Abstract

1
Wind Loads on Single-Family Dwellings in Suburban
Terrain Comparing Field Data and Wind Tunnel
Simulation
P.L. Datina, Z. Liua, D.O. Prevatta, K. Gurleyb,
F.J. Mastersc, and T.A. Reinholdd Departments of
Civil Engineering at Clemson Universitya,
University of Floridab and Florida International
Universityc and the Institute for Business and
Home Safetyd, Tampa, FL
Wind Tunnel Testing
Abstract
House Instrumentation

• Components and cladding on residential structures
  continue to be damaged in high winds despite
  improvements to building codes. Modern wind
  design codes that are helping to prevent
  structural damage to buildings are not as
  effective in preventing building envelope
  failures. Since 1998, a unique collaborative
  effort between Clemson University, the University
  of Florida, and Florida International University
  called the Florida Coastal Monitoring Program
  (FCMP), has been collecting full-scale pressure
  and wind speed data on residential buildings in
  suburban areas. The objectives of the FCMP are
• To increase our limited full-scale data available
  on wind loadings of low-rise buildings none of
  which was obtained from hurricanes until this
  project.
• To determine the validity of current design code
  values that are based on open-country derived
  pressures.
• The issues highlighted above serve as the primary
  motivation to study pressure variation on roofs
  of suburban houses.

• Wind loads in modern building codes are based on
  boundary layer wind tunnel model studies
  conducted in flows that simulate the
  characteristics of winds associated with the
  passage of frontal systems. Questions are
  continually raised concerning the validity of
  these simulations for hurricane wind events.
• The FCMP employs a unique system using pressure
  sensors mounted on pre-selected houses that
  collect wind pressure data on the roofs of
  occupied residential structures along the Florida
  and Carolina coastlines. There are 32 homes in
  Florida participating in the FCMP and six more in
  the Carolinas. Each house is has pressure
  sensors installed on the building envelope (roof,
  walls, soffit) arrayed in areas that experience
  the highest wind suctions.
• Until the success of the FCMP in 2004, wind load
  data for residential structures subjected to
  sustained hurricane level winds has been
  virtually

• All wind tunnel studies were conducted in the
  atmospheric boundary layer wind tunnel at the
  Wind Load Test Facility (WLTF) of Clemson
  University. The upwind terrain used for this
  study was modeled to simulate suburban terrain at
  a 150 scale. Pressure taps are located where
  the actual sensors on the house are located and
  they are also concentrated in the worst loaded
  critical zones near the eaves and ridge corners
  of the roof.

Figure 3 Instrumented house for collecting
full-scale wind pressures. Clockwise from left
Instrumented house with pressure sensors and
computer box wiring under the eaves computer
box installed pressure sensors
FCMP Deployment Record
Figure 7 Wind tunnel arrangement for 150
suburban terrain
non-existent. Consequently, there was very
little basis for supporting or modifying existing
wind load estimates for buildings in hurricanes.
The FCMP collected data in all four landfalling
2004 hurricanes, including roof pressure data
sets on 16 houses. Six of these houses were
subjected to sustained hurricane level winds, a
first in experimental wind engineering.

• This wind tunnel pressure data is then analyzed
  and converted to pressure coefficient (GCp)
  values normalized to 3-second gust wind speed at
  mean roof height. As a result, a direct
  comparison between the full-scale results and
  wind tunnel data can be made.

Introduction

• The current ASCE 7 wind design procedure ASCE
  7-02 effectively assumes that one set of
  pressure coefficients for components and cladding
  systems can be used regardless of the wind
  exposure. For buildings located in a suburban
  area that qualifies as exposure B for all wind
  directions, the wind loads will be calculated
  using the same set of pressure coefficients as
  those used for exposure C locations, but the
  design loads will be 20 45 lower because of
  the lower design wind speeds in exposure B.
  Recent field measurements of wind loads on a
  house subjected to tropical storm winds in
  suburban settings and subsequent boundary layer
  wind tunnel model studies question the validity
  of this approach.
• Central to the FCMP effort is the collection of
  full-scale high-fidelity data sets of the
  turbulent wind behavior near ground at multiple
  locations within a storms path. The three main
  goals of the program are
• To measure in-field hurricane wind velocities
  using portable instrumentation that can be placed
  in the path of the storm.
• To measure the resultant pressures on residential
  structures through instrumented houses and
  evaluate the effectiveness of inexpensive
  retrofit measures
• To compare the full-scale pressure with the
  results of wind tunnel tests.

Figure 8 150 scale model of FL-27
Figure 2 FCMP instrumented houses in FL
Results
Full-Scale Data

• Wind speed data is available on the FCMP website
  in near real-time at www.ce.ufl.edu/fcmp as the
  storm comes ashore. This data includes 10Hz
  sampling rate, 15-minute mean wind speed, and
  3-second gust wind speed.

• A wind tunnel test was conducted on an FCMP house
  (FL-27) located in the panhandle of Florida that
  was hit by Tropical Storm Isidore. Current
  testing is being done on this same house that was
  hit by Hurricane Ivan (2004). The peak 3-second
  gust wind speed measured in Tropical Storm
  Isidore at FL-27 was about 40 mph at a wind
  direction of 135. Pressure coefficients were
  determined from both the full-scale results and
  the wind tunnel analysis.
• The results showed that there is positive
  agreement between the average values of the RMS
  pressure coefficients for the full-scale tests
  and wind tunnel results, indicated in Figure 9 by
  the linear regression analysis indicating a small
  offset between the slope of the line y x and
  the slope of the comparative study. This change
  in slope is minimal, which suggests that the
  fluctuations in pressures are faithfully
  reproduced in the model study. There also
  appeared to be a shift in the full-scale pressure
  coefficients for all the pressure taps, which may
  be due to errors in temperature measurement or
  barometric pressures.
• Establishing a reference pressure for the full
  scale results was difficult, however once
  adjustments were made by matching the maximum and
  mean pressure coefficients (Cp) between
  full-scale and the wind tunnel models, it was
  found that the wind tunnel results consistently
  underestimate the peak minimum Cp values as
  compared to the full-scale values (Figure 10).

Figure 9 Linear regression analysis of
full-scale vs. model RMS Cp values
Tower Deployment
Figure 4 Typical wind speed summary available
on the FCMP website
Figure 5 FCMP Homepage at www.ce.ufl.edu/fcmp
Summary
When a storm approaches land, teams from CU, UF,
and FIU travel to meet the inbound hurricane with
specially designed portable equipment in tow on
trailers. Each of the six trailers unfolds into
a stiff 10 m (30 ft) tall tower designed to
withstand 200 mph gusts. These trailers are
placed in the hurricanes path

• The wind speed information, shown in Figure 6,
  provided wind speeds from Hurricane Ivan
  collected by different groups and is used by
  other researchers and engineers to determine
  failure loads during the storms, as well as to
  better understand the degradation of wind speed
  in a landfalling hurricane.

• In 2004, the Florida Coastal Monitoring Program
  has provided, for the first time, high resolution
  full-scale uplift pressure data on the roofs of
  real, occupied residential structures
  experiencing sustained hurricane force winds, a
  first for wind engineering research.
• The dataset now includes several houses that
  experienced more than one storm, allowing future
  comparisons of the effects of storm intensity and
  other characteristics on measured Cp values.
• The tower and house data provide the first direct
  measure of the wind speed/load/damage chain, and
  it has laid the groundwork for effective
  mitigation.

• several hours before impact. Each tower is
  outfitted with multiple sensors and a data
  acquisition system that measure
• Wind velocities at 5-meter elevation with 3-axis
  gill anemometers
• Wind velocities at 10-meter elevation with vane
  and 3-axis gill anemometers
• Average width of turbulent gusts
• Temperature
• Barometric pressure
• Rate of rainfall
• Relative humidity

Figure 10 FL-27 house results full-scale vs.
model scale (adjusted) pressure coefficients
Acknowledgement
Future Work
The authors would like to acknowledge the
generous support of the Department of Civil
Engineering at Clemson University, the Florida
Department of Community Affairs, FEMA, NOAA, the
SC and FL Sea Grant Consortia, and the Institute
for Business and Home Safety.

• Validation of full-scale results through
  additional tests.
• Wind tunnel and full-scale comparisons of Cp
  values.
• Develop turbulence intensity relationships to
  pressures.

Figure 1 FCMP portable 10-meter tower
The towers also have the ability to provide the
wind data in near real-time to a public access
website (www.ce.ufl.edu/fcmp).
Figure 6 Measured wind speeds during Hurricane
Ivan (2004) converted to 3-second gust at
10-meters