-like vortex were also shown in the
plot. The streamlines in the vertical plane also reveal that air streams near the ground far away from the center of
the tornado-like vortex would flow towards the vortex core and turn upward abruptly before reaching the vortex
core. It indicates a radial and upward vertical flow appearing in the region outside the vortex core, as it is expected.
An interesting flow feature is seen in the vortex core region, where flow is found to be a downdraft jet impacting the
ground. As the downdraft jet approaching the ground in the vortex core, it would move outwards radially and
contact with the radial inflow from outer region, both branches will join and move upward. As show clearly in Fig.
6, a “circulation bubble” structure was found to form near the ground in the interface between the upward outside
flow stream and the downdraft jet in the core of the tornado-like vortex. The downdraft jet flow in the tornado-like
vortex core revealed from the present PIV measurements was found to be well consistent to the findings of Wurman
& Gill (2000), who conducted Doppler Radar measurements of the Dimmitt, Texas (2 June 1995) Tornado and
found the downdraft jet flow in the tornado core penetrating up to 400m above the ground level.
ISU Tornado simulator (Z/RO=0.10)
ISU Tornado simulator (Z/RO=0.25)
ISU Tornado simulator (Z/RO=0.40)
ISU Tornado simulator (Z/RO=0.50)
Mulhall Radar Data (Z/RC=0.50)
Spencer Radar Data (Z/RC=0.50)
R/RC
Fig. 7: The tornado-like vortex vs. Mulhall and Spencer tornadoes found in nature (Doppler Radar Data were taken
from Wurman & Alexander, 2005).
Figure 7 shows quantitative comparison of the measured velocity profiles of the tornado-like vortex at different
elevation levels and the field measurement data of two tornados found in nature to demonstrate the similarity
between the tornado-like vortex generated by using the ISU tornado simulator and the tornados found in nature. The
field measurement data of the tornados used for the comparison were made available to ISU researchers by Mr. J.
Wurman under a subcontract of a NSF-sponsored project, which were also published in Wurman & Alexander
(2005). The field data were acquired using Doppler Radar on Wheels observations from the Spencer, South Dakota
tornado of May 30, 1998 and the Mulhall, Oklahoma tornado of May 3, 1999. It can be seen clearly that, even
though the tornado-like vortex generated by the ISU tornado simulator and the two tornados found in nature have
significantly differences in their core diameters (about 0.3 m for the tornado-like vortex generated by the ISU
tornado simulator vs. approximately 400m and 800m for Spencer and Mulhall tornados, respectively), the overall
flow structures scale with each other reasonably well.
3.2 Flow-structure interactions between the tornado-like winds and building model
A. The effects of the distance between the centers of the tornado-like vortex and the test model
Figure 8 shows the profiles of the measured wind loads (both force and moment coefficients) acting on the
gable-roof building model as a function of the distance between the centers of the tornado-like vortex and the test
model. During the experiments, the orientation angle (OA) of the building model related to the tornado-like vortex
was set to 0.0 deg. (i.e., OA=0.0 deg.). As revealed in Fig. 8, since the X, Y and Z components of the measured
aerodynamic forces are positive for all the components, it indicates that the tornado-like wind would try to push the
gable-roof building model tangentially; pull the building model towards the vortex core; and lift the building model
up from the test ground, which are the most commonly damage patterns to the damaged low-rise buildings observed
in the aftermath of intense tornado attacks in nature (Bluestein and Golden, 1993).
1.5
Moment Coefficient CMx, CMy, CMz
Force Coefficients: CFx, CFy, CFz
3.0
2.5
CFz
CFy
CFx
2.0
1.5
1.0
0.5
0
0
1
2
3
4
R/R0
5
6
7
CMz
1.0
CMy
CMx
0.5
0
-0.5
-1.0
0
1
2
3
4
5
6
7
R/R0
(a). Force coefficients
(b). Moment coefficients
Fig. 8: Measured wind loads vs. the distance between the centers of the tornado-like vortex and the test model
The measured force coefficient profiles given in Fig.8(a) also reveals clearly that all three components of the
aerodynamic force acting on the gable-roof building model would increase with the increasing distance between the
centers of the test model and the tornado-like vortex at first, then reach their peak values in the neighborhood of
R/RO | 1.0. As the building model was mounted in the outer region of the tornado-like vortex core (i.e., R/RO > 1.0),
all three components of the aerodynamic force were found to decrease gradually with the increasing distance
between the centers of the test model and the tornado-like vortex. The measurement results indicates that the gableroof building model would experience significant wind loads as it is mounted near the outer boundary of the
tornado-like vortex core. Since the tangential components of the flow velocity vectors was found to be dominant in
the core region of the tornado-like vortex, the X-component of the aerodynamic force, which is along the tangential
direction of the flow streams approaching the gable-roof building model, was found to be greater than the inward
pushing force (i.e., Y-component). It should be noted that, the Z-component of the aerodynamic force (i.e., the
uplift force) was found to be the largest among the three components, which is more than three times stronger
compared with the other two components. It indicates that the low-rise, gable-roof building model would experience
significant roof uplift force as it is located near the outer boundary of the core region of the tornado-like vortex. The
finding is found to be consistent with the fact that the tornado-induced damages to low-rise, gable-roof buildings are
usually associated with torn-off of the roofs of the gable-roof buildings.
The negative sign of the X-component of the measured moment coefficient, CMX, indicates that the gable-roof
building model would likely bend towards the tornado-like vortex core due to the wind loads induced by tornadolike winds. The positive sign of the Y-component of the measured moment coefficient, CMY, indicates that the gableroof building model would bend towards tangential flow direction, as expected. Such collapsed patterns are found
to agree with those of the damaged buildings observed aftermath of tornado attacks in nature. The magnitude of the
Z-component of the moment coefficient, CMZ, was found to be much smaller than other two components, which is
negligible. Similar to the measured force coefficients, the measured moment coefficients were also found to reach
their peaks at the location of R/RO | 1.0. The force and moment measurement results suggest that low-rise, gableroof buildings would most likely to fail when they are located near the outer boundary of the tornado core (i.e. R/RO
| 1.0), where the wind reaches its maximum speed.
For a gable-roof building model placed in a straight line wind, Holmes (1993) reported a symmetric pattern for
the surface pressure distribution around the gable-roof building model, which differs from the present measurement
results as shown in Fig. 9 significantly. From the measured surface pressure coefficient distributions given in Fig. 9,
it can be clearly seen that the surface pressure distribution around the gable-roof building model is not symmetric
about the central plane (Z=0 cross plane) of the building model due to the swirling motion of the approaching
tornado-like wind. The measured surface pressures on the walls and roof at the windward side were found to be
much higher than those on the walls and roof at the leeward side, which results in the pushing force acting on the
test model toward flow direction. The lower pressures on the roof and the leeward walls are due to the occurrence of
the large-scale flow separations in those areas, which can be seen clearly in the PIV measurements shown in Fig. 10.
The pressure distributions on the walls and roofs in the outer-side regions (i.e., regions farther away from the
tornado-like vortex core) were found to be higher compared to those in the inward side region (region closer to the
vortex center). The unbalanced pressure distribution on the inward side and the outer-side wall results in a net force
acting on the building model, which pulls the building model towards the core of the tornado like vortex, as it is
found in the force measurements. All pressure values on the roofs were found to be negative. Such distribution on
the roofs is consistent with the well-known fact of that there exists a significant low-pressure region in the core
region of a tornado-like vortex.
When the test model was placed inside the core region of the tornado-like vortex at R/RO | 0.5 as shown in
Fig.10(a), a peak pressure coefficient value of -1.6 was found at the wall on the windward side because the incoming
flow impacts on the wall directly. In a straight-line flow, the pressure at the stagnation point on the wall on the
windward side should be positive, but it is not the case in tornado like wind. Negative pressure values were found on
the windward side, because there was considerable low pressure in the core of a tornado-like vortex. A minimum
pressure coefficient value of -4.2 was found on the roof at the leeward side due to the significant separation vortex
near the roof. When the gable-roof building model was placed at out of the core region of the tornado-like vortex at
R/RO | 1.5 as shown in Fig.10(b), the pressure coefficients on the surface of the test model were found to increase
significantly at all the corresponding positions compared with those of the case R/RO | 0.5. The peak pressure
coefficient value of -0.2 was still found on the wall at the windward side. A minimum pressure coefficient value of 3.4 was found at both the roof and lateral wall on the leeward side.
50
100
0
-1.2
- 1.
40
-0.20
Wall
60
0
-0.6
80
100
-100
-50
-0.60
-3.80
-3
-3
.6
.2
0
0
-4.20
-4.20
X (mm)
-0.20
-0.60
-1.00
-1.40
-1.80
-2.20
-2.60
-3.00
-3.40
-3.80
-4.20
-0.40
40
-0.60
0
CP
-0.40
-0.60
-2.
80
-0.20
20
-0.60
-50
0
100
-100
-3.0
80
-0.20
-0.60
-3.80
-3.60
-3
.4
0
Wall
60
60
-0. .60
-0
-0.60
-1.60
0
-0.20
-0.4
0
-4.0
-3 -4.00 0
.6
0
-4.00
0
Y (mm)
-1.8
20
-20
-1.20
0
-0.20
-0.60
-1.00
-1.40
-1.80
-2.20
-2.60
-3.00
-3.40
-3.80
-4.20
-4.00
-2.0
20
-4.
-40
CP
Wall
Roof
-1.00
0
0
40
-3.80
-2.4
-20
-2
.
- 2 80
.6
0
-40
-60
Wall
Roof
-1.40
-0
.
-0 40
.4
0
-60
Tornado-like
Vortex core
-1.40
-80
-0.60
Tornado-like
Vortex core
-3.80
-3.80
Y (mm)
-80
0
50
X (mm)
(b). R/R01.50
(a). R/R00.50
Fig. 9: Measured pressure distributions around the gable-roof building model
100
With the findings derived from the wind load measurements in mind, PIV measurements were conducted to
quantify the behavior of the flow structures around the gable-roof building model to elucidate underlying physics to
improve our understanding about the characteristics of the flow-structure interactions and the wind loads acting on
the test model in tornado-like winds. Fig. 10 shows both the side view and top view of the flow structures around
the gable-roofed building model as the test model was located at the vortex core (Figs. 10a and 10d), the boundary
between the inner core and outer regions of the vortex, i.e., where local wind speed reaches maximum, (Figs.10b
and 10e), and outer region (Figs. 10c and 10f) far away from the center core of the tornado-like vortex.
Compared to those in straight-line winds (Sousa et. al, 2004), the wake vortices and flow structures around the
building model were found to become much more complicated in tornado-like winds. In the vertical plane (i.e., side
views), as shown in Fig. 10 a, two large vortices were found to be formed above the roofs of the test model, which
was located in the core region of tornado-like vortex (R/R00). Even though the flow velocity is relatively low, it is
still visualized clearly that the flow streams near the ground move towards the test model and turn upward abruptly
before reaching the walls. Two large vortices were formed in the central region near the roofs when the downdraft
jet met with the incoming flow from outer region. At the same location, an almost axisymmetric flow pattern on the
horizontal plane (Z=15mm) was visualized in Fig. 10d. The streamlines farther away from test model were found to
be in a spiral flow pattern. Since the local wind speed in the core of the tornado-like vortex was relatively low, the
aerodynamic forces and moments acting on the gable-roof building model were relatively small when the model was
placed inside the core region of the tornado-like vortex, which is confirmed from the wind load measurements
shown in Fig. 8.
Fig. 10: Flow structures around the gable-roofed building model in tornado-like winds. (a), (b) and (c) for side
views; (d), (e) and (f) for top views.
Figs. 10b and 10e show the top view and side view of the flow field around the gabble roof building model
when the test model was mounted at the outer boundary of the tornado-like vortex core (i.e., R/RO | 1.0). As
described above, this is the position where maximum wind speed was observed at the center of the model. The
maximum wind loads were also found at this location. Correspondingly, the large high-speed region appeared above
the top of the test model in the vertical measurement plane. Two vortices were found in the vertical measurement
plane. One was generated near the leeward side of the roof due to the separation there, which corresponded to a low
pressure region at the leeward roof, and is confirmed by the pressure measurements shown in Fig. 9. The other was
generated at the far leeward side near ground. The streamlines in the horizontal plane demonstrated only one large
12
American Institute of Aeronautics and Astronautics
vortex structure instead of two, as one might expect to see at the leeward side of the test model. This is because the
wind speed at the outer side is much higher than the inward side. Air flow at inward side has no chance to turn at the
corner into the leeward region. The incoming streamlines are almost perpendicular to the windward wall of the test
model, and then turn a right angle toward the inward side after passing the model induced by the spiral flow feature
of the tornado-like vortex.
The time-averaged flow fields around the test model for R/RO | 4.6 are shown in Fig. 10c and 10f. As visualized
clearly in the vertical measurement plane, the separation vortex structure for this case is quit different from the other
cases. This discrepancy may be attributed to low Reynolds number effects. For a low Reynolds number flow, the air
flow is more readily to separate when passing over the roof of test model. The separation region, with low speed,
became larger at leeward side. In the horizontal plane, as visualized in the figures, the wake vortex structures at the
leeward side of the model were found to be elongated significantly because of the spiral motion in the outer region
of the tornado-like vortex. As the local wind speed was found to become smaller with the increasing radial distance,
the aerodynamic forces and resultant moments were found to decrease slightly with the increasing distance, and
were revealed from the wind load measurement results shown in Fig. 8.
B-2. The effects of the orientation angle of the test model in tornado-like winds
The effects of the orientation angle of the test building model on the characteristics of the wake vortices and flow
structures around the test model as well as the resultant wind loads (forces and moments) acting on the test model
were also investigated in the present study. The definition of the orientation angle (OA) of the model related to the
tornado-like vortex is given schematically in Fig. 2. During the experiments, the test model was mounted at the
location of R/RO | 1.0.
Fig. 11 shows the measured wind loads (i.e., aerodynamic force and moment coefficients) as functions of the
orientation angle in tornado like wind. The maximum force coefficient, CFy, in the tangential flow direction (Xdirection) and the corresponding moment coefficient, CMy, were found at OA 30.0~45.0 degrees. The force in Ydirection is always positive due to the radial velocity component in tornado-like wind. The maximum force in Ydirection was observed at OA=0~15 degrees. Correspondingly, the maximum moment coefficient in the X-direction
was found at the same orientation angle. The maximum uplift coefficient appears at OA=15~30 degrees. The lowest
force coefficients appeared at around 75 degrees in all three directions. The torsion moment in the Z-direction was a
small and did not vary much with different orientation angles.
2.0
2.5
Force Coefficient
1.6
Moment Coefficient CMx, CMy, CMz
CFz
CFy
CFx
1.2
0.8
0.4
0
-0.4
0
15
30
45
60
Orientation Angle (deg)
75
90
CMz
CMy
CMx
2.0
1.5
1.0
0.5
0
-0.5
-1.0
-1.5
0
15
30
45
60
75
90
Orientation Angle (Deg)
(a). force coefficient
(b). Moment coefficient
Fig. 11: The measured wind loads vs. the orientation angle
The PIV measurement results given in Fig. 11 elucidate more details about the variations of the wake vortex and
flow structures around the test model with different orientation angles, which might be used to explain the
characteristics of the wind loads induced by tornado-like winds. It can be assumed that all wind loads were induced
by the impacting of air flow on the test model. Taking the flow field boundary as a control volume boundary in a 2D plane, the momentum difference in the Y direction would result in the force in the Y direction. Assuming that the
Y-direction flow comes in from the bottom boundary with the same velocity distribution, then the velocity of the
flow going out from the top boundary could indicate the momentum change to some extent. For OA=0deg, the
velocity of the flow leaving the control volume from the top boundary is relatively low compared with other
13
American Institute of Aeronautics and Astronautics
orientation angles, which means this momentum change would lead to a large force in the Y-direction acting on the
test model. This analysis was confirmed with the force measurement results. For OA = 15deg, the largest wake
region was shown near the sidewall at the inward side. Correspondingly, the maximum force coefficients in the Y
and Z directions and the momentum coefficient in the X-direction were also found at this angle. Even though the
relationship between wind load measurements and flow measurements in one plane is not straight forward, one can
still get some clues to understand the wind loads acting on the test model. The wake pattern for OA = 30deg, OA =
45deg, OA = 60deg and OA = 75deg are very similar. The area of wake region for OA=45 deg was the smallest one
and corresponded to the largest force in the X-direction as displayed in Fig. 11. Since the flow structure would be
drastically complicated due to the existence of the roof, the averaged velocity field in only one horizontal plane
cannot give a complete explanation of the force behavior acting on the test model. More comprehensive and detailed
studies would be needed to elucidate the underlying physics of the orientation effect.
Streamwise
Velocity m/s
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9
Streamwise
Velocity m/s
7 m/s
100
7 m/s
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9
100
50
50
OA=15 deg
Y (mm)
Y (mm)
OA=0deg
1:100 Gable-roofed
building model
0
-50
1:100 Gable-roofed
Building model
0
-50
-100
-100
-100
-50
0
50
100
150
-100
-50
X (mm)
Streamwise
Velocity m/s
Streamwise
Velocity m/s
7 m/s
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9
100
150
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9
7 m/s
100
50
50
OA=45deg
Y (mm)
OA=30deg
1:100 Gable-roofed
Building model
0
-50
1:100 Gable-roofed
Building model
0
-50
-100
-100
-100
-50
0
50
100
150
-100
-50
0
X (mm)
50
100
150
X (mm)
(c) OA=30.0 deg.
Streamwise
Velocity m/s 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9
(d) OA=45.0 deg.
Streamwise
Velocity m/s
7 m/s
0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5
7 m/s
100
100
50
50
OA=60deg
OA=75 deg
Y (mm)
Y (mm)
50
(b) OA=15.0 deg.
100
Y (mm)
0
X (mm)
(a) OA=0.0 deg.
1:100 Gable-roofed
Building model
0
1:100 Gable-roofed
Building model
0
-50
-50
-100
-100
-100
-50
0
50
100
X (mm)
(e) OA=60.0 deg.
150
-100
-50
0
50
X (mm)
(f) OA=75.0 deg.
Fig. 12: Flow field around the test model vs. the orientation angles
14
American Institute of Aeronautics and Astronautics
100
150
B-3. The vortex and flow structures around the test model
During the experiments, PIV measurements were also conducted at different cross planes parallel to the ground at
different elevations and different vertical cross planes parallel to the lateral walls. Fig. 13(a) shows typical PIV
measurement results at different elevations and the central vertical plane with the test model mounted at R/RO = 1.0
with the zero orientation angle (OA=0). As revealed clearly from the PIV measurements results, two main vortex
flow structures were observed at the leeward roof and the leeward wall of the model, respectively. In the vertical
plane, after the flow passed the roof top, the streamlines surrounding the vortex at the leeward roof sank down
abruptly into the vortex near the leeward wall. It seems the flow retrieved in that vortex and blew downstream. The
maximum velocity region in the vertical plane appeared above the top of the test model. In the horizontal plane, the
streamlines surrounding the vortex structure turned a right angle toward the tornado core with a high speed after
passing the leeward wall.
Based on the PIV measurements taken at different elevation planes and vertical planes, the flow structures around
the test model in tornado-like winds were reconstructed. Fig. 13 (b, c) show the perspective views of the
reconstructed wake vortex and flow structures around the test model in tornado-like winds in the terms of ISOsurfaces of local swirling strength (Adrian et al. 2000 and Sousa et al. 2002) in the Y and Z directions. The
imaginary part of the complex eigenvalue pair of the velocity gradient tensor matrix was referred to as the local
swirling strength of the vortex. The 3-D features of the wake vortex and the flow structures around the test model in
the tornado-like vortex can be clearly visualized from the 3-D view of the reconstructed flow field.
a.
b.
c.
Fig. 13: Reconstraucted 3-D flow field with streamlines and swirling strength contours in Y- and Z-direction
15
American Institute of Aeronautics and Astronautics
IV. Conclusion
An experimental study was conducted to quantify the characteristics of wake vortex and flow structures around a
low-rise gable-roofed building model as well as the wind loads acting on the test model induced by tornado-like
winds. In addition to measuring wind loads (i.e., aerodynamic forces and moments) and pressure distribution acting
on the test model using a high-sensitivity force-moment sensor and a pressure scanner respectively, a digital Particle
Image Velocimetry (PIV) system was used to conduct detailed flow field measurements to quantify the evolution of
the unsteady wake vortex and flow structures around the test model in tornado-like winds. The detailed flow field
measurements were correlated with the wind load measurement data and pressure data to elucidate underlying
physics to quantify the dynamics of flow-structure interactions between the tested building model and tornado-like
vortex in order to provide more accurate prediction of wind damage potential to built environment with the ultimate
goal of reducing life loss, injury, casualty and economic loss that results from violent tornado-like winds.
The PIV measurement results reveal clearly that a tornado-like vortex is a very complex, highly turbulent, threedimensional vortex flow. In addition to a strong upward flow in the outer region and downdraft jet flow in the vortex
core along vertical direction, axisymmetric flow pattern in the form of a well-defined vortex structure can be seen
clearly in the horizontal planes. The flow field of a tornado-like vortex in each horizontal plane could be divided
into two regions: an inner core region and an outer region. In the inner core region, air streams were found to flow
concentrically with the wind speed increasing linearly with the increasing radial distance away from the center of the
tornado-like vortex. The airflows inside the tornado-like vortex core rotate like a rigid column with the flow velocity
vectors being tangential and the radial components being negligible (i.e., Vr |0). Wind speed was found to reach its
maximum value at the outer boundary of the tornado-like vortex core, and then, begins to decrease with increasing
distance from the center of the vortex core in the outer region. The flow velocity vectors in the outer region of the
tornado-like vortex were found to have significant radial components, which results in a strong spiral motion that
sucks surrounding air flowing towards the vortex core.
The wind loads (both force and moments) acting on the test model induced by tornado-like wind were found to
vary significantly with the position of the test model relative to the center of the tornado-like vortex. The maximum
aerodynamic force in tangential direction and radial direction and corresponding moments were found to reach their
maximum values when the test model was mounted on the outer boundary of the tornado-like vortex core. Unlike
those in straight-line winds, the wake vortex and flow structures around the test model in tornado-like winds were
found to become quite unsymmetrical related to the approaching flow. The orientation angle of the test model
related to the tornado-like vortex was found to have considerable effects on the wake vortex and flow structures
around the test model as well as the resultant wind loads. The lift force acting on the gable-roofed model, which is
the most important force according to the damage pattern induced by tornado, was found to be maximum at
orientation angle OA | 15 ~ 30 degrees. Another significant force in tangential direction reached the maximum
value at orientation angle OA | 30 ~ 45 degrees.
Acknowledgments
This research was funded by National Oceanic and Atmospheric Administration (NOAA)-Award #
NA060AR4600230 with Mr. John Gynor as the program manager. The authors also want to thank Mr. Bill Richard
of Iowa State University for his help in manufacturing the test model.
References
1
Adrian, R.J, Christensen, K.T, Liu, Z.C., “Analysis and interpretation of instantaneous turbulent velocity fields,”
Experiments in Fluids, Vol.29, 2000, pp. 275-290.
2
Bienkiewicz B., Dudhia, P., “Physical modeling of tornado-like flow and tornado effects on building loading,” Proceeding
th
7 US National Conference on Wind Engineering, 1993, pp. 95-106.
3
Bluestein, H.B., Golden, J.H., “A review of tornado observations”, The Tornado: Its Structure, Dynamics, Prediction and
Hazards, Geophys. Monogr., No. 79, Amer. Geophysics, Union, 1993, pp. 19-39.
4
Castro, I.P. and Robins, A.G., “The flow around a surface-mounted cube in uniform and turbulent streams”, J. Fluid Mech.
Vol. 79, 1977, pp. 307–335
5
Cermak, J.E., “Applications of fluid mechanics to wind engineering-a Freeman Scholar lecture”, Journal of Fluids
Engineering, Vol.97, 1975, pp. 9-38.
16
American Institute of Aeronautics and Astronautics
6
Chang, C.C., “Tornado effects on building s and structures with laboratory simulation”, Proceedings of 3rd International
Conference on Wind Effects on Buildings and Structures, Tokyo, Japan, 1971, pp.151-174.
7
Church, C.R., Snow, J.T. and Agee, E.M., “Tornado vortex simulation at Purdue University,” Bull. Amer. Meteor. Soc. Vol.
58, 1977, pp.900-908
8
Davies-Jones R.P., “Laboratory simulations of tornadoes”, Proceedings, Symposium on Tornadoes, Texas Tech. Univ.,
1976, pp. 151-174.
9
Haan, F.L., P.P. Sarkar, W.A. Gallus, “Design, Construction and Performance of a Large Tornado Simulator for Wind
Engineering Applications,” Engineering Structures, Vol.30, 2008, pp.1146-1159.
10
Holmes, J.D., “Wind load on lowrise building - A review,” CSIRO, Division of Building Research, Highett, Victoria,
Australia, 1993.
11
Hunt, J.C.R., Abell, C.J., Peterka, J.A. and Woo, H., “Kinematical studies of the flows around free or surface-mounted
obstacles: applying topology to flow visualization”, Journal of Fluid Mechanics, Vol. 86, No.1, 1978, pp. 179-200.
12
Jischke, M.C. and Light, B.D., “Laboratory simulation of Tornadic Wind Loads on a rectangular Model structure,” Journal
of Wind Engineering and Industrial Aerodynamics, v.13, 1983, pp. 371-382.
13
Kanda, M. and Maruta, E., “Characteristics of fluctuating wind pressure on long low-rise buildings with gable roofs,” J.
Wind Eng. Ind. Aerod., Vol.50, 1993, pp.173–182.
14
Kim, K.C., Ji, H.S. and Seong, S.H., “Flow structure around a 3-D rectangular prism in a turbulent boundary layer,” J. Wind
Eng. Ind. Aerod., Vol.91, 2003, pp. 653-669.
15
Krajnovic, S. and Davidson, L., “Large-eddy simulation of the flow around a surface-mounted cube using a dynamic oneequation subgrid model,” First Intl. Symp. on Turbulence and ShearFlow Phenomena (ed. S. Banerjee & J. Eaton), Begel House,
Inc., New York, 1999.
16
Martinuzzi, R. and Tropea, C., “The flow around surface-mounted, prismatic obstacles placed in a fully developed channel
flow,” J Fluids Eng., Vol.115, 1993, pp. 85–92.
17
Shah, K.B., and Ferziger, J.H., “A fluid mechanicians view of wind engineering: large eddy simulation of flow past a cubic
obstacle,” J. Wind Eng. Ind. Aero., Vol.67/68, 1997, pp. 211–224.
18
Sousa, J.M.M., Freek, C., Pereira, J.C.F., “PIV measurements of turbulence statistics in the three-dimensional flow over a
surface mounted obstacle,” Adrian RJ, Durao DFG, Durst F, Heitor MV, Maeda M, Whitelaw JH (eds) Laser techniques
applied to fluid mechanics, Springer, Berlin Heidelberg New York, 2000, pp. 157–176.
19
Sousa J.M.M., “Turbulent flow around a surface-mounted obstacle using 2D-3C DPIV,” Experiments in Fluids, Vol.33,
2002, pp. 854-862.
20
Sousa, J.M.M. and Pereira, J.C.F., “DPIV study of the effect of a gable roof on the flow structure around a surface-mounted
cubic obstacle,” Experiments in Fluids, Vol.37, 2004, pp.409-418.
21
Wurman J., and Gill S., “Finescale Radar Observations of the Dimmitt, Texas (2 June 1995), tornado”, AMS Monthly
Week Review, vol.128, 2000, pp. 2135-2164.
22
Wurman J, Alexander CR, “The 30 May 1998 Spencer, South Dakota, Storm. Part II: Comparison of observed damage and
radar-derived winds in the tornadoes”. Ames Monthly Weather Review, Vol.133, 2005, pp. 97-118.
23
Yakhot, A., Anor, T., Liu, H.and Nikitin, N., “Direct numerical simulation of turbulent flow around a wall-mounted cube:
spatio-temporal evolution of large-scale vortices,” J. Fluid Mech., Vol.566, 2006, PP.1-9.
24
Yang, Z., Sarkar, P. and Hu, H., “Flow around a high-rise building model in tornado-like wind”, 46th AIAA Aerospace
Sciences Meeting and Exhibit, AIAA-2009-0749, Orlando, Florida, 2009, pp. 1-16.
17
American Institute of Aeronautics and Astronautics