The Seventh Asia-Pacific Conference on

Wind Engineering, November 8-12, 2009,

Taipei, Taiwan

WIND LOADS ON LOW-RISE BUILDINGS: UPSTREAM

EXPOSURE EFFECT
Ioannis Zisis1 and Ted Stathopoulos2
1
Research Assistant, Building Aerodynamics Laboratory, Concordia University
Montreal, Quebec, Canada, i_zisis@live.concordia.ca
2
Professor and Associate Dean, Concordia University
Montreal, Quebec, Canada, statho@bcee.concordia.ca
ABSTRACT
An experimental test house has been constructed in Fredericton (NB, Canada) and both envelope pressures and
its wind-induced structural response have been monitored during the last three years. In addition, a 1:200-scaled
model of the house was tested at the boundary layer wind tunnel of Concordia University. Acquired field data
have been analyzed to evaluate the exposure characteristics for different wind angles of attack. Upstream terrain
characteristics variations with respect to wind direction are presented. These data are also considered to properly
simulate the upstream terrain in the wind tunnel. Mean and peak pressure coefficient distributions are evaluated
and the impact of a varying exposure on the overall structural wind-induced load is discussed.
KEYWORDS: FULL SCALE WIND MONITORING, WIND TUNNEL TESTS, UPSTREAM EXPOSURE,
PRESSURE DISTRIBUTION, OVERALL WIND LOAD

Introduction
Wind-induced loads on low-rise buildings are evaluated based on a set of parameters,
such as dynamic velocity (q), exposure factor (Ce) and gust pressure coefficient (CpCg). The
importance of accurate representation of upstream terrain characteristics during wind tunnel
tests has been widely addressed (Stathopoulos 1984, Levitan et al. 1990, Mehta and Smith
2004). The success of wind tunnel experiments is directly related to the availability of
accurate upstream terrain characteristics such as wind speed, direction, power spectra,
roughness length etc. The impact of different upstream exposures on the gust pressure values
was until recently addressed by various codes through the conservative approach of open
terrain simulation properly factored for directionality effects. Current interest of the American
and Canadian code committees initiated a new potential for studies where a more efficient and
rigorous definition of upstream exposure and roughness is intended (Stathopoulos et al 2009).
Currently both the American (ASCE 7-05) and the Canadian (NBCC 2005) provisions use the
suburban velocity profile for cases in which the exposure is classified in accordance with the
code definition as suburban.
Of special interest to the present study is the experimental work conducted by Wang
and Stathopoulos (2006). A low building model was tested for a total of 66 fetch cases and 13
wind angles of attack to evaluate the effect of upstream exposure on the wind loading. The
different upstream configurations revealed that peak wind loads are basically affected by short
distance roughness characteristics rather than further terrain properties. In most cases these
loads can be determined by considering a fetch of 300-400 meter since lengths greater than
that seem not to affect the wind-induced response. Moreover, the study carefully addresses
the impact of “rougher” patches located on the proximity of the examined building and the
importance of such variations to both mean and peak wind loads. The findings are closely
The Seventh Asia-Pacific Conference on Wind Engineering, November 8-12, 2009, Taipei, Taiwan

related to the objectives of the current study where full-scale findings in conjunction to wind
tunnel experiments are coupled to evaluate upstream terrain characteristics and their influence
on local pressures and total response of the building.
Field Facilities
The experimental house is located in Fredericton (New Brunswick, Canada), in a
relatively open-suburban area. The test building is a single storey typical North-American
residential house with a rectangular layout, external dimensions of 8.6x17.2x5.6 meter
(WxLxH) and a duo-pitch roof of 4/12 slope. More details about the test house and the
instrumentation can be found in Doudak (2005) and Zisis (2006).
A 12-meter high mast located approximately 20 meters west of the house was used to
mount three propeller anemometers at heights of 5.0, 6.5 and 10.0 meters. A second smaller
mast was added on the east side of the test house equipped with one anemometer at a height
of 5.0 meters. An aerial view of the area with the actual building and the meteorological
towers is shown in Figure 1. The test house was equipped with 27 foundation and 6 roof load
cells. Moreover, 40 pressure taps were distributed on the wall and roof surface, as shown in
Figure 2.

Figure 1: Test House and Figure 2: Load Cell and Pressure Tap Location on the
Meteorological Tower Location Experimental House
Wind Tunnel Simulation
The wind tunnel tests were conducted at the Building Aerodynamics Laboratory
located in the Engineering Complex at Concordia University. A geometric scale of 1:200 was
selected for the wind tunnel study. The metallic model has external dimensions of 86.5 mm
by 42.5 mm (length – width) and a total height of 24.35 mm (ridge height – excluding
concrete foundation wall). The model is equipped with 126 pressure taps located on the wall
and roof surface respecting the full-scale tap distribution. In addition to the building model, a
proximity model of was constructed at the same 1:200 scale. Considering the scale of the test
model and the size of the wind tunnel test section, a circular wood base of 1.60-meter
diameter and 3.0 mm thickness was used to place all surrounding structures and tree elements
on it, as shown in Figure 3.
The Seventh Asia-Pacific Conference on Wind Engineering, November 8-12, 2009, Taipei, Taiwan

Figure 3: Wind Tunnel Building Model, Surroundings and Examined Wind Directions
Results and Discussion

Full-scale Terrain Data

The reliability of the weather tower results was validated through comparison to the
closest weather stations operated by Environment Canada, i.e. Fredericton Airport and CDA
station. Wind speed and direction comparison charts are shown in Figures 4a and 4b
respectively. Records acquired from the 6.5 and 10.0-meter anemometers in the period
October 21-31, 2008 were used for this comparison. The agreement between the two sources
validates the accuracy of the test facility instrumentation results.

Figure 4a: Wind Speed Comparison (5-minute averages)

Figure 4b: Wind Direction Comparison (5-minute averages)

The Seventh Asia-Pacific Conference on Wind Engineering, November 8-12, 2009, Taipei, Taiwan

Of great interest was to evaluate the basic exposure characteristics (power law
exponent, turbulence intensity and roughness length) using the available field data. Wind data
collected from two anemometers (at 6.5 and 10.0-meter height) during October to December
2008 and April to June 2009 were analyzed and power law exponent, turbulence intensity and
roughness length values were calculated. Due to a sensor malfunction the third anemometer
(5.0 meter height) could not be used reducing, to some extent, the accuracy of the derived
characteristics. The results are based on 10-minute averaged statistical values (mean and
standard deviation) and were filtered for wind speeds over 10 km/h (at 6.5-meter height). The
distribution with respect to direction is shown in Figures 5a, 5b and 5c. The power law
exponent ranges from 0.05 to 0.50, the turbulence intensity from 20 to 50% and the roughness
length from a few millimeters up to 1.2 meters. It is quite interesting how these properties
vary considering that the test house is located in a relatively open area with only few low-rise
buildings and medium height trees in the proximity. Following current wind provision
guidelines and “common” wind engineering sense, the terrain would be classified in the open
to sub-urban region expecting a power law exponent in the range of 0.20. The higher
variations should be clearly attributed to the influence of adjacent buildings (north and south
sides) and forestry area (east side) located inside the radius of the 300-400 meter fetch. These
deviating terrain properties indicate that complex terrains need to be examined carefully in
order to properly conduct scaled model tests and successfully compare full-scale to wind
tunnel results.
N E S W
1.0
0.9
Power Law Exponent .

0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction

Figure 5a: Power Law Exponent for Field Data

N E S W
1.0
0.9
Turbulence Intensity

0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction

Figure 5b: Turbulence Intensity for Field Data

The Seventh Asia-Pacific Conference on Wind Engineering, November 8-12, 2009, Taipei, Taiwan

N E S W
1.4
Roughness Length 1.2
1.0
0.8
0.6
0.4
0.2
0.0
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction
Figure 5c: Roughness Length for Field Data

Wind Tunnel Experiments

Currently, two sets of wind tunnel tests have been completed. These tests have been
conducted using a power law exponent of 0.16 and 0.21. The turbulence intensity levels at the
roof height were approximately 17 and 20% respectively. The complete properties for the two
cases are presented in Table 1. In both tests a total of 36 wind directions have been examined
(see Figure 3). A more detailed discussion and presentation of each individual wind tunnel
experimental approach will be presented as soon as the remaining terrain categories (sub-
urban and urban) will be completed.
Table 1: Upstream Terrain Properties for Wind Tunnel Experiments
Power Law Exponent Turbulence Intensity Roughness Length (zo)
(a) (%) Wind Tunnel (m) Full-Scale (m)
Test A 0.16 17 0.00023 0.04637
Test B 0.21 20 0.00045 0.08984
Pressure Coefficient Comparison
Representative full-scale pressure data from two records (May 14, 2009 and June 1,
2009) were used to verify the agreement of each of the two available wind tunnel tests (Test
A and B). For this comparison two pressure taps were selected, one on the wall (Figure 6a)
and one on the roof surface (Figure 6b). The mean and peak pressure coefficient was plotted
with respect to direction using Eq. (1):
p −p
c p ,mean / peak = mean / peak 2 a (1)
1/ 2 ρVBH
where ρ=air density (kg/m3); VBH =mean wind speed at the building height (m/s); pa=ambient
atmospheric pressure (Pa), and pmean/peak=actual surface pressure (Pa). It should be noted that
for the full-scale calculations the mean values where based on a 10-minute average and the
instantaneous peak on a 3-second average (full-scale time scale). The dynamic pressure was
always averaged on a 10-minute basis and was referenced to the mean roof height. Moreover
field data were integrated over a wind angle of attack of 10-degree range to account for the
higher standard deviation values and to be directly compared to wind tunnel tests carried out
using intervals of 10 degrees. To better represent the varying characteristics of the full-scale
results the minimum and maximum integrated values of each set of data were considered in
addition to the mean values.
The agreement for both mean and peak pressure coefficients is good. Wind tunnel
values are in most cases within the range of the field results. The discrepancies are somewhat
The Seventh Asia-Pacific Conference on Wind Engineering, November 8-12, 2009, Taipei, Taiwan

Wind Tunnel (a=0.21) Wind Tunnel (a=0.16) Full-Scale (mean)

0.8
0.6 Pressure Tap WW10
0.4

CP,mean
0.2
0.0
-0.2
-0.4
-0.6
-0.8
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction

Wind Tunnel - Cp,min (a=0.21) Wind Tunnel - Cp,max (a=0.21)

Wind Tunnel - Cp,min (a=0.16) Wind Tunnel - Cp,max (a=0.16)
Full-Scale (minimum) Full-Scale (maximum)
4.0

3.0 Pressure Tap WW10

2.0

1.0
Cp

0.0

-1.0

-2.0

-3.0
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction

Figure 6a: Mean and Peak Pressure Coefficient Comparison (Wall Pressure Tap)

Wind Tunnel (a=0.21) Wind Tunnel (a=0.16) Full-Scale (mean)

0.2
0.0
-0.2
CP,mean

-0.4
-0.6
-0.8
-1.0 Pressure Tap R26
-1.2
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction

Wind Tunnel - Cp,min (a=0.21) Wind Tunnel - Cp,max (a=0.21)

Wind Tunnel - Cp,min (a=0.16) Wind Tunnel - Cp,max (a=0.16)
Full-Scale (minimum) Full-Scale (maximum)
2.0

1.0

0.0

-1.0
Cp

-2.0

-3.0

-4.0 Pressure Tap R26

-5.0
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction

Figure 6b: Mean and Peak Pressure Coefficient Comparison (Roof Pressure Tap)
The Seventh Asia-Pacific Conference on Wind Engineering, November 8-12, 2009, Taipei, Taiwan

higher for the peak pressure coefficient for which the effect of the different simulation profile
is also pronounced (a=0.16 versus a=0.21). These discrepancies also coincide with the region
where the upstream terrain data were disturbed resulting into higher power law exponent and
roughness length (240-290 degrees). This particular case also demonstrates the impact and
importance of varying upstream terrain properties to the resulting wind-induced local
pressures.
Force Coefficient Comparison
A unique characteristic of the experimental house is that it is structurally isolated from
the foundation. The total uplift force was recorded and normalized in accordance to the
approaching wind characteristics. Moreover data from the wind tunnel experiments were used
to assess the equivalent total uplift force coefficient using Eq. (2):
Fz , mean / peak
c f , mean / peak = 2
(2)
(1/ 2 ρVBH ) A
where ρ=air density (kg/m3); VBH=wind speed at the building height (m/s); Fz,mean/peak=total
uplift force (N), and A=building area (m2). The full-scale force coefficient was based on field-
recorded data, whereas the wind tunnel force coefficient was the result of the integration of
the vertical component of all pressures. Similar to the pressure coefficient analysis, both mean
and peak components were considered (Figure 7). The field data were collected during May
and June of 2009 and were filtered for wind speeds over 10 km/h. The mean values were
averaged over a period of 10 minutes and the peaks on a 3-second basis (full-scale time scale).

Wind Tunnel (a=0.21) Wind Tunnel (a=0.16) Full-Scale (mean)

0.0

-0.2
Cf,mean

-0.4

-0.6

-0.8

-1.0
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction

Wind Tunnel - Cf,min (a=0.21) Wind Tunnel - Cf,max (a=0.21)

Wind Tunnel - Cf,min (a=0.16) Wind Tunnel - Cf,max (a=0.16)
Full-Scale (minimum) Full-Scale (maximum)
1.0

0.5

0.0

-0.5
Cf

-1.0

-1.5

-2.0

-2.5

-3.0
0 30 60 90 120 150 180 210 240 270 300 330 360
Wind Direction

Figure 7: Mean and Peak Force Coefficient Comparison (Total Uplift)

The Seventh Asia-Pacific Conference on Wind Engineering, November 8-12, 2009, Taipei, Taiwan

The effect of adequate upstream simulation in the wind tunnel experiments is even
more pronounced on this case. The two different wind tunnel tests (Test A and B) resulted
into two discrete trends which for some critical wind directions differ up to 25% for the mean
and 40% for the peak values. Results from both wind tunnel tests are within the range of the
field data but seem to underestimate the averaged (integration of different full-scale records)
peak component, which can be partially justified by the lower level of turbulence. Higher
fluctuations of the wind direction should also be addressed and considered accountable for the
above mentioned discrepancies.
Conclusions
Field data from two anemometers were used to define the exposure characteristics of a
full-scale testing facility. The data interpretation showed higher than expected variations for
power law exponents and roughness lengths; the variation for wind turbulence intensity was
relatively small. Two sets of data from wind tunnel tests have been compared with the field
results in the form of pressure and total uplift force coefficients. The comparisons indicate
discrepancies between the two different wind tunnel terrain configurations, which in some
cases could be considered critical for the agreement between the field and the simulation
studies.
References

ASCE/SEI 7-05, 2006, “Minimum Design Loads for Building and other Structures”, American Society of Civil
Engineers, ASCE, New York, USA.
Doudak, G., 2005, Field Determination and Modeling of Load Paths in Wood Light-Frame Structures, PhD
Thesis, Department of Civil Engineering and Applied Mechanics, McGill University, Montreal, Canada.
Environment Canada Weather Office, Natural Resources Canada (NRC), http://www.climate.weatheroffice.ec.gc.ca.
Levitan, M., L., Mehta, K., C., Chok, C., V. & Millsaps, D., L., 1990, “An Overview of Texas Tech's Wind
Engineering Field Research Laboratory”, Journal of Wind Engineering and Industrial Aerodynamics, 36, Part
2, 1037-1046.
Mehta, K. C., & Smith, D. A., 2004, “Full-Scale Measurements for Wind Effects on Buildings”, ASCE
Conference Proceedings 153 (27), Houston, Texas.
NBCC, 2005, “National Building Code of Canada”, National Research Council of Canada (NRC), Ottawa,
Canada.
Stathopoulos, T., 1984, “Wind Loads on Low-Rise Buildings: A Review of the State of the Art”, Engineering
Structures, Volume 6, Issue 2, Pages 119-135.
Stathopoulos, T., Wang, K., & Zisis, I., 2009, “Wind Loading and Building Exposure: Are we Still on A, B, C?”,
SEI/ASCE Structures Congress, April 2009, Austin, Texas, USA.
Wang, K., & Stathopoulos, T., 2006, “The Impact of Exposure on Wind Loading of Low Buildings”, SEI/ASCE
Structures Congress, May 2006, St. Louis, Missouri, USA.
Zisis, I., 2006, Structural Monitoring and Wind Tunnel Studies of a Low Wooden Building, MASc Thesis, Civil,
Building and Environmental Engineering Department, Concordia University, Montreal, Canada.