Re f

Ser
TH1 National Research Conseil national
N21 d 1*1 council ~anacia
de recherches Canada
no, 3130
I
BL DG
- - - -
Institute for lnstitut de
Research in recherche en
Construction construction
Wind Engineering Studies on
Tall Buildings - Transitions in
Research
by Appupillai Baskaran
Appeared in
Building and Environment
Volume 28, Number 1
p. 1-19, 1993
(IRC Paper No. 3130)
Reprinted with permission from
Pergamon Press
NRCC 35497
ANALYZED
MAR 8 4 1993
Building andEnvironmenf, Val. 28, No. 1 , pp. 1-19, 1993
Printed in Great Britain.
0j6&1323/93 55.00+0.00
Pergamon Press Ltd.
Wind Engineering Studies on Tall
~ui l di n~s-~ranGt i ons in Research
APPUPILLAI BASKARAN*
Development of new building materials and advances in architectural concepts have led to light
weight and more unconventional buildings. Consequently, unexpected wind forces may act on these
structures. Experimental results from wind-tunnel studies are considered as one reliable source for
the wind loading information. An extensive literature survey has been conducted; this paper reviews
some of the selected studies to show the research transitions in wind engineering studies of tall
buildings. Ten dzferent stages and an establishedpattern have been identzfied in wind engineering
studies of tall buildings.
1. INTRODUCTION entire building. Today most partitions are removable,
TALL buildings are the unique North American con-
tribution to world architecture [I]. The ten tallest build-
ings in the world are listed in Fig. 1 ; most of these are in
two major American cities, New York and Chicago. Tall
buildings were constructed in response to social and econ-
omic needs of a particular place and time. Their unique-
ness attracts tourists and increases a city's image. More
usefully, they satisfy the increasing office space demands
of major cities.
More and more developers are erecting tall buildings
in comparison to the past., Architects and engineers plan
buildings of irregular shape with extremely light exterior
skins or large H/B ratios. Any increase in building height
increases the importance of wind loading. ~ u i l d i i ~ codes
and wind standards are formulated to provide design
information for wind engineering practice. However,
they provide very little design guidance for wind loads
on buildings of unusual geometrical shapes or structural
properties. This inadequacy in building codes and stan-
dards redirects designers and engineers to rely on wind-
tunnel model tests for evaluating wind effects on very tall
buildings.
Advances in construction technology and development
of modern building materials also enhance the growth of
tall buildings and affect the conventional design pro-
cedures. Thus there are significant differences in the con-
struction of tall buildings today, in comparison to build-
ings constructed during the early thirties ; a few of these
changes are grouped below :
A 20 feet column spacing was found to be adequate
for office spacing half a century ago. Today, a mini-
mum of about 40-50 feet is considered adequate.
Partitions were generally made of solid masonry from
floor to floor, adding considerably to the rigidity of the
*Institute for Research in Construction, National Research
Council of Canada, Montreal Road, Ottawa, Ont., Canada KIA
0R6.
and are therefore lighter and more.flexible.
a Exterior wall detail was generally made of solid
masonry or stone, with the opening a small percentage
of the total wall surface. In contrast, the modern glass
curtain wall systems significantly reduces structural
stiffness of the outer skin.
a Use of mechanical systems, HVAC facilities and pro-
vision for modern equipment such as computers are
noticeably increased in comparison to the past.
The external shape of the building is more irregular
today than in the past, to maximize the working space
as well as to express the architectural dignity of one
building from others.
Along with the changes in construction, evaluation
methodologies have also changed. In this paper a sys-
tematic attempt has been made to show the changes in
wind engineering research on tall buildings by reviewing
a few carefully selected studies. Wind effects on buildings
are mostly quantified by using wind tunnel measure-
ments, and research transitions in wind tunnel studies are
presented in the following section. Efforts are also made
in identifying the various research stages in full scale
monitoring of wind effects on buildings and this is pre-
sented in Section 3. Section 4 summarizes ten different
stages in wind engineering research, along with an estab-
lished pattern for tall building studies in the wind tunnel.
2. REVIEW OF SELECTED WIND TUNNEL
STUDIES
General reviews of wind effects on tall buildings in
relation to structural design factors can be found else-
where [2, 31. Similarly, the research developments on tall
buildings are periodically compiled and presented by the
Council on Tall Buildings and Urban Habitat. So far,
the Council has released a five volume monograph on the
planning and design of tall buildings, published from
1978 to 1981 [4], followed by developments in tall build-
ings [5] and advances in high-rise buildings [6]. In this
2 A. Baskaran
1 1. Sears Tower 443 Chicago 1974 1
- -
3. World Trade 415 New York 1973
Center Souh
4. Empire State 381 NewYork 1931
5. Central Plaza 372 Hong Kong 1992
16. Bank olchlna 367 Chicago 1988 1
18. John Hancxldr 343 Chicaao 1968 1
10 Firs! tnteratale 310 Los Angeles 1 ~ 9 0
I
1 1. Miglin Beitler 594 Chicago 19m 1
( 2. Tour Sans Fin 400 Paris M A 1
1 3. Taiwan Tower 331 Koahsiuno t993 1
--
(Source: Engineering News Review - Nov. 15,1990)
Fig. 1. The ten tallest buildings in the world.
paper, after an extensive literature survey, a few studies
have been selected and reviewed to emphasize the tran-
sition of methodology employed for wind engineering
studies of tall buildings. The most typical buildings of
the century are considered, such as the Empire State
Building (New York), the World Trade Center Towers
(New York), the Sears Tower (Chicago) and the newest
example, the Bank of China (Hong Kong). These studies
may represent the design methodologies of 1930s, 1960s,
1970s and 1980s.
2.1 Empire State Building
Dryden and Hill [q undertook the first significant wind
tunnel study on the Empire State Building. A 1:250
model made of rolled aluminum plates 114 in. thick was
constructed to represent the 1250 ft (381 m) high building.
Both wind-induced pressures and overturning moments
on the building were examined in a 10 ft. wind tunnel at
the National Bureau of Standards. Pressure on the model
was measured at three different elevations (36th, 55th
and 75th floors) by connecting a pressure gauge to exter-
nal holes with rubber tubing. In total there were 34 pres-
sure taps on each floor level and the model was rotated
through 180 degrees to study the effect of wind azimuth
angle. The test was repeated at three wind speed levels :
40, 60 and 80 ft/sec (approximately 12, 18 and 24 m/s).
Pressure coefficient distributions at three different
levels are shown in Fig. 2, for two typical wind directions.
Positive pressure was measured for the windward walls,
whereas a more or less constant suction was found for
other walls. The situation becomes more complicated
when the wind arrives at an oblique angle to the build-
ings. In addition to the measurement of external pressure
distributions, the base overturning moments were also
measured and are presented in Fig. 3. Coefficients for
two principal sway directions are shown. The measured
moments are normalized by the velocity pressure, rep-
resentative area and arm length which is taken as 4.4 ft.
and 2.0 ft. (model scale) for x and y directions, respec-
tively. This study, which was the first of this kind, shows
an appreciation of the effect of wind loads on the building
design.
2.2 World Trade Center Towers
The twin World Trade Center Towers of New York
attracted significant attention from wind engineers
before, and even after, their construction. The wind effect
on the towers were examined at Colorado State Uni-
versity (CSU) and confirmation tests were carried out at
the National Physical Laboratory (NPL). Wind effects
on the plaza level environment were measured at the
University of Western Ontario (UWO). This was the first
major tall building project in which the simulation of
natural wind turbulence was introduced.
2.2.1. Windloadon towers. A model of the twin towers,
including the low-rise plaza level buildings and the sur-
roundings, was tested at CSU with the shear flow tur-
bulence as a simulation of natural wind [8]. A geometric
scale of 1 : 500 was used for the model simulation. About
250 pressure taps were connected to a scanivalve pressure
measuring system. The distance between the towers was
varied to provide a guideline for placing the twin towers
relative to each other. Pressure measured at CSU was
confirmed by the NPL study. A static wind load of 55
psf (1 psf = 48 Pa) for the top 100 ft. and 45 psf for the
remaining portion of the tower was recommended from
the wind tunnel test results for the 100 year wind of
Wind Engineering Studies on Tall Buildings
A. Baskaran
Angle of face to wind Angle of faoe to wind
Fig. 3. Measured overturning moment for x and y directions for different wind direction [7].
140 mph (62.6 m/s). Visco-elastic damping units were cal of the wind coming across the Hudson River from
suggested to limit the maximum deflection to 318 in. (9.5 Jersey City, and
mm) per story at the same bench-mark wind speed. Exposure 111-representing the Manhattan fetch,
typical of the wind coming over heavily built-up ter-
2.2.2 Dynamic response of towers. The study at NPL rain.
consisted of two parts: a pressure test, which more or
less confirmed the results of CSU, and the prediction
of wind-induced dynamic response using a 1 : 400 scale
aeroelastic model [9]. The building model was con-
structed of a light timber frame covered with thin
plywood, designed as a rigid body with proper mass
simulation and the required stiffness. Variable damping
was prdvided by a system under the wind tunnel floor via
an,extended aluminum tube from the model connected
to electromagnets (Fig. 4).
Its response to wind was observed in an idealized
smooth flow and in two kinds of turbulent flow, homo-
geneous turbulence and shear turbulence, both created
by grids installed at the front end of the wind tunnel test
section. Configurations of an isolated tower and twin
towers were tested and it was concluded that the twin
towers were unlikely to undergo any adverse wind effect
from aerodynamic instability for wind speeds below 100
mph (160 km/li or 45 m/s) on either configuration. How-
ever, in order to limit amplitudes at the tower top to less
than 10 ft. for wind speeds up to 150 mph, a very high
damping (approximately 12% of critical) is required.
All three exposure conditions were physically modelled
in the wind tunnel with the surrounding topography to
a radius of 1600 ft. in order to include the local flow
characteristics. Such a precise terrain simulation was
one of the novel points of this particular study. Isolated
pressure signals were collected by tubing with a scani-
valve pressure transducer. Mean, RMS and peak pres-
sure coefficients were obtained based on the wind speed
at the top of the main towers. For the design of window
panels and exterior cladding elements, gust factors were
obtained. A summary of the measured peak factors is
given in Fig. 6 which shows an average value of about
4.5 for all building elements. The positive peak pressure
factors were about 4 to 5, whereas the main peak suction
factor was typically in excess of 7 for some locations. The
largest pressures, suctions and their fluctuations were
observed when the wind came from the SW quadrant,
which is over the Exposure I.
Flow visualization and velocity measurement were car-
ried out to establish the acceptable pedestrian level wind
conditions. The flow visualization was performed by gen-
erating smoke in the wind tunnel, whereas the velocity
measurement was done by using a hot wire anemometer
2.2.3 Study of plaza level buildings. The wind engin-
system. All three flow regimes as discussed in the previous
eering study for the plaza level buildings consisted of
section were considered. For each wind azimuth angle,
two parts : measurement of wind-induced pressure on the
20 observation points were chosen at a full-scale elevation
plaza level buildings for the design of exterior cladding,
of 6 to 12 ft. Results indicate generally greater wind
and pedestrian level environmental wind conditions
speeds near the main towers. The passageways, especialy
around the towers. These experiments were carried out
between the U.S. Custom Building, the Towers and the
at the Boundary Layer Wind Tunnel Laboratory of the
University of Western Ontario [lo].
Hotel building (see Fig. 5) show the highest mean speed
ratios, particularly for WNW to SSW winds. The peak
Using a linear scale of 1 : 400, the four main buildings
values of the wind speed ratio vary from 0.4 to 1.2 for
and the surroundings were modelled. There were 45 pres-
the positions examined.
sure taps on each building model. The upstream terrain
conditions of the site vary depending on the wind direc-
tion. As shown in Fig. 5, three different exposure con-
2.3 Sears Tower
ditions were simulated in the wind tunnel. They are :
To this date, the 443 m tall Sears Tower holds the
title of the world's tallest office building. The proposed
Exposure I-representing the open water fetch, typi- Miglin-Beitler Tower (585 m) upon its completion will
cal of the wind coming across the Upper Bay and move the Sears Tower to the second place [ll]. A corn-
along the Hudson River ; prehensive wind engineering study was performed at the
Exposure 11-representing the Jersey City fetch, typi- University of Western Ontario [12]. The Sears Tower
Wind Engineering Studies on Tall Buildings
-
NPL 7' x 7' wind tunnel
working section
-
model constructed
frame mvered with
It-
2" dia aluminum tube
coil springs provide
required stiffness I
Fig. 4. General arrangement of the 11400 scale Model tested at NPL 191.
project comprised a study of local wind climate, measure-
ment of wind-induced pressure loads and determination
of wind induced dynamic response of the building and
the prediction of wind loads based on these results [13].
This particular project more or less established a pat-
tern for a wind engineering study of tall buildings and was
adapted for a number of tall buildings tested thereafter.
These include : First National City Corporation Building,
New York [14], John Hancock Tower, Boston [IS, 16,
171, Columbia Seafirst Center, Washington [18, 191, and
the OUB Center, in Singapore [20].
2.3.1 Local wind climate study. Local wind climate at
the site of the Sears Tower was established based on two
approaches. First, the meteorological data from surface
and upper level observations in the Chicago area were
used to establish the general wind climate of the area.
Second, using a 1 : 2000 scale topographical model of the
Chicago area in the wind tunnel, details of the wind
condition for the site were measured. For the topo-
graphical modelling, the surrounding area extended over
a circle with a full scale radius of 400 m centered approxi-
mately at the tower site. Two types of upstream terrain,
open water and urban terrain, were considered. Vertical
profiles of mean and rms wind speeds at the site were
established by normalizing with the gradient wind speed.
A 1 : 400 scale wind tunnel study was also performed
to examine further details of the upstream flow regimes.
Based on this study, three flow conditions were identified
and represented by power law exponents of 0.56, 0.40
and 0.13. These conditions correspond to winds coming
from the NE from NW or SW and from the SE, respec-
tively.
Full scale wind data from six locations were used to
evaluate the probability of exceeding a given wind speed
from a particular direction. The macro-scale spectra were
also established ; these provide the time domain variation
of mean wind speed averaged over intervals of time long
enough, compared to time scales associated with tur-
bulent velocity fluctuations. The effective cycling rate for
the Chicago area was found to be 0.11 cycles/hour ; i.e.,
the number of events becomes about 960 per annum.
This was based on the frequencies associated with the
macro-scale variations in velocity spectra.
2.3.2 Pressure study. Two different models with the
linear scales of 1: 400 and 1: 2000 were fabricated to
evaluate the wind-induced external pressure distribution
on the Sears Tower. The 1:2000 model was used for
finding the scale effect on the measured pressures and
also to correlate the local wind statistics influenced by
the local topography. Detailed pressure measurements
were performed using the 1 : 400 model with 183 pressure
taps. The model was tested for various wind directions
using all three exposures discussed in the previous
A. Baskaran
9.
NORTH
Fig. 5. Three kinds of exposure conditions based on wind directions [12].
section. A typical distribution of pressure for a west wind
condition is presented in Fig. 7. The figure contains the
mean and dynamic pressure values, typically at four levels
on the building. Irregular building shapes, which are
often introduced in modem tall buildings, affect the wind
pressure distribution; this is evident from the figure,
which shows a significant change in pressure distribution
from one level to another. This emphasizes the need for
wind tunnel testing of unusual buildings.
At the time of this project, the importance of fluc-
tuating wind loads on buildings were better realized and
incorporated in the design procedure through peak
factors. These factors are very useful for the design of
window glass panels and other exterior cladding elements
which are subject to the fluctuating wind gusts. Cal-
culated factors for selected tap locations are shown in
Fig. 8. The measured results suggest a peak factor of
approximately 3.5 to 4 whereas the negative peak suction
factors were sometimes found to exceed 10. Figure 9
compares the mean pressure coefficients at various levels
of the building obtained from 1 : 2000 and 1 : 400 models.
This comparison confirms that the scale effect, if any, is
negligibly small. Thus the pressure results obtained from
the 1 :400 model were extensively used with the topo-
graphical wind speed data obtained from the 1:2000
model for the wind load predictions. For the design, a
peak pressure of 25 psf and 60 to 70 psf for peak suction
were recommended by considering a 100 year return
wind.
2.3.3 Aeroelaslic study. A multi-degree of freedom
aeroelastic model of the Sears Tower was constructed to
a scale of 1 : 400. The model was mounted on a flexible
base designed to represent the rotational flexibility of the
foundation. The model consisted of seven rigid floor
plates, a base plate, and columns to simulate the building
stiffness. Including the three degrees of freedom at the
base, the model has a total of 24 degrees of freedom. At
the full scale height of 1165 ft., the top floor acceleration
was monitored. Structural damping was assumed to be
0.5 and 1.0% of critical. Measurements were carried out
in three different flow regimes which are developed in the
wind climate study.
Figure 10 shows a typical aeroelastic response with
two damping values for a benchmark gradient wind speed
of 100 mph. Results are shown for different wind azimuth
angles tested. Discontinuities occur due to the changes in
the upstream terrain conditions both for the mean and
dynamic response. The dynamic response was found to
be preliminary in the fundamental sway modes of
vibration. Increasing the damping from 0.5% to 1.0%
generally causes buffeting response for all wind direc-
tions. This may be due to the turbulence action of wind.
As shown in Fig. 11 the measured mean base moment
coefficients from the aeroelastic test agreed well with the
calculated values from the pressure study and this has
been found true for the three exposure conditions and
two building sway motions considered. In the case of the
pressure study, the measured mean pressure values were
Wind Engineering Studies on Tall Buildings
u
HOUSE
EXPOSURE 3
/-GTI 2 = o p E l
MPOSURE 3
MPOSURE 3
55.7 24.3 TOWER 0
- + 2 6 6 AZIMUTH 1 4 4
u NOTATION:
3-4.0 X - Y
x wall level
lave. hourly extreme - mean!
RMS pressure
Fig. 6. Peak pressure factors on the Plaza buildings for different wind directions [12].
integrated over the building surface and the base
moments were taken at the level of 105 ft. below the plaza
level. Comparisons of this nature provide confirmation
of the different measurement techniques and may reveal
the experimental errors or uncertainty, if any.
2.4 First National City Corporation Building
The study of the FNCC [14] is considered to be unique
because of the building's height, unusual geometrical
shape and the installation of Tuned Mass Damper
(TMD) system to suppress its possible dynamic motion.
The construction site was a heavily built-up area with
high turbulence intensity. The study consists of the
following: the wind climate at the project site, the pres-
sure study, the aeroelastic study of the tower, and the
pedestrian level wind environment.
The wind tunnel flow regime was established based on
records from the U.S. National Weather Record Center
in Ashveille, N.C. These observations were taken at the
John F. Kennedy Airport, N.Y., during the period of
1960 to 1969. The prevailing wind in the New York area
is westerly, particularly in winter months. However, four
different upstream roughnesses were established for the
wind tunnel testing.
2.4.1 Aeroelastic study. Only two fundamental sway
modes of vibrations were modelled in the aeroelastic
study. Any contributions of the torsional mode and
higher sway modes of vibration were neglected. A build-
ing model was fabricated using a scale of 1 : 500. It con-
sisted of seven lumped masses interconnected with elastic
columns. The natural periods of vibration in two sway
A. Baskaran
N
Level 7
Level 6
L ~ w ! 4
Level 2
-
--------...---
CP
I ~ PI + C~ r ms
-.-.-*-
."""-
I ~ P I + aCprm,
.......
....... -- Dynamic pressure
.r
Fig. 7. Variation of external pressure for a West wind on the Sears Tower [12].
modes were 6.66 and 6.25 seconds. Measurement of the
wind-induced response was carried out with three differ-
ent values of structural damping: 0.5, 1.0 and 2.0% of
critical. Figure 12 shows the dynamic response of the
building for different gradient wind speeds with a damp-
ing ratio of 0.020. It is evident that the correlation
between the x and y displacements was very small and
any increase in the wind speed increased the response.
2.4.2 Pressure study. A rigid model equipped with 147
taps was constructed using a geometric scale of 1: 500
and pressure was measured under simulated wind con-
ditions to evaluate wind-induced external pressure loads.
Results were normalized by the reference dynamic pres-
sure at the gradient height. These coefficients were then
integrated with the wind climate statistics of the site to
obtain the peak exterior wind-induced pressures and suc-
tions for a given return period. Examples are given in
Fig. 13, which shows the pressure contours for a return
period of 50 years. The largest suction was found to be
about 35 psf (1 psf = 48 Pa). The maximum pressure of
about 25 psf was predicted on the south face of the
building, whereas all other walls have more or less the
same suction. Comparison of the mean base moments
obtained by integrating the pressure data with those from
the aeroelastic test gave good agreement.
2.4.3 Other studies. Another interesting feature of this
study is the use of Tuned Mass Dampers (TMDs) in the
building. In building, the total damping consists of the
structural damping and aerodynamic damping compon-
ent. The latter can be evaluated from the autocorrelation
function that can be obtained from the model free
vibration. For FNCC, the TMDs were added to reduce
the peak acceleration values and it was found that a
combination of 0.5% structural damping and 1.0%
TMD damping would suppress the peak acceleration
down to an acceptable level for human comfort. A similar
approach was also followed for the wind tunnel study of
the John Hancock Tower, in Boston [15]. In addition,
for the determination of pedestrian level wind environ-
ment, the local wind condition was also observed at eight
different locations. The results were then integrated with
the statistics of reference wind climate and predictions of
local extreme wind conditions were made for various
seasons.
2.5 Bank of China building
When completed, the Bank of China building will
become the tallest structure in Hong Kong and also the
tallest building outside of North America. Its unusual
geometry and the local high incidence of typhoon winds
pointed to the need for a wind engineering study. An
extensive study of typhoon conditions in Hong Kong has
been reported elsewhere [21, 221; the following infor-
mation is gathered from Davenport et al. [23].
First, the wind records were synthesized to obtain the
profile of the hourly mean wind speed. The Hong Kong
wind climate can be divided into two types of winds:
those associated with typhoons and those which are free
Wind Engineering Studies on Tall Buildings
0.0 8.0
TAP- 530
6.0
4.0
20 -
e '
0
90.
a-Mmum angle (-) a-muth angh (-)
3 n o -2.0 4 0
-4.0 -4.0
l ml n
4. 0 4.0
8. 0 4.0
Fig. 8. Peak pressure factors at selected taps on the Sears Tower [12].
from typhoons. The structural design for safety and
strength is always governed by typhoon winds, typically
about 48 m/s, whereas the occupants' comfort and ser-
viceability are designed based on the non-typhoon wind
climate of around 28 m/s. Both values are considered at
the gradient height and they correspond to a return
period of 50 years.
Second, a 1 : 500 pressure model was tested at the UWO
to predict the wind loads for various return periods. The
largest suction for the 100 year wind was about 6.6 kPa
occurring on a joint corner of the building. Generally,
the east exposure has higher peak suction values than
other exposures. This is not only because of the prevailing
wind direction but is also due to the unusual building
shape. The 50 year suction of 5.9 kPa was observed, as
opposed to the Hong Kong building code value of 5.3 kPa.
This is a case in which the conservativeness of the code
was not enough to cover the high loads caused by an
unusual building configuration. For the final design of
cladding and other external elements, the code has been
generally used, except at those locations where it was
exceeded by the wind tunnel predictions.
Another approach taken in this study was the use of
the force balance technique developed by Tschanz [24]
for the measurement of wind loads and for response
prediction. It is a simple approach compared with con-
ventional aeroelastic modelling, as it does not include
the details of the structural dynamic properties. Con-
struction of simpler models reduces the model cost. Also
the structural properties are not vitally important during
initial design of the building. A comparison of the results
obtained by using the new force balance technique to the
conventional aeroelastic testing is shown in Table 1.
Base bending moments calculated from the Hong
Table 1. Comparison of the results derived from aeroelastic modelling and force balance
technique [24]
Moments (50 yr.) Acceleration (100 yr.)
(lo6 kN-m) (milli g)
X Y T X Y T
Force balance method 5.18 4.86 0.28 6.8 5.5 8.4
Aeroelastic technique 3.42 3.00 3.16 5.1 4.4 10.6
Hong Kong Code 14.6 10.2 - -
-
A. Baskaran
?
1 : 400 scale model
I Pressure
----I Suction
o 1 : 2000 scale model
b
A
Cp - scale
0 0.5 1.0 1.5
17
C,
WEST WIND
Level 7
Level 4
Level 2
NORTH - EAST WIND
a = 180' a - 4 5 O
Fig. 9. Comparison of mean pressure coefficients obtained with 1 : 2000 and 1 : 400 scale models [12].
Kong Code are also included as a reference. In general,
the moments obtained from the aeroelastic model study
are smaller than those derived by the force balance tech-
nique. However, the major structural components were
designed to meet the requirements of the code.
3. FULL SCALE STUDIES AND
COMPARISON WITH WIND TUNNEL
RESULTS
The only way to verify the wind tunnel test results is
to compare them with the behavior of the real buildings.
Since this information can be obtained only after con-
struction of the structure, it cannot be used during the
design stage of that building. However, full scale data
have vital importance for the validation of physical mod-
elling and numerical simulation. Unfortunately, full scale
measurements are relatively costly and they may often
provide obscure outputs, which do not allow straight-
forward comparison, due to various reasons. Thus, only
a few studies have been made so far [25].
Some of these rare and yet important measurements
are summarized here. The buildings considered are:
Empire State Building, Commerce Court Building and
the Allied Bank Plaza. These three buildings may typ-
ically represent the construction of the 1930s, the 1960s
and the 1980s, respectively. Moreover, their full scale
data were used for validating the wind tunnel measure-
ments on pressures, aeroelasic response and design loads
respectively.
3. 1 Empire State Building
Full scale measurement of wind-induced pressure on
the Empire State Building was camed out by Rathbun
[26]. In this experiment, one anemometer, 30 mano-
meters, 28 cameras with operating mechanisms, 22 exten-
someters, 1 collimator with its target and 1 plumb-bob
were used. Pressure signals were measured at 10 stations
on each of three floors using manometer boards and flash
cameras. As mentioned previously, Dryden and Hill [7]
performed wind tunnel measurements for the same build-
ing configuration. However, no attempts were made by
Rathbun to compare his full scale values with the wind
tunnel test results, presumably because they appeared to
agree very little.
In 1969, Dalgliesh [27] made some comparisons using
the results of the above two studies ; an example is shown
in Fig. 14. Only few points (solid points) are available
from the full scale study. Generally speaking, the wind
tunnel values are higher than the full scale data and this
may be due to the differences in reference pressure used.
More seriously, Dryden and Hill assumed that the wind
flow would be uniform at 200 ft. or more above ground
and based on this assumption they used an aeronautical
type wind tunnel for the measurement. The results could
have been, of course, significantly different if one con-
Wind Engineering Studies on Tall Buildings 11
Exposure 1 -+( Exposure 3
r
Expowre 1 -q + Exposure 3
I' Expowre 2 -1
Azimuth angle a (degrees) Arimuh angle a (degrees)
Fig. 10. Typical variation of aeroelastic response with azimuth angle [12].
siders the boundary layer wind tunnel profile, which var-
ies with height. In any event, the change in wind direction
causes significant changes in pressure reading for both
cases. The agreement seems to be better when the wind
is normal to the building walls rather than when the wind
strikes the building at an angle.
Another comparison was attempted by Davenport [28]
in terms of the overturning base moments as a result of
the recent boundary layer wind tunnel tests on the Empire
State Building model ; the comparison was carried out
using the base balance technqiue. A model of the building
machined from a stiff foamed plastic was mounted on a
sensitive high frequency balance to measure the base
shears, moments and torques. Figure 15 compares the
measured values of mean moment coefficients with their
full scale counterparts. The agreement between them is
remarkable. This provides a very important full scale
confirmation of a model test.
3.2 Commerce Court Building
Full scale measurements were undertaken by Dalgliesh
and other members of the Division of Building Research,
National Research Council Canada, during the period
1973-1980. Surface wind-induced pressure was measured
simultaneously at 32 points on the building. The building
internal pressure was also measured at one point and
used as the reference for the calculation. Pressures were
collected for all points at a sampling rate of 120 samples
per minute over a period of 5 minutes.
Extensive comparisons between full-scale results and
wind tunnel experiments were reported in [29, 301 and
[31]. Figure 16 depicts the mean and rms pressure mea-
sured at two different levels of the building. The solid line
indicates the full-scale estimates and the open circles are
from the wind tunnel model data. The mean pressure
coefficients are in better agreement than the rms values,
particularly for the south winds. These discrepancies were
attributed to the fact that winds from the south had
not been frequent enough or strong enough to provide
sufficient reliable rms data.
Wind tunnel studies of the Commerce Court building
were first carried out in the UWO and this study included
measurement of mean and fluctuating pressures, dynamic
response of a two degree of freedom aeroelastic model
and a synthesis with metrological data [32]. Later the
National Aeronautical Establishment of the National
Research Council of Canada also performed an extensive
wind tunnel study on the Commerce Court building.
A. Baskaran
Fig. I I.
0.5
0.4
0.9
0.2
0.1
0
-0.1
-0.2
- from pressure measurements
-0.3
Aemelastic results:
o exposure 1
-0.4
A exposure2
-0.5 0 exposure3
Comparison of mean base bending moment obtained from the aerolastic and pressure measure-
ments [12].
There are many significant differences between the NRC
study and that of the UWO [33, 34, 351. The pressure
study at NRC concentrated on occurrence of the peaks
[31] for the design of cladding elements and glass and
window panels. Also, in the aeroelastic study [31], seven
mass levels were chosen by placing one at each of the five
instrumented levels in the full-scale building and then
dividing the remaining level into two modules to make
all the modules approximately of the same height. More-
over, taking advantage of the 9 m x 9 m, NRC wind
tunnel, models were fabricated using a geometric scale of
1 : 200 in comparison to the 1 : 500 scale of UWO.
To demonstrate the measured aeroelastic response,
Fig. 17 compares the model and full scale acceleration
power spectra of the first mode for two typical wind
directions (North-South and East-West). In general, the
agreement is quite satisfactory. However, in the North-
South acceleration, sharper peaks and greater fall-off of
contribution by the second mode are evident in the model
than in the full-scale results. This may be due to the mode
stiffness, which was based on several practical con-
siderations. This resulted in a frequency scale of 1 : 53
rather than a full-scale value of 1 : 58. Other noticeable
factors from the full-scale measurements were the highest
5 min mean reference speed of 33 m/s and the largest
peak pressure difference of 640 N/m2. For the displace-
ment, the building experienced about 220 mm at its 234 m
level along with a peak acceleration of 10 to 15 milli g.
3.3 Allied Bank Plaza
Full scale observation of the top floor acceleration of the
Allied Bank Plaza in Houston, Texas has been reported by
Halvarson and Isyurnov [36] as a comparison with the wind
tunnel test results. The measurement was done using two
kinemetric Model VM-1 accelerometers, with a range of 0.1
mg to 1.0 g. The measurement was not successful in the
beginning, when the wind speed was in the range of 35 to
45 mph. However, when the area was later hit
by a tropical storm with wind gusts of 56 mph and also
by Hurricane Alicia, with the fastest mile speed of
Wind Engineering Studies on Tall Buildings
12
11
N
X RMS Response b
0 40 80 120 160 200 240 280 320 3M)
Azimuth (degrees)
4 2
E Y RMS Response Smchrral damping t
ratio 5, = 0.020
-
"g
- - - 125 mph
--- lOOmph
- ....... 75 mph
- 50mph
Azimuth (degrees)
Fig. 12. Variation of aeroelastic rms x and y response with 4 wind speed and wind direction for different
values of structural damping [14].
1 90 mph, some data were obtained. The response at the
71st floor level is compared with the wind tunnel test
results in Fig. 18. A possible fluctuation of wind speed
and wind yaw angle is expected to be less than 2.5 m/s
and 5", respectively.
In addition to the direct comparison of the wind-
induced parameters, it would be useful if the design wind
loading criteria were validated. This was attempted for
this building using the acceleration measurement. The
lateral force acting on each floor can be estimated by the
product of the weight at each floor, the top floor peak
acceleration and the mode shape factor, normalized at
the top floor. Using this concept, the estimated base shear
and moment from the wind tunnel test results can be
compared (Table 2) with the observations, recorded dur-
ing Hurricane Alicia. The recommended values of the
Houston Building Code (C, = 1.4 assumed) are included
for comparison. In general the agreement was satis-
factory. The full scale monitoring program started after
the structure was competed ; however, the interior con-
struction and windows still remained to be finished. The
calculation, on the other hand, assume a fully occupied,
completed building. With this in mind, the agreement
between full scale observation and wind tunnel data is
acceptable. This comparison shows that the code values
are conservative and overestimate both the moment and
shear, typically by a factor of two.
4. SUMMARY OF TRANSITION IN
RESEARCH
Review of the wind tunnel studies and full scale
measurements are presented in the previous sections. As
mentioned before, only a few studies were selected as
representative of the major changes in the research
approach. A summary for the transitions of wind engin-
eering study of tall buildings can be listed as follows :
Appreciation of wind loads in design
Pressure measurements using aeronautical wind tun-
nel
Measurements using aeroelastic models
Full scale measurements of wind pressures
Better simulation of turbulence conditions
r
-
14 A. Baskaran
I
I
I
I
r
-.
-
Wind Engineering Studies on Tall Buildings
75th FLOOR 75th FLOOR
0
36th FLOOR 36th FLOOR
Fig. 14. Comparison of the mean pressure coefficients measured in a model and full scale study of Empire
State Building [27].
Integrating with local wind climate data
Introduction of peak factors in structural design
Cross checking of force (pressure) and response
Full scale measurements of dynamic response
Introduction of high frequency force balance
These stages may overlap and the above grouping is not
in chronological order ; rather it identifies major changes
in the research activities. Appreciation of wind loads in
design started nearly 100 years ago, when the Eiffel Tower
was completed to mark the occasion of the Paris exhi-
bition in 1889 [28]. On the other hand, research for the
wind effects on tall buildings started only during the
design of the Empire State Building.
Static wind loads are evaluated by fabricating and
testing scale models in aeronautical wind tunnels. Once
the building has been erected, full scale measurements
are carried out to validate the results obtained from the
aeronautical wind tunnels. These comparisons sig-
nificantly helped the wind engineering community to
NORTH - SOUTH EAST - WEST
c i y c i x
a - AZIMUTH (degrees)
0-0 Full scale measurement, Rathbun (1940)
- Wind tunnel measurement, Davenport (1988)
Fig. 15. Comparison of the base overturning moment for the Empire State Building as measured full scale
and in the wind tunnel [28].
BAE 28:l-0
16 A. Baskaran
-1 .o -1 .o
West North East South West West North East South West
Wind direction Wind direction
Fig. 16. Comparison of the pressure coefficients measured and full scale study of Commerce Court Building [29].
overcome the misconception of uniform velocity field
around buildings.
A new area of research simulates wind turbulence con-
$itions in the wind tunnel by constructing atmospheric
boundary layer wind tunnels. The surrounding top-
ography is also considered and included in the study
of the World Trade Center. Metrological wind climate
records are integrated with the wind tunnel results for
the probabilities method of design.
In the early 1980s, pressure measurements and aero-
elastic measurements were regarded as equally important
for the evaluation of wind effects on tall buildings and
the results from the measurements have been compared
for cross checking the experimental techniques. Next, the
full scale'dynamic wind effects on the Commerce Court
Building were monitored and compared with the wind
tunnel results. This will help in validating the frequency
and fluctuating nature of wind conditions in the wind
tunnel simulations. To reduce the design and model cost
of the aeroelastic testing, the high frequency balance tech-
nique was introduced in the Bank of China project. Dur-
ing the course of all these processes, the wind tunnel
results are also compared with values from the Building
codes and wind standards in order to transfer the new
information to the end users by updating the codes and
standards [37, 381.
At present a majority of wind engineering studies on
tall buildings follow a pattern as shown in Figure 19.
Complete analysis of the wind effects on buildings can be
obtained by following the four-fold experimental
approach, namely, local wind climate study, aeroelastic
modelling, pressure measurements and wind environ-
Iv'
N-s E-W I
0 0.2 0.4 0.6 0 0.2 0.4
Frequency (Hz) Frequency (Hz)
Fig. 17. Comparisons of the model and full-scale acceleration power spectrum [30].
Wind Engineering Studies on Tall Buildings
FunTseale
maw-
0 1 " " " " " " " ' ~
0300 0600 1000 1400 1800 GMT
2200 01 00 0500 0900 1300 CDT
hme - Augus118.1983
Fig. 18. Peak resultant accelerations on the plaza top level and comparisons with the wind tunnel data [36].
mental study. The local wind climate conditions can be
obtained from the meteorological data of the location
and can also be used to determine probability dis-
tribution of the wind speed ; this information will help in
probabilities methods of design. However, the meteoro-
logical data may not represent the local surroundings,
which cause most of the turbulence effects on such build-
ings.
The aeroelastic measurements or base balance tech-
niques are vitally important in identifying wind-induced
dynamic effects on buildings. Damping evaluation and
top floor accelerations will provide better serviceability
criteria for a building. Pressure measurements are equally
important, since they are useful not only for the design
of structural elements and cladding, they also play a
major role in the energy calculations for buildings.
Finally, the pedestrian level velocity measurements will
provide information on the local wind environmental
conditions and help in town planning.
No doubt, the state of the art for tall buildings will be
different tomorrow from today. Currently, in the wind
engineering research activities on tall buildings, three
main areas are in-progress, as listed below :
Time domain treatment of wind loads
Computer modelling of wind effects
Winds induced internal pressures
Effects of wind on tall buildings have usually been
analyzed in the frequency domain because of its station-
ary random characteristics over a considerably long per-
iod of time as opposed to the earthquake response cal-
culations which are usually done in the time domain. For
places where earthquake and wind have equal magnitude,
a common approach will not only make the design pro-
cess economical, it also helps the designer in selecting
the optimum conditions. Studies have been initiated to
represent wind load conditions in the time domain.
Advancements of computer software and hardware
technology provide a new direction for analyzing engin-
eering problems. The field of wind engineering is gaining
significant momentum in computer modelling processes.
Table 2. Comparison of the base moment and shear for the Allied Bank Plaza as measured
in full scale and in a wind tunnel [36]
Wind tunnel Full scale Houston
(Alicia) Code
100 yrs 50 yrs
First mode
Base shear (kips) 5,600 4,900 4,500 12,500
Base moments (ft - K) x lo6 3.5 3.1 2.7 7.1
Second mode
Base shear (kips) 4,200 3,500 3,800 9,500
Base moments (ft . K) x lo6 2.6 2.2 1.8 5.4
18 A. Baskaran
WlND ENOWmNO S T WV OFTAU
BUILDINGS
anplw
Smy
d.nOFnoh.
Top mW
.-Im#
-r
DESION O F T U BWLDIMGS
Fig. 19. Established pattern for wind engineering study of tall
buildings.
Attempts have been made in predicting wind flow con-
ditions around buildings and in computing the wind-
induced effects on buildings [39]. Research efforts have
also been extended in modelling time dependent wind
flow conditions around buildings in super computers [40]
by applying sophisticated numerical algorithms [41].
Until recently less emphasis has been placed on wind-
induced internal pressures, particularly for tall buildings.
Internal building pressure is not only caused by external
wind pressure ; the mechanically operated ventilation sys-
tems in the building and the stack effect developed by the
thermal difference across the envelope also contribute to
the changes in internal pressure. The effect of these fac-
tors on tall buildings were reviewed by Tanaka [42] and
Tanaka and Lee [43]. Field monitored results [44] on a
20 storey building show that the wind and stack effect
has predominant influence on the pressure difference
across the envelope. Developments in wind tunnel
measurement techniques and comparison of the internal
pressure coefficient with those of North American Stan-
dards are respectively reported in [44] and [45,46].
Acknowledgement-It is indeed a great pleasure to acknowledge
Dr Hiroshi Tanaka for providing continuous advice during the
course of this project.
REFERENCES
1. A. C. Erickson and A. Steele, The tall buildings: Is America's past East Asia's future?, Proc. 4th
International Conf. on Tall Buildings, Hong Kong, 2-7 (1988).
2. B. S. Taranath, Structural Analysis and Design of Tall Buildings. McGraw-Hill(1988).
3. A. Coull and B. Stafford-Smith, Tall Buildings, Pergamon Press (1966).
4. Monograph on planning and Design of Tall Buidings, Council on Tall Buildings and Urban Habitat,
Lehigh university, Bethlehem, Pennsylvania, U.S.A., 1978 to (1981).
5. Developments in Tall Buildings, Council on Tall Buildings and Urban Habitat, Lehigh university,
Bethlehem, Pennsylvania, U.S.A. (1983).
6. High-Rise Buildings : Recent Progress, Council on Tall Buildings and urban Habitat, Lehigh univer-
sity, Bethlehem, Pennsylvania, U.S.A. (1986).
7. H. L. Dryden and G. C. Hill, Wind pressures on a model of the Empire State Building. Bureau oJ
Standards Journal of Research, 10,493-523 (1933).
8. L. S. Feld, Superstructure for 1,350 feet, World Trade Center. Civ. Engng, ASCE, 6670 (1971).
9. R. E. Whitebread and C. Scrnton, An investigation of the aerodynamic stability of the World Trade
Center. New York (Part I), NPL Aero Rept. 1156, National Physical Laboratory (July 1965).
10. A. G. Davenport, N. Isyumov, D. J. Fader and C. F. P. Bowen, Study of wind effect on World Trade
Center, New York; Exterior Pressure on Plaza Buildings and Airflow in Plaza. BLWT-6-70, University
of Western Ontario (1970).
11. ASCE, World's tallest tower approved in Chicago. Civ. Engng, ASCE, 18 (1990).
12. A. G. Davenport, N. Isyumov and T. Jandali, A Study of wind effects for the Sears project. BLWT-
5-71, University of Western Ontario (1971).
13. A. G. Davenport, On the Statistical Prediction of structural performance in the wind environment.
Meeting Preprint 1420, ASCE National Structural Meeting, Baltimore, Maryland (1971).
14. N. Isyumov, J. D. Holmes, D. Surry and A. G. Davenport, A study of the wind effect for the First
National City Corporation project. New York, U.S.A., BLWT, University of Western Ontario (1976).
15. D. Surry, H. Tanaka and A. G. Davenport, The effect of proposed tuned mass damper system on the
wind induced response of the John Hancock Tower, Boston, BLWT, University of Western Ontario
(1976).
16. A. G. Davenport, D. Surry and H. Tanaka, A study of wind induced response of the John Hancock
Tower, Boston, BLWT-SS2-76, University of Western Ontario (1975).
17. D. Surry, R. B. Kitchen and A. G. Davenport, Wind induced exterior pressures on the John Hancock
Tower. Boston, BLWT-SS2-76, University of Western Ontario (1976).
18. J. B. Skilling, T. Tschanz, N. Isyumov, P. Loh and A. G. Davenport, Experimental studies, structural
design and full scale measurements for the Columbia Seafirst Center. Building Motion in Wind, ASCE,
1-22 (1986).
19. C. J. Keel and P. Mahmoodi, Design of viscoelastic dampers for the Columbia Center building.
Building Motion in Wind, ASCE, 67-82 (1986).
20. S. Sasaki, M. Shiko and Y. Atsukawa, Superstructure of OUB center. Singapore, Second Century of
the Skyscrapers, Van Nostrand Reinhold Co., 975-987 (1988).
21. G. R. Lythe, D. Surry and A. G. Davenport, Wind proses over selected sites in the Hong Kong area.
BLWT-SS22-81, University of Western Ontario (1981).
22. P. N. Georgiou, M. J. Mikitiuk, D. Surry and A. G. Davenport, Wind climate for Hong Kong.
BLWT-SS2-84, University of Western Ontario (1984).
Wind Engineering Studies on Tall Buildings
A. G. Davenport, N. Isyumov, G. R. Lythe, L. E. Robertson, A. Steckley and D. Surry, The wind
engineering study for the bank of China. Proc. 4th International Conference on Tall Buildings, Hong
Kong, 143-147 (1988).
T. Tschanz, The base balance measurement technique and applications to dynamic wind loading of
structures. Ph.D. Thesis, Faculty of Engineering Science, University of Western Ontario (1982).
A. G. Davenport, Perspectives on the full scale measurement of wind effects. J. Ind. Aerod. 1,23-54
(1975).
C. J. Rathbun, Wind forces on a tall building. Trans. ASCE (105) Paper No. 2056, 1-82 (1940).
W. A. Dalgliesh, Experiences with Wind Pressure Measurements on a Full-scale Building, Proc.
Technical Meeting concerning Wind Loads on Buildings and Structures. NBS Bldg. Sc. (30) Gaithers-
burg, Md., 61-71 (1979).
A. G. Davenport, The response of supertall buildings to wind. Second Century of Skyscraper, Van
Nostrand Reinhold Ltd., New York, 705-725 (1988).
W. A. Dalgliesh, Comparison of Model/Full-scale Wind Pressure on a High-rise Building. J. Ind.
Aerod. 1, 55-66 (1975).
W. A. Dalgliesh, Comparison of the model and full scale test of the commerce court building in
Toronto. Proceedings of International Workshop on Wind Tunnel Modelling Criteria of Techniques in
Civil Engineering. National Bureau of Standards, Gaithersburg, 515-527 (1982).
W. A. Dalgliesh, J. T. Templin and K. R. Cooper, Comparisons and wind tunnel and full-scale building
surface pressures with emphasis on peaks, Wind Engineering. Proc. 5th International Conference on
Wind Engineering, July 1979 (Ed. by J . E. Cermak (1, 553-565 (1980).
A. G. Davenport, M. Hogan and N. Isyumov, A study of wind effects on the commerce court tower.
BLWT-7-69, University of Western Ontario (1969).
J. T. Templin and K. R. Cooper, Design and performance of a multi-degree-of-freedom aeroelastic
building model, J. Wind Engng Ind. Aerod. 8, 157-175 (198 1).
J. T. Templin and K. R. Cooper, Torsional effect on the wind-induced response of a high-rise building.
Presented 4th U.S. International Conference on Wind Engineering Research, Seattle, Wash~ngton, 2 6
29 (1981).
W. A. Dalgliesh and J. H. Rainer, Measurements of wind induced displacements and accelerations
of a 57-storey building in Toronto, Canada, Proc. 3rd Colloquium on Industrial Aerodynamics,
Buildings Aerodynamics, Part 2, Aachen, Germany, 1, 67-78 (1978).
R. Halvarson and N. Isyumov, Comparison of the predicted and measured dynamic behavior of
Allied bank plaza. Building Motion in Wind, ASCE, 2341 (1986).
M. R. Wiliford and A. J. Fitzpatrick, The integration of structural analysis and wind tunnel testing
for the New Hong Kong and Shanghai Banking Corporation headquarters in Hong Kong-Part I.
Proc. 3rd International Conf. on TUN Buildings, Hong Kong, 243-249 (1984).
A. G. Davenport, D. Surry and G. R. Lythe, The integration of structural analysis and wind tunnel
testing for the New Hong Kong and Shanghai Banking Corporation headquarters in Hong Kong-
Part 11, Proc. 3rd International Con$ on Tall Buildings, Hong Kong, 256256 (1984).
A. Baskaran, Computer simulation of 3D turbulent wind effects on buildings, Ph.D. thesis, Concordia
University, Montreal, P.Q., Canada (1990).
F. Baetke, Numerische Berechnung der Turbulenten umstrongum eines kubisher korpers, Ph.D. Thesis,
Technische Universitat Munchen, Germany (1986).
S. Murakarni, A. Mochida and K. Hibi, Three dimensional numerical simulation of air flow around
a cube model by means of a large eddy simulation, J. Wind En,qng Ind. Aerod., 25, 291-305 (1987).
H. Tanaka, Building internal pressures and air infiltration, Proceedings of 4th Canadian workshop on
Wind Engineering, Toronto, Canada, 88-1 12 (1984).
H. Tanaka and Y. Lee, Stack effect and building internal pressure, Proceedings of 7th International
Conf. on wind engineering and industrial aerodynamics, Aachen, Germany, July 1987,293-302 (1988).
Y. Lee, H. Tanaka and C. Y. Shaw, Distribution of wind and temperature induced pressure difference
across the walls of a twenty storey compartmentalized buildng, J. Wind Engng Ind. Aerod. 10, 287-
301 (1982).
T. Stathopoulos and H. D. Luchion, Transient wind induced internal pressures, J. Struct. Engng,
ASCE, 115,7, 16501-1514 (1987).
A. G. Davenport and D. Surry, The estimation of internal pressures due to wind with application to
cladding pressures and infiltration, published in Proc. Wind Pressure Workshop, Belgium (1984).
T. Stathopoulos and R. Kozutsky, Wind-induced internal pressures in buildings, J. Struct. Engng,
ASCE, 112,9,2012-2026 (1986).