<%BANNER%>

Determination of Wind Uplift Forces using Database-Assisted Design (DAD) Approach for Light Framed Wood Structures

Permanent Link: http://ufdc.ufl.edu/UFE0041841/00001

Material Information

Title: Determination of Wind Uplift Forces using Database-Assisted Design (DAD) Approach for Light Framed Wood Structures
Physical Description: 1 online resource (134 p.)
Language: english
Creator: Mensah, Akwasi
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Civil Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: During major hurricanes, damages to light framed wood structures (LFWS) represented the largest proportion of monetary losses. The absence of wind load transfer mechanism in wood structures was identified as a major cause of their structural failures. Wind load paths in LFS are not well understood. This study aims to develop a better approach for determining wind design loads on LFWS. The study was part of an on-going National Science Foundation (NSF) funded project titled, Performance Based Wind Engineering: Interaction of Hurricane Forces with Residential Structures , which has a primary objective of investigating the relationship between spatially varying wind loads and structural load paths on LFWS. This study was accomplished in two phases. In Phase 1, a Database-Assisted Design (DAD) methodology was used to combine time histories of wind tunnel pressure coefficients with experimentally determined influence functions for a wood framed structure. From this analysis, structural reactions at roof-to-wall and wall-to-foundation connections were developed. Peak reactions were compared to wind design loads based on ASCE-7 (2005) provisions for main wind force resisting systems (MWFRS) and components and cladding (C & C). Whereas, peak reactions estimated, using DAD methodology, were higher than maximum reactions obtained using the MWFRS provisions, they were lower than C & C based maximum reactions. In Phase 2 of the project, an experimental study was conducted to validate the DAD methodology. A 1/3-scale LWFS instrumented, with surface pressure transducers and load cells, was the immersed in wind flow. Structural reactions were developed from measured roof pressures using the DAD methodology. A comparison of developed reactions with directly measured reactions showed a good agreement between their mean and peak values.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Akwasi Mensah.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Prevatt, David O.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041841:00001

Permanent Link: http://ufdc.ufl.edu/UFE0041841/00001

Material Information

Title: Determination of Wind Uplift Forces using Database-Assisted Design (DAD) Approach for Light Framed Wood Structures
Physical Description: 1 online resource (134 p.)
Language: english
Creator: Mensah, Akwasi
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Civil Engineering thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: During major hurricanes, damages to light framed wood structures (LFWS) represented the largest proportion of monetary losses. The absence of wind load transfer mechanism in wood structures was identified as a major cause of their structural failures. Wind load paths in LFS are not well understood. This study aims to develop a better approach for determining wind design loads on LFWS. The study was part of an on-going National Science Foundation (NSF) funded project titled, Performance Based Wind Engineering: Interaction of Hurricane Forces with Residential Structures , which has a primary objective of investigating the relationship between spatially varying wind loads and structural load paths on LFWS. This study was accomplished in two phases. In Phase 1, a Database-Assisted Design (DAD) methodology was used to combine time histories of wind tunnel pressure coefficients with experimentally determined influence functions for a wood framed structure. From this analysis, structural reactions at roof-to-wall and wall-to-foundation connections were developed. Peak reactions were compared to wind design loads based on ASCE-7 (2005) provisions for main wind force resisting systems (MWFRS) and components and cladding (C & C). Whereas, peak reactions estimated, using DAD methodology, were higher than maximum reactions obtained using the MWFRS provisions, they were lower than C & C based maximum reactions. In Phase 2 of the project, an experimental study was conducted to validate the DAD methodology. A 1/3-scale LWFS instrumented, with surface pressure transducers and load cells, was the immersed in wind flow. Structural reactions were developed from measured roof pressures using the DAD methodology. A comparison of developed reactions with directly measured reactions showed a good agreement between their mean and peak values.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Akwasi Mensah.
Thesis: Thesis (M.S.)--University of Florida, 2010.
Local: Adviser: Prevatt, David O.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0041841:00001


This item has the following downloads:


Full Text





DETERMINATION OF WIND UPLIFT FORCES USING DATABASE-ASSISTED DESIGN
(DAD) APPROACH FOR LIGHT FRAMED WOOD STRUCTURES




















By

AKWASI FRIMPONG MENSAH


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2010
































2010 AkwasiFrimpongMensah



































To my parents Mr. & Mrs. P. K. Mensah









ACKNOWLEDGMENTS

The completion of this research and report would not have been successful without the

support and encouragement of a number of persons. I want to first thank the Almighty God

through Jesus Christ for being my Lord and sufficiency.

I also wish to express my sincere gratitude to my thesis committee chair, Dr. David O.

Prevatt for his continual support, time and guidance in this endeavor. I am grateful to my

committee members: Dr. Kurtis R. Gurley and Dr. Forrest Masters, and Dr. Gary Consolazio,

who have each contributed immensely to the success of this research. I am also grateful to the

faculty and staff of the Civil and Coastal Engineering Department for their tutorage and

assistance during my stay in University ofFlorida. I am again appreciative of the financial

support provided by National Science Foundation (NSF) through grant #080023 "Performance

Based Wind Engineering (PBWE): Interaction of Hurricanes with Residential Structures".

To my parents, Mr. and Mrs. P. K. Mensah, I am indebted to you for all the love,

encouragement, care and support you have given me. God richly bless you. I also wish to thank

all the friends I met in Gainesville. Of mention especially are Patrick Bekoe and Joyce Dankyi

whose presence in this chapter of my live cannotbe overemphasized. My deep appreciation also

goes to my entire family and friends for their timely encouragements.

Lastly, I wish to express my profound gratitude to all my office colleagues and the

hurricane group of University of Florida especially Peter L. Datin, Jason Smith and Scott Bolton

for their assistance and contributions.









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S .................................................................................... .....................4

LIST OF FIGURES .................................. .. .. .... .... ................. 10

A B S T R A C T .............................................................................................13

CHAPTER

1 INTRODUCTION ................. ............................ ........................ .... 15

B ack ground and M motivation ......................................................................... ....................15
O bjectiv e ................... ...................1...................6..........
S co p e o f W o rk ...............................................................................17
O organization of Report .................................................... ............... ......... ... 17

2 LITER A TU R E R E V IEW ....................................................................... .......................... 19

W ind F low over Low -R ise B buildings .............................................................................. ... 19
Current Design Provisions of ASCE 7 for Wind Loads on Low-Rise Buildings .................20
Background on ASCE 7 Wind Load Provisions ................................ .................20
Analytical Procedure for Wind Design Loads on a Low-Rise Building .......................21
Limitations of Current Design Provisions......... ........................ .. ...............23
Database-Assisted Design (DAD) Methodology for a Low-Rise Building .........................25
Background of the DAD methodology....................................... ......................... 25
DAD Concept and Software D evelopm ent ........................................ .....................26
Limitation to the Application of DAD Approach......... ..................... .............. 27
Design and Construction of Light Framed Wood Structures and their Performance to
W in d F o rce s ............................................................................ 2 7
C construction M methods ................................................ .. ............. ......... 27
C critical C om ponents and System s ..................................................._........................ .28
Structural Failures of LFW S in Hurricane Events ....................... ........................... 29
Wind-Induced Pressures and Structural Responses on Light Wood Framed Structures........29

3 ANALYSIS OF WIND TUNNEL DATA TO GENERATE PRESSURE
COEFFICIEN TS ..................................... .................. ................ .......... 36

Wind Tunnel Data................... .................................36
House M ode and Pressure Tap Layout...................................... ......................... 36
Wind Simulation and Pressure Measurements ...................................... ............ 36
Aerodynam ic Data Processing ..........................................................................37
Tubing Response Correction ................ ................ ......... ............ ......................... 37
Determining Pressure Coefficients .......... .... ........... .................. 38
Re-referencing ofPressure Coefficients...................................... ........................ 38
W ind T unnel R results and A nalysis.............................................................. .....................40









Wind Pressure Coefficients Time Histories ....................................... ............... 40
Observed Statistical Values of Wind Pressure Coefficients................. ............... 41
Extreme Value Analysis of Pressure Coefficients....................................42
Area-Averaged Pressure Coefficients ........................................ ........................ 44

4 APPLICATION OF DAD METHODOLGY ........................................ ..................... 57

Structural Influence Function ............................................. ..........................................57
Evaluating Vertical Reactions Based on DAD methodology...............................................59
Velocity Pressure ..................................................................... ........ 59
Pressure Taps and Influence Functions ........................................ ....................... 59
R eactio n L oads ................................................................60
D A D Results and A analysis ............................................................................................ 60
Observed Statistical Values of Structural Reactions..................... ..... ............... 60
Extreme Value Analysis of Vertical Loads................................ ...............61
Vertical Reaction Based on ASCE 7-05 Standard.................................... ............... .. ............62
Velocity Pressure ...................................................................... ........ 63
A SCE 7-based D design Loads ........................................ ..........................................63
Comparing Uplifts Reactions Predicted Based on DAD vs. ASCE 7-05.............................64

5 WIND FLOW CHARACTERIZATION USING TFI COBRA PROBE .............................78

T he C obra P robe ............................... ........................ 78
Preliminary Experiments Using the Cobra Probe.............................................................79
Comparing Wind Flow Measurements by the Cobra Probe and Hotwire
A nem om eter ..................................................................................................... ...... 79
Wind Tunnel Model ...................................... ........ ..... ...80
Mapping of Wind Field Generated by UF Wind Generator ....................................... 80

6 WIND INDUCED PRESSURE AND STRUCTURAL LOAD MEASUREMENTS ...........89

M materials and M ethod s ................................................................................ ....................89
Scale H house M odel .................... ........ ..................... 89
Pressure and Load Sensors on the Building ........................................ .....................89
Test Arrangement ............... ............................. .. ........ .. ............... 90
W ind G eneratio n .............................. ......................................................................... 9 1
Experimental Procedure and Measurements ....................................... ............... 91
Experimental Results, Analysis and Discussion................................... ...................... 92
W ind P pressure M easurem ents .............................................................. .....................92
W ind-Induced Structural Loads ........................................................... ................ ..93
Structural L oad C om prison ......................................... .............................................94

7 CONCLUSION AND RECOMMENDATIONS............................... .........................111

S u m m ary ............. .. ...............111................. .........
C o n c lu sio n s ............. ..... ............ ................. ..................................................1 12
Recommendation ............................. ...................................113



6









APPENDIX

A MEAN, RMS AND EXTREME VALUES OF WIND TUNNEL PRESSURE
CO EFFICIEN TS ........................................ ................ .................. ........... ..114

B MEASURED STATISTICAL VALUES OF VERTICAL REACTIONS DERIVED
FROM W IND TUNN EL DA TA ........................ ................... .................................116

C PRESSURE COEFFICIENTS MEASURED IN THE 1/3-SCALE HOUSE TEST ............124

D STRUCTURAL REACTIONS MEASURED IN THE 1/3-SCALE HOUSE TEST...........127

L IS T O F R E F E R E N C E S .................. ......... ........................................................................... 130

BIO G RAPH ICA L SK ETCH .............................................................. 134









LIST OF TABLES
Table page

2-1 Comparison of bending moments (KNm) determined using ASCE 7-98 and DAD
(S im iu et al. 2 0 0 3) ........................................................................ 3 3

3-1 W ind tunnel study configuration and parameters ................................... .................46

3-2 L ieb lein B L U E estim ato rs ........................................................................ ...................46

3-3 Comparison of wind tunnel and ASCE 7-05 peak pressure coefficients........................46

3-4 Measured peak, mean and RMS reactions for sample 1 ....................................... 66

4-2 A averaged m ean and RM S reactions........................................................ ............... 67

4-3 Parameters for Type I Extreme Value Distribution of peak negative reactions ...............68

4-4 Expected peak negative reactions and standard deviation estimated from BLUE fitted
probability distribution................ ........... .. .. .... .. ..... ........... 69

4-5 Pressures based on MWFRS for different building surface.............................................69

4-6 Pressures based on C& C for different zones .....................................................................69

4-7 Comparison of uplift reaction estimates based on DAD and ASCE 7-05.........................70

5-1 Comparison of flow measurements by Cobra Probe and hot-wire anemometer ..............82

6-1 M manufactures and specifications ofpressure sensors ................................ ............... 97

6-2 Statistical values of measured pressure coefficients ......................................................98

6-3 Coefficients of BLUE for Type 1 Extreme-Value Distribution (Lieblein 1974)...............99

6-4 Mean, RMS and BLUE estimated peak values of measured structural loads (lbs) ..........99

6-5 Correlation coefficients measured between time histories of measured and estimated
reaction n s ........................................................................................ 10 0

6-6 Comparison of directly measured and DAD-based estimated reactions ......................100

A- Peak, mean and RMS pressure coefficients of selected pressure taps.............................114

B-l Measured peak, mean and RMS reactions (lbs) for sample 2..........................................117

B-2 Measured peak, mean and RMS reactions (lbs) for sample 3 ............... ................118

B-3 Measured peak, mean and RMS reactions (lbs) for sample 4 .............. ... ................119









B-4 Measured peak, mean and RMS reactions (lbs) for sample 5 ...............................120

B-5 Measured peak, mean and RMS reactions (lbs) for sample 6......................................121

B-6 Measured peak, mean and RMS reactions (lbs) for sample 7.......................................122

B-7 Measured peak, mean and RMS reactions (lbs) for sample 8.......................................123

C-1 Measured peak, mean and RMS pressure coefficients for test repeat 1 ........................124

C-2 Measured peak, mean and RMS pressure coefficients for test repeat 2 ........................125

C3 Measured peak, mean and RMS pressure coefficients for test repeat 3 ........................126

D-1 Measured peak, mean and RMS reactions for test repeat 1 ..........................................127

D-2 Measured peak, mean and RMS reactions for test repeat 2............................................128

D-3 Measured peak, mean and RMS reactions for test repeat 3 ..........................................129









LIST OF FIGURES


Figure page

2-1 Separation and reattachment pattern of wind flow over a low-rise building ...................33

2-2 Typical building surfaces for ASCE 7-05 MWFRS external pressure coefficients ..........34

2-3 ASCE 7-05 provision for determining external pressure coefficients for the design of
com ponents and cladding ................................................................................ .... ..........34

2-4 Isometric view of the steel portal frame structure..........................................................35

2-5 Comparison of vertical reaction records measured by load cells and estimated based
on envelope roo f p ressures............................................................................... ...... ...35

3-1 1:50 Scale house model used in the wind tunnel study ........................................... 47

3-2 Test section arrangement for 1:50 suburban terrain................................... ... ..................47

3-3 Wind flow characteristics for 1:50 suburban wind tunnel study. ......................................48

3-4 Frequency response characteristics of the pressure tubing system.............. .................48

3-5 Typical wind pressure coefficient time histories .................................... ............... 49

3-6 Format of MATLAB files ofpressure coefficient data ........................................... 49

3-7 Spatial distributions of mean wind pressure coefficients ................................ ..........50

3-8 Spatial distributions of RMS of pressure coefficients ................................................51

3-9 Spatial distributions of expected negative peak wind pressure coefficients...................52

3-10 Spatial distributions of expected positive peak wind pressure coefficients.....................53

3-11 Area-averaged pressure coefficients for regions corresponding to zone 1 ......................54

3-12 Area-averaged pressure coefficients for regions corresponding to zone...........................55

3-13 Area-averaged pressure coefficients for regions corresponding to zone 3........................56

4-1 1/3 Scale house model for determining influence functions............... ...............70

4-2 Locations of load cells and wind direction ................................... .......... ...... .. ....... 71

4-3 Grid points for experimental determination of influence coefficients.............................71

4-4 Influence lines for vertical reactions at a support of an internal truss ............. ...............72









4-5 Typical influence surfaces for vertical reaction determined on 1/3-scale house model ....72

4-6 Typical wind-induced reaction time histories ....................................................... 73

4-7 Estimated mean and peak vertical reactions at load cells for wind azimuth 000.............73

4-8 Estimated mean and peak vertical reactions at load cells for wind azimuth 045 ............74

4-9 Estimated mean and peak vertical reactions at load cells for wind azimuth 090 .............74

4-10 Estimated mean and peak vertical reactions at load cells for wind azimuth 1350.............75

4-11 Estimated mean and peak vertical reactions at load cells for wind azimuth 180.............75

4-12 Building surfaces for determining wind pressure for each truss based on ASCE 7
provisions for M W F R S ............................................................................ ....................76

4-13 Zones for determining wind pressure for each truss based on ASCE 7 provisions for
components and cladding .................................................................. ............... 76

4-14 Uplift reactions at roof-to-wall load cells based on DAD approach and ASCE 7-05
provisions ......... ......................................77

5-1 C obra P rob e.....................................................82

5-2 Cobra Probe and Hot-wire anemometer setup for simultaneous measurements ...............82

5-3 W ind tunnel model used in pilot studies ................................................. ..... .......... 83

5-4 Measuring points for mapping flow measurements at of the wind tunnel.........................83

5-5 Flow measurements at exit of w ind tunnel m odel ........................................ ..................84

5-6 Spatial variations of long tudinal velocity and turbulence intensity across exit section
of w ind tunnel m odel ................... .... .... ...................... .. .. ...... ............... 84

5-6 Spatial variations of lateral velocityand turbulence intensity across exit section of
wind tunnel model........... .... ................... ............. ....84

5-8 Spatial variations of lateral velocity and turbulence intensity across exit section of
w ind tunnel m od el ........................................................................ 85

5-9 Spectral contents of wind speed at 0.5 in. downstream of the exit of the wind tunnel......85

5-10 Position of traverse frame in test section ..................................................................... 86

5-11 Location of measurements points for flow mapping ............................... ............... .86

5-12 Variation of longitudinal wind speed across 2D measurement surface.............................87









5-13 Variations of lateral and vertical wind speed across 2D measurement surface.................87

5-14 Spectral contents of w ind speed............................................... .............................. 88

6-1 Locations ofpressure taps and load cells on the scale house model. ............................100

6-2 A sample pressure tap and layout of pressure taps on the roof....................................101

6-3 Interior of the house m odel ................................................................... ..................... 10 1

6-4 Pressure Sensors A) Omega PX 138 B) Setra 265 C) Dwyer 616.............................. 102

6-5 Futek load cells A) Roof-to-wall connection B) Wall-to-foundation connection ...........102

6-6 Fluke pressure calibrator ........................................... ....................... ............... 102

6-7 Dwyer Transducer and p itot-tube for wind velocity measurements............................103

6-8 Layout of experim mental set-up ..................................................................... ........... 104

6 -9 T est setup ...............................................................10 5

6-10 UF W ind Generator................... ............... ........ ................ ......... 106

6-11 Spatial distributions of peak pressure coefficient .....................................................106

6-12 Spatial distributions of mean pressure coefficient ...........................................................107

6-13 Location of load cells on the house........................................................... ............... 107

6-14 Mean and expected peak values of measured structural loads for wind direction 00 ......108

6-15 Mean and expected peak values of measured structural loads for wind direction45 ....108

6-16 Mean and expected peak values of measured structural loads for wind direction 90 ....109

6-17 Comparison of measured and estimated reactions for wind direction 0.........................109

6-18 Comparison of measured and estimated reactions for wind direction 45 .......................110

6-19 Comparison of measured and estimated reactions for wind direction 90.......................110









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science


DETERMINATION OF WIND UPLIFT FORCES USING DATABASE-ASSISTED DESIGN
(DAD) APPROACH FOR LIGHT FRAMED WOOD STRUCTURES


By

Akwasi Frimpong Mensah

August 2010

Chair: David Prevatt
Major: Civil Engineering

During major hurricanes, damages to light framed wood structures (LFWS) represented the

largest proportion of monetary losses. The absence of wind load transfer mechanism in wood

structures was identified as a major cause of their structural failures. Wind load paths in LFS are

not well understood.

This study aims to develop a better approach for determining wind design loads on LFWS.

The study was part of an on-going National Science Foundation (NSF) funded project titled,

"Performance Based Wind Engineering: Interaction of Hurricane Forces with Residential

Structures", which has a primary objective of investigating the relationship between spatially

varying wind loads and structural load paths on LFWS.

This study was accomplished in two phases. In Phase 1, a Database-Assisted Design

(DAD) methodology was used to combine time histories of wind tunnel pressure coefficients

with experimentally determined influence functions for a wood framed structure. From this

analysis, structural reactions at roof-to-wall and wall-to-foundation connections were developed.

Peak reactions were compared to wind design loads based on ASCE-7 (2005) provisions for

main wind force resisting systems (MWFRS) and components and cladding (C&C). Whereas,









peak reactions estimated, using DAD methodology, were higher than maximum reactions

obtained using the MWFRS provisions, they were lower than C&C based maximum reactions.

In Phase 2 of the project, an experimental study was conducted to validate the DAD

methodology. A1/3-scale LWFS instrumented, with surface pressure transducers and load cells,

was the immersed in wind flow. Structural reactions were developed from measured roof

pressures using the DAD methodology. A comparison of developed reactions with directly

measured reactions showed a good agreement between their mean and peak values.









CHAPTER 1
INTRODUCTION

Background and Motivation

Wind flow over low-rise buildings is characterized by patterns of flow separation and

reattachment which creates spatially and temporally varying pressure fields on building surfaces.

Generally, peak wind suction forces occur on the leeward walls and at roof edge areas while

positive pressure are created on the windward walls and interior roof areas. Such forces in recent

hurricanes caused substantial damage to wood-framed residential structures.

Hurricane damage to light framed wood structures (LFWS) are by far the largest

contributor to the monetary losses associated with hurricane disaster (Rosowsky et al. 2003).

Post-hurricane inve stigations report widespread structural damage to wood structures due to loss

of roof sheathing, and failure of load transfer at joints and mechanical connections (FEMA 2005;

van de Lindt et al. 2007). Clearly, there is a lack of understanding of structural load paths in

wood structures.

Wind design of light framed wood residential structures is problematic because of their

complex geometric shapes. Current wind design provisions lack codified pressure values for

typical residential buildings. i.e. pressure coefficients are only provided only for simple shaped

building. Moreover, high variability in material properties of wood introduces greater uncertainty

in wind resistance estimates. Selection of materials and connections for LFWS has been mainly

based on prescriptive building guidelines which increase their susceptibility to wind damages.

In the United States, wind load design provisions are included in the ASCE 7 (2005),

which codifies information on wind flow characteristics (obtained from meteorological data) and

aerodynamic pressures (developed on scaled models in boundary layer wind tunnels). Pressure

coefficients and climatological data, which are used by structural engineers for wind load









designs, are presented in reductive figures and tables suitable for hand calculations. However, a

research conducted by Simiu & Stathopoulos (1997) suggests that such design standards can

produce risk inconsistent results. Simiu & Stathopoulos (1997) asserted that the current code

provides insufficient information for designers to realistically account for the spatial and

temporal variation of wind load effects.

To address these codification deficiencies, a new wind analysis approach, in which utilizes

large aerodynamic and climatological databases are used to define wind design loads, was

proposed by Simiu & Stathopoulos (1997). This analysis methodology, called database assisted

design (DAD) was used by Simiu et al. (2003) to estimate internal forces in a steel portal framed

building. They used wind tunnel databases of surface pressure time-histories and analytically

derived influence functions to determine bending moments at knee and ridge joints of the portal

frames. When the DAD results were compared with ASCE 7 design values, they concluded that

the ASCE provisions produced risk inconsistent designs and errors in excess of 50 % in peak

load estimations.

With the availability of powerful computations and proven usefulness of the DAD

methodology, its application has been extended to light-framed wood structures. The hypothesis

of this research is that, the DAD methodology will provide better accuracy in predicting wind

load effects on LFWS than using the current codes. The validation of this hypothesis would lead

to a better understanding of structural load paths on LFWS and improved engineering design

models for LFWS.

Objective

The specific objectives of this investigation are to:

1. Apply the DAD methodology to predict structural reactions in a LFWS system.









2. Validate the DAD approach by experimentally determining structural loads on a 1/3 wood
building model.

Scope of Work

Wind tunnel data developed on a 1/50 scale house model were analyzed to generate spatial

distributions of wind pressure coefficients on the roof ofa gable roofbuilding. The DAD

methodology was used to combine the pressure time histories and experimentally-derived

influence coefficients for vertical reactions on a third scale wood-structure. The data was

analyzed to determine peak values of roof-to-wall and wall-to-foundation connection loads for

LFWS. Results of the analysis were compared to wind design loads based on the ASCE 7-05

standard.

In phase two of this study, the validity of the DAD approach was evaluated experimentally

by subjecting the third-scale house to fluctuating wind forces while simultaneously measuring

surface pressures and structural reactions. The DAD methodology was applied by utilizing the

measured pressure distributions and the influence coefficients to determine reaction loads at

roof-to-wall connections. The results and directly measured structural loads were compared.

Organization of Report

Literature reviews of relevant topics to this project are presented in Chapter 2. The chapter

discusses wind load effects on low-rise buildings, the current provisions for wind load designs

and the concept and de velopment of the DAD approach. Finally, a review of previous

experimental studies in which structural responses and wind pressures were simultaneously

monitored on LFWS is presented.

In Chapter 3, the wind tunnel study, which produced the aerodynamic pressure data

utilized in this project, is introduced. Analysis of wind tunnel derived pressures to generate a

pressure coefficient database for this study is described. Extreme value analysis based on the









Lieb lien BLUE (best linear unbiased estimators) estimation procedure to obtain the expected

peak pressure distributions is explained also in Chapter 3. Lastly, area-averages of pressure

coefficients from wind tunnel analysis are compared to ASCE 7-05 external pressure coefficients

for components and cladding.

The DAD-based procedure for evaluating wind load reactions is described in Chapter 4.

Chapter 4 also contains an overview of experimental derivation of structural influence functions

for this study as well as experimental results. Analysis of results to estimate peak reactions is

discussed in this chapter. Peak reactions based on the DAD approach are compared to results

based on ASCE 7-05.

In Chapter 5, the TFI Cobra Probe, which is used in the experimental study, is introduced.

Tests undertaken to validate and understand the operations of this equipment are reported as

well. Finally, characterization of the wind field for the experimental study is also described.

In Chapter 6, the 1/3-scale model house experiment is described. The chapter contains

descriptions of materials and equipment used in the tests, such as the 1/3 scale house, UF wind

generator and load and pressure sensors. The test arrangement and procedures are also described.

Finally, analysis of experimental results and correlation of structural loads derived from pressure

measurements and directly measured structural loads are reported in this chapter.

A summary of the entire project is contained in Chapter 7. The usefulness of this research

is also discussed in this chapter. Lastly, recommendations for future work are presented.









CHAPTER 2
LITERATURE REVIEW

Wind Flow over Low-Rise Buildings

Wind loading on a building depends upon the flow pattern around the building which, in

turn, depends on building geometry, dimensions, surroundings, upstream terrain and wind flow

characteristics.

Wind flow over a low-rise building is characterized by separation and reattachment pattern

(shown in figure 2-1) which together with its velocity fluctuations generate a spatially and

temporally varying pressure field on the surface (Ginger et al. 2000). Ginger and Letchford

(1993) observed that large fluctuating suction pressures are generated in flow separation regions

close to the leading edges of the roof of low rise buildings. They explained that the flow

mechanisms that generate these pressures are the 2D separation bubble for flow perpendicular to

the edge discontinuity and the 3D conical vortex for flow at oblique angles to the edge

discontinuity and that the largest suction pressures are generated close to the leading corner for a

wind orientation of approximately 30.

Ginger et al. (2000) determined wind loads on the roof of a typical low-rise house for

approach wind directions of 00 to 900 by carrying out a wind tunnel model study at a 1/50

geometric scale. They observed that the second truss from the windward gable end was subjected

to the largest wind load

Stathopoulos et al. (2000) conducted and presented a wind tunnel study which provided

detailed extreme local and area-averaged pressure coefficients for low-building roofs exposed to

open-country upstream terrains. They observed that when the wind flow is normal to the

ridgeline of a gable roof building quasi-flat roofs in the range of 0-300 create a similar flow

pattern of separation, entrainment, and reattachment; a high suction prevails, especially at the









windward edges and corners. They noted however that, if the roof angle is greater than 300, wind

flow generally strikes on the windward roof prior to separating from the windward edge or ridge

which induces a positive pressure region on part of the windward slope and a negative region on

the leeward slope. They concluded that these flow patterns and pressure distributions may vary

with the wind direction, but remain comparable in respective roof slope ranges.

Current Design Provisions ofASCE 7 for Wind Loads on Low-Rise Buildings

Background on ASCE 7 Wind Load Provisions

The provisions ofASCE 7-05 for wind loads on low buildings are largely based on wind

tunnel study works conducted in the late 1970s at the boundary wind tunnel in the University of

Western Ontario (UWO)(Davenport et al. 1978; Stathopoulos 1979). Researchers at UWO used

an approach that consisted essentially of permitting the building model to rotate in the wind

tunnel through a full 360 in increments of 450 while simultaneously monitoring the loading

conditions on each surfaces. Both open and suburban exposure conditions were considered.

Wind pressure coefficients which represent "pseudo" loading conditions, that when applied to a

building envelope the desired structural actions (bending moment, shear, thrust), and the

maximum induced force components to be resisted for all possible directions and exposures were

developed from the studies (see C6.5.11 (ASCE/SEI. 2005)).

The current edition of the ASCE 7 standard (2005) has refined pressure and force

coefficients to reflect the latest boundary-layer wind tunnel and full-scale research findings. This

research has been however only limited to gable-roofbuildings, and a rational method of

applying the coefficients to hip roofs based on experience, intuition and judgment has been

developed and presented in ASCE 7-05.









Three methods are provided in the ASCE 7 standard for determining wind design loads.

These are the "simplified method" (method 1), the analytical procedure (method 2) and a wind

tunnel procedure (method 3).

Analytical Procedure for Wind Design Loads on a Low-Rise Building

The main wind-force resisting system (MWFRS) of a building consists of a structural

frame or an assemblage of structural elements such as roof trusses, cross-bracing shear walls and

roof diaphragms that work together to transfer wind load action on the entire structure to the

ground (ASCE/SEI. 2005). MWFRS provides support and stability for the overall structure and

generally receives wind loading from more than one surface. ASCE 7-05 defines components

and cladding as elements of the building envelope that do not qualify as part of the MWFRS.

Cladding receives wind loads directly. Components receive wind loads directly or from cladding

and transfer the load to the MWFRS. Members which are categorized as components and

cladding included fasteners, purlins, girts, studs, roof ducking, and rooftrusses.

In the determination of design wind loads on all buildings, a velocity pressure, qz, is

evaluated at a height z above the ground using the equation below:

q, = 0.00256K,K,,KdV2I (lb ft2) (2-1)

where K, is velocity pressure exposure coefficient, obtained from table 6-3 of ASCE 7-05,

which modifies the design wind speed to account for terrain exposure condition and the height z;

Kz is a topographic factor which accounts the wind speed-up (topographic) effect; Kd is the wind

directionality factor, which is 0.85 for buildings to account for reduced probability of maximum

winds coming from any direction and the reduced probability of the maximum pressure

coefficient occurring for any wind direction; V is the basic wind speed determined from figure 6-

1 in ASCE 7-05 and its value is a nominal 3-second gust wind speed in miles per hour at 33 ft









above ground for an open exposure; and I is importance factor of the building determined from

table 6-1 inASCE 7-05 which is used to adjust the level of reliability of building or structure to

be consistent with the building classifications indicated in the standard.

Design wind pressures, for both MWFRS and components and cladding, are determined as

the product of the velocity pressure and the sum of internal and external pressure coefficients.

The internal pressure coefficients, GC,, are provided in figure 6-5 ofASCE 7-05 in terms of the

building enclosure classification (i.e. open, partially enclosed or enclosed building). The external

pressure coefficients are given separately for MWFRS and components and cladding for

different scenarios but generally in terms of pressure zones. Pressure zones specified in the

ASCE standard for both MWRS and Components and Cladding are in terms of a dimension

denoted by a (Simiu and Miyata 2006). The dimension a is 10% of the least horizontal building

dimension or 0.4 h (h=mean roof height), whichever is smaller, but not less than 4% of the least

horizontal building dimension or 3ft.

Design wind pressures on the MWFRS of low-buildings are determined by the equation

below:

p= qh[(GCp )-(GC,)] (lb /ft2) (2-2)

Where: qh is velocity pressure evaluated at mean roof height using equation 2-1, GCp,

internal pressure coefficient (obtained from Figure 6-5 in ASCE 7-05) and GCfis an external

pressure coefficient combined with a gust effect factor. Values for GCpare provided in figure 6-

10 in ASCE 7-05 as a function of the building roof angles. Roof overhangs are to be designed for

a positive pressure on the bottom surface of windward roof overhangs corresponding to C, = 0.8

in combination with the pressures determined from Figure 6-10. For determining eternal pressure

coefficients, eight loading patterns are to be considered to design the building for all wind









directions. The loading patterns have the walls and roofs zoned into several building surfaces

which envelope wind load distributions on the building. Figure 2-2 shows typical load patterns

in the ASCE 7-05 for wind design loads on a MWFRS of a building.

Design wind pressures on component and cladding of low-buildings are determined by the

equation below:

p = qh[(GC, (GCp,)](lb /ft2) (2-3)

where GC, are the external pressure coefficients and the other terms are as defined

previously. Values of GC, are selected from Figures 6-11 through 6-16 of the ASCE 7-05 based

the type of roof and angle roofs. Figure 2-3 shows the typical pressure zones of a gable roof

building and external pressure coefficient provision for roof angles between 70 and 270. External

pressure coefficients for deign of component and cladding are specified for the wall, roof and

overhang as a function of effective wind area. The effective wind area is defined by the ASCE 7-

05 as the span length multiplied b y an effective width that need not be less than one-third the

span length. It is worth noting that the resulting induced wind load is however applied over the

actual tributary area to the component been designed.

Limitations of Current Design Provisions

Several investigations have over the years been conducted and results compared with

ASCE 7-05 wind load provisions. Issues have been raised by researches on the standard

provisions and this section discusses some of them.

Simiu et al. (2003) illustrated the practical effects of simplifications inherent in the ASCE

7-05 provisions by evaluating moments in steel portal frames of a building(shown in Figure 2-4)

by using ASCE 7-05 standard provisions on one hand and the DAD procedure (discussed later in

this chapter) based on wind tunnel database on the other. Table 2-1 shows values obtained. Simiu









et al. (2003) demonstrated that the use of the tables and plots in wind load design provisions can

entail errors that can exceed 50% in the estimation of wind effects.

Furthermore, Whalen et al. (2002) assert that the accuracy in the definition of wind loads

inherent in such tables and plots are lower than that inherent in current methods for stress

computation. There is so much complexity with geometries and shapes of low-rise buildings and

hence high accuracy in predicting design loads based on tables and plots cannot be achieved.

Wind directionality effects on low-rise buildings are accounted for in the ASCE 7 standard

by a reduction factor of 0.85. Simiu et al. (2003) observed that this approach is inadequate as

wind effect reductions due to directionality effects are less significant as the mean recurrence

interval of the wind effects increase, rending the use of this factor potentially unsafe, particularly

for MWFRS.

Design parameters such as building geometry, building orientation, proximity of adjacent

structures and, the spatial and temporal variation of wind loads are not realistically and

comprehensively accounted for when a designer uses the conventional standard provisions

(Simiu and Stathopoulos 1997).

Recently, wind tunnel test data on generic low buildings were obtained at UWO to

contribute to the National Institute of Standards and Technology (NIST) aerodynamic database

(Ho et al. 2005b). St. Pierre et al. (2005) compared the NIST aerodynamic database to the

historical data obtained by Stathopoulos in the late 1970s, from which the current ASCE 7

provisions were developed. They observed that for the exterior bay of the test building model,

ASCE 7 generally underestimates the response coefficients significantly. For the interior bays,

the ASCE 7 overestimates the response coefficients. They also observed that generally, there was

10-85% underestimation of peak response coefficients in the suburban terrain by ASCE 7.









Attempts by writers of the standard provisions on wind loads to reduce the limitations of

earlier versions of the standard (example ASCE 7-05) resulted in bulky and complex provisions

(Simiu and Stathopoulos 1997).

Database-Assisted Design (DAD) Methodology for a Low-Rise Building

Background of the DAD methodology

With the backdrop of the above mentioned limitations of the wind design load provisions,

it was necessary to work on an alternative approach which offers the potential for significantly

more risk-consistent, realistic, safer and economical design by using adequate aerodynamic

databases and information. Owning to current information storage and computational

capabilities, Simiu and Stathopoulos (1997) proposed a new generation of standard with

provisions on wind loads that are no longer based on reductive and distorting tables and plots,

but can be structured on knowledge-based systems drawing the requisite information from large

databases.

Their postulation was that, wind loads evaluated via the new methodology would be

functions of design parameters, which includes building geometry, building orientation, position

with respect to and geometry of neighboring buildings, built-up terrain roughness, etc. They

intimated that their proposal would allow the designer to target specific situations, rather than

providing blanket coverage for a broad range of situations. They explained that this methodology

would furthermore allow the designer to account for the specific linear or non-linear structural

characteristics of the building or structure (eg. influence function).

Subsequently, Whalen et al. (1998) conducted a pilot project on the estimation of wind

effects in low-rise building frames using this methodology. Whalen et al. (1998) used records of

pressure time histories measured at large number of taps on a building surface at the UWO

boundary layer wind tunnel. Time histories of bending moments in a frame were obtained by









summing up pressures time histories tributary to that frame multiplied by the respective tributary

areas and frame influence coefficients. They compared results with results based on ASCE 7

standard provisions. Their comparison suggested that, significantly more risk-consistent, safer

and economical designs could be achieved using this approach than using conventional standard

provisions.

The approach of using electronic aerodynamic and climatological databases to define wind

loads was coined "database-assisted design" (DAD) and was accepted by the ASCE 7-98

standard (Rigato et al. 2001).

DAD Concept and Software Development

The first generation DAD application called WiLDE-LRS Wind Load Design

Environment for Low-rise Structures was developed by NIST (Whalen et al. 2000). WiLDE-

LRS, a MATLAB-based software, adopted interactive graphical user interfaces (GUI) to give a

visual, user-friendly design environment. MATLAB scripts were used in the software to analyze

the behavior of rigid portal frames and other components under high winds and to produce time

histories of wind load effects in these structural members. The software had its origins in a

prototype application called Frameloads, used to study wind effects on moment resisting frames

in low-rise buildings designed by the Standard Metal Building Manufactures Association

methodology. A latter version (2.7) ofWiLDE-LRS (Whalen et al. 2002) greatly enhanced the

GUIs that directly accepted input of influence coefficients accounting for frame properties. Post-

processing was incorporated in this version to calculate realistic and robust statistical estimates

of the peak load effect values based upon the entire time history.

Subsequently, the DAD approach has been extended to consider nonlinear static response

of low buildings (Jang et al. 2002) and also to account for the probability distribution of the

peaks of time histories of wind effects and of sampling errors in the estimation of that









distribution (Sadek and Simiu 2002; Sadek et al. 2004). A scheme to interpolate existing data in

available database to other configurations in a reliable, accurate and simple way, without

resorting to further wind tunnel experiments, has been incorporated in DAD applications (Kopp

and Chen 2006).

In 2006, NIST released software packages developed using the MATLAB language to

fully implement the DAD approach and all its improvements (Main and Fritz 2006). Two

separate software packages are available through the internet at http://www.nist.gov/wind for

rigid, gable-roofed buildings and for tall, flexible buildings.

Limitation to the Application of DAD Approach

To the be st of author's knowledge:

1. Application of the DAD approach and its software has been limited to steel portal frame
buildings.

2. Structural influence functions used by researches so far in DAD applications have been
analytically derived using hand-calculations or 2-D models in structural analysis software

3. .The validity of the approach has not been demonstrated experimentally.

Design and Construction of Light Framed Wood Structures and their Performance to
Wind Forces

Wood-frame construction forms the majority of residential and other low-rise structures. A

number of these structures are located along hurricane-prone zones in the United States. This

section discusses the construction methods prevalent in the wood-frame industry and their

performance during hurricane events. The literature presented here is based on studies done by

Rosowsky and Schiff (2003) and van de Lindt et al. (2007).

Construction Methods

Three construction methods have been identified by van de Lindt et al. (2007). These are

the conventional, engineered and prescriptive. The conventional method consists of following









documents such as the International Residential Code outlining certain exceptions and

limitations. Most wood constructions are based on conventional methods. For engineered

construction, structures are specifically designed by a design professional to meet jurisdictional

requirements. Interestingly, very few residential buildings are engineered. Prescriptive

construction involves the use of basic material strength level and tabulated values obtained from

construction manuals.

Rosowsky and Schiff (2003) referred to designs based on the conventional method as

deemed-to-comply design, which is largely derived from traditional rules of thumb for building

light-frame wood structures (LFWS). They observed that most of the rules focused on building

structures to safely resist gravity loads, ignoring geographic considerations. Until recently, most

buildings, including those located in high-wind environments, were constructed using

conventional methods which did not meet wind-resistant design requirements. This caused these

structures to have the greatest vulnerability to extreme wind events.

Critical Components and Systems

According to Rosowsky and Schiff(2003), the three most important areas to consider in

designing a wind-resistant wood-frame structure are:

1. The building envelope: This forms the first line of defense against wind and water
intrusion. Traditionally, the building envelope is considered to be architecture in nature
and therefore not de signed by engineers. However, studies have shown that a direct
correlation exists between the performance and damage (losses) sustained by wood-frame
buildings. Structural engineers are becoming actively involved in the building envelope
designs.

2. Attachment of roof and wall sheathing: this component is critical in keeping structures
enclosed, preventing infiltration and providing critical links in the structural load path.
Removal of roof sheathing is the second largest failure mode observed in post-hurricane
investigation after removal of roof cover. Significant highlight has been given to the need
to provide more and larger fasteners around roof edges to resist high wind uplift pressures.









3. Structural systems to transfer the applied loads to foundation: In most wood-frame
construction, complicated load paths exist because of conventional framing techniques and
irregular floor and roof plans in residential buildings.

Structural Failures of LFWS in Hurricane Events

Structural observation made by van de Lindt et al. (2007) during a reconnaissance trip after

Hurricane Katrina are discussed as follows:

1. In many of the houses examined, there was absence of continuous load path for the transfer
of wind loads from the roof down to the foundation.

2. Loss of roof sheathing at corners, which typically experience the highest uplift pressure
during wind storms. In most of these cases, the current code minimum nail spacing
requirements were not met.

3. Gable end wall losses as a result of loss of vinyl siding and failure of the foam sheathing.

4. The prevalent use of conventional construction in high wind regions.

Light framed wood structures (LFWS) have generally not performed well when subjected

to high wind loads due to design/construction practices. Rosowsky and Schiff (2003) remarked

that better understanding of the wind loading on buildings and behavior of wood-frame

structures under sever wind events must be sought. This, they noted, will lead to improvements

in both prescriptive and engineered design methodologies for new and retrofit construction.

Wind-Induced Pressures and Structural Responses on Light Wood Framed Structures

The final stage of this project is to validate the DAD methodology for its application to

LFWS by simultaneously monitoring pressures and structural loads on a 1/3 scale house

subjected to wind forces. This section presents experimental studies done by researchers

whereby wind-loads and structural responses were simultaneously measured on full-scale

buildings. These experiments are generally aimed at investigating whether observed structural

responses correspond to predictions by numerical models.









Doudak et al. (2005) monitored a single story industrial shed building to determine its

displacement response to wind and snow loads. He attempted to correlate the observed

displacements with real-time estimates using SAP 2000 of these environmental loads. Wind

speed and direction was measured as well as displacements on the building during the typical

wind storm season. Doudak et al. (2005) however did not take pressure measurements on the

house. Wind pressures for numerical simulations were estimated from archived pressure

coefficients and the measured wind speeds. They achieved a quite reasonable agreement between

measured and predicted displacements. Discrepancies ranged from as low as 6 % in most cases

to as high as 90 % for all four incident wind directions.

In a follow-up to the experiment done by Doudak et al. (2005), Zisis and Stathopoulos

(2009) undertook an experimental study whereby they monitored and collected full-scale

pressure and force data on a light framed wood building. A total of 51 load cells were installed at

roof-to-wall and wall-to-foundation interfaces of the building while ensuring that the stiffness of

the building was unaltered. The building was also equipped with 27 pressure taps. All acquired

data were analyzed and converted into dimensionless coefficients based on the following

equations:

Pressure measurements

P -P
C pea ean /peak a(2-4)
p,mean/ peak 1,/ 2pU H
/2pUBH

Force measurements


Cm R mean/ peak (2-5)
f,mean/peak (1/2pUSH)LW









where p = air density; UBH = wind speed at the building height; Pa = ambient atmospheric

pressure; P= actual surface pressure; R= reaction at the load cell location; L = length of the

building and W= width of the building.

Zisis and Stathopoulos (2009) also conducted wind tunnel experiments on a 1-200 scale

model and obtained pressure data which were also converted into pressure coefficients. Force

coefficients were also derived by area averaging the measured local pressure coefficients from

the wind tunnel studies on the building. These pressure coefficients were input into a numerical

(finite element analysis) model of the test building which computed reaction forces at each of the

27 foundation load cell locations. These forces were transformed into dimensionless coefficients.

They observed good a agreement of pressure distribution comparison between the wind

tunnel and full-scale data. They also concluded that the comparison between the full-scale load

cell readings and the base reactions computed by the finite element analysis made in the form of

force coefficients shows good agreement as far as mean values are concerned.

Zisis and Stathopoulos (2009) also conducted a 2-dimensional structural analysis of two

main frames of the building. In the analysis, the individual roof pressures tap records (full scale)

acting on each frame were used to evaluate the total expected vertical reaction due to wind

pressure on each frame. They compared the estimated results to the actual total reactions

measured by the respective load cells of each frame in terms of force coefficients. Their

comparison as well as the layout of their test building is shown in Figure 2-5. Zisis and

Stathopoulos (2009) found excellent agreement as far as the mean values were concerned. They

also observed significantly more fluctuating signals with higher peak forces were obtained using

the measured pressure coefficients on the building envelope in comparison with recorded signal

by load cells placed in the building foundation. They explained that this observation may be









partly attributed to the dynamic load attenuation effect due to structural and material damping of

the building components hence lower reactions measured than computed.

Both experimental studies discussed above exposed the test buildings to natural wind

forces. Consequently, wind pressure data collected from the field was highly affected by

fluctuations of wind directions during the test. Moreover, the structure does not experience winds

that would cause the worst load effects or design level events.










Table 2-1. Comparison of bending moments (KNm) determined using ASCE 7-98 and DAD
(Simiu et al. 2003)

Frame 6.1 m eave height 9.75 meave height
Knee Ridge Knee Ridge
ASCE DAD % ASCE DAD % ASCE DAD % ASCE DAD %

Outer 339 330 3 118 136 -13 463) 631 -27 86 137 -37

1 520 401 30 180 168 7 724 723 0 134 179 -25

2 471 301 56 163 97 68 624 799 -22 115 150 -23

3 471 310 52 163 101 61 624 782 -20 115 145 -21

4 471 327 44 163 106 54 624 586 6 115 112 3


Reatachmrent


layer


Wind
Forces


Figure 2-1. Separation and reattachment pattern of wind flow over a low-rise building (After
Simiu & Miyata (2006))























Refemce Reference
Corner A Comer B

Figure 2-2. Typical building surfaces for ASCE 7-05 MWFRS external pressure coefficients


y-:a- -1-W'A A .;..N
^-.tm-.". .t.L


10 10D


-2.02I -
-14

_ _4_ 1.1

0 ---- -- -
4 1 -







I C10 20 50 100 200 5001000
Eecv-- Wi-nd Aft i 2n M2
I 1<3 W 50 A 00f200 5M1002

EfficttroWind Arera, 1m2),


Figure 2-3. ASCE 7-05 provision for determining external pressure coefficients for the design of
components and cladding.


























Left Sde Right Sde FraneP4 (o e



Figure 2-4. Isometric view of the steel portal frame structure (Simiu et al. 2003) (End frame not
shown)


Mean Wind Speed 7 8 ms
Mean Wind Direction 201 doq
Value measured instantaneous Mean
(field)
Foundation
Load Cells
Rodf
Pressure Tps


Time i-c


Time (sec)


Figure 2-5. Comparison of vertical reaction records (in terms of force coefficients) measured by
load cells and estimated based on envelope roof pressures (Zisis and Stathopoulos
2009)









CHAPTER 3
ANALYSIS OF WIND TUNNEL DATA TO GENERATE PRESSURE COEFFICIENTS

Wind Tunnel Data

For this study, pressure coefficients were derived from wind tunnel data developed by

Datin and Prevatt (2007) on a 1:50 scale model house. The experiments were carried out in an

atmospheric boundary layer wind tunnel at the Wind Load Test Facility (WLTF) at Clemson

University. An overview of the experiment is discussed below.

House Model and Pressure Tap Layo ut

The tests were conducted on a 1/50 house model called Clemson standard model (CSM) 4-

12 which is shown in Figure 3-1A. The house model has length of 14.4 in. and width of 7.2 in

with a mean roof height of 3.4 in. CSM 4-12 has a gable roof with a slope of 18.4 (4 in 12). The

model was configured for a 60 ft X 30 ft full scale building with a mean roof height ofl4.3 ft.

The model has 387 pressure taps installed on its roof The pressure taps are evenly spaced

along the length of the roof at a nominal distance of 1 in except around the edges of the roof

where they are densely grid at a nominal distance of 0.2 in. Figure 3-1B shows the pressure tap

layout. These pressure taps were constructed with 0.063 outside diameter metal tubes glued to

Plexi-glass sheets and which are connected to Scanivalve electronic pressure scanners by 12 in

long vinyl tubes.

Wind Simulation and Pressure Measurements

A suburban expo sure was simulated upstream of the wind tunnel (shown in figure 3-2).

The velocity profile and turbulence intensity profile of the created exposure condition plotted

against the log law profiles for suburban terrain are shown in Figure 3-3A. Figure 3-3B shows

the longitudinal wind speed normalized power spectrum taken in the wind tunnel at equivalent

full scale height of 10 m (33 ft) as well as vonKarman spectrum.









Near simultaneous pressure time-histories were recorded using a scanivalve ZOC 33

system. Tests were repeated for five wind directions; 00, 450, 900, 1350 and 1800, as defined in

Figure 3-1B. Eight test repeats were done for each wind direction. Data was sampled at 300 Hz

and recorded for 120 seconds for each test repeat. Table 3-1 summarizes the common test

parameters used.

These stored files were used as the raw wind tunnel data for this project. There were forty

text files each containing 389 columns and 36000 rows.

Aerodynamic Data Processing

The raw data was low-pass filtered at 150 Hz prior to analyzing them. Pressure coefficients

were developed from the raw data as follows:

* It was corrected for tubing response to remove any effects of tube length and size on the
data

* Pressures were normalized by mean hourly pressure measured at 33 ft full scale height to
obtain pressure coefficients.

* Pressure coefficients were re-referenced to 3-second gust mean velocity measured at the
mean roof height of the building (14.2 ft)

Tubing Response Correction

The effect of the tubing system, used in the wind tunne 1 study, on the measured wind

pressure data was eliminated using a tubing frequency response shown in Figure 3-4. This

response was reported in Liu et al. (2009).

The raw pressure signal measured at each tap was first converted to the frequency domain

using a Fast Fourier Transformation. This provided a frequency (power) spectrum of the pressure

signal. The frequency spectrum was then divided by the frequency response to remove the

distortion caused by the volume and length of the tube. The corrected spectrum was then

converted to time domain using a n Inverse Fast Fourier Transformation.









Determining Pressure Coefficients

Pressure coefficients were derived from the measured local pressure time series as follows:


Cp,ref,i(t, ) = (3-1)
P ref (0)

where, Cp,ref,(t,O) is the pressure coefficient at Pressure Tap i, referenced to the dynamic pressure

at reference height at time t for wind angle 0; P,(t,O) is the measured wind pressure at tap i at

time t for wind angle 0; Pref(0) is the mean hourly reference dynamic pressure recorded by a

Pitot tube at the reference height of full height of 33 ft for wind angle 0. Pressure coefficients

were referenced at that height because flow is uniform with low turbulence levels at that height.

This ensures accurate speed control of the wind tunnel and accurate calibration of the pressure

scanners (Ho et al. 2005a).

Re-referencing of Pressure Coefficients

It is widely accepted that aerodynamic data referenced to mean roof height dynamic

pressure produce the least variability and therefore all low building pressure data sets, including

those in the building code s, follow this convention (Ho et al. 2005a). It is intended that the wind

tunnel results should be comparable to those in ASCE 7 and other aerodynamic database. For

this reason, the wind pressure coefficients were normalized to a 3-second gust mean wind speed

at the mean roof height (14.3 ft full-scale), U3secmrh. The wind pressure coefficients Cp,ref,(t,O)

were converted to the equivalent coefficient as follows.

cp,, (t, 8) = C, x Cpre, (t, 8) (3-2)

where, Cp,,(t,O) is the wind pressure coefficients at pressure tap i, referenced to a 3-second gust

wind speed at the mean roof height, at time t for wind angle 0; and Ca is an adjustment factor









which is given by the squared ratio of the mean wind speed at reference height Uref to the

equivalent 3-second gust wind speed at mean roof height U3sec,mrh (Shown in Equation 3-3).

-2
Ca -- (3-3)
U 3sec,mrh

The mean hourly wind speeds at the reference height Uef (13.03 m/s) and mean roof

height Umrh (6.54 m/s) were determined from the velocity profile for the wind tunnel testing. The

ASCE 7-05 provides the Durst curve which relates the wind speed averaged over gust duration, t

(in seconds), Ut to hourly mean speed, U360o. However, the curved corresponds to open terrain

conditions and an elevation, z= 10 m (Simiu and Scanlan 1996). As already stated, the wind

tunnel data was developed for a suburban terrain condition and pressure coefficients are intended

to be referenced to mean roof height of 14.3 ft (4.2 m). Hence, conversion of the mean hourly

speed to 3-second gust was done using Equation 3-4 provided by Simiu & Scanlan (1996).



U (z) = U3600 (z) 1+C(t) (3-4)



where, C(t) is the time averaging constant for a given time averaging interval /f is the squared

ratio of the friction velocity to the longitudinal turbulence fluctuations; zo is the roughness length;

z is the height at which the wind speed is to be evaluated. The calculation of 3-second gust wind

speed at mean roof height, which was based on C(3sec) =2.85; zo = 0.22m, z = 4.2m (14.3ft) and

S= 5.25, is as follows:



U3sec,mrh = Umrh +C(3sec) (3-5)
z2.51n1
\ o}











U3sec,mrh = (6.54m / s) 1+2.85 = 12.33m / s (3-6)
4f .2m \
2.51n( 4.2
S 0.22m)

The resulting adjustment factor, Ca for re-referencing the pressure coefficients is 1.1168.

Wind Tunnel Results and Analysis

Wind Pressure Coefficients Time Histories

The resulting pressure coefficient time histories were converted to equivalent full-scale

pressure coefficients using the reduced frequency relationship shown in Equation 3-7.


l = (3-7)
V I V

where f, L and V are respectively sampling frequency, characteristic dimension, and wind speed

referenced at mean height over a specified duration. Subscripts m and p denote model (1:50

scale) and prototype (full scale) buildings respectively. Based on mode 1 frequency and 3-second

gust wind speed at mean roof height of 300 Hz and 29 mph respectively, and 3-second gust full

scale wind speed, at mean roof height of prototype for suburban terrain, of 80.27 mph, the

prototype frequency is calculated as:

300Hz x 1 f, x 50
fh f = 17.46 Hz (3-8)
29mph 80.27mph

Using equality of non-dimensional time, the equivalent full-scale duration is given by:



fm

300Hzx 120s
TP =3H T = 34.36minutes (3-10)
S 17.46Hz p

The equivalent full-scale time step for the time histories is:









1 1
t -- =0.0573s (3-11)
S fp 17.46Hz

Pressure coefficient time histories in duration of 34 minutes in full scale of the 387

pressure taps were generated for each sample of a wind direction. Calculation for determining

the equivalent full-scale duration of the 120 seconds of test period is presented in the appendix.

Figure 3-5 shows time series plots pressure coefficients at pressure taps 1 and 387 for wind

direction 0.

Pressure coefficient time series, which are useful for analyzing dynamic responses of low-

rise buildings; have been saved in a MATLAB data format, as shown in Figure 3-3. There are 40

binary files with filenames structured as "CSM4-12_Suburban Cp_data dir XXX_Y.mat",

where "CSM4-12" identifies that the "Clemson Standard Model" with a roof slope of 4-12 was

used in the wind tunnel study; "XXX" denotes the wind direction; and "Y" denotes the sample

number. Each file also contains information of full scale geometric properties of the building

model tested, sampling frequency and period, time-step, locations of the pressure taps on model,

data sample number, wind azimuth, etc.

Observed Statistical Values of Wind Pressure Coefficients

The sample mean, root mean square (RMS) and peak local pressure coefficients were

computed for the eight samples of each wind azimuth. The statistical values of pressure

coefficients, which are useful for the design of cladding and components such as roof fasteners,

purlins and panes, have been evaluated for 34-minute equivalent full-scale aerodynamic pressure

coefficient time histories. These values were also saved in the forty MATLAB files containing

the pressure coefficient time histories. The mean and RMS pressure coefficient values were

averaged values of the eight samples:










Cp, j) = (n,0) (3-12)


Cp,,(0)= Cp,,(n,0) (3-13)
8 ,

where, Cp,, (0) and C,, (0) are respectively the mean and RMS pressure coefficients for at


pressure tap i, for wind angle 0 of the entire experiment; and C,, (n, 0) and C,, (n, 0)are

respectively the mean and RMS values of time series of the nth sample i, for wind angle 0.

Contour plots of mean and RMS pressure coefficients measured for each direction are shown in

Figures 3-7 and 3-8.

Extreme Value Analysis of Pressure Coefficients

Peak values estimated based on a probability distribution function are generally more

statistically stable quantities than the observed peaks from individual samples (Ho et al. 2005a).

The extreme negative and positive pressure coefficients measured from the eight samples of each

wind direction were fitted to an Extreme Type 1 Value Distribution. The probability density

function (PDF) and the cumulative distributive function (CDF) of the Extreme value Type 1

(also referred to as Gumbel distribution) are given by:


f(x) = e(x-)/pe (x-)/p (3-14)


F(x)= e-(x)p (3-15)

where, u is the location parameter (mode); and /f is the scale parameter (NIST 2003). The

parameters were calculated using the Best Linear Unbiased Estimators (BLUE) (Lieblein 1974).

There are three methods proposed by Lieblein (1974) based on sample sizes for the estimation of

the location and scale parameters. Method one is for an analysis with sample size less than









sixteen. The second method s should b e used for a study with sample size larger than sixteen but

generally smaller than about fifty. For an analysis with larger same size, method three is to be

used. The first method is adopted in this study since the sample size is eight. Furthermore, for

this analysis, the peak negative pressure coefficients were multiplied by negative one to make

them positive since BLUE analysis was developed for maximum values of the Type I extreme

Value distribution. The positive values were then sorted in the ascending order to p lace them in

the following order:

X1 <, < ...
The loc ation parameter, u and the scale parameter, /f were then estimated as follows:

8 8
P = a= xf = b, x (3-16)
n=l n=l

where, x, is the ith value of the ascending array of maximum values of the eight samples and a,

and b, are given by Table 3-2.

The "best" expected (mean) peak pressure coefficient measured at each pressure tap in a

given wind direction is given by:


C, = + 0.57728 (3-17)

Figures 3-9 and 3-10 show the spatial variations of the expected extreme pressure

coefficients on the roof of the building for different wind directions. The roof corners and gable

edges experience spatial variations at close distances and higher magnitude of suctions for all

wind directions except for wind direction 900. A nearly even distribution is observed away from

roof edges and corners. A similar pattern is observed be tween the mean, RMS and extreme

pressure coefficient distributions of wind azimuths 450 and 1350 are at opposing angles. The

same observation is made between the distributions of 0 and 1800. Appendix A provides mean,









RMS and extreme pressure coefficients of selected pressure taps for wind azimuths 0, 45and

900.

Area-Averaged Pressure Coefficients

Area-averaged pressure coefficients have been derived from pressure coefficient time

histories for regions of different size as follows:



CF, (j, t) ( (t,i)x A)/ A (3-18)
1=1 1=1

where CF(t) is a area-averaged wind pressure coefficient on region at time t; C,(i,t) is the wind

pressure coefficient at pressure tap i at time t; A, is the tributary area of pressure tap i and N, is

the total number of pressure taps on region.

The mean, RMS and extreme values of area-averaged pressure coefficients for the eight

samples of each direction were determined. The average mean and RMS of were calculated as

discussed above, while an extreme value analysis was done to determine the peak negative and

positive area-averaged pressure coefficients. Figures 3-11 to 3-13 display the area-averaged

pressure coefficients as a function of wind azimuth for corner, ridge corner, eave, ridge, interior,

and gable edge, corner, ridge corner, eave, ridge, interior, and gable edge. These pressure

coefficient values measured in the different regions on the surface of the building are compared

to ASCE 7-05 external wind pressure coefficients for components and cladding provided in

Figure 6-11C of the ASCE 7-05. Table 3-3 provides a summary of peak local pressure

coefficients observed within each of the three zones defined in ASCE 7-05, the area averaged

pressure coefficients and the C&C external pressure coefficients corresponding to the zones.

It is observed that, the peak local (tap) negative pressures suctionss) are generally higher in

magnitude as compared to the ASCE 7-05 provisions for the design of components and cladding.









However, the peak area-averaged pressure coefficients measured for each zone fall within the

provisions ofASCE 7 for the various zones. Also, it is observed that, the peak area averaged

pressure coefficients for two wind directions (i.e. 0 and 1800; and 450 and 1350) are identical.









Table 3-1. Measurement configuration and parameters
Model scale
Sampling frequency
Sampling period
Test angles
Upstream expo sure
3-second gust nominal wind tunnel speed at mean roof height


1:50
300 Hz
120 s
0000, 0450, 0900, 1350 and 1800
Suburban
12.33 m/s


Table 3-2. Coefficients of BLUE for Type 1 Extreme-Value Distribution (Lieblein 1974)
i 1 2 3 4 5 6 7 8
a, 0.274 0.190 0.150 0.121 0.097 0.076 0.056 0.036
b, -0.394 -0.06 0.011 0.059 0.087 0.103 0.108 0.102


Table 3-3. Comparison of wind tunnel and ASCE 7-05 peak pressure coefficients
Zones Wind Tunnel ASCE 7-05
Zones
Local Peak Area-averaged C&C
1 -2.73 -0.81 -0.9
2 -4.06 -1.84 -1.7
3 -4.39 -2.77 -2.6










x
-8.4-
























Figure 3-1. 1:50 Scale house model (CSM 4-12) used in the wind tunnel study


Figure 3-2. Test section arrangement for 1:50 suburban terrain





47












Wind tunnel, U/Uref log law, U/Uref
O Wind tunnel, Iu Log law, Iu

ol *
01















0 0.3 0.6 0.9 1.2 1.5
U/U--, Iu
U/U, ,f, Iu


0 i0


10 100


Figure 3-3. Wind flow characteristics for 1:50 suburban wind tunnel study. A) Mean wind speed
and turbulence intensity profiles B) Longitudinal spectrum of wind speed at full scale
height of 32.8 ft


200 mm lone-1 37 mm ID


300 mm long-0.06 mm ID


10 mm long-1.37 mm ID
tube connector


Figure 3-4. Frequency response characteristics of the pressure tubing system


--WLTF
----- von Karman


model


-scanner


---------------I----------------L --------------j---------------t---------------I---------------t--------------

............................................................................................................................ --........................-"-........................






-------------- 4 -------------- I -------------- -I------------- -------------------------


200


Frequency (Hz)


300


M.


I


I







































100 200 300
Time (sec)


400 500 600


Time (sec)


Figure 3-5. First 10 minutes time Series of wind pressure coefficient measured in direction 00 at

A) Tap 1 B) Tap 2


# .,,.-,- ..- P- i- T. r- 1 ,! 4- 1
File Edit Vew Graphc Debug Desktop Wndow Help


Shortcuts L1J Howto Add L1 What's New
Variable Editor Cp_data

W Cp ata < ~60 -3 7 dSubltak:
[B Cp_data <36000C3B7 double>


p ] Cpminmeanmaxrms x


0 New to MATLAB? Watch this Vdeo see Demos or read Gettn Sta
degree-granting institution Professional and cor

Af EDU>>

[ Start Ready


Sample_number
Wind azimuth
FS_Building_Height_
FS_Building_Lengthft
FSBuilding_Width_ft
Model Scale
Sampling_Period_s
Sampling_frequency_Hz
Slope
FS
FS_Pressure_tap_tributary_are...
FS Pressuretaplocations ft
Cpmin mean max rms
FS_timestep
Cp_data


Bytes Class
2 char
6 char
8 double
8 double
8 double
8 double
8 double
8 double
8 char
20 char
6192 double
9288 double
12384 double
288000 double
111456000 double


Command Histoy

Sx 0-- 10131,'0 10:23 PM
rated. x __ 12 V17' 0 12:41 PM
ImercialL L -- 12'17 09 5:os PM
S i-- 12,19 O9 12:2~ PM
S clear all
S-- 1/19/05 4:06 PM


Figure 3-6. Format of MATLAB files of pressure coefficient data


-I I


Sx rI ,IM't ,l, ,~,


1 2 3 4 5
1 -0 4210 -0 3510 -0464 -0 3990 -0 2210
2 -05210 -07050 -04230 -047 0 -02250
3 -03720 -03630 -04610 -0 3100 -04230
4 0399 3520 4360 04170 20
5 0 3460 -04140 -02480 -04000 -03310
-0384
7 -0240 -03010 -02960 -0 -00680
S -03940 -030 -0 3750 -04240 -0 4240
S -04120 -04340 -0 3310 -0 3060 -0 3520
1 -02730 -0 2470 -0 3440 -0 3940 -03140
11 0 2490 0 2630 0 187-0 3380 -0 3870
12 -0 3200 0 2770 0 3060 -0 2270 0 5020
13 -01760 -0 2210 -2070 2 -0 810 -1880
14 -02580 1970 -02450 -0 3250 -0 3450
15 -02570 -02370 -20 -02190 -0 2360
16 1290 1370 -0 1710 -0 2460 270
17 0 1870 -0 1580 -0 1550 0 1980 -02960


Size
1xl
1x3
1xl
1xl
1xl
1xl
1xl
lxl
1x4
1x10
387x2

387x4
1x36000
36000x387


































Wind Azimuth = 000









I----------------------------------
I

















Wind Azimuth = 090
I




I
I
I
I
I
I














L r --------------------- ------
IW
I




















Wind Azimuth = 180


LL~ ~ ~ L -~ ~ C ~- -~ ~ L -~ ~ )-( ~~-P


-1 -0.8 -0.6 -0.4 -0.2 0


Figure 3-7. Spatial distributions of mean wind pressure coefficients


I-------------------------
I
I


I
I
I
I
I
I

I
I





Wind Azimuth = 135


Wind Azimuth = 045


*\




















Wind Azimuth = 000


"~"`~ ~ -'~ ` ~'` ~ `'~ ` ~`~'"'


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8


Figure 3-8. Spatial distributions ofRMS of pressure coefficients


L 1-
I__j .


Wind Azimuth = 180


Wind Azimuth = 045


\


, -- --- -- -----





















Wind Azimuth = 000
-------- ------


LL~ ~ ~ L -~ ~ C ~- -~ ~ L -~ ~ )-( ~~-P


,, --I 75 m C,____.

I -- Q I---- -----
I4-,------- -- --------- ------


Wind Azimuth = 180


Wind Azimuth = 135
,--*-"-_---- --------------7-

'I
*1, 1)n.


-4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5
ff ::


Figure 3-9. Spatial distributions of expected negative peak wind pressure coefficients


Wind Azimuth = 090


/ind Azimuth = 045


VW


























Wind Azimuth = 000





I '
r -------------^-------------1--- -J







19



L --------------------------------------------------I

Wind Azimuth = 090



L
r----------------------------










Wind Azimuth =180
-^--^ ---^ -------------------------------------.
I-
!I
!I













I ^-



















l~s- ------------------------------------ --i
!
!
!
!
!

I !




Wind Azimuth = 809







I-


Wind Azi


i*- H 2 H 2 H 2 1 -r2 t


L--------------------------------

Wind Azimuth = 135


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
i i i


Figure 3-10. Spatial distributions of expected positive peak wind pressure coefficients


muth = 045















6-1
U- -1

-2

-3


45 90
Azimuth(8)


1

0

S-1

2

-3


135 180


0 45 90
Azimuth (8)


-- Min -a- Mean -- Max -- RMS


Figure 3-11. Area-averaged pressure coefficients for regions corresponding to zone 1 (Figure 6-
11C in ASCE 7-05).


135 180


LL-----~~
_-i



tf~











gion



S--- -


0 45 90
Azimuth (8)


0 45 90
Azimuth (8)

Gable End


Ridge Re
1

0

a -1

-2

-3


135 180


0 45 90
Azimuth (8)


S-1

-2

-3


135 180


45 90 135 18(
Azimuth (8)






. i


0 45 90
Azimuth (8)


0 45 90
Azimuth (8)


1

0

S-1

-2

-3


135 180


45 90
Azimuth (8)


--- Min -- Mean -- Max -- RMS
Figure 3-12. Area-averaged pressure coefficients for regions corresponding to zone 2 (Figure 6-
11C in ASCE 7-05).


u -1

2

-3


Eave


135 180


45 90
Azimuth (8)


135 180


_zzzz~zz4-ccc


0

U 1

-2

-3


135 180


135 180


I I


~c It I _--r



------~---~
I------- ---"



i~B


I


------C~












Ridge Corner


1

0

S-1


-2


-3


45 90
Azim th (eI


U 1

-2


-3


S 45 90 135 18C
Azimuth ()











E I ~


135 180


45 90
Azimuth ()


Eave Corner


1

0


S-1

-2

-3


0 45 90
Azimuth(9)


0 45 90
Azimuth ()


135 180


S1

-2

-3


135 180


0 45 90
Azimuth ()


M- Mi -e- Mean -- Max RMS


Figure 3-13. Area-averaged pressure coefficients for regions corresponding to zone 3 (Figure 6-
11C in ASCE 7-05).


0 45 90
Azimuth ()


135 180


135 180


45 90 135 18C
Azi' m .jthi (O)


135 180


I-~



----,~~__

ill









CHAPTER 4
APPLICATION OF DAD METHODOLGY

The database-assisted design (DAD) methodology utilizes wind-tunnel derived

aerodynamic database of wind pressures in combination with structural influence functions to

predict the wind load effects on a structural system.

In this chapter, pressure coefficient time histories developed in Chapter 3 are used with

experimentally derived structural influences to estimate structural reactions at roof-to-wall and

wall-to-foundation connections. This chapter opens with an overview of the structural influence

functions which were determined by Datin et al. (2010). A detailed explanation of the application

of the DAD approach in this study is provided. Finally, estimated structural reactions, analysis of

results and their comparison with design wind loads based on ASCE 7-05 are presented.

Structural Influence Function

In structural analysis, the variation of shear, reaction, b ending moment or deflection in a

structure subjected to live load or moving load is generally best described using an influence line

(Hibbeler 2006). An influence line therefore represents the variation of any structural response at

a specific point in a member as a concentrated force moves over the member. Usually, the

numerical values of a function for an influence line are determined using a dimensionless unit

load. Sub sequently the value of a response due to a force applied at any position can be obtained

by multiplying the ordinate of the influence line at that position.

This concept of influence functions is also adopted in the use of the DAD methodology to

predict wind-induced structural responses. For steel framed structures, influence functions used

in DAD-based software provided by NIST(2008) are derived analytically from simple 2D model

of a single frame. However, due to the high variability in wood properties, this study utilizes









influence functions determined experimentally on a 1/3-scale wood house model (shown in

Figure 4-1A).

In the experiment, a point load was applied at grid points (shown in figure 4-1C) on the

roof surface of the model house. The reactions inducedby the point loads at 8 roof-to-wall and 9

wall-to-foundation connections were measured by load cells (locations shown in figure 4-1B).

Influence coefficients at each load cell were determined by dividing the measured reactions by

the load applied at the grid points. The resulting fraction (percentage) of the applied load

provides a relative measure of forces transferred to the connections. Structural influence surfaces

over of vertical reactions were generated for each load cell.

To justify the experimental procedure used, simple 2-D and 3 -D models of the building

roof system were analyzed in structural analysis software, 'Visual Analysis 4.0'. Truss members

were modeled as simply supported beams and were pinned together. For the 3-D model, the roof

sheathing was modeled as a plate element. In both cases, a unit load was applied at points on a

truss of the system and an influence line was drawn for reaction at the truss support. Figure 4-3

shows a comparison of the results with the experimentally derived influence coefficients at grid

points on a truss for a roof-to-wall connection of the truss. Influence function determined by 2D

analytical model decreased linearly with distance between loading point and the support.

However, while the 3D analytical model produced a close agreement with respect to the

variation of influence function as compared to the experimental work, its influence coefficients

differ from experimental results by as much as 30%.

Typical experimentally derived influence surfaces are shown in Figure 4-4.









Evaluating Vertical Reactions Based on DAD methodology


Velocity Pressure

Estimates of vertical reactions at the roof-to-wall and wall-to-foundation connections are

based on wind speeds estimates without regard for direction as in the ASCE 7-05 provision.

Velocity pressure, qz (psf) was evaluated at the mean roof height of 14.3 ft as follows:

1 2
q 2 P-air V14ft,sub (4-1)


V14ft,sub = KVz33ft,open (4-2)

where pair is the density of air; Kz is an velocity pressure exposure coefficient whose square root

transforms the 3-second wind speed in miles per hour at 33 ft above ground over open terrain,

V33ftopen into 3-second gust wind speed at mean roof height over a suburban terrain, V4ft,sub.

Kz=0.70 (Table 6-3 of ASCE 7-05) was used in the DAD approach in order to be consistent with

the ASCE 7 procedure. Also the design wind speed, V33ft,open=130 mph (figure 6-1 of ASCE 7-05

(2005) was chosen, for consistency with vertical reactions evaluated on the basis of ASCE 7-05.

Hence,


V14ft,sub = 7(130 mph) = 108.8 mph (4-3)

q, = 0.00256 (108.8 mph)2 = 30.29 psf (4-4)

Pressure Taps and Influence Functions

The roof ofthe prototype house was divided into tributary areas of the pressure taps with

the assumption that pressure coefficient measured by a tap was uniform within its geometric

tributary area. Pressure taps were identified for the influence function grid points which fell

within the tributary of the pressure taps. Consequently, two or more influence coefficients were









assigned to a pressure tap and load calculations were based on the tributary area of the influence

function grid point.

Reaction Loads

Reaction time histories at eight roof-to-wall (gable end connection removed) and eleven

wall-to-foundation connections were elevated by "UFDAD" using the wind-tunnel pressure

coefficient time histories and the structural influence functions. The structural loads were

determined by:

R (t) = q, [(N, ), C,, (t)A] (4-5)
i7=1

where R,(t) is structural load estimated atjth load cell at time, t; q is the velocity pressure

evaluated at the mean roof height; C,,(t) is pressure coefficient for ith grid point at time t; (Nz) is

the influence coefficient at the ith grid point forjth load cell; and A, is the tributary area ofith

grid point.

DAD Results and Analysis

Observed Statistical Values of Structural Reactions

Again, forty files were generated in this analysis each containing structural reaction time

histories of the 17 load cells. These times histories are useful in determining peak loads and other

statistical values of interests. They can furthermore be used effectively in probabilistic and

structural reliability studies. Figures 4-6, 4-7 and 4-8 show the time-histories of the vertical

reaction at connections 5, 10 and 15 for wind directions 0 and 45. The sample mean RMS and

extreme reactions were measured from the time histories for the forty samples of data. The

measured statistical values of reactions for sample 1 are provided in Table 4-1.The statistical

values for the other samples are provided in Appendix B.









Average values of RMS and mean reactions were calculated for each wind direction from

the eight sample mean and RMS reactions. These quantities are not directly used in structural

design but are useful for reliability analysis. Tables 4.2 shows computed RMS and mean

reactions. The mean reactions are also shown in Figures 4-7 to 4-11.

Extreme Value Analysis of Vertical Loads

The peak or extreme uplift reactions are important statistical values that are used in

structural design. Generally, Lieblein-fitted peak values are more statistically stable quantities

than the measured peaks (Ho et al. 2005a). The measured peak reactions from the eight samples

for each wind direction were fitted to Extreme Value Type 1 distributions. The location and scale

parameters of these distributions were estimated using Best Linear Unbiased Estimators (BLUE)

as proposed by Lieblein (1974). The same procedure explained in Chapter 3 of this report was

followed.

The location parameter, i and the scale parameter, P which were estimated for the uplift

reactions are provided in Table 4-3. These parameters were used to obtain the probability density

functions (PDF) and the cumulative density functions (CPF) of the peak reactions from which

the mean peak reactions were estimated..

The "best" expected (mean) peak reaction load and corresponding standard deviation at the

jth connection, in a given wind direction, were respectively determined as follows:


R = /j + 0.5772 j s =pc / (4-6)

The estimated peak reactions and standard deviations from the Type I Extreme Value

analyses are provided in Table 4-4. Figures 4-7 to 4-11 show the expected peak and mean

reactions at the connections for all the wind directions considered in this study.









It is observed from the plots that, a greater percentage of the uplift loads were transferred

through the roof-to-wall connections (load cells 15 and 19) at the gable end. The highest uplift

reaction (1750 lb s) estimated was transferred through load cell 15 for wind directions 0 and 45.

It is worth nothing that, the influence coefficients for this connection are higher at the gable edge

and the ridge corners, where high suctions were observed. Again its influence surface was wide

spread than in the case of the other roof-to-wall connections.

Again, though the connections between the gable end roof and wall were discontinued

except at the end connections, reasonable uplift loads are transferred through foundation load

cells 11, 12 and 13 for the directions. This is explainedby the widespread of their influence

surfaces as a result of the diaphragm action in the wall (similar to deep beam action). However,

very small loads are transferred through load cell 14.

Relatively higher uplift loads were estimated at all connections for wind direction 0 than

the other azimuths. The loads generally decreased as the direction of the incident increased.

Higher positive reactions were transferred through the load cells for wind azimuth 900 because

peak positive pressures enveloped the building for that wind direction.

Vertical Reaction Bas ed on ASCE 7-05 S standard

ASCE 7-05 provides separate provisions for wind design using loads for either the main

wind force resisting system (MWFRS) or for components and cladding (C&C) members. Major

members of a building, which work together with other members in an assemblage to provide

support and stability for the overall building, are designed using wind loads provided for

MWFRS. Other members, which are directly loaded, are designed for localized wind load effects

on relatively small areas using provisions for C&C.

Roof-to-wall connections (truss reactions) which serve as media for transfer of wind

loading on the roof system to walls have in the past (Datin and Prevatt 2007) been designed as









part of the building's MWFRS. However, roof-to-wall connection failures may be likely due to

localized wind effects acting on the roof which necessitate estimating wind load reactions at

these connections based on C&C pressures. Moreover, ASCE 7-05 lists roof trusses under both

C&C and MWFRS, which complicates the interpretation of the code for estimating reactions.

For this reason, two separate analyses were done using both wind load provisions.

Velocity Pressure

Vertical reactions loads at roof-to-wall connections of the building were obtained using the

Analytical Procedure (Method 2) for low-rise buildings (ASCE/SEI. 2005) with the design wind

speed of V33ft,open, 3c=130 mph (figure 6-1 ofASCE 7-05) and a suburban terrain. A velocity

pressure exposure coefficient ofKz=0.7 (obtained from table 6-3 ofASCE 7-05) to modify the

design wind speed to account for the suburban terrain condition and a mean roof height of 14.3

ft. This is very consistent with the wind speed used in the DAD estimations. A reduction factor,

Kd= 0.85 to account for wind directionality effect was used in the ASCE 7-05 calculations even

though no reduction factor was used in the DAD analysis. The importance factor I and the

topographic effect factor Kzt were assumed to be unity. The velocity pressure, qz for ASCE 7-05

calculation is as follows:

q, = 0.00256K,K,,KdV3 ft,open,3 ec (4-7)

q, = 0.00256(0.7)(1.0)(0.85)(130mph)2 (1.0) = 25.741b/ ft2 (4-8)

ASCE 7-based Design Loads

Wind pressures on the building were determined as the just product of computed velocity

pressure and external pressure coefficients because internal coefficients were not considered in

the studies. In both MWFRS and C&C analysis, the study external pressure coefficients were

determined based on the dimension, a = 3ft (10% of the smallest horizontal dimension).









For ASCE 7-05 MWFRS, external pressure coefficients are defined with respect to roof

angles and building surfaces. The building has a roof angle of 18.4 and therefore linear

interpolations were used to determine external pressure coefficients. Partitioning of the roof

surface into four building surfaces as defined in Figure 6-10 of ASCE 7-05 is shown in Figure 4-

12. Partitions were used to determine the pressures on the individual trusses. Table 4-5 shows

MWFRS based wind pressures for the different building surface based on the velocity pressure.

In the case of C&C, external pressure coefficients were obtained from Figure 6-11C ofthe

ASCE 7-05. Selection of the coefficients was based on an effective wind area of 300 ft2, which is

a product of 30 ft span length and one-third the span length. The building roof surface again, was

partitioned into pressures zones as defined in ASCE 7 (see Figure 4-13). Wind Pressures based

on C&C provisions for the zones including the roof overhang as given in Table 4-6.

Wind load intensity on each truss was calculated by multiplying pressure on the truss

tributary area by its tributary width. Each truss was loaded with its pressure intensity and

analyzed separately in "Visual Analysis 4.0" to determine the structural reaction at its support,

for both MWFRS and C&C cases. For MWFRS case, the procedure was repeated for all the eight

load patterns defined in Figure 6-10 in ASCE 7 to obtain the maximum uplift reaction.

Comparing Uplifts Reactions Predicted Based on DAD vs. ASCE 7-05

For comparison, the DAD-based reactions results were multiplied by 0.85 to account for

the reduced probability of maximum winds coming from any direction, as it was in the case of

the ASCE 7-based estimations. The worst uplift loads estimated at each of the roof to wall

connections were selected and compared to the peak estimates based on the ASCE 7 Standard for

MWFRS and Components and Cladding (C&C).

The comparison of the results based on ASCE 7-05 and DAD methodology are presented

in Table 4-7. The results are graphically presented in Figure 4-13.









It is observed that the DAD-based reactions are generally higher than reactions based on

ASCE 7-05 MWFRS provisions. At the gable end supports, ASCE 7 MWFRS underestimate the

reaction at that support by as much as 28 % as compared to the DAD estimates.

Underestimations of reactions by ASCE 7 MWFRS provisions are high at truss supports closer to

the gable end of the house. Very close results are however observed between the results at truss

suppo rts which are far away from the gable end.

Significantly high reactions (by 6% to 62%) are observed for reactions based on ASCE 7-

C&C than in the case of the DAD. Using discrepancies observed in the two comparisons above,

the DAD results have a better agreement with ASCE 7 MWFRS estimates than C&C based

estimates.












Table 4-1. Measured peak, mean and RMS reactions (lbs) for sample 1
Azimuth 0 450 90 1350 1800

Load Neg. Mean Pos. RMS Neg. Mean Pos. RMS Neg. Pos. Ne. Mean Pos. RMS Neg. Mean Pos. RMS
Cells Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak

1 -828 -171 86 199 -408 -96 180 115 -460 -61 287 90 -378 -88 235 106 -245 -23 139 48
2 -939 -201 70 229 -434 -116 137 134 -445 -61 276 87 -346 -83 223 99 -242 -24 132 47
3 -1067 -256 63 287 -749 -140 142 161 -414 -77 327 102 -364 -98 173 113 -259 -30 127 53
4 -1187 -294 67 328 -988 -141 184 167 -410 -82 403 110 -386 -104 139 120 -262 -33 134 56
5 -1099 -278 40 307 -987 -177 93 204 -442 -94 230 112 -327 -103 50 115 -212 -29 115 48
6 -400 -96 22 107 -245 -62 55 70 -178 -35 102 44 -143 -44 49 50 -109 -12 55 22
7 -218 -51 13 57 -123 -33 33 37 -103 -19 57 24 -81 -24 29 28 -61 -7 31 12
8 -335 -82 18 90 -222 -50 43 57 -131 -27 89 34 -112 -34 38 38 -82 -10 41 17
9 -766 -190 37 210 -559 -116 90 135 -275 -63 198 78 -242 -76 75 85 -179 -22 89 38
10 -1149 -291 47 320 -887 -191 110 218 -440 -102 259 122 -354 -117 93 131 -260 -33 132 56
11 -1562 -396 63 436 -1210 -260 146 298 -602 -142 344 168 -483 -161 127 180 -352 -45 181 77
12 -1391 -353 57 389 -1080 -230 134 264 -536 -125 310 150 -432 -142 115 159 -312 -40 161 68
13 -404 -103 16 114 -316 -69 37 79 -159 -38 86 44 -127 -42 32 47 -90 -12 47 20
14 -52 -14 2 15 -42 -10 4 11 -21 -5 11 6 -18 -6 4 6 -11 -2 6 3
15 -1668 -446 67 494 -1754 -308 158 373 -757 -157 308 184 -509 -164 74 182 -345 -49 182 77
19 -1388 -315 50 354 -1340 -328 9 359 -660 -169 50 185 -453 -139 86 153 -215 -36 111 55
20 -1343 -291 55 322 -1067 -310 10 339 -629 -171 60 187 -474 -143 101 157 -234 -32 133 54











Table 4-2. Averaged mean and RMS reactions (Ibs)

Load 00 450 90 1350 1800


Cells

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
19
20


Mean RMS Mean RMS
-168 196 -103 121
-197 226 -123 140
-250 282 -149 168
-288 321 -150 174
-272 302 -187 212
-94 105 -66 73
-50 56 -35 39
-80 89 -53 60
-186 206 -123 140
-285 315 -201 227
-388 429 -275 310
-346 382 -243 275
-101 112 -73 82
-13 15 -10 12
-437 486 -324 385
-309 349 -344 372
-286 318 -326 352


Mean
-63
-62
-78
-83
-95
-36
-19
-28
-64
-103
-143
-127
-38
-5
-159
-169
-171


RMS Mean
93 -89
89 -84
103 -99
111 -105
113 -105
45 -45
24 -24
35 -34
79 -77
124 -118
171 -163
151 -144
45 -43
6 -6
186 -167
184 -139
186 -143


RMS
107
100
114
121
116
50
28
39
86
132
181
161
47
6
184
154
158


Mean
-32
-32
-39
-43
-38
-16
-9
-13
-29
-44
-60
-53
-15
-2
-64
-48
-45


RMS
53
52
58
62
54
24
13
19
42
63
86
77
22
3
88
63
63











Table 4-3. Parameters (Ibs) for Type I Extreme Value Distribution of peak negative reactions
Load 0 450 90 1350 1800
Cells t P 3 P P 33 P

1 758 61 441 101 380 29 379 28 249 23
2 856 50 463 91 355 35 352 26 241 22
3 1037 40 608 68 391 33 374 21 259 22
4 1168 112 791 82 412 34 399 23 268 22
5 1053 98 875 55 380 38 344 27 225 24
6 369 14 232 21 160 14 148 12 106 7
7 199 11 120 14 88 9 84 6 60 4
8 304 16 206 15 125 9 116 7 82 5
9 698 49 510 32 278 20 256 18 179 11
10 1061 78 824 47 428 30 380 33 261 18
11 1446 105 1123 64 587 42 522 47 355 26
12 1289 96 1000 58 522 38 465 40 315 24
13 377 29 295 17 154 11 137 12 91 7
14 50 4 40 2 21 1 19 1 12 1
15 1608 171 1665 75 630 84 538 46 354 27
19 1257 106 1136 96 598 70 491 39 235 25
20 1125 86 1022 95 563 64 496 27 239 23











Table 4-4. Expected peak negative (uplift) reactions (lbs) and standard deviation (lbs) estimated
from BLUE fitted probability distribution


Load 0


Cells

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
19
20


450 90 1350 1800
Std Mean Std Mean Std Mean Std


Mean Std Mean
-793 78 -500
-885 65 -516
-1061 52 -647
-1232 144 -838
-1110 126 -907
-377 18 -244
-206 14 -128
-314 21 -215
-726 63 -528
-1106 100 -851
-1506 134 -1160
-1344 123 -1033
-393 37 -304
-52 5 -42
-1707 219 -1708
-1318 135 -1192
-1175 110 -1077


-397
-375
-410
-432
-402
-168
-93
-130
-289
-446
-611
-543
-160
-22
-679
-638
-599


-394
-366
-386
-413
-360
-155
-87
-121
-267
-399
-549
-489
-144
-20
-564
-513
-511


Table 4-5. Pressures based on MWFRS for different building surface
Building Surface 2 3 2E 3E
GCpf -0.69 -0.47 -1.07 -0.67
Pressure (lb/ft2) -17.76 -12.10 -27.54 -17.25


Table 4-6. Pressures based on C&C for different zones
Roof Overhang
Zones 1 2 3 2 3
GCp -0.8 -1.2 -2 -2.2 -2.5
Pressure (lb/ft2) -20.6 -30.9 -51.5 -56.6 -64.4


-262
-254
-271
-281
-239
-110
-62
-84
-185
-271
-370
-329
-96
-13
-369
-249
-252









Table 4-7. Comparison of uplift reaction estimates based on DAD and ASCE 7-05
Connection
Connection 1 2 3 4 5 15 19 20
Number
DAD -674 -752 -902 -1047 -944 -1452 -1120 -999
ASCE MWFRS -763 -763 -763 -830 -830 -1043 -840 -659
7-05 C&C -956 -956 -956 -956 -1435 -1794 -1794 -1435


Figure 4-1. 1/3 Scale house model for determining influence functions






















14
13 6
12 8 7





Figure 4-2. Locations of load cells and wind direction


_-------_21 Trusses @ 8" ----
S.. ... .. .. .. I

I ....... ....................I



I I


.+ .
:.:': : : : : :. .



.t ". ". .
I............. ... ... ..I

*"t "'"".......... ...........
I.. ...... I
.... ........................I
I.......... ............I
.- ---.-.- -....- -.-.-..-- I


Figure 4-3. Grid points for experimental determination of influence coefficients






























-5 0 5 10 15 20 25
Distance from Support, x (ft)


Figure 4-4. Influence lines for vertical reactions
internal truss


at a support (roof-to-wall connection) of an


-02 0 02 04 06 08 1 12


Figure 4-5. Typical influence surfaces for vertical reaction determined on 1/3-scale house model
A) load cell 15 B) load cell 5 C) load cell 4 D) load cell 11


3D Analytically Derived
-w-- 2D Analytically Derived
- Experimentally Derived


30 35













500




-500

C
0J

.9 -1000


r -1500


-2000
0 100 200 300 400 500 600
Time (sec)


500


0 0

- -500
0
o

. -1000
0
2 -1500


-2000
0


100 200 300 400 500 600
Time (sec)


Figure 4-6. First 10 minutes load time Series at roof-to-wall load cell 5 A) wind azimuth 0 B)
wind azimuth 450


600

400

200

0

-200

-400

-600

-800

-1000

-1200

-1400

-1600

-1800


1 2 3 4 5 15 19 20 6 7 8 9 10 11 12 13 14
Load Cell


Figure 4-7. Estimated mean and peak vertical reactions at load cells for wind azimuth 0000


1 .


-


































1 2 3 4 5 15


19 20 6 7 8 9 10 11
Load Cell


Figure 4-8. Estimated mean and peak vertical reactions at load cells for wind azimuth 0450


Roof-to-Wall
Connections


11.1


1 2 3 4 5 15


19 20 6 7
Load Cell


Wall-to-Foundation
Connections


I.I.'I


8 9 10 11 12 13 14


Figure 4-9. Estimated mean and peak vertical reactions at load cells for wind azimuth 0900


Roof-to-Wall Wall-to-Foundation
Connections Connections



El El I 7




[ I-





I I Negative Extreme
Positive Extreme
E Mean


-1000

-1200

-1400

-1600

-1800


12 13 14


fl-hf51fEFll2HEECD1Ellfi


I INegative Extreme
Positive Extreme
E Mean


-1000
-1200
-1400
-1600
-1800


I I


mmm m


1 1


I I


I











Roof-to-Wall


Roof-to-Wall
Connections

I.E.-.


E D -


wwD


Wall-to-Foundation
Connections


HLi LJ


D Il


I I Negative Extreme
S Positive Extreme
E Mean


1 2 3 4 5 15 19 20 6 7 8 9 10 11 12 13 14
Load Cell
Figure 4-10. Estimated mean and peak vertical reactions at load cells for wind azimuth 1350


Roof-to-Wall


Roof-to-Wall
Connections

Eu..-m.


77MNO O


Wall-to-Foundation
Connections


I I Negative Extreme
S Positive Extreme
E Mean


1 2 3 4 5 15 19 20 6 7 8 9 10 11 12 13 14
Load Cell
Figure 4-11. Estimated mean and peak vertical reactions at load cells for wind azimuth 1800


-400
-600
-800
-1000
-1200
-1400
-1600
-1800


-400
-600
-800
-1000
-1200
-1400
-1600
-1800


m


7 El-


11111,,11









j- 2a


- Trusses @ 2 oc, -
{q9i;


2






IA 1(1 I IS 321 ) B
Figure 4-12. Building surfaces for determining wind pressure for each truss based on ASCE 7
provisions for MWFRS A) Transverse direction B) Longitudinal direction


.- Trusses @ 24" o.c.-


i 9 ". I i I I


I I I I I I I I

I I I I I Zone 1

II I Zone 2
I I I I I I l
i I i F J i
I I I I-IIIIII IZone3
SZone 2 Overhang

IZone 3 O erhan'

I I I I I a= 3'-0"


I I I I I I I I


m a ti, I


Figure 4-13. Zones for determining wind pressure for each truss based on ASCE 7 provisions
for components and cladding


r.
r


I


I
li
i
L


a


















0 -1ow

a -1200



-N- DAD
-1600 ----ASCE 7-O0 MWFRS
ASCE 7-05 C&C
-100
1 2 3 4 5 15 19 20
Roof-to-Wall Load ellss

Figure 4-14. Uplift reactions at roof-to-wall load cells based on DAD approach and ASCE 7-05
provisions









CHAPTER 5
WIND FLOW CHARACTERIZATION USING TFI COBRA PROBE

The flow characteristics of wind largely influence the intensity and fluctuation of wind

forces on structures. The experimental component of the study includes a scale house model

immersed within a wind flow generated by a UF Wind Generator. A multi-hole pressure probe

called the Cobra Probe was used to map the wind flow characteristics to provide the necessary

information for analyzing the data.

This chapter discusses the Cobra Probe, its principle of operation and pilot flow

measurements conducted prior to using it in the main experimental study. Flow characterization

of the wind field produced for the experiment is also presented in this chapter.

The Cobra Probe

The Cobra Probe (shown in Figure 5-1A) is a robust and compact multi-hole pressure

probe designed by Turbulent Flow Instrumentation (TFI), Australia, for measuring turbulent

wind flows. The Cobra probe was first proposed by Shepherd (1981) for mean flow

measurements and further developed by Hooper and Musgrove (1997) for resolving turbulence

structure. The device has the following advantages:

1. It resolves highly turbulent wind flows into three orthogonal components at high frequency
(up to 10000 Hz).

2. It is a robust self contained device.

3. It can withstand significant levels of instrument vibration and maintain accuracy (Watkins
et al. 2004).

The high sampling frequency of the probe allows for continuous data streaming, with

virtually unlimited data to be recorded while avoiding aliasing and reducing measurement of

noise. The Probe has a truncated triangular pyramid-shaped head with four faces ground flat to

450 (Chen et al. 2000). Each face has a 0.5 mm diameter pressure tap located at its center that is









connected to piezo-resistive bridge pressure transducers, and a preamplifier system located

within the body of the probe. This arrangement was necessary to minimize the length of tubing

so as to achieve a high frequency response as well as mechanically protecting the transducers.

The probe has an overall length of 160 mm and a body diameter of 14 mm. The design of the

probe is shown in Figure 5-1B.

The principle of operation of the probe is to relate the pressure field detected by four

pressure transducers to the magnitude of the instantaneous local velocity vector, yaw and pitch

angles and the instantaneous static pressure. Pressure signals measured are corrected for

transmission effects, using a predetermined transfer function (frequency response function) of

pressure tubing (Chen et al. 2000).

A Series 100 Cobra Probe (Serial ID 193) was used in the study. It has a velocity range of

3 m/s to 60 m/s and an accuracy of0.5+m/s and was supplied with an interface unit housing an

integrated data acquisition and device control software.

Preliminary Experiments Using the Cobra Probe

Pilot experiments were conducted using the Cobra Probe aimed at (1) comparing flow

measurements by the probe and hotwire anemometer (b) understanding the use and principle of

the probe and (3) evaluating methods in modeling flow field conditions for the study's main

experiment.

In all the preliminary experiments, flow measurements were usually sampled by the probe

at 5000 Hz and down sampled at a rate of 2500 Hz for 120 seconds per each test run. Before

each test, the probe was zeroed to remove offset voltages from its pressure transducers.

Comparing Wind Flow Measurements by the Cobra Probe and Hotwire Anemometer

Flow measurements were simultaneously taken by the cobra probe and a hot-wire

anemometer in a generated wind field. Table 5-1 compares the mean axial velocities, U, and









turbulence intensities, Iu, measured by the Cobra Probe and the hot-wire anemometer. Good

agreement is observed for the mean velocities. However, significant differences are observed for

the turbulence intensities, which may in part be due to errors as a result of relative positions of

the instruments.

Wind Tunnel Model

Several tests were done in a small wind tunnel (shown in Figure 5-3) to understand the

principle and usage of the probe. The wind tunnel consists of a 10 ft long test section, a

contraction area, and a plenum and flow straighter section. It is powered by a 2 ft diameter axial

fan. The test section is 24 in wide and 8 in high.

Results from one test where flow measurements were taken at points (shown in Figure 5-

4) within a 2D surface at 0.5 in downstream of the wind tunnel exit are presented. Profiles for all

the orthogonal mean velocities measured are shown in Figure 5-5. Corresponding turbulence

intensity distributions for longitudinal (Iuu), transverse (Ivv) and vertical (Iww) directions of flow

are also displayed in Figure 5-5B. It is seen from the plots that very low transverse and vertical

wind velocities are recorded away from the walls as compared to the high axial velocities.

Relatively low axial velocities with corresponding high turbulent intensities are observed near

the wall surfaces of the test section. This is due to a high frictional effect near the wall surfaces.

Figures 5-6 to 5-8 show the 2D variation of the mean velocities and turbulent intensities across

the test section. The spectral contents of wind speeds at the exit of the wind tunnel are displayed

in Figure 5-9.

Mapping of Wind Field Generated by UF Wind Generator

Velocity and turbulent intensity measurements of the wind flow generated by the UF wind

generator were taken prior to this study's experimental work. The main purpose was to

investigate if the wind field was uniform within the test area.









The probe was mounted on a computer-controlled traverse to control the horizontal

position. Vertical adjustments of the probe were done manually. The traverse frame was set up in

two locations laterally to map the wind flow across the section. Figure 5-10 shows the traverse

frame within the test area.

The Cobra Probe was used to measure flow characteristics at 90 locations (shown in

Figure 5-11) within a cross-section at 16 ft downstream of the wind generator. Each

measurement was taken for 60 seconds at a sampling frequency of 10000 Hz and output to file at

5000 Hz. A mean wind velocity of 50 mph (22 m/s) was used for all the tests.

Contour plots of longitudinal, transverse and vertical wind fields are shown in Figures 5-

12 and 5-13. Relatively low longitudinal wind speeds at locations outside the jet field were

observed. Longitudinal wind speed was generally uniform across the cross-section but decrease

near the wall.









Table 5-1. Comparison of flow measurements by Cobra Probe and hot-wire anemometer
Height Cobra Probe Hot-wire %Difference
(in) U (m/s) luu (%) U (m/s) Iuu (%) U Iuu
2 5.47 14.33 5.13 13.38 6.2% 6.6%
6 6.26 13.83 6.36 10.68 1.6% 22.8%
10 5.83 10.15 6.28 11.38 7.7% 12.1%


Figure 5-1. Cobra Probe


Figure 5-2. Cobra Probe and Hot-wire anemometer setup for simultaneous measurements




























Figure 5-3. Wind tunnel model used in pilot studies


r-60 cm




22cm




-5 cm Grid points @ i ..-.


Figure 5-4. Measuring points for mapping flow measurements at of the wind tunnel

















m Axial Component
STransverse Component
SVertical Component



A & A -
L = r


-


-0.5 0 0.5


..............._


Figure 5-5.


Flow measurements at exit of wind tunnel model A) Mean velocity (m/s)
distributions. B) Turbulence intensity distributions (y is distance of grid point
relative to the midpoint of test section; D is half the width of test section)


8 10 12 14 16 4 6 8 10 12
Figure 5-6. Spatial variations of longitudinal velocity (m/s) and turbulence intensity (%) across
exit section of wind tunnel model


-1 -0.5 0 0.5


-1.5 -1 -0.5 0 0.5 1 1.5 2


Figure 5-7. Spatial variations of lateral velocity (m/s) and turbulence intensity (%) across exit
section of wind tunnel model






84


Sa


luu
SI0 0.5 1











-0.5 0 0.5


I : :


I : :


















-I -1 -5 0
-1.5 -1 -0.5 0 0.5


1 1.5 2


2 4 6 8 10


Figure 5-8.


fS,(f) 0.1







0.01
0.


1







fs ,(f) 0.1
0.1
crJ'~-w j


Spatial variations of lateral velocity (m/s) and turbulence intensity (%) across exit
section of wind tunnel model


NI,


-Data
on Karman
0.01
0.01 0.1


1 -







fs,(f) 0.1
-


,I C'.4 I


U.U I
0.01


Figure 5-9. Longitudinal, transverse and vertical spectral contents of wind speed at 0.5 in.
downstream of the exit of the wind tunnel


- Data
---von Karman


f : W
Ci Iil




































Figure 5-10. Position of traverse frame in test section



z



'T s srisiu
erlijrd. Indpinr iaIonr












// I I






Hurricane Simulator behind


Figure 5-11. Location of measurements points for flow mapping
I/ I _1.



/ -- -.;* -----------------(_,rt,jl f ..' '"* ,'-----------

Hurricane Simulator behind


Figure 5-11. Location of measurements points for flow mapping































8 10 12


14 16 18 20


Figure 5-12. Variation of longitudinal wind speed across 2D measurement surface


-2 -1 0 1 2 3 4


Figure 5-13. Variation of lateral and vertical wind speed across 2D measurement surface


I : : :


















fS, (f)
C 0.1








0.01
0.01











a 0.1








0.01-
0.01


- Data
--von Karman
0.1
fLU
U


0.01 4
0,.


- Data
---von Karman


1 0.1 1


Figure 5-14. Spectral contents of wind speed A) longitudinal B) Transverse C) Vertical





















88


f S(f)
2 0.1


D ata .----. p i.. .. ...... .

K a m a i i | :;i| -\-\\


Data
--von Karman


0









CHAPTER 6
WIND INDUCED PRES SURE AND S TRUC TURAL LOAD MEASUREMENTS

The obj ective of the experimental component of the research is to evaluate the validity of

using the DAD methodology to predict wind-induced structural loads as discussed in Chapter 5.

This chapter presents the experimental study, results and analysis performed to validate the DAD

approach.

Materials and Methods

Scale House Model

The 1/3 scale house (shown in Figure 6-9) was modeled and constructed by Datin et al.

(2009) based on non-dimensional geometrical scaling laws. The building model has a rectangular

floor plan which is 10 ft wide by 13 ft 4 in. long (30 ft by 40 ft at full scale) and a mean roof

height of 4 ft 1 in. (12 ft 3 in at full scale) with a 6 in. (1 ft 6 in at full scale) overhang around the

perimeter. It has a gable roof sloping at 180, with Fink style wood trusses installed at 8 in.

spacing (24 in o.c. at full scale). The building model is mounted on a 3 ft high steel frame clad

with oriented standard boards (OSB) sheets. To be representative of typical residential

constructions, the gable end trusses were connected only at the truss reaction points and not

through bottom chords with the end wall.

Pressure and Load Sensors on the Building

Figure 6-1 shows the layout of pressure taps and load cells. Twenty-nine pressure taps

were installed on the building model, 25 on the roof and 4 on the walls (See Figure 2 for the

distribution of the taps on the roof). The nearest pressure taps to the gable end edge of the roof

are installed 3.5 in. from the edge. Each pressure tap was made using 3 in. long 0.124 in inside

diameter (ID) brass tube soldered to a thin metal disc as shown in Figure 6-2A. Figure 6-2B

shows the distribution of pressure taps over the house roof 27 Omega PX 138 pressure









transducers and 2 Setra model 265 transducers were used to capture pressure distributions on the

roof and walls of the house. The high port of each transducer was connected to the taps via a 6

in. long 3/16 in. ID vinyl clear tube. The internal pressure was measured using a Dwyer 616-20B

transducer. The low ports of the transducers were connected to a single manifold and held at

ambient pressure (see Figure 6-3). The specifications of these transducers are given in Table 6-1.

Pictures of the three different transducers used are shown in Figure 6-4.

Structural reactions on the building were monitored using twenty-one 300 single axes

compression/tension load cells (Futek model LRF 350). Twelve load cells were installed at roof

to wall connections and 9 at the wall-to-foundation connections (see Figure 6-5). The building

was isolated at the instrumented end to capture the total load reactions through this load cell

array.

All these pressure transducers were factory-calibrated. However, a Fluke 718 1G -

pressure calibrator with serial number of9916005(shown in Figure 6-6) was used to confirm the

calibration charts. The data acquisition system consisted ofa National Instruments (NI)

CompacDAQ chassis (NI cDAQ 9172) with modules NI 9205 and NI 9219. This data

acquisition system was controlled by Measurement and Automation Explorer and LabView

software version 8.5. Prior to the experiment, the instruments were left overnight to measure

room condition signals to check for possible drift effects in their performances.

Test Arrangement

Two 16 ft tall by 34 long wood framed walls were constructed at the exit the UF wind

generator to enclose the test building. The windward face of the 1/3 scale model was placed at

approximatelyl8 ft away from the exit to provide a flow development length of at least four

times the fan diameter (4.5 ft). The house was oriented in three separate directions (00, 450 and

900) to the wind flow. Figure 6-8 shows a layout for the experimental set-up. Figures 5-9 shows









the different orientations of the house in the experimental setup. The house was anchored to the

ground to resist all horizontal and vertical movements as well as overturning during the test.

The Cobra Probe was positioned centrally within the test section 6 ft downstream of the

contraction of the wind generator to record wind speeds. It was installed at 8 ft above ground to

match the mean roof height of the house. A pitot tube, connected to a Dwyer pressure transducer

(shown in Figure 6-7), was installed 2 ft above the roof ridge.

Wind Generation

The UF wind generator (shown in Figure 6-10) situated at the Powel Family Structures

Lab was developed for testing window and door panels. It generates wind forces by using an 8-

fan array of 4.5 ft diameter fans hydraulically powered by four marine diesel engines with a

combined 2800 hp. The generator can produce fluctuating winds at speeds up to 120 mph.

Further description of its develop ment and ope ration can be found in Masters et al.(2008).

A mean wind velocity of 50 mph (22 mph) was used for all the tests. The wind speed was

chosen based on scaling considerations for equivalent 130 mph full scale mean speed. To

produce this speed, the wind generator engines were run at 1000 rpm turning the eight fans at

700 rpm. The wind generator produced approximately 6 % turbulence intensity, measured at the

mean roof height, for the experiment.

Experime ntal Procedure and Measurements

Tests were conducted with the building in three orientations to incident wind flow: 00, 900

and 45. Zero degree was taken as wind flowing normal to the gable end wall of the building.

Three repeats were done at each house orientation. Data was sampled at 200 Hz lasting 10

minutes for each test. The LabView program synchronized measurements by all the pressure and

load sensors and saved them into test file output for each test. The Cobra Probe, which was









controlled by TFI device software, was timed using a digital trigger connected to the LabView

program. Sampling by all the measuring instruments started at the same time for all the tests.

Prior to each test, the Cobra Probe was zeroed and initial readings from the transducers

and load cells were taken for 120 seconds. These initial records were subtracted from pressure

and load measurements before data analysis remove any initial offset of the instruments. The

pressure and structural load measurements were low-passed filtered at 100 Hz and their

statistical values were measured using a MATLAB-based program.

Experimental Results, Analysis and Discussion

Wind Pressure Measurements

Positive peak pressures (7 11 psf) were recorded on the windward walls whereas suctions

were recorded on the leeward walls for 0 and 900 azimuths. For 450 degrees, both the

instrumented gable end wall and the side wall experienced positive pressures. The internal

pressure recorded during all the experiments was zero psf The wind distribution over the roof of

the building was generally characterized by peak suctions especially for the 0 and 450 azimuths.

Wind suctions were highest along the second interior truss (with load cells 4 and 16) for the 00

azimuth. High suctions were recorded along the gable edge of the roof while relatively low

suctions were measured at the interior roof areas of the roof for the 450 azimuth. For the 900

azimuth however, relatively low suctions enveloped the roof of the house except at the eave end

area. The wind pressure distributions observed in all these experiments were in good agreements

with what is reported in literature and observed from the wind tunnel data.

Wind pressures measured were normalized into pressure coefficients by the velocity

pressure at the mean roof height of the building. The pressure coefficients were evaluated using

Equation 5-1.










C mrh, (t)= -2 (5-1)
1/ 2pU,,r

where Cp,mrh,2(t) pressure coefficient at pressure tap i at time t, referenced to the mean wind speed

measured at the mean roof height of the building; P, is the measured wind pressure at tap i; p is

the air density; Umrh is mean wind speed (mph) recorded at mean roof height by the Cobra Probe.

Pressure coefficients obtained for the nine test repeats are given in Appendix C. Table 6-2

shows the mean, peak and root mean square (RMS) pressure coefficients averaged from results

of the three datasets for each wind direction. The peak negative and mean pressure coefficients

contour plot results are also shown in Figure 6-11.

Wind-Induced Structural Loads

Measured statistical values of structural loads obtained from each test of the experimental

study are provided in Appendix D. The three sample data obtained for each direction were

augmented. Each augmented time history was then divided into six segments to obtain six 10-

minute (full scale) samples. Mean, peak and RMS reactions were measured from these segments

and an extreme value analysis (discussed in Chapter 3 and Chapter 4) was used to estimate the

expected positive and negative peak reactions. The Lieblein-BLUE estimators (Lieblein 1974)

used in the analysis of the six samples in each direction are provided in Table 5-3. The mean,

RMS and expected peak values of structural loads measured by the twenty-one load cells are

provided in Table 5-4. These statistical values are average loads from the three datasets for each

wind azimuth. T he locations of the load cells on a 3D drawing of the house are shown in Figure

6-13. Figures 5-14 to 5-16 show the mean, extreme positive and negative structural reactions.

The plots are such that the gable end load cells (15 & 19) are at the center of the roof-to-wall

zone. Load cells on the same truss are mirrored about load cells 15 and 19 on the plots for easy

comparison of measurements.









Even though there was no connection between the gable end truss and the gable end wall

except at end supports of the truss, there was considerable load transferred from the gable end

wall to the foundation.

At the 900 azimuth, significantly high reactions were recorded at the gable end wall-to-

foundation load cells than the other load cells. Another unusual observation made for this

direction was that load cell 20 recorded high positive reactions whereas adjacent load cells (19

and 16) recorded relatively high negative reactions. Perhaps, these observations could be

explained by fact that the portion of the building where these load cells were located was not

completely immersed in the wind flow jet (as shown in Figure 5-8) and therefore subjected to

unrealistic wind loads.

For incident 0 azimuth, it is ob served that fairly symmetrical loads were recorded at the

roof-to-wall connections on side walls. Also, load cells 4 and 16 recorded highest uplift load (26

psf& 22 psf respectively) on either sidewall suggesting that high suctions experienced along the

second interior truss were mainly transferred through its support. Load cells located at the gable

end measured equally high uplifts as observed in the DAD-based loads of wind tunnel data.

Generally, the structural loads measured at the roof-to-wall connections are highest at the

0 azimuth and lowest at the 900 azimuth. This observation was also made in the case of loads

computed based on wind tunnel data where the highest loads were estimated for the 00 azimuth

and decreased with wind direction increment of 150.

Structural Load Comparison

The envelope pressures over the house were used to estimate the structural loads being

transferred through roof-to-wall load cell numbers 4, 5 and 15, and foundation load cell number

11. Predictions were made based on the DAD methodology using the experimentally derived

influence coefficients (described in Chapter 4). For comparison with the directly measured









structural loads, actual pressure time histories (not pressure coefficients) based on the wind flow

characteristics of the various tests were utilized in the DAD approach as follows:

R (t) = [(N), x A x P (t))] (5-2)
1=1

where R,(t) is load reaction in pounds at thejth connection at time t; (Nz), is influence coefficient

at the ith grid point for thejth connection; A, is tributary area ofi th grid point; and P,(t) is

pressure measured at the ith grid point at time t.

Figures 6-17 to 6-19 show 1 minute records of measured and evaluated reactions. High

correlations are observed between the reaction time histories of the two records. The highest

correlation coefficient of 0.9 was measured between the records of load cells 15 for the 0 wind

azimuth. Measured correlation coefficients for the records are given in Table 6-5. Even though

the estimated and measured reactions are not in excellent agreement as far as magnitude is

concerned for some the load cells considered, their records show good trends. It is observed that

in almost all the comparisons, the peaks of time histories of the two records seem to be occurring

at the same time. Furthermore, significantly more fluctuating signals were obtained using the

roof pressures relative to the directly measured load cell signals. It is worth noting that a similar

observation was made by Zisis & Stathopoulos (2009) in a parallel study. Zisis & Stathopoulos

(2009) attributed this phenomenon to the attenuation effect due to structural and material

damping of the building components thereby lowering the fluctuations in the measured signal.

Peak and mean uplift structural loads measured and estimated for the four load cells are

provided in Table 6-6. Good agreements are observed comparing the mean and peak reaction in

almost all cases. Small discrepancies (lowest is 2%) are seen between the statistical values for

load cell 15 as compared to the others. Generally, the DAD methodology underestimated

structural loads expected to be transferred through the other load cells considered. This may be









due to the fact that the effective influence surfaces of these load cells were not completely

covered with pressure taps. Hence, pressure loads which were transferred through these load

cells were not fully captured.

In summary, the DAD methodology adequately predicted the expected loads at roof-to-

wall and wall-to-foundation connections based on realistic wind load paths. The ability to

estimate structural loads on light framed wood structures using the DAD approach has been

demonstrated in this experimental study.










Table 6-1. Manufactures and specifications of pressure sensors
Transducer Number Response
Model Manufacturer Used Range Output Frequency
Model Used Frequency
PX 138-001 Omega 27 1 psi (144 psf) 1 to 6 VDC 1000 Hz
D5V Engineering
Setra Systems,
265 Setraystems, 2 10 in. H20 (52 psf) 0 to 5 VDC 50 Hz
Inc.
616 -20B Dwyer 1 10 in. H20 (52 psf) 4 to 20mA 2.5 Hz
Instruments, Inc.











Table 6-2. Statistical values of measured pressure coefficients


0 450 90
Press. Neg. n Pos. R Neg. n Pos. R Neg. n Pos. R
taps Peak Peak Peak Peak Peak Peak
Roof pressure taps


-1.59 -0.56 -0.06 0.60 -0.15 0.04
-2.70 -0.51 1.08 0.56 -1.97 -0.05
-2.74 -0.48 1.77 0.52 -2.70 -0.80
-2.73 -0.45 1.35 0.50 -2.31 -0.56
-2.26 -0.52 0.13 0.56 -1.64 0.06
-1.91 -0.53 0.25 0.57 -0.05 0.11
-2.95 -0.47 0.80 0.50 -0.18 0.12
-2.41 -0.42 1.26 0.45 -0.69 0.10
-1.16 -0.42 0.05 0.44 -3.14 -0.56
-1.30 -0.44 0.05 0.46 -0.41 0.01
-2.23 -0.47 0.16 0.49 -0.17 0.04
-1.86 -0.47 0.54 0.49 -0.18 0.02
-0.89 -0.34 0.09 0.35 -3.16 -0.80
-1.02 -0.39 -0.06 0.40 -1.25 -0.07
-1.36 -0.42 0.05 0.44 -0.53 -0.10
-1.60 -0.47 0.27 0.49 -0.33 -0.10
-0.85 -0.34 0.07 0.35 -3.28 -1.08
-1.05 -0.36 -0.07 0.37 -2.30 -0.42
-2.08 -0.45 0.18 0.47 -0.76 -0.39
-1.68 -0.40 0.33 0.41 -1.15 -0.35
-0.88 -0.34 -0.01 0.35 -6.34 -1.72
-1.02 -0.36 -0.04 0.37 -3.48 -0.64
-1.56 -0.40 -0.04 0.41 -2.55 -0.54
-1.31 -0.34 0.03 0.35 -3.21 -0.92
-1.59 -0.41 0.10 0.43 -1.63 -0.23


0.21 0.06 -0.93 -0.14
1.56 0.31 -2.98 -0.33
1.24 0.88 -3.01 -0.45
0.98 0.60 -2.90 -0.51
1.38 0.20 -1.75 0.01
0.27 0.11 -1.68 0.07
0.41 0.12 -1.71 0.02
0.77 0.13 -1.81 -0.02
1.94 0.67 -0.86 0.10
0.40 0.06 -0.80 0.10
0.27 0.06 -0.57 0.15
0.22 0.05 -0.86 0.38
1.41 0.87 -1.02 -0.10
1.03 0.17 -0.53 0.10
0.22 0.11 -0.44 0.06
0.12 0.11 -0.46 0.08
0.82 1.13 -0.47 -0.01
1.30 0.49 -0.40 0.04
0.10 0.40 -0.39 0.04
0.39 0.36 -0.37 0.05
2.53 1.87 -1.09 -0.05
1.68 0.77 -1.14 -0.03
1.07 0.59 -0.86 -0.02
1.37 0.96 -0.91 -0.04
1.26 0.30 -0.89 -0.05


Wall pressure taps
26 -0.63 -0.19 0.15 0.20 0.60 0.79 1.04 0.79 -0.43 0.45
27 -1.21 -0.31 0.19 0.32 0.39 0.53 0.72 0.54 0.33 0.92
28 0.45 0.77 1.01 0.78 0.55 0.73 0.90 0.73 -1.02 -0.41
29 0.86 1.07 1.25 1.07 0.12 0.35 0.56 0.35 -0.61 0.05
Internal pressure
30 -0.01 0.00 0.01 0.00 -0.01 0.00 0.01 0.00 -0.03 0.00


0.47 0.24
1.79 0.48
1.94 0.54
1.66 0.58
1.46 0.20
1.56 0.24
1.51 0.24
2.00 0.23
1.03 0.16
0.70 0.13
0.70 0.16
1.26 0.60
0.51 0.14
0.55 0.11
0.39 0.08
0.56 0.09
0.64 0.07
0.43 0.07
0.33 0.07
0.41 0.07
0.46 0.09
0.29 0.07
0.20 0.07
0.30 0.09
0.18 0.09


1.64 0.46
1.36 0.92
-0.07 0.42
0.65 0.12


0.02 0.01











Table 6-3. Coefficients of BLUE for Type 1 Extreme-Value Distribution (Lieblein 1974)
i 1 2 3 4 5 6
a, 0.355 0.225 0.166 0.121 0.083 0.048
b, -0.459 -0.036 0.073 0.127 0.150 0.146


Table 6-4. Mean, RMS and BLUE estimated peak values of measured structural loads (lbs)


Load 450 90
Load


Cells


Neg. Mean Pos RMS Neg. Mean Pos RMS Neg. Mean Pos RMS
Peak Peak Peak Peak Peak Peak
Roof-to-wall load cells
-14.3 -8.0 -0.7 8.3 -6.9 -3.9 -1.1 4.0 0.1 3.7 6.5 3.8
-10.1 -7.1 -3.1 7.1 -8.6 -5.9 -3.1 5.9 -3.9 0.1 3.8 1.3
-13.0 -8.1 -4.6 8.2 -7.5 -5.1 -2.7 5.2 -2.0 1.3 4.4 1.5
-25.6 -18.6 -11.8 18.7 -16.1 -12.5 -8.9 12.6 -2.4 1.4 5.0 1.7
-19.7 -15.7 -11.5 15.8 -13.0 -9.9 -7.5 9.9 -16.2 -9.0 -4.9 9.1
-22.9 -15.3 -9.1 15.5 -17.5 -12.1 -6.8 12.2 -3.4 1.9 6.9 2.5
-20.7 -16.3 -12.3 16.3 -20.3 -15.8 -11.5 15.9 -7.2 -3.0 -0.5 3.1
-9.7 -5.6 -2.7 5.6 -5.7 -2.3 -1.5 2.4 -3.4 -1.1 0.1 1.3
-10.7 -7.1 -4.3 7.2 -13.5 -9.2 -5.2 9.3 -6.4 -1.0 1.9 1.6
-20.1 -13.9 -8.8 14.0 -19.0 -13.0 -7.2 13.1 -16.3 -8.5 -2.5 8.8
-19.9 -14.2 -9.9 14.3 -19.8 -14.2 -7.7 14.3 5.0 9.3 13.2 9.3
-11.0 -6.5 -2.7 6.6 -7.3 -3.7 -1.1 3.9 -6.0 -0.7 1.6 1.3
Wall-to-foundation load cells
-7.6 -3.4 0.2 3.6 -14.3 -9.9 -5.8 10.0 -6.7 -0.3 6.3 2.3
-17.5 -11.8 -5.7 11.9 -5.1 -1.1 3.2 1.8 6.4 12.7 19.4 12.9
-23.3 -17.7 -11.1 17.8 -4.8 -0.5 3.8 1.5 7.3 14.7 21.4 14.9
-22.3 -18.1 -13.3 18.1 -10.7 -7.3 -3.9 7.4 -5.3 2.3 8.8 3.2
-21.8 -18.1 -13.7 18.1 -16.7 -13.7 -10.8 13.7 -26.9 -17.4 -10.5 17.5
-18.5 -13.7 -8.8 13.8 -19.4 -14.5 -10.2 14.6 -32.3 -21.0 -11.8 21.2
-18.8 -14.6 -10.4 14.7 -17.6 -13.5 -9.5 13.6 -44.7 -29.8 -19.0 30.1
-10.7 -8.1 -5.5 8.2 -9.7 -7.4 -4.9 7.4 -18.2 -11.7 -6.7 11.8
2.7 4.7 6.3 4.7 -2.0 -0.2 1.8 0.7 -5.7 -1.8 1.5 2.1










Table 6-5. Correlation coefficients measured between time histories of measured and estimated
reactions
Azimuth 00 450 900
Load Cell 4 5 11 15 4 5 11 15 4 5 11 15
Coefficient 0.83 0.85 0.56 0.90 0.48 0.30 0.30 0.41 0.47 0.43 0.43 0.45



Table 6-6. Comparison of directly measured and DAD-based estimated reactions
Wind Peak Uplift (lbs) Mean (lbs)
Direction Load 4 5 11 15 4 5 11 15
Cell
S DAD -19.6 -14.7 -16.4 -22.0 -9.7 -8.0 -9.6 -12.3
00
Measured -25.9 -19.7 -18.5 -22.9 -18.6 -15.7 -13.7 -15.3
'450 DAD -11.3 -11.4 -11.4 -25.3 -5.0 -5.6 -7.2 -10.3
Measured -15.9 -13.0 -19.4 -17.5 -12.5 -9.9 -14.5 -12.1
DAD -7.8 -5.7 -5.3 -8.8 0.5 0.2 0.5 0.4
900
Measured -3.3 -16.0 -32.3 -3.36 1.4 -9.0 -21.0 1.9


- Load Cell
SPressure Sensor
3.0 Internal Pressure Sensor


13'-4" k


2'-10"


Figure 6-1. Locations of pressure taps and load cells on the scale house model.


f 21 Trusses @ 8" o.c.
19 20 16 17 18 21
T -- -



25


Z 22 23 .



1J 14 15 16

21 1 11 12.


6 7 8
j;I*4.3-2- _I -- --- --


10 9 8 7 6
























Figure 6-2. A) A pressure tap sample B) Layout of pressure tap over the building roof


Figure 6-3. Interior of the house model









L -r
-- ~C:I~Y -14 ~~
6 ~ ~4 .


Figure 6-4. Pressure Sensors A) Omega PX 138 B) Setra 265 C) Dwyer 616











SA

Figure 6-5. Futek load cells A) Roof-to-wall connection B) Wall-to-foundation connection

... .7 I


Figure 6-6. Fluke pressure calibrator























Figure 6-7. Dwyer Transducer and pitot-tube for wind velocity measurements









































410


-scale Model House :-scale Model House
.j 4.- r., wind flow @ 900 to wind flow


Figure 6-8. Layout of experimental set-up



104


2'1-2,











































C
Figure6-9. A) Completed setup B) House at 00 orientation C) House at 450 orientation D) House at 900 orientation






















Figure 6-10. UF Wind Generator


Azimuth = 000


Azimuth = 045


Azimuth = 090


Figure 6-11. Spatial distributions of peak pressure coefficient


i,











Azimuth = 090
,


i U


Figure 6-12. Spatial distributions of mean pressure coefficient


Figure 6-13. Location of load cells on the house













107


0.2

-0

S-0.2

S-0.4


Azimuth = 045


Azimuth = 000










Wind Direction 000


30

20

10


0


- -10
0
-j
-20


-30


-40

-50


Figure 6-14. Mean and expected peak values of measured structural loads for wind direction 0


Wind Direction 045


7- -10
0
-j
-20


-30

-40


-50


Figure 6-15.


1 2 3 4 5 6 7 8 9 101112 1 2 3 4 5 6 7 8 9
Load Cells
Mean and expected peak values of measured structural loads for wind direction 450


1 2 3 4 5 6 7 8 9 101112 1 2 3 4 5 6 7 8 9
Load Cells











Wind Direction 090


Roof-to-Wall
Connections






El O
H0


30


20


10


0


- -10
O
0
-20


-30


-40


-50


CU S


Wall-to-Foundation
11 connections


I INegative Extreme
i Positive Extreme
E Mean

8 9 101112 1 2 3 4 5
Load Cells


6789


Figure 6-16. Mean and expected peak values of measured structural loads for wind direction 900


0 10 20 30
Time (Sec)


40 50 60


40 50 60


0 10 20 30
Time (Sec)


10

0
vw
0 -10
o
0
_j
-20


Load Cell 5








I I I
) 10 20 30 40 50 61
Time (Sec)

Load Pell 15 |
. -..-.. .. ...... .......... -





i--- DAD estimated
SMeasured


0 10 20 30
Time (Sec)


40 50 60


Figure 6-17. Comparison of measured and estimated reactions for wind direction 00


1234567


a


Load Cell 11
0 .. ... ......... .... ............ .-- .............. ........ .... ............. -- ............. ....... ..... .............


-20

-20 .----- -------------....-. ..-....

zin _____ _________


III II I I I I I l--




















10


-20


-30


lu
LoadiCell 4




-10 .


-2 0 ........... .......... ... ...................... .


-30 ----- ----------
0 10 20 30 40 50 60
Time (Sec)

10
Load Cell 11

0I I I


10 30 40 50 60


-20-


-30
0 10 20 30 40 50 60


In


LoadiCell 5












) 10 20 30 40 50 61
Time (Sec)


10
0 r-i- i.


U)

-o

_1


SLoad Cell 15

SI i i j


0 10 20 30 40


50 60


Time (Sec) Time (Sec)
Figure 6-18. Comparison of measured and estimated reactions for wind direction 450


Iu-





-1 0 .. ......... .... ..... ............. ............. ........ ... ............. ............. ............
0f





Load Cell 4
-20 ---- -----------


-3 0 I J .. L .. .


0 10 20 30
Time (Sec)


10


0


-10
-20
0

-20


40 50 60


0 10 20 30
Time (Sec)

10


0


-20
0 Load Cell 1E
_i
-20 --- --- -- =


40 50 60


0 10 20 30 40 50 60 0 10 20 30 40
Time (Sec) Time (Sec)

Figure 6-19. Comparison of measured and estimated reactions for wind direction 900


50 60


- I I- L I


A









CHAPTER
CONCLUSION AND RECOMMENDATIONS

This chapter provides a summary of work undertaken and key findings made. Significant

conclusions which can be drawn from the study as well as recommendations for future research

in this study area are discussed.

Summary

Light framed wood structures (LFWS) are by far the largest contributor to monetary losses

associated hurricane damages. Inadequate load transfer mechanism provided in wood

constructions account significantly to structural failures of such buildings. The need for studies

to provide better understanding of wind load paths on LFWS buildings cannot be overemphasize.

This research was aimed at estimating wind-induced structural loads transferred through

roof-to-wall and wall-to-foundation connections on LFWS. The estimations were based on

Database-Assisted Design (DAD) methodology, developed by the National Institute of Standards

and Technology (NIST). This methodology, which utilizes large databases of aerodynamic

pressures and climatological information, has been used by other researchers to predict structural

responses due to wind forces on steel portal frame buildings. This current study, apart from

extending the application of the DAD approach to LFWS, also demonstrates the validity of this

method logy to adequately evaluate structural reactions on LFWS.

The project was accomplished in two phases: a hybridized analytical approach and an

experimental approach. In the former, spatially distributed pressure coefficients were derived

from wind tunnel data for a 1/50 scale model house. These pressure coefficients (time histories)

were combined with structural influence coefficients (developed on a 1/3 scale model wood

house) through computer-based analysis to generate structural reactions. Peak uplift reactions

were estimated from a Lieblein BLUE-fitted distribution of the measured peak reactions. These









were then compared to wind design loads based on ASCE-7 provisions for both Main Wind

Force Resisting Systems (MWFRS) and Components and Cladding (C&C). It was realized that

using MWFRS pressures underestimated the reactions by up to 32% while C&C provisions

resulted in highly conservative estimates (up to 60% overestimation).

In the experimental study, the 1/3-scale wood house model was instrumented with pressure

and load sensors. The house was subjected to wind forces while load and pressure measurements

were simultaneously taken. Based on the DAD approach, structural reactions were estimated

using measured roof pressures and results were compared to directly-measured structural loads.

Even though significantly fluctuating reaction records were obtained using the DAD in

comparison to the directly measured reaction records, the two quantities were highly correlated.

Again, good agreements were found comparing mean and peak values of the estimated and

measured reactions.

Conclusions

The conclusions of this study can be summarized as follows:

1. Local peak pressure coefficients derived from wind tunnel analysis are considerably
higher, in most cases, than ASCE 7-05 component and cladding external pressure
coefficients. However, an excellent agreement is observed between the wind tunnel area-
averaged pressure coefficients and the ASCE 7 standard provisions.

2. ASCE 7-05 MWFRS provisions produced lower peak reactions ( average of21% lower)
than predictions based on DAD methodology while predictions based on ASCE 7-05
components and cladding were generally higher.(average of 33 % higher) than DAD-based
reactions.

3. Despite limited match of the wind generated in the experiment to realistic wind flows,
overall the roof pressure distributions on the 1/3-scale house were reasonably matched to
the wind tunnel pressure spatial distributions in the literature.

4. The structural reaction time history obtained using the DAD method had greater dynamic
content than the time history of directly measured structural reactions. Despite this fact
there was still good a agreement between the peak values and the mean values of DAD
results and the measured reactions.









5. DAD-based reactions were highly correlated with directly measured structural reactions.

6. Both the analytical and experimental components of this study confirmed that maximum
uplifts loads are transferred through the gable end support ts.

7. Lastly, a wider spread of load sharing through wall-to-foundation connection was observed
in both analytical and experimental studies. This may be due to diaphragm (deep beam)
actions of the building walls.

Finally, the study has not just extended the application of DAD methodology to LFWS but

has evaluated its validity to adequately evaluate structural reactions on LFWS through an

experimental study.

Recommendation

Notwithstanding the significant findings reported in this study and its possible influence on

design practices for LFWS, further research is needed in order to fully comprehend wind load

paths in LFWS. Recommendations for future research in this discipline are discussed below.

1. The study house is a simple rectangular gable roof structure. Most typical residential
structures have complex shapes which may result in different pressure distributions on
houses and subsequently different wind load paths. It is recommended that studies are done
on LFWS with complex building shapes and different configurations.

2. In future experimental studies, load cells should be installed around the whole perimeter of
the 1/3-scale building at roof-to-wall and wall-to-foundation interfaces. This will ensure
that wind load paths are not unduly affected by the stiffness of the load cells.

3. Also, the entire roof of the house should be equipped with pressure sensors so that the
DAD methodology could be validated at all critical connections.

4. A better boundary layer condition should be created for future experiments so that
comparison can be made between field and wind tunnel measurements.

5. Reliability studies should be carried out to quantify the uncertainty parameters for load
estimations on LFWS based on DAD approach.











APPENDIX A
MEAN, RMS AND EXTREME VALUES OF WIND TUNNEL PRESSURE COEFFICIENTS

Table A-1. Peak, mean and RMS pressure coefficients of selected pressure taps
0 450 90


Press.
Taps
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
262
263
264
265
266
267
268
269


Neg. Pos.
Peak Peak
-2.74 -0.37 0.16 0.43
-2.43 -0.37 0.21 0.43
-2.22 -0.37 0.35 0.43
-2.16 -0.36 0.34 0.42
-2.10 -0.32 0.43 0.39
-1.84 -0.22 0.49 0.28
-1.79 -0.15 0.48 0.21
-1.42 -0.13 0.46 0.18
-1.20 -0.18 0.33 0.21
-1.13 -0.11 0.39 0.15
-0.86 -0.09 0.28 0.13
-0.86 -0.07 0.29 0.11
-0.72 -0.08 0.31 0.11
-0.70 -0.08 0.24 0.11
-0.65 -0.06 0.27 0.09
-0.60 -0.05 0.25 0.09
-0.55 -0.05 0.24 0.08
-0.56 -0.08 0.21 0.10
-0.58 -0.08 0.20 0.10
-0.56 -0.09 0.18 0.11
-0.53 -0.07 0.19 0.09
-0.78 0.84 2.34 1.04
-0.40 -0.10 0.13 0.11
-0.44 -0.09 0.16 0.11
-0.45 -0.07 0.18 0.09
-0.44 -0.07 0.20 0.09
-0.45 -0.07 0.22 0.09
-0.50 -0.06 0.23 0.08
-0.57 -0.05 0.25 0.08
-0.63 -0.06 0.26 0.09


Neg. Pos Neg. Pos
g Mean RMS Neg. Mean RMS
Peak Peak Peak Peak
-4.14 -0.81 0.17 0.92 -1.74 -0.25 0.12 0.28
-4.13 -0.83 0.15 0.94 -1.67 -0.22 0.11 0.25
-4.39 -0.78 0.22 0.90 -1.60 -0.23 0.10 0.26
-3.80 -0.79 0.19 0.89 -2.02 -0.22 0.21 0.25
-3.56 -0.71 0.23 0.80 -1.19 -0.20 0.12 0.23
-2.65 -0.60 0.23 0.67 -1.02 -0.20 0.10 0.23
-2.17 -0.52 0.16 0.58 -0.88 -0.20 0.09 0.22
-1.86 -0.49 0.06 0.53 -0.93 -0.22 0.06 0.24
-1.72 -0.50 -0.03 0.54 -1.06 -0.30 0.00 0.31
-1.60 -0.45 0.01 0.48 -1.01 -0.23 0.04 0.25
-1.46 -0.42 0.01 0.44 -1.11 -0.25 0.03 0.27
-1.40 -0.33 0.09 0.36 -0.96 -0.20 0.08 0.22
-1.27 -0.31 0.08 0.34 -0.94 -0.21 0.08 0.23
-1.19 -0.28 0.09 0.31 -0.94 -0.21 0.06 0.23
-1.22 -0.24 0.09 0.27 -0.97 -0.20 0.08 0.22
-1.16 -0.22 0.10 0.25 -1.04 -0.20 0.09 0.22
-1.20 -0.18 0.16 0.21 -1.09 -0.20 0.10 0.23
-1.12 -0.22 0.14 0.24 -1.19 -0.23 0.08 0.25
-1.22 -0.21 0.12 0.23 -1.35 -0.23 0.09 0.25
-1.36 -0.21 0.12 0.24 -1.41 -0.24 0.08 0.26
-1.50 -0.20 0.14 0.24 -1.43 -0.23 0.10 0.26
-1.22 0.51 2.06 0.65 -1.06 0.62 2.10 0.76
-0.52 -0.13 0.12 0.14 -0.74 -0.15 0.36 0.18
-0.50 -0.11 0.21 0.13 -0.64 -0.11 0.43 0.14
-0.46 -0.10 0.23 0.12 -0.57 -0.08 0.45 0.12
-0.48 -0.10 0.24 0.12 -0.55 -0.08 0.44 0.12
-0.48 -0.08 0.27 0.10 -0.45 -0.05 0.44 0.10
-0.42 -0.08 0.29 0.10 -0.43 -0.04 0.46 0.09
-0.48 -0.08 0.28 0.11 -0.46 -0.05 0.47 0.10
-0.45 -0.09 0.29 0.11 -0.46 -0.06 0.45 0.10











Table A-1. (Cont'd)
0 450 90


Press.
Taps
270
273
274
275
276
277
278
279
280
281
282
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387


Neg. n Pos. R
Peak Peak
-0.56 -0.07 0.24 0.09
-0.77 -0.08 0.33 0.11
-0.86 -0.13 0.34 0.16
-1.02 -0.14 0.33 0.17
-1.09 -0.12 0.44 0.17
-1.30 -0.20 0.46 0.25
-2.15 -0.32 0.36 0.38
-2.15 -0.37 0.40 0.42
-2.29 -0.37 0.28 0.43
-2.22 -0.36 0.26 0.42
-2.13 -0.37 0.21 0.42
-0.57 -0.09 0.21 0.11
-0.63 -0.08 0.20 0.10
-0.63 -0.07 0.20 0.10
-0.63 -0.07 0.21 0.09
-0.69 -0.07 0.23 0.09
-0.66 -0.07 0.25 0.09
-0.65 -0.07 0.24 0.10
-0.69 -0.06 0.26 0.09
-0.83 -0.07 0.32 0.10
-0.92 -0.07 0.34 0.10
-0.97 -0.15 0.26 0.17
-0.96 -0.12 0.30 0.15
-1.02 -0.18 0.24 0.21
-1.17 -0.12 0.41 0.17
-1.26 -0.15 0.39 0.20
-1.45 -0.18 0.36 0.24
-1.60 -0.24 0.41 0.29
-2.01 -0.28 0.42 0.34
-2.25 -0.32 0.43 0.39
-2.52 -0.39 0.44 0.47
-3.71 -0.47 0.39 0.59


Neg. n Pos. RMS
Peak Peak
-0.45 -0.08 0.29 0.11
-0.47 -0.09 0.28 0.11
-0.52 -0.13 0.26 0.15
-0.50 -0.12 0.27 0.14
-1.15 -0.05 0.47 0.08
-1.34 -0.04 0.50 0.08
-1.90 -0.07 0.40 0.12
-2.28 -0.09 0.53 0.23
-2.29 -0.16 0.58 0.33
-2.08 -0.20 0.66 0.35
-1.90 -0.21 0.73 0.34
-1.26 -0.16 0.28 0.20
-1.28 -0.17 0.33 0.20
-1.18 -0.16 0.38 0.20
-1.31 -0.16 0.44 0.21
-1.37 -0.17 0.43 0.21
-1.31 -0.17 0.41 0.22
-1.47 -0.17 0.47 0.22
-1.43 -0.17 0.47 0.22
-1.35 -0.18 0.46 0.23
-1.52 -0.18 0.50 0.23
-1.40 -0.21 0.46 0.26
-1.50 -0.21 0.45 0.25
-1.43 -0.25 0.37 0.29
-1.29 -0.14 0.50 0.20
-1.38 -0.13 0.53 0.19
-1.32 -0.11 0.56 0.18
-1.20 -0.11 0.63 0.19
-1.23 -0.11 0.70 0.20
-1.73 -0.08 0.83 0.21
-1.86 -0.02 0.85 0.19
-3.91 -0.04 0.89 0.33


Neg. n Pos. R
Peak Peak
-0.43 -0.05 0.46 0.10
-0.48 -0.07 0.47 0.11
-0.53 -0.12 0.41 0.14
-0.49 -0.10 0.43 0.13
-0.39 -0.01 0.50 0.08
-0.40 -0.02 0.52 0.09
-0.58 -0.05 0.47 0.10
-0.48 -0.04 0.53 0.10
-0.54 -0.04 0.56 0.10
-0.64 -0.06 0.54 0.12
-0.81 -0.12 0.51 0.17
-1.59 -0.25 0.60 0.31
-1.61 -0.22 0.67 0.29
-1.58 -0.20 0.69 0.28
-1.59 -0.20 0.76 0.28
-1.50 -0.20 0.80 0.27
-1.37 -0.20 0.73 0.27
-1.51 -0.20 0.75 0.27
-1.43 -0.20 0.78 0.27
-1.43 -0.21 0.72 0.27
-1.48 -0.20 0.73 0.27
-1.56 -0.31 0.67 0.36
-1.58 -0.25 0.71 0.30
-1.60 -0.31 0.71 0.35
-1.32 -0.18 0.86 0.25
-1.45 -0.18 0.86 0.25
-1.38 -0.16 0.90 0.25
-1.44 -0.18 0.91 0.26
-1.47 -0.19 0.85 0.27
-1.62 -0.18 0.90 0.27
-1.75 -0.18 0.85 0.27
-1.45 -0.18 0.73 0.26









APPENDIX B
MEASURED STATISTICAL VALUES OF VERTICAL REACTIONS DERIVED FROM
WIND TUNNEL DATA











Table B-1. Measured peak, mean and RMS reactions (lbs) for sample 2
Azimuth 0 450 90 1350 1800

Load Neg. Mean Pos. Neg. Mean Pos. Neg. Mean Pos. Neg. Mea Pos. Neg. Pos.
Cells Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak

1 -812 -178 101 204 -375 -99 170 117 -411 -69 352 99 -439 -107 151 124 -272 -28 177 51
2 -878 -207 72 234 -400 -119 120 136 -376 -69 318 95 -403 -100 140 116 -254 -28 159 49
3 -1031 -262 58 292 -577 -146 99 164 -408 -87 300 112 -440 -117 145 131 -256 -35 134 56
4 -1220 -301 50 333 -730 -146 155 170 -423 -95 310 121 -468 -124 157 139 -266 -38 134 59
5 -1008 -285 19 313 -869 -183 47 207 -382 -106 176 122 -365 -123 104 134 -225 -35 104 51
6 -374 -99 25 109 -212 -64 31 71 -163 -40 115 48 -176 -53 50 58 -111 -14 56 23
7 -204 -52 16 58 -107 -34 20 38 -92 -21 68 26 -100 -29 29 32 -62 -8 33 13
8 -300 -84 15 92 -192 -51 25 58 -120 -31 91 38 -137 -40 39 44 -85 -11 39 18
9 -688 -195 24 214 -484 -120 50 136 -268 -71 187 85 -296 -90 86 99 -185 -26 87 40
10 -1044 -299 30 327 -785 -197 60 221 -414 -115 245 134 -433 -139 121 152 -273 -39 127 60
11 -1423 -407 41 446 -1071 -268 80 302 -570 -159 331 185 -596 -191 165 209 -374 -54 173 82
12 -1270 -363 36 397 -952 -238 72 268 -508 -141 298 164 -531 -170 147 186 -332 -48 154 73
13 -372 -106 10 116 -280 -71 19 80 -151 -42 84 49 -155 -50 42 55 -97 -14 44 21
14 -49 -14 2 15 -38 -10 2 11 -21 -6 12 7 -20 -7 5 7 -13 -2 6 3
15 -1476 -458 61 504 -1644 -315 119 374 -669 -175 220 201 -558 -195 127 212 -343 -58 162 83
19 -1286 -323 98 360 -1024 -344 -58 369 -698 -183 14 198 -507 -163 64 176 -243 -44 102 60
20 -1303 -300 100 331 -910 -325 -31 348 -647 -186 19 200 -507 -168 74 182 -229 -40 129 59











Table B-2. Measured peak, mean and RMS reactions (Ibs) for sample 3
Azimuth 0 450 90 1350 1800

Load Neg. Pos. RMS Neg. load Neg. Mean Pos. RMS Neg. load Neg. Pos. RMS Neg. load Neg.
Mean Mean ean RMS Mean RMS Mean
Cells Peak Peak Peak cells Peak Peak Peak cells Peak Peak Peak cells Peak

1 -771 -179 104 205 -399 -100 184 119 -384 -72 495 101 -356 -77 184 95 -317 -33 153 55
2 -858 -206 96 233 -494 -123 164 140 -405 -72 440 97 -342 -73 163 89 -287 -33 147 53
3 -1061 -259 85 289 -601 -151 156 171 -478 -90 437 114 -357 -86 148 101 -315 -41 142 60
4 -1327 -297 54 328 -779 -153 170 179 -519 -98 477 123 -385 -91 157 107 -320 -44 144 64
5 -1227 -281 19 308 -1041 -192 95 219 -454 -109 243 125 -336 -91 66 103 -262 -40 135 56
6 -361 -98 26 109 -249 -66 51 74 -180 -41 159 49 -146 -39 51 45 -133 -17 64 25
7 -200 -52 17 58 -128 -35 31 39 -98 -22 95 27 -82 -21 31 25 -75 -9 35 14
8 -293 -83 17 91 -219 -53 44 61 -144 -32 121 38 -114 -30 40 34 -100 -13 49 19
9 -696 -193 27 212 -553 -125 100 143 -328 -73 244 87 -249 -67 79 76 -216 -30 107 44
10 -1087 -295 20 324 -903 -205 134 233 -498 -118 313 136 -371 -103 101 117 -320 -45 159 65
11 -1479 -402 22 441 -1237 -280 178 318 -681 -164 421 188 -511 -142 138 160 -438 -62 218 89
12 -1321 -358 20 393 -1104 -248 161 282 -606 -145 378 167 -457 -125 125 142 -390 -55 193 79
13 -389 -105 4 115 -324 -75 44 84 -177 -43 106 50 -135 -37 35 42 -114 -16 56 23
14 -53 -14 1 15 -43 -11 5 12 -24 -6 14 7 -19 -5 5 6 -15 -2 7 3
15 -1723 -450 31 497 -1980 -334 231 402 -720 -180 349 204 -507 -145 79 163 -408 -66 214 90
19 -1400 -321 49 359 -1139 -352 -19 379 -667 -186 34 200 -542 -122 36 136 -275 -49 113 65
20 -1160 -299 97 330 -1104 -335 -9 361 -657 -189 42 203 -505 -125 44 139 -278 -46 140 64











Table B-3. Measured peak, mean and RMS reactions (lbs) for sample 4
Azimuth 0 450 90 1350 1800

Load Neg. Pos. Neg. PosNeg. Pos. Neg. Pos. Neg. Pos.
Load Neg. Mean RMS Mean RMS Neg. Mean Pos. RMS Neg. Mean RMS N Mean RMS
Cells Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak

1 -787 -174 86 201 -873 -98 257 117 -353 -74 349 101 -358 -86 200 103 -291 -34 120 53
2 -912 -202 73 229 -979 -120 228 137 -324 -73 307 97 -331 -81 168 97 -289 -34 108 52
3 -1162 -256 68 287 -718 -147 237 167 -361 -90 309 113 -390 -95 143 110 -295 -41 112 59
4 -1471 -295 67 328 -926 -150 264 174 -418 -97 324 122 -431 -101 139 116 -317 -45 122 63
5 -1291 -279 47 308 -874 -186 148 211 -365 -107 151 123 -333 -101 98 112 -317 -41 91 55
6 -409 -97 21 107 -267 -64 85 72 -148 -41 115 49 -138 -43 52 49 -117 -17 45 24
7 -212 -51 13 57 -142 -34 48 38 -81 -22 69 27 -78 -23 33 27 -66 -9 26 13
8 -354 -82 15 91 -220 -52 70 59 -117 -32 89 38 -112 -33 39 37 -90 -13 34 19
9 -837 -191 35 211 -518 -121 154 138 -264 -73 179 87 -249 -74 74 83 -198 -31 73 43
10 -1266 -293 48 322 -847 -199 203 224 -408 -117 226 135 -367 -114 107 127 -292 -46 107 64
11 -1720 -399 61 438 -1159 -271 273 306 -557 -162 304 187 -501 -157 145 175 -399 -63 146 87
12 -1536 -355 54 391 -1028 -240 245 271 -496 -144 273 166 -449 -139 128 155 -354 -56 130 78
13 -448 -104 14 114 -305 -72 67 81 -146 -43 76 49 -131 -41 37 46 -103 -16 38 23
14 -59 -14 2 15 -44 -10 7 12 -20 -6 10 7 -19 -6 5 6 -14 -2 5 3
15 -2008 -449 94 497 -1757 -322 293 382 -616 -177 205 201 -535 -161 134 178 -458 -67 135 89
19 -1378 -318 100 357 -1132 -340 5 366 -590 -181 11 194 -463 -135 80 149 -309 -51 75 65
20 -1073 -295 71 326 -1012 -323 21 348 -560 -185 10 198 -470 -138 91 153 -301 -49 89 64











Table B-4. Measured peak, mean and RMS reactions (lbs) for sample 5
Azimuth 0 450 90 1350 1800

Load Neg. Pos. Neg. PosNeg. Pos. Neg. Pos. Neg. Pos.
Load Neg. Mean RMS Mean RMS Neg. Mean Pos. RMS Neg. Mean RMS N Mean RMS
Cells Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak

1 -745 -158 121 185 -499 -101 151 120 -396 -45 335 80 -383 -88 213 106 -242 -33 142 54
2 -851 -188 102 216 -516 -124 134 141 -373 -44 286 76 -345 -84 171 99 -241 -33 128 53
3 -1016 -240 97 270 -628 -152 156 171 -429 -56 239 86 -372 -98 136 113 -271 -41 132 59
4 -1100 -276 102 310 -802 -153 176 177 -449 -59 255 92 -400 -105 118 120 -270 -44 136 63
5 -970 -261 65 291 -850 -190 117 215 -380 -73 147 93 -327 -104 51 116 -220 -40 118 55
6 -360 -89 34 100 -266 -66 62 74 -175 -27 95 37 -151 -44 53 50 -102 -17 59 25
7 -199 -47 20 53 -139 -35 35 39 -97 -14 59 21 -86 -24 34 27 -57 -9 34 14
8 -298 -76 27 85 -226 -53 49 60 -134 -21 69 29 -117 -34 38 38 -80 -13 44 19
9 -665 -178 59 198 -542 -125 109 141 -292 -48 132 65 -254 -76 66 85 -173 -30 97 43
10 -1006 -273 78 302 -877 -204 152 229 -441 -79 168 102 -370 -117 76 131 -253 -46 142 64
11 -1373 -371 103 412 -1191 -279 205 313 -605 -109 228 141 -510 -161 105 180 -346 -62 195 88
12 -1227 -331 92 367 -1057 -247 183 277 -537 -97 204 125 -453 -143 94 160 -308 -55 173 78
13 -360 -97 26 107 -314 -74 51 83 -159 -29 58 37 -133 -42 26 47 -90 -16 50 23
14 -48 -13 4 14 -44 -11 6 12 -22 -4 8 5 -19 -6 4 6 -12 -2 6 3
15 -1489 -420 111 470 -1607 -327 220 386 -611 -125 259 154 -526 -165 74 183 -349 -66 182 89
19 -1255 -296 82 335 -1298 -354 12 382 -597 -141 61 158 -549 -138 51 153 -255 -50 105 65
20 -1053 -272 71 304 -1213 -335 38 361 -545 -141 67 158 -558 -141 56 157 -249 -47 140 64











Table B-5. Measured peak, mean and RMS reactions (lbs) for sample 6
Azimuth 0 450 90 1350 1800

Load Neg. Pos. Neg. PosNeg. Pos. Neg. Pos. Neg. Pos.
Load Neg. Mean RMS Mean RMS Neg. Mean Pos. RMS Neg. Mean RMS N Mean RMS
Cells Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak

1 -683 -159 95 189 -639 -106 168 124 -366 -56 281 89 -427 -88 294 107 -237 -34 127 54
2 -795 -189 88 220 -479 -125 146 142 -345 -55 285 85 -405 -84 265 100 -223 -35 127 53
3 -1055 -242 81 275 -715 -148 140 168 -379 -68 311 96 -382 -98 233 114 -239 -42 120 60
4 -1301 -279 90 315 -928 -147 173 172 -376 -72 348 102 -390 -104 214 121 -255 -46 114 64
5 -1180 -264 73 295 -962 -185 71 210 -343 -85 209 102 -387 -105 90 117 -213 -42 88 56
6 -362 -90 36 102 -263 -66 49 74 -156 -32 103 42 -156 -44 88 51 -101 -17 50 25
7 -186 -48 20 54 -152 -35 29 39 -87 -17 59 23 -88 -24 54 28 -57 -9 29 14
8 -316 -77 26 87 -236 -53 39 60 -120 -25 86 32 -122 -34 65 39 -79 -14 36 19
9 -753 -180 56 201 -576 -123 82 140 -265 -57 184 73 -274 -76 128 86 -174 -31 77 44
10 -1169 -275 80 307 -906 -201 105 227 -406 -92 244 113 -423 -118 159 132 -255 -47 111 65
11 -1596 -375 111 418 -1229 -275 140 310 -555 -128 330 156 -582 -163 216 182 -347 -65 153 89
12 -1424 -334 96 373 -1099 -243 127 274 -493 -113 297 139 -516 -144 194 161 -309 -58 135 79
13 -419 -98 27 109 -322 -73 34 82 -146 -34 84 41 -153 -43 54 48 -89 -17 39 23
14 -56 -13 4 14 -43 -10 3 12 -20 -5 11 6 -21 -6 6 6 -11 -2 5 3
15 -1883 -425 119 475 -1649 -325 160 385 -547 -142 310 169 -590 -168 133 186 -335 -69 136 91
19 -1124 -300 139 342 -1192 -342 1 370 -531 -157 44 172 -536 -140 61 155 -226 -52 82 66
20 -1162 -276 133 310 -1072 -325 -4 351 -494 -157 58 172 -531 -144 72 160 -220 -49 105 65











Table B-6. Measured peak, mean and RMS reactions (lbs) for sample 7
Azimuth 0 450 90 1350 1800

Load Neg. Pos. Neg. PosNeg. Pos. Neg. Pos. Neg. Pos.
Load Neg. Mean RMS Mean RMS Neg. Mean Pos. RMS Neg. Mean RMS N Mean RMS
Cells Peak Peak Peak Peak Peak Peak Peak Peak Peak Peak

1 -826 -159 142 190 -453 -108 176 125 -416 -62 285 92 -416 -91 221 109 -232 -33 136 54
2 -899 -188 147 219 -481 -127 142 144 -395 -61 217 88 -416 -86 164 102 -229 -34 125 52
3 -1003 -240 152 273 -544 -151 122 170 -434 -76 222 102 -373 -100 118 116 -251 -41 132 59
4 -1068 -276 172 312 -729 -150 139 174 -446 -80 266 109 -393 -107 125 122 -251 -44 134 63
5 -1020 -261 95 292 -835 -187 77 212 -409 -93 186 111 -380 -106 57 117 -214 -40 106 55
6 -369 -90 54 102 -219 -67 44 75 -187 -35 79 44 -154 -45 48 51 -108 -17 55 25
7 -205 -47 31 54 -122 -36 27 40 -104 -19 47 24 -86 -25 31 28 -60 -9 31 14
8 -295 -76 42 86 -193 -54 36 60 -144 -27 62 35 -118 -35 32 39 -79 -13 42 19
9 -659 -178 92 200 -484 -125 75 141 -318 -63 142 78 -265 -78 57 87 -171 -30 90 43
10 -998 -273 121 305 -791 -204 99 229 -482 -102 203 122 -410 -120 75 134 -248 -46 130 64
11 -1361 -371 161 416 -1075 -278 135 312 -664 -141 273 169 -564 -165 101 184 -333 -62 177 88
12 -1208 -331 143 370 -958 -246 122 277 -590 -125 246 150 -500 -146 90 163 -294 -56 157 78
13 -351 -97 40 108 -284 -74 34 83 -174 -37 71 44 -148 -43 25 48 -85 -16 46 23
14 -45 -13 5 14 -40 -11 4 12 -24 -5 10 6 -21 -6 3 6 -12 -2 6 3
15 -1549 -420 141 471 -1672 -327 166 387 -650 -157 290 185 -613 -169 64 186 -351 -67 171 89
19 -1259 -293 73 336 -1199 -343 5 371 -576 -166 72 181 -540 -140 39 155 -252 -50 107 64
20 -1118 -272 114 306 -1033 -326 13 353 -570 -168 69 183 -522 -144 49 159 -276 -47 134 64











Table B-7. Measured peak, mean and RMS reactions (lbs) for sample 8


Azimuth 0 450 90 1350 1800


Load
Cells

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
19
20


Neg. Pos. R Neg. M
Mean RMS Mean
Peak Peak Peak

-846 -165 109 194 -432 -115
-917 -195 100 224 -431 -134
-1096 -248 98 279 -635 -157
-1185 -285 108 318 -832 -156
-1085 -269 61 297 -859 -193
-389 -93 35 104 -225 -70
-214 -49 20 55 -112 -37
-327 -79 29 88 -203 -56
-750 -184 63 204 -504 -130
-1130 -282 83 311 -803 -211
-1538 -384 107 423 -1093 -288
-1371 -342 97 377 -975 -255
-400 -100 27 110 -286 -76
-53 -13 4 14 -39 -11
-1863 -432 114 478 -1647 -336
-1372 -308 115 349 -1172 -351
-1189 -283 92 314 -1158 -334


Pos. S Neg. M
RMS Mean
Peak Peak

164 131 -386 -62
130 149 -339 -61
131 175 -377 -76
150 179 -411 -82
77 216 -433 -95
38 77 -155 -36
24 41 -83 -19
34 62 -132 -28
67 145 -309 -64
91 234 -483 -103
123 320 -656 -142
111 283 -584 -126
31 84 -172 -38
4 12 -24 -5
132 393 -853 -159
-16 376 -779 -168
5 357 -662 -170


Pos. Neg. Pos. Neg.
Pos. RMS Neg. Mean RMS Mean
Peak Peak Peak Peak

318 93 -398 -88 210 105 -265 -33
299 89 -349 -83 176 99 -266 -33
301 103 -413 -98 169 113 -284 -41
337 110 -461 -104 202 120 -309 -45
173 113 -440 -104 162 116 -272 -41
97 45 -179 -44 63 50 -105 -17
59 24 -98 -24 38 27 -60 -9
79 35 -139 -34 50 38 -83 -13
164 79 -309 -76 111 85 -188 -31
216 123 -471 -117 162 131 -278 -46
294 170 -647 -161 222 180 -379 -63
264 151 -576 -143 197 160 -339 -56
75 45 -170 -42 57 47 -98 -16
10 6 -24 -6 7 6 -13 -2
240 187 -702 -166 233 184 -383 -68
49 184 -497 -138 120 153 -224 -52
56 186 -514 -142 136 158 -235 -49


Pos.
RMS
Peak

137 55
126 54
121 61
120 66
104 57
50 25
29 14
36 20
80 44
121 66
167 91
148 81
43 23
6 3
160 92
117 67
144 67











APPENDIX C
PRESSURE COEFFICIENTS MEASURED IN THE 1/3-SCALE HOUSE TEST

Table C-1. Measured peak, mean and RMS pressure coefficients for test repeat 1
0 450 90
Press. Neg. Pos. Neg. Pos. Neg. Pos.
Mean RMS Mean RMS Mean RMS
Taps Peak Peak Peak Peak Peak Peak
Roof Pressure Taps


-1.47 -0.52 -0.03 0.55 -0.15 0.05
-2.42 -0.50 1.05 0.54 -1.90 -0.06
-2.90 -0.47 1.85 0.51 -2.77 -0.82
-2.62 -0.44 0.92 0.49 -1.94 -0.56
-2.05 -0.49 0.08 0.52 -1.54 0.07
-1.85 -0.50 0.22 0.53 -0.04 0.12
-2.48 -0.45 0.71 0.48 -0.15 0.13
-1.86 -0.41 1.48 0.44 -0.63 0.11
-0.98 -0.40 -0.01 0.41 -3.30 -0.57
-1.31 -0.42 0.31 0.44 -0.42 0.03
-2.32 -0.45 0.43 0.47 -0.14 0.05
-2.05 -0.46 0.47 0.48 -0.15 0.03
-0.79 -0.33 0.20 0.34 -3.62 -0.82
-1.02 -0.37 -0.06 0.38 -1.38 -0.06
-1.39 -0.41 0.11 0.42 -0.47 -0.08
-1.52 -0.46 0.21 0.48 -0.32 -0.09
-1.03 -0.32 0.01 0.33 -3.18 -1.12
-1.27 -0.36 -0.05 0.37 -2.53 -0.43
-2.21 -0.45 0.03 0.47 -0.75 -0.38
-1.54 -0.40 -0.09 0.41 -1.28 -0.34
-0.92 -0.34 0.02 0.35 -6.61 -1.71
-1.20 -0.36 -0.04 0.37 -3.70 -0.67
-1.86 -0.40 -0.08 0.42 -2.46 -0.54
-2.02 -0.34 -0.01 0.36 -3.23 -0.89
-1.53 -0.43 0.27 0.44 -1.58 -0.26


0.22 0.07
1.55 0.34
1.29 0.90
1.73 0.60
1.24 0.20
0.26 0.12
0.53 0.14
0.81 0.15
1.91 0.68
0.39 0.06
0.24 0.06
0.24 0.05
1.00 0.88
1.04 0.16
0.18 0.10
0.11 0.10
1.08 1.17
1.64 0.50
0.04 0.39
0.42 0.36
2.91 1.86
1.57 0.80
1.04 0.59
2.05 0.93
1.19 0.33


Wall Pressure Taps
26 -0.58 -0.19 0.17 0.20 0.64 0.83 1.07 0.83
27 -1.14 -0.30 0.00 0.32 0.39 0.56 0.73 0.56
28 0.46 0.76 0.98 0.76 0.54 0.72 0.88 0.72
29 0.93 1.08 1.26 1.08 0.11 0.36 0.60 0.36
Interior Pressure Taps
30 -0.01 0.00 0.01 0.00 0.00 0.01 0.02 0.01


-0.90 -0.16
-3.50 -0.36
-3.26 -0.49
-3.00 -0.54
-1.86 0.01
-1.63 0.06
-1.92 0.01
-2.05 -0.05
-0.81 0.10
-0.76 0.10
-0.59 0.14
-0.08 1.15
-1.01 -0.11
-0.59 0.09
-0.44 0.05
-0.56 0.07
-0.41 -0.01
-0.32 0.04
-0.39 0.04
-0.29 0.04
-0.96 -0.05
-0.65 -0.03
-0.55 -0.02
-0.82 -0.04
-0.66 -0.05


-0.41 0.46
0.23 0.92
-0.93 -0.43
-0.64 0.04


0.00 0.01


0.44 0.26
1.66 0.50
1.61 0.57
1.31 0.60
1.51 0.21
1.47 0.25
1.51 0.25
2.17 0.24
1.09 0.16
0.71 0.12
0.70 0.16
1.98 1.15
0.46 0.15
0.47 0.11
0.42 0.07
0.71 0.09
0.77 0.07
0.47 0.07
0.36 0.07
0.40 0.06
0.36 0.08
0.21 0.06
0.21 0.05
0.24 0.08
0.15 0.08


1.73 0.47
1.26 0.92
-0.14 0.44
0.57 0.11


0.02 0.01











Table C-2. Measured peak, mean and RMS pressure coefficients for test repeat 2


0 450 90
Press. Neg. Pos. Neg. Pos. Neg. Pos.
Mean RMS Mean RMS Mean RMS
Taps Peak Peak Peak Peak Peak Peak
Roof Pressure Taps


1 -1.66 -0.53 -0.06 0.57 -0.16 0.03
2 -2.43 -0.49 1.08 0.53 -1.98 -0.05
3 -2.92 -0.46 1.82 0.51 -2.64 -0.80
4 -2.63 -0.44 1.40 0.48 -2.30 -0.56
5 -2.46 -0.50 0.34 0.53 -1.66 0.05
6 -1.96 -0.50 0.18 0.54 -0.06 0.10
7 -4.54 -0.45 0.80 0.48 -0.29 0.10
8 -2.80 -0.41 1.08 0.44 -0.67 0.09
9 -1.13 -0.40 -0.02 0.42 -3.27 -0.55
10 -1.19 -0.43 -0.11 0.45 -0.36 0.01
11 -1.84 -0.45 -0.05 0.47 -0.20 0.03
12 -1.51 -0.46 0.34 0.48 -0.20 0.01
13 -0.76 -0.34 0.00 0.35 -2.97 -0.80
14 -0.98 -0.38 -0.07 0.39 -1.16 -0.08
15 -1.36 -0.41 0.08 0.43 -0.47 -0.10
16 -1.52 -0.46 0.38 0.48 -0.33 -0.10
17 -0.75 -0.35 -0.04 0.36 -3.47 -1.06
18 -1.05 -0.35 -0.06 0.36 -2.10 -0.42
19 -1.97 -0.44 0.42 0.46 -0.75 -0.39
20 -1.89 -0.39 1.14 0.41 -1.04 -0.35
21 -0.95 -0.33 -0.04 0.34 -5.94 -1.69
22 -0.91 -0.36 -0.03 0.36 -3.67 -0.64
23 -1.59 -0.39 0.02 0.41 -2.40 -0.54
24 -0.78 -0.33 0.06 0.34 -3.00 -0.91
25 -1.60 -0.41 -0.08 0.42 -1.76 -0.23


0.20 0.05 -0.99 -0.12 0.56 0.23
1.58 0.30 -2.35 -0.29 1.69 0.44
1.11 0.87 -2.55 -0.40 1.96 0.49
0.56 0.60 -2.65 -0.48 1.60 0.55
1.56 0.20 -1.51 0.02 1.55 0.18
0.26 0.10 -1.54 0.07 1.67 0.22
0.35 0.11 -1.35 0.01 1.39 0.23
0.81 0.12 -1.85 -0.02 1.85 0.21
2.02 0.67 -0.75 0.10 0.94 0.15
0.36 0.06 -0.74 0.10 0.81 0.12
0.28 0.05 -0.56 0.14 0.69 0.15
0.21 0.04 -1.49 -0.20 0.73 0.47
1.88 0.87 -0.92 -0.10 0.50 0.13
1.16 0.17 -0.37 0.09 0.62 0.11
0.20 0.12 -0.34 0.06 0.34 0.08
0.13 0.12 -0.43 0.08 0.57 0.09
0.67 1.11 -0.50 -0.02 0.43 0.07
1.12 0.49 -0.36 0.03 0.35 0.06
0.19 0.40 -0.35 0.04 0.31 0.06
0.25 0.36 -0.39 0.05 0.49 0.07
2.75 1.85 -0.73 -0.04 0.23 0.08
1.71 0.76 -1.54 -0.03 0.36 0.06
1.10 0.59 -0.67 -0.01 0.20 0.06
0.98 0.95 -0.99 -0.05 0.24 0.09
1.44 0.30 -0.97 -0.04 0.18 0.08


Wall Pressure Taps
26 -0.53 -0.18 0.18 0.20 0.59 0.77 1.10 0.77 -0.41 0.40 1.89 0.42
27 -1.08 -0.30 0.18 0.32 0.40 0.52 0.74 0.52 0.27 0.89 1.30 0.90
28 0.41 0.74 0.99 0.75 0.55 0.72 0.91 0.72 -1.01 -0.38 0.00 0.39
29 0.71 1.04 1.23 1.05 0.09 0.33 0.56 0.34 -0.53 0.06 0.67 0.12
Interior Pressure Taps
30 -0.01 0.00 0.01 0.00 -0.01 0.00 0.01 0.00 -0.08 0.00 0.02 0.01











Table C-3. Measured peak, mean and RMS pressure coefficients for test repeat 3


0 450 90
Press. Neg. Pos. Neg. Pos. Neg. Pos.
Mean RMS Mean RMS Mean RMS
Taps Peak Peak Peak Peak Peak Peak
Roof Pressure Taps


1 -1.63 -0.62 -0.09 0.66 -0.14 0.05 0.21 0.06 -0.90 -0.13
2 -3.24 -0.55 1.12 0.60 -2.03 -0.03 1.55 0.28 -3.10 -0.33
3 -2.41 -0.49 1.62 0.54 -2.68 -0.78 1.32 0.86 -3.23 -0.46
4 -2.95 -0.46 1.74 0.52 -2.70 -0.55 0.65 0.60 -3.06 -0.52
5 -2.28 -0.59 -0.03 0.62 -1.73 0.05 1.36 0.20 -1.87 0.02
6 -1.93 -0.59 0.36 0.63 -0.05 0.11 0.27 0.11 -1.87 0.09
7 -1.84 -0.51 0.89 0.54 -0.10 0.12 0.36 0.12 -1.87 0.05
8 -2.58 -0.43 1.24 0.47 -0.78 0.10 0.71 0.13 -1.54 0.00
9 -1.35 -0.45 0.18 0.47 -2.86 -0.54 1.91 0.66 -1.01 0.11
10 -1.41 -0.48 -0.05 0.50 -0.44 0.01 0.46 0.06 -0.89 0.11
11 -2.53 -0.50 0.10 0.52 -0.18 0.04 0.30 0.06 -0.55 0.16
12 -2.03 -0.50 0.81 0.52 -0.20 0.01 0.21 0.04 -1.02 0.18
13 -1.13 -0.36 0.06 0.38 -2.89 -0.79 1.34 0.86 -1.13 -0.11
14 -1.05 -0.41 -0.04 0.42 -1.21 -0.07 0.89 0.17 -0.61 0.10
15 -1.33 -0.45 -0.03 0.46 -0.64 -0.10 0.28 0.12 -0.54 0.06
16 -1.75 -0.50 0.21 0.52 -0.33 -0.10 0.12 0.11 -0.38 0.08
17 -0.77 -0.35 0.24 0.36 -3.18 -1.06 0.71 1.12 -0.50 -0.01
18 -0.82 -0.38 -0.11 0.38 -2.28 -0.41 1.14 0.48 -0.53 0.04
19 -2.05 -0.47 0.08 0.49 -0.78 -0.39 0.08 0.40 -0.42 0.04
20 -1.63 -0.41 -0.04 0.43 -1.15 -0.35 0.50 0.37 -0.42 0.05
21 -0.77 -0.35 -0.02 0.36 -6.48 -1.75 1.92 1.90 -1.57 -0.05
22 -0.94 -0.37 -0.05 0.38 -3.06 -0.62 1.77 0.74 -1.22 -0.04
23 -1.23 -0.41 -0.06 0.42 -2.79 -0.54 1.06 0.60 -1.38 -0.02
24 -1.11 -0.34 0.04 0.35 -3.41 -0.95 1.07 0.99 -0.92 -0.05
25 -1.65 -0.41 0.10 0.42 -1.57 -0.21 1.17 0.27 -1.04 -0.05
Wall Pressure Taps
26 -0.77 -0.19 0.11 0.20 0.57 0.77 0.96 0.77 -0.48 0.48
27 -1.41 -0.31 0.39 0.33 0.38 0.52 0.68 0.52 0.50 0.95
28 0.49 0.82 1.06 0.82 0.56 0.74 0.90 0.74 -1.11 -0.42
29 0.93 1.08 1.27 1.08 0.14 0.34 0.53 0.35 -0.64 0.06
Interior Pressure Taps
30 -0.01 0.00 0.01 0.00 -0.01 0.00 0.01 0.00 -0.01 0.00


0.40 0.25
2.03 0.51
2.26 0.56
2.07 0.59
1.32 0.21
1.54 0.26
1.64 0.26
1.97 0.23
1.05 0.17
0.59 0.13
0.70 0.17
1.08 0.20
0.56 0.15
0.58 0.12
0.42 0.09
0.41 0.10
0.72 0.08
0.48 0.08
0.33 0.08
0.34 0.08
0.79 0.11
0.30 0.10
0.19 0.09
0.42 0.10
0.21 0.10


1.31 0.50
1.52 0.95
-0.07 0.43
0.72 0.13


0.01 0.00











APPENDIX D
STRUCTURAL REACTIONS MEASURED IN THE 1/3-SCALE HOUSE TEST

Table D-1. Measured peak, mean and RMS reactions for test repeat 1


0 450 90


Load Neg. Pos. Neg. Pos. Neg. Pos.
Mean RMS Mean RMS Mean RMS
Cell Peak Peak Peak Peak Peak Peak
Roof-to-wall load cells
1 -13.5 -7.7 -0.9 7.9 -6.2 -3.7 -1.3 3.8 1.4 5.0 9.1 5.1
2 -10.9 -6.9 -3.1 7.0 -6.6 -4.5 -1.7 4.6 -5.1 -1.0 3.3 1.4
3 -13.5 -8.0 -4.1 8.1 -6.8 -4.8 -2.6 4.8 -2.7 0.8 4.2 1.2
4 -26.4 -19.5 -11.1 19.6 -13.9 -11.1 -7.9 11.1 -2.9 1.1 4.8 1.4
5 -22.3 -18.2 -14.6 18.3 -11.4 -8.7 -6.6 8.7 -17.8 -9.8 -4.5 9.9
15 -20.5 -13.4 -7.4 13.6 -15.3 -10.7 -6.1 10.7 -4.2 0.8 6.6 1.7
16 -22.8 -18.3 -14.5 18.3 -19.4 -15.1 -10.9 15.2 -8.6 -4.6 -2.3 4.7
17 -11.4 -5.8 -2.8 5.9 -5.1 -2.0 -1.5 2.1 -3.6 -1.4 0.0 1.4
18 -11.8 -7.1 -4.1 7.2 -12.2 -8.1 -4.2 8.2 -7.0 -1.4 1.9 1.8
19 -21.2 -13.6 -8.7 13.8 -17.2 -12.0 -5.8 12.1 -16.0 -7.9 -1.0 8.2
20 -20.8 -14.3 -10.1 14.4 -17.4 -13.1 -6.7 13.2 3.9 8.3 13.4 8.4
21 -10.3 -6.3 -2.8 6.4 -5.9 -2.8 -0.7 2.9 -7.4 -1.0 1.7 1.4
Wall-to-foundation load cells
6 -7.3 -2.6 1.5 2.9 -12.5 -8.0 -4.1 8.1 -9.5 -2.2 5.4 2.8
7 -17.3 -11.7 -5.6 11.8 -4.1 -0.3 3.8 1.2 4.5 12.5 20.9 12.7
8 -23.3 -18.3 -11.9 18.3 -3.8 0.3 4.3 1.2 5.1 14.1 21.6 14.3
9 -24.2 -19.7 -15.1 19.8 -9.6 -6.8 -4.0 6.9 -6.8 1.4 9.1 2.6
10 -24.7 -20.8 -15.2 20.9 -15.1 -12.4 -9.7 12.4 -29.3 -18.6 -10.0 18.8
11 -20.6 -15.1 -8.8 15.3 -16.5 -12.2 -7.6 12.3 -37.3 -24.6 -13.7 24.8
12 -21.7 -16.6 -11.2 16.6 -16.0 -12.6 -8.8 12.7 -50.3 -31.8 -19.1 32.0
13 -11.2 -7.2 -3.7 7.3 -8.3 -6.5 -4.6 6.6 -20.1 -12.6 -7.1 12.7
14 3.1 5.8 8.1 5.8 -1.2 0.4 2.0 0.6 -6.7 -2.6 1.5 2.8











Table D-2. Measured peak, mean and RMS reactions for test repeat 2


0 450 90


Load
Cell


Pos.
Mean RMS
Peak


Neg. Pos. Neg. Pos. Neg.
Peak Mean Peak Peak Mean Peak Peak
Roof-to-wall load cells
-15.0 -8.3 1.8 8.6 -7.1 -3.7 -0.4 3.8 -0.4
-10.5 -7.3 -2.0 7.4 -9.3 -6.5 -3.1 6.5 -3.3
-12.9 -8.1 -4.6 8.2 -7.7 -5.3 -2.1 5.3 -1.5
-25.5 -18.4 -10.1 18.6 -17.0 -13.3 -9.3 13.4 -2.3
-19.6 -14.9 -9.6 15.0 -14.5 -10.5 -7.5 10.5 -13.5
-23.2 -14.7 -8.5 15.0 -18.9 -13.1 -7.1 13.2 -4.0
-23.3 -16.1 -8.2 16.2 -21.3 -16.2 -10.9 16.3 -6.2
-11.7 -5.7 -2.2 5.8 -6.3 -2.4 -1.4 2.6 -3.6
-11.2 -7.4 -3.5 7.4 -13.9 -9.7 -4.8 9.8 -6.9
-19.6 -13.2 -8.2 13.3 -19.8 -13.6 -6.9 13.7 -15.4

-21.2 -15.0 -9.0 15.1 -21.9 -15.3 -7.5 15.4 4.2
-11.6 -6.4 -1.7 6.5 -8.0 -4.2 -0.8 4.3 -6.7
Wall-to-foundation load cells
-8.0 -3.5 1.4 3.7 -15.3 -10.7 -5.3 10.7 -5.9
-17.8 -11.9 -4.4 12.0 -6.0 -1.3 4.7 2.0 6.0
-23.9 -17.7 -9.3 17.8 -5.6 -0.7 4.8 1.6 6.8
-22.6 -17.8 -11.4 17.8 -11.8 -7.7 -2.8 7.8 -3.4
-21.5 -17.8 -10.8 17.9 -17.9 -14.4 -11.3 14.4 -23.9
-19.0 -14.0 -7.8 14.1 -20.9 -15.7 -11.0 15.8 -28.7
-16.6 -12.6 -9.0 12.7 -19.5 -13.8 -9.0 13.9 -43.2
-10.0 -7.2 -4.9 7.2 -10.2 -7.5 -2.4 7.5 -18.9
2.8 4.6 6.1 4.7 -2.5 -0.4 3.9 0.8 -6.4


3.1
0.6
1.6
1.6
-8.1
2.5
-1.8
-0.9
-0.6
-8.2

9.6
-0.5


0.7
13.0
14.9
2.7
-16.0
-18.8
-28.3
-11.2
-1.6


6.2
4.3
4.7
6.1
-4.2
7.9
0.6
0.4
2.3
-2.6

14.7
2.0


7.4
20.3
22.7
10.4
-9.5
-9.2
-13.9
-4.9
2.2











Table D-3. Measured peak, mean and RMS reactions for test repeat 3


0 450 90


Load
Cell


Neg. Pos. Neg. Pos. Neg. Pos.
Peak Mean Peak Peak Mean Peak Peak Mean Peak
Roof-to-wall load cells
-14.6 -8.0 -0.2 8.3 -7.9 -4.2 -1.2 4.3 -2.4 2.9 5.4 3.1
-10.0 -7.0 -1.9 7.1 -9.4 -6.6 -3.6 6.6 -5.1 0.7 4.6 1.4
-13.9 -8.1 -3.3 8.2 -8.0 -5.3 -2.7 5.4 -2.9 1.4 4.6 1.7
-25.8 -17.9 -9.0 18.0 -16.7 -13.2 -9.0 13.2 -4.7 1.4 5.6 1.9
-18.6 -14.1 -10.0 14.2 -13.8 -10.4 -7.7 10.5 -22.4 -9.1 -4.6 9.3
-26.5 -17.8 -10.9 17.9 -18.2 -12.5 -6.3 12.6 -4.2 2.3 7.3 2.8
-18.6 -14.4 -10.1 14.5 -21.2 -16.0 -11.4 16.1 -10.7 -2.6 0.2 2.8
-9.3 -5.1 -2.0 5.2 -6.6 -2.4 -1.5 2.5 -4.7 -1.1 0.5 1.3
-10.8 -6.9 -4.0 7.0 -15.1 -9.9 -5.4 10.0 -8.5 -1.0 2.6 1.8
-21.9 -14.9 -8.2 15.0 -19.6 -13.5 -6.8 13.6 -23.1 -9.5 -3.2 9.8
-18.7 -13.4 -8.0 13.5 -20.6 -14.3 -7.0 14.5 6.2 9.9 13.2 9.9
-12.1 -6.7 -3.0 6.8 -8.8 -4.2 -1.1 4.3 -9.2 -0.5 2.2 1.5
Wall-to-foundation load cells
-8.4 -4.1 -0.6 4.2 -16.1 -11.1 -6.8 11.2 -8.9 0.7 7.8 2.2
-18.3 -11.8 -5.8 11.9 -6.3 -1.8 2.7 2.2 3.5 12.8 19.8 12.9
-23.5 -17.1 -10.9 17.2 -6.0 -1.0 3.4 1.7 5.3 15.1 21.8 15.2
-20.8 -16.8 -11.8 16.8 -11.4 -7.4 -3.3 7.5 -8.6 2.8 8.9 3.5
-19.3 -15.6 -11.4 15.6 -17.4 -14.3 -10.7 14.4 -30.8 -17.6 -10.4 17.7
-17.1 -12.0 -7.1 12.1 -20.7 -15.6 -10.3 15.7 -34.2 -19.8 -11.3 20.0
-19.0 -14.6 -10.2 14.7 -18.0 -14.1 -10.3 14.2 -52.3 -29.5 -18.1 29.7
-12.5 -10.1 -7.7 10.1 -10.7 -8.2 -5.9 8.2 -21.4 -11.3 -6.1 11.4
1.3 3.5 5.8 3.6 -2.3 -0.5 1.0 0.7 -6.0 -1.3 2.5 1.8









LIST OF REFERENCES


ASCE/SEI. (2005). "Minimum design loads for buildings and other structures." 7-05, American
Society of Civil Engineers, Reston, VA.

Chen, J., Haynes, B. S., and Fletcher, D. F. (2000). "Cobra probe measurements of mean
velocities, Reynolds stresses and higher-order velocity correlations in pipe flow."
Experimental Thermal and Fluid Science, 21(4), 206-217.

Datin, P. L., and Prevatt, D. O. (2007). 'Wind Uplift Reactions at Roof-to-Wall Connections of
Wood-Framed Gable Roof Assembly." 12th International Conference on Wind
Engineering Australasian Wind Engineering Society, Cairns, Australia.

Datin, P. L., Prevatt, D. O., and Mensah, A. "Performance Based Wind Engineering: Interaction
of Hurricanes with Residential Structures." Engineering Research and Innovation
Conference, Honolulu, Hawaii.

Davenport, A. G., Surry, D., and Stathopoulos, T. (1978). "Wind loads on low-rise buildings."
Universtiy of Ontario, London, Ontario, Canada.

Doudak, G., McClure, G., Smith, I., Hu, L., and Stathopoulos, T. (2005). "Monitoring Structural
Response of a Wooden Light-Frame Industrial Shed Building to Environmental Loads."
Journal of Structural Engineering, 131(5), 794-805.

FEMA. (2005). "Mitigation Assessment Team Report: Hurricane Ivan in Alabama and Florida -
Observations, Recommendations, and Technical Guidance." FEMA 489, Federal
Emergency Management Agency.

Ginger, J. D., and Letchford, C. W. (1993). "Characteristics of large pressures in regions of flow
separation." Journal of Wind Engineering and Industrial Aerodynamics, 49, 301-310.

Ginger, J. D., Reardon, G. F., and Whitbread, B. J. (2000). "Wind load effects and equivalent
pressures on low-rise house roofs." Engineering Structures, 22(6), 638-646.

Hibbeler, R. C. (2006). StructuralAnalysis, Pearson Prentice Hall, New Jersey.

Ho, T. C. E., Surry, D., Morrish, D., and Kopp, G. A. (2005a). "The UWO contribution to the
NIST aerodynamic database for wind loads on low buildings: Part 1. Archiving format
and basic aerodynamic data." Journal of Wind Engineering and Industrial Aerodynamics,
93(1), 1-30.

Ho, T. C. E., Surry, D., Morrish, D., and Kopp, G. A. (2005b). "The UWO contribution to the
NIST aerodynamic database for wind loads on low buildings: Part 1. Archiving format
and basic aerodynamic data." 93(1), 1-30.

Hooper, J. D., and Musgrove, A. R. (1997). "Reynolds stress, mean velocity, and dynamic static
pressure measurement by a four-hole pressure probe." Experimental Thermal and Fluid
Science, 15(4), 375-383.









Jang S., Lu, L.-W., Sadek, F., and Simiu, E. (2002). "Database-assisted wind load capacity
estimates for low-rise steel frames. "Journal of Structural Engineering, 128(12), 1594-
1603.

Kopp, G. A., and Chen, Y. (2006). "Database-assisted design of low-rise buildings:
Aerodynamic considerations for a practical interpolation scheme." Journal of Structural
Engineering, 132(6), 909-917.

Lieblein, J. (1974). "Efficient Methods fo Extreme-Value Methodology." National Bureau of
Standards, Washington, D.C.

Liu, Z., Prevatt, D. O., Gurley, K. R., Aponte-Bermudez, L., and Reinhold, T. A (2009). "Field
Measurement and Wind Tunnel Simulation of Hurricane Wind Loads on a Single Family
Dwelling." Engineering Structures, (in press).

Main, J. A., and Fritz, W. P. (2006). "Database-Assisted Design for Wind: Concepts, Software,
and Examples for Rigid and Flexible Buildings." National Institute of Science and
Technology, Gaithersburg MD.

Martin, K. G., Gupta, R., Prevatt, D. O., Datin, P. L., and van de Lindt, J. W. (2010). "Evaluation
of System Effects and Structural Load Paths in a Wood-Framed Structure." Journal of
Architectural Engineering, Submitted for review 25 Feb. 2010.

Masters, F., Gurley, K., and Prevatt, D. 0. (2008). 'Full-Scale Simulation of Turbulent Wind-
Driven Rain Effects on Fenestration and Wall Systems." 3rd International Symposium on
Wind Effects on Buildings and Urban Environment, Tokyo, Japan.

National Institute of Standards and Technology, N. (2008). "windPressure-DAD software for
rigid, gable roof buildings."
http://www.itl. nist. gov/div898 /winds/windp ressure/windp pressure. htm, Date accessed:
12/20/09.

NIST. (2003). "NIST/SEMATECH e-Handbook of Statistical Methods." National Institute of
Standards and Technology. Retrieved March 8, 2007, from
http://www. itl. nist. gov/div898/handbook/.

Rigato, A., Chang P., and Simiu, E. (2001). "Database-assisted design, standardization, and
wind direction effects. "Journal of Structural Engineering, 127(8), 855-860.

Rosowsky, D., and Schiff, S. (2003). "What are our expectations, objectives, and performance
requirements for wood structures in high wind regions?" Natural Hazards Review, 4(3),
144-148.

Rosowsky, D. V., Walsh, T. G., and Crandell, J. H. (2003). "Reliability of residential woodframe
construction from 1900 to present." Forest Products Journal, 53(4), 19-28.









Sadek, F., Diniz, S., Kasperski, M., Gioffre, M., and Simiu, E. (2004). "Sampling Errors in the
Estimation of Peak Wind-Induced Internal Forces in Low-Rise Structures." Journal of
Engineering Mechanics, 130(2), 235-239.

Sadek, F., and Simiu, E. (2002). "Peak Non-Gaussian Wind Effects for Database-Assisted Low-
Rise Building Design. Journal of Engineering Mechanics, 128(5), 530-539.

Shepherd, I. C. (1981). "Four-Hole Pressure Probe for Fluid Flow Measurements in Three
Dimensions." Journal of Fluids Engineering, Transactions of the ASM E, 103(4), 590-
594.

Simiu, E., and Miyata, T. (2006). Design of Buildings and Bridges for Wind, A Practical Guide
for ASCE-7 Standard Users and Designers of Special Structures, John Wiley & Sons,
Inc., New Jersey.

Simiu, E., Sadek, F., Whalen, T. M., Jang, S., Lu, L.-W., Diniz, S. M. C., Grazini, A., and Riley,
M. A. (2003). "Achieving safer and more economical buildings through database-
assisted, reliability-based design for wind." Journal of Wind Engineering and Industrial
Aerodynamics, 91(12-15), 1587-1611.

Simiu, E., and Scanlan, R. H. (1996). Wind Effects on Structures Fundamentals and
Applications to Design, John Wiley & Sons, Inc., New York.

Simiu, E., and Stathopoulos, T. (1997). "Codification of wind loads on buildings using bluff
body aerodynamics and climatological data bases." Journal of WindEngineering and
Industrial Aerodynamics, 69-71, 497-506.

St. Pierre, L. M., Kopp, G. A., Surry, D., and Ho, T. C. E. (2005). "The UWO contribution to the
NIST aerodynamic database for wind loads on low buildings: Part 2. Comparison of data
with wind load provisions." Journal of Wind Engineering and Industrial Aerodynamics,
93(1), 31-59.

Stathopoulos, T (1979). "Turbulent Wind Action on Low-rise Buildings," Universtiy of Ontario,
London, Ontario, London.

Suresh Kumar, K., and Stathopoulos, T. (2000). 'Wind loads on low building roofs: a stochastic
perspective." Journal of Structural Engineering, 126(8), 944-956.

van de Lindt, J. W., Graettinger, A., Gupta, R., Skaggs, T., Pryor, S., and Fridley, K. J. (2007).
"Performance of Wood-Frame Structures during Hurricane Katrina." Journal of
Performance of Constructed Facilities, 21(2), 108-116.

Watkins, S., Mousley, P., and Vino, G. "The Development and Use of Dynamic Pressure Probes
With Extended Cones of Acceptance (ECA)." 15th Australasian Fluid Mechanics
Conference, The University of Sydney, Sydney, Australia.

Whalen, T., Simiu, E., Harris, G., Lin, J., and David, S. (1998). "The use of aerodynamic
databases for the effective estimation of wind effects in main wind-force resisting









systems:: application to low buildings." Journal of WindEngineering andIndustrial
Aerodynamics, 77-78, 685-693.

Whalen, T. M., Sadek, F., and Simiu, E. (2002). "Database-assisted design for wind: Basic
concepts and software development." Journal of Wind Engineering and Industrial
Aerodynamics, 90(11), 1349-1368.

Whalen, T. M., Shah, V., and Yang, J.-S. (2000). "APilot Project For Computer-Based Design
of Low-Rise Buildings for Wind Loads The WiLDE-LRS User's Manual." Purdue
University, West Lafayette, IN.

Zisis, I., and Stathopoulos, T. (2009). 'Wind-Induced Cladding and Structural Loads on Low-
Wood Building." Journal of Structural Engineering, 135(4), 437-447.









BIOGRAPHICAL SKETCH

Akwasi Frimpong Mensah was born and raised in Takoradi, Ghana. He graduated with a

Bachelor of Science degree in civil engineering from the Kwame Nkrumah University of Science

and Technology in May 2006. He served as a teaching assistant for a year with the same

institution after graduating and later on worked as a Civil/ Structural Engineer with Comptran

Engineering and Planning Associates, Accra for another year. He joined University of Florida in

pursuit of a master degree in August 2008. He anticipates receiving a degree of Master of

Science in Civil Engineering in August 2010. The author hopes to practice as an engineer and

also lecture in the discipline.





PAGE 1

1 DETERMINATION OF WIND UPLIFT FORCES USING DATABASE ASSISTED DESIGN (DAD) APP ROACH FOR LIGHT FRAMED WOOD STRUCTURES By AKWASI FRIMPONG MENSAH A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIA L FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Akwasi Frimpong Mensah

PAGE 3

3 To my parents Mr. & Mrs. P. K. Mensah

PAGE 4

4 ACKNOWLEDGMENTS The completion of thi s research and report would not have been successful without the support and encouragement of a number of persons. I want to first thank the Almighty God through Jesus Christ for being my Lord and sufficiency. I also wish to express my sincere gratitude t o my thesis committee chair, Dr. David O. Prevatt for his continual support, time and guidance in this endeavor. I am grateful to my committee members: Dr. Kurtis R. Gurley and Dr. Forrest Masters, and Dr. Gary Consolazio, who have each contributed immense ly to the success of this research. I am also grateful to the faculty and staff of the Civil and Coastal Engineering Department for their tutorage and assistance during my stay in University of Florida. I am again appreciative of the financial support prov ided by National Science Foundation (NSF) through grant #080023 Performance Based Wind Engineering (PBWE): Interaction of Hurrica nes with Residential Structures To my parents, Mr. and Mrs. P. K. Mensah, I am indebted to you for all the love, encourageme nt, care and support you have given me. God richly bless you. I also wish to thank all the friends I met in Gainesville. Of mention especially are Patrick Bekoe and Joyce Dankyi whose presence in this chapter of my live cannot be overemphasized My deep ap preciation also goes to my entire family and friends for their timely encouragements. Lastly, I wish to express my profound gratitude to all my office colleagues and the hurricane group of University of Florida especially Peter L. Datin, Jason Smith and Sc ott Bolton for their assistance and contributions.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF FIGURES .......................................................................................................................10 ABSTRACT ...................................................................................................................................13 CHAPTER 1 INTRODUCTION ..................................................................................................................15 Background and Motivation ...................................................................................................15 Objective .................................................................................................................................16 Scope of Work ........................................................................................................................17 Organization of Report ...........................................................................................................17 2 LITERATURE REVIEW .......................................................................................................19 Wind Flow over Low Rise Buildings .....................................................................................19 Current Design Provisions of ASCE 7 for Wind Loads on Low Rise Buildings ...................20 Background on ASCE 7 Wind Load Provisions .............................................................20 Analytical Procedure for Wind Design Loads on a Low Rise Bui lding .........................21 Limitations of Current Design Provisions .......................................................................23 Database Assisted Design (DAD) Methodology for a Low Rise Building ...........................25 Background of the DAD methodology............................................................................25 DAD Concept and Software Development .....................................................................26 Limitation to the Application of DAD Approach ............................................................27 Design and Construction of Light Framed Wood Structures and their Performance to Wind Forces ........................................................................................................................27 Construction Methods .....................................................................................................27 Critical Components and Systems ...................................................................................28 Structural Failures of LFWS in Hurricane Eve nts ..........................................................29 Wind Induced Pressures and Structural Responses on Light Wood Framed Structures ........29 3 ANAYSIS OF WIND TUNNEL DATA TO GENER ATE PRESSURE COEFFICIENTS .....................................................................................................................36 Wind Tunnel Data ...................................................................................................................36 House Model and Pressure Tap Layout ...........................................................................36 Wind Simulation and Pressure Measurements ................................................................36 Aerodynamic Data Processing ................................................................................................37 Tubing Response Correction ...........................................................................................37 Determining Pressure Coefficients ..................................................................................38 Re referencing of Pressure Coefficients ..........................................................................38 Wind Tunnel Results and Analysis .........................................................................................40

PAGE 6

6 Wind Pressure Coefficients Time Histories ....................................................................40 Observed Statistical Values o f Wind Pressure Coefficients ............................................41 Extreme Value Analysis of Pressure Coefficients ...........................................................42 Area Averaged Pressure Coefficients .............................................................................44 4 APPLICATION OF DAD METHODOLGY .........................................................................57 Structural Influence Function .................................................................................................57 Ev aluating Vertical Reactions Based on DAD methodology .................................................59 Velocity Pressure .............................................................................................................59 Pressure Taps and Influence Functions ...........................................................................59 Reaction Loads ................................................................................................................60 DAD Results and Analysis .....................................................................................................60 Observed Statistical Values of Structural Reactions .......................................................60 Extreme Value Analysis of Vertical Loads .....................................................................61 Vertical Reaction Based on ASCE 7 05 Standard ..................................................................62 Velocity Pressure .............................................................................................................63 ASCE 7 based Design Loads ..........................................................................................63 Comparing Uplifts Reac tions Predicted Based on DAD vs. ASCE 7 05 ...............................64 5 WIND FLOW CHARACTERIZATION USING TFI COBRA PROBE ...............................78 The Cobra Probe .....................................................................................................................78 Preliminary Experiments Using the Cobra Probe ...................................................................79 Comparing Wind Flow Measurements by the Cobra Probe and Hotwire Anemometer .................................................................................................................79 Wind Tunnel Model ........................................................................................................80 Mapping of Wind Field Generated by UF Wind Generator ...................................................80 6 WIND INDUCED PRESSURE AND STRUCTURAL LOAD MEASUREMENTS ...........89 Materials and Methods ...........................................................................................................89 Scale House Model ..........................................................................................................89 Pressure and Load Sensors on the Building ....................................................................89 Test Arrangement ............................................................................................................90 Wind Generation .............................................................................................................91 Experimental Procedure and Measurements ...................................................................91 Experimental Results, Analysis and Discussion .....................................................................92 Wind Pressure Measurements .........................................................................................92 Wind Induced Structural Loads ......................................................................................93 Structural Load Comparison............................................................................................94 7 CONCLUSION AND RECOMMENDATIONS .................................................................111 Summary ........................................................................................................................111 Conclusions ...................................................................................................................112 Recommendation ...........................................................................................................113

PAGE 7

7 APPENDIX A MEAN, RMS AND EXTREME VALUES OF WIND TUNNEL PRESSURE COEFFICIENTS ...................................................................................................................114 B MEASURED STATISTICAL VALUES OF VERTICAL REACTIONS DERIVED FROM WIND TUNNEL DATA ..........................................................................................116 C PRESSURE COEFFICIENTS MEASURED IN THE 1/3 SCALE HOUSE TEST ............124 D STRUCTURAL REACTIONS MEASURED IN THE 1/3 SCALE HOUSE TEST ...........127 LIST OF REFERENCES .............................................................................................................130 BIOGRAPHICAL SKETCH .......................................................................................................134

PAGE 8

8 LIST OF TABLES Table page 21 Comparison of bending moments (KNm) determined using ASCE 798 and DAD (Simiu et al. 2003) ..............................................................................................................33 31 Wind tunnel study configuration and parameters ..............................................................46 32 Lieblein BLUE estimators .................................................................................................46 33 Comparison of wind tunnel and ASCE 705 peak pressure coefficients ...........................46 34 Measured peak, mean and RMS reactions for sample 1 ....................................................66 42 Averaged mean and RMS reactions ...................................................................................67 43 Parameters for Type I Extreme Value Distribution of peak negative reactions ................68 44 Expected peak negative reactions and standard deviation estimated from BLUE fitted probability distribution .......................................................................................................69 45 Pressures based on MWFRS for different b uilding surface ...............................................69 46 Pressures based on C&C for different zones .....................................................................69 47 Comparison of uplift reaction estimates based on DAD an d ASCE 7 05 .........................70 51 Comparison of flow measurements by Cobra Probe and hot wire anemometer ...............82 61 Manufactures and specifications of pressure sensors ........................................................97 62 Statistical values of measured pressure coefficients ..........................................................98 63 Coefficients of BLUE for Type 1 Extreme Value Distribution (Lieblein 1974) ...............99 64 Mean, RMS and BLUE estimated peak values of measured structural loads (lbs) ...........99 65 Cor relation coefficients measured between time histories of measured and estimated reactions ...........................................................................................................................100 66 Comparison of directly measured and DAD based estimated reactions .........................100 A 1 Peak, mean and RMS pressure coefficients of selected pressure taps .............................114 B 1 Measured peak, mean and RMS reactions (lbs) for sample 2 ..........................................117 B 2 Measured peak, mean and RMS reactions (lbs) for sample 3 ..........................................118 B 3 Measured peak, mean and RMS reactions (lbs) for sample 4 ..........................................119

PAGE 9

9 B 4 Measured peak, mean and RMS reactions (lbs) for sample 5 ..........................................120 B 5 Measured peak, mean and RMS reactions (lbs) for sample 6 ..........................................121 B 6 Measured peak, mean and RMS reactions (lbs) for sample 7 ..........................................122 B 7 Measured peak, mean and RMS reactions (lbs) for sample 8 ..........................................123 C 1 Measured peak, mean and RMS pressure coefficients for test repeat 1 ..........................124 C 2 Measured peak, mean and RMS pressure coefficients for test repeat 2 ..........................125 C3 Measured peak, mean and RMS pressure coefficients for test repeat 3 ..........................126 D 1 Measured peak, mean and RMS rea ctions for test repeat 1 .............................................127 D 2 Measured peak, mean and RMS reactions for test repeat 2 .............................................128 D 3 Measured peak, mean and RMS rea ctions for test repeat 3 .............................................129

PAGE 10

10 LIST OF FIGURES Figure page 21 Separation and reattachment pattern of wind flow over a low rise building .....................33 22 Typical building surfaces for ASCE 7 05 MWFRS external pressure coefficients ..........34 23 ASCE 7 05 provision for determining external pres sure coefficients for the design of components and cladding. ..................................................................................................34 24 Isometric view of the steel portal frame structure ..............................................................35 25 Co mparison of vertical reaction records measured by load cells and estimated based on envelope roof pressures .................................................................................................35 31 1:50 Scale house model used in the wind tunnel study .....................................................47 32 Test section arrangement for 1:50 suburban terrain...........................................................47 33 Wind flow characteristics for 1:50 suburban wind tunnel study. ......................................48 34 Frequency response characteristics of the pressure tubing system ....................................48 35 Typical wind pressure coefficient time histories ...............................................................49 36 Format of MATLAB files of pressure coefficient data .....................................................49 37 Spatial distributions of mean wind pressure coefficients ..................................................50 38 Spatial distributions of RMS of pressure coefficients .......................................................51 39 Spatial distributions of expected negative peak wind pressure coefficients ......................52 310 Spatial distributions of expected positive peak wind pressure coefficients .......................53 311 Area averaged pressure coefficients for regions cor r esponding to zone 1. .......................54 312 Area averaged pressure coefficients for regions corresponding to zone ..........................55 313 Area averaged press ure coefficients for regions corresponding to zone 3. .......................56 41 1/3 Scale house model for determining influence functions ..............................................70 42 Lo cations of load cells and wind direction ........................................................................71 43 Grid points for experimental determination of influence coefficients ...............................71 44 Infl uence lines for vertical reactions at a support of an internal truss ...............................72

PAGE 11

11 45 Typical influence surfaces for vertical reaction determined on 1/3 scale house model ....72 46 Typical wind induced reaction time historieso...................................................................73 47 Estimated mean and peak vertical reactions at load cells for wind azimuth 000o .............73 48 Estimated mean and peak vertical reactions at load cells for wind azimuth 045o .............74 49 Estimated mean and peak vertical reactions at load cell s for wind azimuth 090o .............74 410 Estimated mean and peak vertical reactions at load cells for wind azimuth 135o .............75 411 Estimated mean and peak vertical reactions at load cells for wind azimuth 180o .............75 412 Building surfaces for determining wind pressure for each truss base d on ASCE 7 provisions for MWFRS ......................................................................................................76 413 Zones for determining wind pressure for each truss based on ASCE 7 provisions for components and cladding ...................................................................................................76 414 Uplift reactions at roof to wall load cells based on DAD approach and ASCE 705 provisions ...........................................................................................................................77 51 Cobra Probe ........................................................................................................................82 52 Cobra Probe and Hot wire anemom eter setup for simultaneous measurements ...............82 53 Wind tunnel model used in pilot studies ............................................................................83 54 Measuring points for mapping flow measurements at of the wind tunnel .........................83 55 Flow measurements at exit of wind tunnel model .............................................................84 56 Spatial variations of longitudi nal velocity and turbulence intensity across exit section of wind tunnel model .........................................................................................................84 56 Spatial variations of lateral velocityand turbulence intensity across exit section of wind tunnel model ..............................................................................................................84 58 Spatial variations of lateral velocity and turbulence intensity across exit section of wind tunnel model ..............................................................................................................85 59 S pectral co ntents of wind speed at 0.5 in. downstream of the exit of the wind tunnel ......85 510 Position of traverse frame in test section ...........................................................................86 511 Location of measurements points for flow mapping .........................................................86 512 Variation of longitudinal wind speed across 2D measurement surface .............................87

PAGE 12

12 513 Variation s of lateral and vertical wind speed across 2D measurement surface .................87 514 Spectral contents of wind speed .........................................................................................88 61 Locations of pressure taps and load cells on the scale house model. ..............................100 62 A sample pressure tap and layout of pressure taps on the roof ........................................101 63 Interior of the house model ..............................................................................................101 64 Pressure Sensors A) Omega PX 138 B) Setra 265 C) Dwyer 616 ...................................102 65 Fut ek load cells A) Roof to wall connection B) Wallto foundation connection ...........102 66 F luke pressure calibrator ..................................................................................................102 67 Dwyer Transd ucer and pitot tube for wind velocity measurements ................................103 68 Layout of experimental setup .........................................................................................104 69 Test setup .........................................................................................................................105 610 UF Wind Generator ..........................................................................................................106 611 Spatial distribution s of p eak pressure coefficient ............................................................106 612 Spatial distribution s of m ean pressure coefficient ...........................................................107 613 Location of load cells on the house ..................................................................................107 614 Mean and ex pected peak values of measured structural loads for wind direction 0o ......108 615 Mean and expected peak values of measured structural loads for wind direction 45o ....108 616 Mean and expected peak values of measured structural loads for wind direction 90o ....109 617 Comparison of measured and estimated reactions for wind dir ection 0o.........................109 618 Comparison of measured and estimated reactions for wind direction 45o.......................110 619 Comparison of measured and est imated reactions for wind direction 90o.......................110

PAGE 13

13 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science D ETERMINATION OF WIND UPLIFT FORCES USING DATABASE ASSISTED DESIGN (DAD) APP ROACH FOR LIGHT FRAME D WOOD STRUCTURES By Akwasi Frimpong Mensah August 2010 Chair: David Prevatt Major: Civil Engineering During major hurricanes, damages to light frame d wood structures (LFWS) represented the largest proportion of monetary losses The absence of wind load transfer mechanism in wood structures was identified as a major cause of their structural failures Wind load paths in LFS are not well understood. T his study aims to develop a better approach for determ in ing wind design loads on LFWS. The study was part of an on going National Science Foundation (NSF) funded project titled, Performance Based Wind Engineering: Interaction of Hurricane Forces with Residential S tructures, wh ich has a primary objective of investigating the relationship between spatially varying wind loads and structural load paths on LFWS. This study was accomplished in two phases. In P hase 1, a Databa se Assisted Design (DAD) methodology was use d to combine time histories of wind tunnel pressure coefficients with experimentally determined influence functions for a wood framed structure From this analysis, structural reaction s at roof to wall and wallto foundation connections were developed Pea k reactions were compared to wind design loads based on ASCE 7 (2005) provisions for main wind force resisting systems (M WFRS) and components and cladding (C&C) Whereas,

PAGE 14

14 peak reactions estimated using DAD methodology, were higher than maximum reactions obtained using the MWFRS provisions, they were lower than C&C based maximum reactions. In P hase 2 of the project an exper imental study was conducted to validate the DAD methodology. A1/3scale LWFS instrumented, with surface pressure transducers and load cells, was the immersed in wind flow. Structural reactions were developed from measured roof pressures using the DAD metho dology. A comparison of developed reactions with directly measured reactions showed a good agreement between the ir mean and peak values

PAGE 15

15 CHAPTER 1 INTRODUCTION Background and Motivation Wind flow over low rise buildings is characterized by patterns of flo w separation and reattachment which creates spatially and temporally varying pressure fields on building surfaces. Generally, peak wind suction forces occur on the leeward walls and at roof edge areas while positive pressu re are created on the windward wal ls and interior roof areas Such forces i n recent hurricanes caused substantial damage to woodframed residential structures. Hurricane damage to l ight frame d wood s tructur es (LFWS) are by far the largest contributor to the monetary losses associated with hurricane disaster (Rosowsky et al. 2003) Post hurricane investigations report widespread structural damage to wood structures due to loss of roof sheath ing, and failure of load transfer at joints and mechanical connections (FEMA 2005; van de Lindt et al. 2007) Clearly, th e re is a lack of understanding of structural load paths in wood structures Wind design of light framed wood residential structures is problematic because of their complex geometric shapes. Current wind design provisions lack codified pressure values for t ypical residential buildings. i.e. pressure coefficients are only provided only for simple shaped building. Moreover, high variability in material properties of wood introduces greater uncertainty in wind resistance estimates. Selection of materials and co nnections for LFWS has been mainly based on prescriptive building guidelines which increase their susceptibility to wind damages. In the United S tates wind load design provisions are included in the ASCE 7 (2005) which codifies information on wind flow characteristics (obtained from meteorological data) and aerodynamic pressures (developed on scaled models in boundary laye r wind tunnels). Pressure coefficients and climatological data, which are used by structural engineers for wind load

PAGE 16

16 designs, are presented in reductive figures and tables suitable for hand calculations. However, a research conducted by Simiu & Stathopoulo s (1997) suggests that such design standards can produce risk i nconsistent results. Simiu & Stathopoulos (1997) asserted that the curren t code provides insufficient information for designers to realistically account for the spatial and temporal variation of wind load effects. To address these codification deficiencies, a new wind analysis approach, in which utilizes large aerodynamic and c limatological databases are used to define wind design loads, was proposed by Simiu & Stathopoulos (1997) This analysis methodology, called database assisted design (DAD) was used by Simiu et al. (2003) to estimate internal forces in a steel portal framed building. They us ed wind tunnel databases of surface pressure time histories and analytically derived influence functions to determine bending moments at knee and ridge joints of the portal frames. When the DAD results were compared with ASCE 7 design values, they conclude d that the ASCE provisions produced risk inconsistent designs and errors in excess of 50 % in peak load estimations. With the availability of powerful computations and proven usefulness of the DAD methodology, its application has been extended to light fra med wood structures. The hypothesis of this research is that, the DAD methodology will provide better accuracy in predicting wind load effects on LFWS than using the current codes. The validation of this hypothesis would lead to a better understanding of s tructural load paths on LFWS and improved engineering design models for LFWS. Objective T he specific objectives of this investigation are to : 1. Apply the DAD methodology to predict structural reactions in a LFWS system.

PAGE 17

17 2. Valida t e the DAD approach by experime ntally determining structural loads on a 1/3 wood building model Scope of Work Wind tunnel data developed on a 1/50 scale house model were analyzed to generate spatial distributions of wind pressure coefficients on the roof of a gable roof building. The D AD methodology was used to combine the pressure time histories and experimentally derived influence coefficients for vertical reactions on a third scale wood structure. The data was analyzed to determine peak values of roof to wall and wall to foundation c onnection loads for LFWS. Results of the analysis were compared to wind design loads based on the ASCE 705 standard. In phase two of this study, the validity of the DAD approach was evaluated experimentally by subjecting the thirdscale house to fluctuat ing wind forces while simultaneously measuring surface pressures and structural reactions. The DAD methodology was applied by utilizing the measured pressure distributions and the influence coefficients to determine reaction loads at roof to wall connectio ns. The results and directly measured structural loads were compared. Organization of Report Literature reviews of relevant topics to this project are presented in Chapter 2. The chapter discusses wind load effects on low rise buildings, the current provis ions for wind load designs and the concept and development of the DAD approach Finally, a review of previous experimental studies in which structural responses and wind pressures were simultaneously monitored on LFWS is presented. In Chapter 3, t he wind t unnel study, which produced the aerodynamic pressure data utilized in this project, is introduced. Analysis of wind tunnel derived pressures to generate a pressure coefficien t database for this study is described. Extreme value analysis based on the

PAGE 18

18 Liebli en BLUE (best linear unbiased estimators) estimation procedure to obtain the expected peak pressure distributions is explained also in Chapter 3. Lastly, area averages of pressure coefficients from wind tunnel analysis are compared to ASCE 705 external pr essure coefficients for components and cladding. The DAD based procedure for evaluating wind load reactions is described in Chapt er 4. Chapter 4 also contains an overview of experimental derivation of structural influence functions for this study as well as experimental results. Analysis of results to estimate peak reactions is discussed in this chapter Peak reactions based on the DAD approach are compared to results based on ASCE 705. In Chapter 5, the TFI Cobra Probe, which is used in the experimental study, is introduced. Tests undertaken to validate and understand the operations of this equipment are reported as well. Finally, characterization of the wind field for the experimental study is also described. In Chapter 6, the 1/3scale model house exper iment is described. The chapter contains descriptions of materials and equipments used in the tests, such as the 1/3 scale house, UF wind generator and load and pressure sensors. The test arrangement and procedures are also described. Finally, analysis of experimental results and correlation of structural loads derived from pressure measurements and directly measured structural loads are reported in this chapter. A summary of the entire project is contained in Chapter 7. The usefulness of this research is a lso discussed in this chapter. Lastly, recommendations for future work are presented.

PAGE 19

19 CHAPTER 2 LITERATURE REVIEW Wind Flow over Low Rise Buildings Wind loading on a building depends upon the flow pattern around the building, which, in turn, depends on building geometry, dimensions, surroundings, upstream terrain and wind flow characteristics. Wind flow over a low rise building is characterized by separation and reattachment pattern (shown in figure 2 1) which together with its velocity fluctuations g enerate a spatially and temporally varying pressure field on the surface (Ginger et al. 2000) Ginger and Letchford (1993) observed that l arge fluctuating suction pressures are generated in flow separation regio ns close to the leading edges of the roof of low rise buildings. They explained that the flow mechanisms that generate these pressures are the 2D separation bubble for flow perpendicular to the edge discontinuity and the 3D conical vortex for flow at oblique angles to the edge discontinuity and that the largest suction pressures are genera ted close to the leading corner f or a wind orientation of approximately 30. Ginger et al. (2000) determined wind loads on the roof of a typical low rise house for approach wind directions of 0o to 90o by carrying out a wind tunnel model study at a 1/50 geometric scale. T hey observed that the second truss from the windward gable end was subjected to the largest wind load Stathopoulos et al. (2000) conducted and presented a wind tunnel study which provided detailed extreme local and area averaged pressure coefficients for low building roofs exposed to open country upstream terrains. They observed that when the wind flow is normal to the ridgeline of a gable roof building, quasi flat roofs in the range of 0o30o create a similar flow pattern of separation, entrainment, and reattachment; a high suction prevails, especially at the

PAGE 20

20 windward edges and corners. They noted however that, if the roof angle is greater than 30o, wind flow generally strikes on the windward roof prior to separating from the windward edge or ridge which induces a positive pressure region on part of the windward slope and a negative region on the leeward slope. They concluded tha t these flow patterns and pressure distributions may vary with the wind direction, but remain comparable in respective roof slope ranges. Curre nt Design Provisions of ASCE 7 f or Wind Loads on Low Rise Buildings Background on ASCE 7 Wind Load P rovisions Th e provisions of ASCE 705 for wind loads on low buildings are largely based on wind tunnel study works conducted in the late 1970s at the boundary wind tunnel in the University of Western Ontario (UWO) (Davenport et al. 1978; Stathopoulos 1979) Researchers at UWO used an approach that consisted essentially of permitting the building model to rotate in the wind tunnel through a full 360o in increments of 45o w hile simultaneously monitoring the loading conditions on each surfaces. Both open and suburban exposure conditions were considered. Wind pressure coefficients which represent pseudo loading conditions, that when applied to a building, envelope the desire d structural actions (bending moment, shear, thrust), and the maximum induced force components to be resisted for all possible directions and exposures were developed from the studies (see C6.5.11 (ASCE/SEI. 2005) ). The current edition of the ASCE 7 standard (2005) has refined pressure and force coefficients to reflect the latest boundary layer wind tunnel and full scale research findings. This research has been however only limited to gable roof buildings, and a rational method of applying the coefficients to hip roofs based on experience, intuition and judgment has been developed and presented in ASCE 705.

PAGE 21

21 Three methods are provided in the ASCE 7 standard for determining wind design loads. These are the simplified method (method 1), the analytical procedu re (method 2) and a wind tunnel procedure (method 3). Analytical Procedure for W ind Design L oads on a Low Rise B uilding The main wind force resisting system (MWFRS) of a building consists of a structural frame or an assemblage of structural elements such a s roof trusses, crossbracing, shear walls and roof diaphragms that work together to transfer wind load action on the entire structure to the ground (ASCE/SEI. 2005) MWFRS provides support and stability for the overall structure and generally receives wind loading from more than one surface. ASCE 7 05 defines components and cladding as elements of the building envelope that do not qualify as part of the MWFRS. Cladding receives wind loads directly. Components receive wind loads directly or from cladding and transfer the load to the MWFRS. Members which are categorized as components and cladding included fasteners, purlins, girts, studs, roo f ducking, and roof trusses. In the determination of design wind loads on all buildings, a velocity pressure qz, is evaluated at a height z above the ground using the equation below: ) ft / lb ( I V K K K 00256 0 q2 2 d zt z z (2 1) where Kz is velocity pressure exposure coefficient, obtained from table 6 3 of ASCE 705, which modifies the design wind speed to account for terrain exposure condition and the height z ; Kzt is a topographic factor which accounts the wind speed up (topographic) effect; Kd is the wind directionality factor, which is 0.85 for buildings to account for reduced probability of maximum winds coming from any direction and the reduced probability of the maximum pressure coefficient occurring for any wind direction; V is the basic wind speed dete rmined from figure 6 1 in ASCE 7 05 and its value is a nominal 3 second gust wind speed in miles per hour at 33 ft

PAGE 22

22 above ground for an open exposure; and I is importance factor of the building determined from table 6 1 in ASCE 7 05 which is used to adjust the level of reliability of building or structure to be consistent with the building classifications indicated in the standard. Design wind pressures, for both MWFRS and components and cladding, are determined as the product of the velocity pressure and th e sum of internal and external pressure coefficients. The internal pressure coefficients, GCpi are provided in figure 65 of ASCE 7 05 in terms of the building enclosure classification (i.e. open, partially enclosed or enclosed building). The external pres sure coefficients are given separately for MWFRS and components and cladding for different scenarios but generally in terms of pressure zones. Pressure zones specified in the ASCE standard for both MWRS and Components and Cladding are in terms of a dimensi on denoted by a (Simiu and Miyata 2006) The dimension a is 10% of the least horizontal building dimension or 0.4h (h=mean roof height), whichever is smaller, but not less than 4% of the least horizontal building dimension or 3ft. Design wind pressures on the MWFRS of low buildings are determined by the equation be low: ) ft / lb ( ) GC ( ) GC ( q p2 pi pf h (2 2) Where: qh is velocity pressure evaluated at mean roof height using equation 21, GCpi internal pressure coefficient (obtained from Figure 65 in ASCE 7 05) and GCpf is an external pressure coefficient combined with a gust effect factor. Values for GCpf are provided in figure 610 in ASCE 705 as a function of the building roof angles. Roof overhangs are to be designed for a positive pressure on the bottom surface of windward roof overhangs corresponding to Cp = 0.8 in combination with the pressures determined from Figure 6 10. For determining eternal pressure coefficients, eight loading patterns are to be considered to design the building for all wind

PAGE 23

23 directions. The loading patterns have the walls and roofs zone d into several building surfaces which envelope wind load distributions on the building. Figure 22 shows typical load patterns in the ASCE 7 05 for wind design loads on a MWFRS of a building. Design wind pressures on component and cladding of low buildin gs are determined by the equation below: ) ft / lb ( ) GC ( ) GC ( q p2 pi p h (2 3) where GCp are the external pressure coefficients and the other terms are as defined previously. Values of GCp are selected from Figures 6 11 through 616 of the ASCE 705 based the type of roof and angle roofs. Figure 23 shows the typical pressure zones of a gable roof building and external pressure coefficient provision for roof angles between 7o and 27o. External pressure coefficients for deign of component and cladding are specified f or the wall, roof and overhang as a function of effective wind area. The effective wind area is defined by the ASCE 7 05 as the span length multiplied by an effective width that need not be less than one third the span length. It is worth noting that the r esulting induced wind load is however applied over the actual tributary area to the component been designed. Limitations of Current Design Provisions Several investigations have over the years been conducted and results compared with ASCE 705 wind load pr ovisions. Issues have been raised by researches on the standard provisions and this section discusses some of them. Simiu et al. (2003) illustrated the practical effects of simplifications inherent in the ASCE 705 provisions by evaluating moments in s teel portal frame s of a building (shown in Figure 2 4) by using ASCE 705 standard provisions on one hand and the DAD procedure (discussed later in this chapter) based on wind tunnel database on the other. Table 21 shows values obtained. Simiu

PAGE 24

24 et al. (2003) demonstrated that the use of the tables and plots in wind load design provisions can entail errors that can exceed 50% in the estimation of wind effects. Fur thermore, Whalen et al. (2002) assert that the accuracy in the definition of wind loads inherent in such tables and plots are lower than that inherent in current methods for stress computation. There is so much complexity with geometries and shapes of low r ise buildings and hence high accuracy in predicting design loads based on tables and plots cannot be achieved. Wind directionality effects on low rise buildings are accounted for in the ASCE 7 standard by a reduction factor of 0.85. Simiu et al. (2003) observed that this approach is inadequate as wind effect reductions due to direct ionality effects are less significant as the mean recurrence interval of the wind effects increase, rending the use of this factor potentially unsafe, particularly for MWFRS. Design parameters such as building geometry, building orientation, proximity of a djacent structures and, the spatial and temporal variation of wind loads are not realistically and comprehensively accounted for when a designer uses the conventional standard provisions (Simiu and Stathopoulos 1997) Recently, wind tunnel test data on generic low buildings were obtained at UWO to contribute to the National Institute of Standards and Technology (NIST) aerodynamic database (Ho et al. 2005b) St. Pierre et al. (2005) compared the NIST aerodynamic database to the historical data obtained by Stathopoulos in the late 1970s, from which the current ASCE 7 provisions were developed. They observed that for the exterior bay of the test buildi ng model, ASCE 7 generally underestimates the response coefficients significantly. For the interior bays, the ASCE 7 overestimates the response coefficients. They also observed that generally, there was 1085% underestimation of peak response coefficients in the suburban terrain by ASCE 7.

PAGE 25

25 Attempts by writers of the standard provisions on wind loads to reduce the limitations of earlier versions of the standard (example ASCE 7 05) resulted in bulky and complex provisions (Simiu and Stathopoulos 1997) Database Assisted Design (DAD) Methodology f or a Low Rise Building Background of the DAD methodology With the backdrop of the above mentioned limitations of the wind design load provisions, it was necessary to wor k on an alternative approach which offers the potential for significantly more risk consistent, realistic, safer and economical design by using adequate aerodynamic database s and information. Owning to current information storage and computational capabili ties, Simiu and Stathopo ulo s (1997) pro posed a new generation of standard with provisions on wind loads that are no longer based on reductive and distorting tables and plots, but can be structured on knowledge based systems drawing the requisite information from large databases. Their postulation was that, wind loads evaluated via the new methodology would be functions of design parameters, which includes building geometry, building orientation, position with respect to and geometry of neighboring buildings, built up terrain roughness, etc. They intimated that their proposal would allow the designer to target specific situations, rather than providing blanket coverage for a broad range of situations. They explained that this methodology would furthermore allow the designer to account for the specific linear or non linear structural characteristics of the building or structure (eg. influence function). Subsequently, Whalen et al. (1998) conducted a pilot project on the estimation of wind effects in low rise build ing frames using this methodology. Whalen et al. (1998) used rec ords of pressure time histories measured at large number of taps on a building surface at the UWO boundary layer wind tunnel. Time histories of bending moments in a frame were obtained by

PAGE 26

26 summing up pressures time histories tributary to that frame multiplied by the respective tributary areas and frame influence coefficients. They compared results with results based on ASCE 7 standard provisions. Their comparison suggested that, significantly more riskconsistent, safer and economical designs could be achiev ed using this approach than using conventional standard provisions. The approach of using electronic aerodynamic and climatological databases to define wind loads was coined database assisted design (DAD) and was accepted by the ASCE 7 98 standard (Rigato et al. 2001) DAD Concept and S oftware D evelopment The first ge neration DAD application called WiLDE LRS Wind Load Design Environment for Low rise Structures was developed by NIST (Whalen et al. 2000) WiLDE LRS, a MATLAB based software, adopted interactive graph ical user interfaces (GUI) to give a visual, user friendly design environment. MATLAB scripts were used in the software to analyze the behavior of rigid portal frames and other components under high winds and to produce time histories of wind load effects in these structural members. The software had its origins in a prototype application called Frameloads, used to study wind effects on moment resisting frames in low rise buildings designed by the Standard Metal Building Manufactures Association methodology A latter version (2.7) of WiLDE LRS (Whalen et al. 2002) greatly enhanced the GUIs that directly accepted input of influence coefficients accounting for frame properties. Post processing was incorporated in this version to calculate realistic and robust statistical estimates of the peak load effect values based upon the entire time history. Subsequently, the DAD approach has been extended to consider nonlinear static response of low buildings (Jang et al. 2002) and also to account for the probability distribution of the peaks of time histories of wind effects and of sampling e rrors in the estimation of that

PAGE 27

27 distribution (Sadek and Simiu 2002; Sadek et al. 2004) A scheme to interpolate existing data in available database to other configurations in a reliable, accurate and simple way, without resorting to further wind tunnel experiments, has been incorporated in DAD applications (Kopp and Chen 2006) In 2006, NIST released software packages developed using the MATLAB language to fully implement the DAD approach and all its improvements (Main and Fritz 2006) Two separate software packages are available through the internet at http://www.nist.gov/wind for rigid, gable roofed buildings and for tall, flexible buildings. Limitation to the Application of DAD A pproach To the best of authors knowledge: 1. Application of the DAD approach and its software has been limited to steel portal frame buildings. 2. Structural influence functions used by researches so far in DAD applications have been analytically derived using hand calculations or 2 D models in structural analysis software 3. .The validity of the approach has not been demonstrated experimentally. Design and Constr uction of Light Framed Wood Structures and their Performance to Wind Forces Woodframe construction forms the majority of residential and other low rise structures. A number of these structures are located along hurricane prone zones in the United States. This section discusses the construction methods prevalent in the wood frame industry and their performance during hurricane events. The literature presented here is based on studies done by Rosowsky and Schiff (2003) and van de Lindt et al. (2007) Construction Methods Three construction methods have been identified by van de Lindt et al. (2007) These are the conventional, engineered and prescriptive. The conventional method consists of following

PAGE 28

28 documents such as the International Residential Code outlining certain exceptions and limitations. Most wood constructions are based on conventional methods. For engineered construction, structures are specifically designed by a design professional to meet jurisdictional requirements. Interestingly, very few residenti al buildings are engineered. Prescriptive construction involves the use of basic material strength level and tabulated values obtained from construction manuals. Rosowsky and Schiff (2003) referred to designs based on the conventional method as deemed to comply design, which is largely derived from traditional rules of thumb for building light frame woo d structures (LFWS). They observed that most of the rules focused on building structures to safely resist gravity loads, ignoring geographic considerations. Until recently, most buildings, including those located in high wind environments, were constructed using conventional methods which did not meet windresistant design requirements. This caused these structures to have the greatest vulnerability to extreme wind events. Critical Components and Systems According to Rosowsky and Schiff (2003) the three most important areas to consider in designing a windresistant wood frame structure are: 1. The building envelope : This forms the first line of defense against wind and water intrusion. Traditionally, the building envelope is considered to be architecture in nature and therefore not designed by engineers. However, studies have shown that a direct correlatio n exists between the performance and damage (losses) sustained by woodframe buildings. Structural engineers are becoming actively involved in the building envelope designs. 2. Attachment of roof and wall sheathing: this component is critical in keeping struc tures enclosed, preventing infiltration and providing critical links in the structural load path. Removal of roof sheathing is the second largest failure mode observed in post hurricane investigation after removal of roof cover. Significant highlight has b een given to the need to provide more and larger fasteners around roof edges to resist high wind uplift pressures.

PAGE 29

29 3. Structural systems to transfer the applied loads to foundation: In most woodframe construction, complicated load paths exist because of conv entional framing techniques and irregular floor and roof plans in residential buildings. Structural Failures of LFWS in Hurricane Events Structural observation made by van de Lindt et al. (2007) during a rec onnaissance trip after Hurricane Katrina are discussed as follows: 1. In many of the houses examined, there was absence of continuous load path for the transfer of wind loads from the roof down to the foundation. 2. Loss of roof sheathing at corners, which typic ally experience the highest uplift pressure during wind storms. In most of these cases, the current code minimum nail spacing requirements were not met. 3. Gable end wall losses as a result of loss of vinyl siding and failure of the foam sheathing. 4. The preval ent use of conventional construction in high wind regions. Light frame d wood structures (LFWS) have generally not performed well when subjected to high wind loads due to design/construction practices. Rosowsky and Schiff (2003) remarked that better understanding of the wind loading on buildings and behavior of woodframe structures under sever wind event s must be sought. This, they noted, will lead to improvements in both prescriptive and engineered design methodologies for new and retrofit construction. Wind Induced Pressures and Structural Responses on Light Wood Framed Structures The final stage of th is project is to validate the DAD methodology for its application to LFWS by simultaneously monitoring pressures and structural loads on a 1/3 scale house subjected to wind forces. This section presents experimental studies done by researchers whereby wind loads and structural responses were simultaneously measured on full scale buildings. These experiments are generally aimed at investigating whether observed structural responses correspond to predictions by numerical models.

PAGE 30

30 Doudak et al. (2005) monitored a single story industrial shed building to determine its displacement response to wind and snow loads. He at tempted to correlate the observed displacements with real time estimates using SAP 2000 of these environmental loads. Wind speed and direction was measured as well as displacements on the building during the typical wind storm season. Doudak et al. (2005) however did not take pressure measurements on the house. Wind pressures for numerical simulations were estima ted from archived pressure coefficients and the measured wind speeds. They achieved a quite reasonable agreement between measured and predicted displacements. Discrepanc ies rang ed from as low as 6 % in most cases to as high as 90 % for all four incident wi nd directions. In a follow up to the experiment done by Doudak et al. (2005) Zisis and Stathopoulos (2009) undertook an experimental study whereby they monitored and collected full scale pressure and force data on a light frame d wood building. A total of 51 load cells were installed at roof to wall and wall to foundation interfaces of the building while ensuring that the s tiffness of the building was unaltered. The building was also equipped with 27 pressure taps. All acquired data were analyzed and converted into dimensionless coefficients based on the following equations: Pressure measurements 2 BH a peak / mean peak / mean pU 2 / 1 P P C (2 4) Force measurements LW ) U 2 / 1 ( R C2 BH peak / mean peak / mean f (2 5)

PAGE 31

31 w here = air density; UBH = wind speed at the building height; Pa = ambient atmospheric pressure; P = actual surface pressure; R = reaction at the load cell location; L = length of the building and W = width of the building. Zisis and Stathopoulos (2009) also conducted wind tunnel experiments on a 1200 scale model and obtained pressure data which were also converted into pressure coefficients. Force coefficients were also derived by area averaging the measured local pressure coefficients f rom the wind tunnel studies on the building. These pressure coefficients were input into a numerical (finite element analysis) model of the test building which computed reaction forces at each of the 27 foundation load cell locations. These forces were tra nsformed into dimensionless coefficients. They observed good agreement of pressure distribution comparison between the wind tunnel and full scale data. They also concluded that the comparison between the full scale load cell readings and the base reactions computed by the finite element analysis made in the form of force coefficients shows good agreement as far as mean values are concerned. Zisis and Stathopoulos (2009) also conducted a 2 dimensional structural analysis of two main fr ames of the building. In the analysis, the individual roof pressures tap records (full scale) acting on each frame were used to evaluate the total expected vertical reaction due to wind pressure on each frame. They compared the estimated results to the act ual total reactions measured by the respective load cells of each frame in terms of force coefficients. Their comparison as well as the layout of their test building is shown in Figure 2 5. Zisis and Stathopoulos (2009) found excellen t agreement as far as the mean values were concerned. They also observed significantly more fluctuating signals with higher peak forces were obtained using the measured pressure coefficients on the building envelope in comparison with recorded signal by lo ad cells placed in the building foundation. They explained that this observation may be

PAGE 32

32 partly attributed to the dynamic load attenuation effect due to structural and material damping of the building components hence lower reactions measured than computed. Both experimental studies discussed above exposed the test buildings to natural wind forces. Consequently, wind pressure data collected from the field was highly affected by fluctuations of wind directions during the test. Moreover, the structure does not experience winds that would cause the worst load effects or design level events.

PAGE 33

33 Table 21. Comparison of bending moments (KNm) determined using ASCE 798 and DAD (Simiu et al. 2003) Frame 6.1 m eave height 9.75 m eave height Knee Ridge Knee Ridge ASCE DAD % ASCE DAD % ASCE DAD % ASCE DAD % Outer 339 330 3 118 136 13 463) 6 31 27 86 137 37 1 520 401 30 180 168 7 724 723 0 134 179 25 2 471 301 56 163 97 68 624 799 22 115 150 23 3 471 310 52 163 101 61 624 782 20 115 145 21 4 471 327 44 163 106 54 624 586 6 115 112 3 Figure 21. Separation and reat tachment pattern of wind flow over a low rise building (After Simiu & Miyata (2006) )

PAGE 34

34 A B Fig ure 2 2. Typical building surfaces for ASCE 7 05 MWFRS external pressure coefficients A B Fig ure 2 3. ASCE 7 05 provision for determining external pressure coefficients for the design of components and cladding.

PAGE 35

35 Figure 24. Isometric view of the steel portal frame structure (Simiu et al. 2003) (End frame not shown) Fi gure 25. Comparison of vertical reaction records (in terms of force coefficients) measured by load cells and estimated based on envelope roof pressures (Zisis and Stathopoulos 2009)

PAGE 36

36 CHAPTER 3 ANAYSIS OF WIND TUNNEL DATA T O GENERATE PRESSURE COEFFICIENTS Wind Tunnel Data For this study, pressure coefficients were derived from wind tunnel data developed by Datin and Prevatt (2007) on a 1:50 scale model house. The experiments were carried out in an atmospheric boundary layer wind tunnel at the Wind Load Test Facility (WLTF) at Clemson University An overview of the experiment is discussed below. House Model and Pressure Tap Layout Th e tests were conducted on a 1/50 house model called Clemson standard model (CSM) 4 12 which is shown in Figure 31A The house model has length of 14.4 in. and width of 7.2 in with a mean roof height of 3.4 in. CSM 4 12 has a gable roof with a slope of 18.4o (4 in 12). The model was configured for a 60 ft X 30 ft full scale building with a mean roof height of14.3 ft. The model has 387 pressure taps installed on its roof. The pressure taps are evenly spaced along the length of the roof at a nominal distance of 1 in except around the edges of the roof where they are densely grid at a nominal distance of 0.2 in. Figure 31B shows the pressure tap layout. These pressure taps were constructed with 0.063 outside diameter metal tubes glued to P lexi glas s sheets an d which are connected to Scanivalve electronic pressure scanners by 12 in long vinyl tubes. Wind Simulation and Pressure Measurements A suburban exposure was simulated upstream of the wind tunnel (shown in figure 3 2) The velocity profile and turbul ence i ntensity profile of the created exposure condition plotted against the log law profiles for suburban terrain are shown in Figure 33A Figure 33B shows the longitudinal wind speed normalized power spectrum taken in the wind tunnel at equivalent full scale height of 10 m (33 ft) as well as von Karman spectrum

PAGE 37

37 Near simultaneous pressure time histories were recorded using a scan ivalv e ZOC 33 sy stem. Tests were repeated for fi ve wind directions; 0o, 45o, 90o, 135o and 180o, as defined in Figure 31B. Eight t est repeats were done for each wind direction. Data was sampled at 300 Hz and recorded for 120 seconds for each test repeat. Table 3 1 summarizes the common test parameters used. These stored files were used as the raw wind tunnel data for this project. There were forty text files each containing 389 columns and 36000 rows. Aerodynamic Data Processing The raw data was low pass filtered at 150 Hz prior to analyzing them. Pressure coefficients were developed from the raw data as follows: It was corrected f or tubing response to remove any effects of tube length and size on the data Pressures were normalized by mean hourly pressure measured at 33 ft full scale height to obtain pressure coefficients. Pressure coefficients were re referenced to 3 second gust me an velocity measured at the mean roof height of the building (14.2 ft) Tubing Response Correction The effect of the tubing system, used in the wind tunnel study, on the measured wind pressure data was eliminated usi ng a tubing frequency response shown in F igure 34. This response was reported in Liu et al. (2009) T he raw pressure signal measured at each tap was first converted to the frequency domain using a F ast F ourier T ransformation This pr ovided a frequency (power) spectrum of the pressure signal. The frequency spectrum was then divided by the frequency response to remove the distortion caused by the volume and length of the tube. The corrected spectrum was then converted to time d omain using an Inverse Fast Fourier Transformation.

PAGE 38

38 Determining Pressure Coefficients Pressure coefficient s were derived from the measured local pressure time series as follows: ) ( P ) t ( P ) t ( Cref i i ref p (3 1) w here Cp,ref ,i is the pressure coefficient at Pressure T ap i referenced to the dynamic pressure at reference height at time t for wind angle ; Pi is the measured wind pressure at tap i at time t for wind angle ; ) ( Pref is the mean hourly reference dynamic pressure recorde d by a P itot tube at the reference height of full height of 33ft for wind angle Pressure coefficients were referenced at that height because flow is uniform with low turbulence levels at that height. This ensures accurate speed control of the wind tunne l and accurate calibration of the pressure scanners (Ho et al. 2005a) Re referencing of Pressure Coefficients It is widely accepted that aerodynamic data referenced to mean roof height dynamic pressure produce the least variability and therefore all low building pressure data sets, including those in the building codes, follow this convention (Ho et al. 2005a) It is intended that the wind tun nel results shoul d be comparable to those in ASCE 7 and other aerodynamic database. For this reason, the wind pressure coefficients were normalized to a 3 second gust mean wind speed at the mean roof height (14.3 ft full scale), mrh sec, 3U Th e wind pressure coefficients Cp,ref,i were converted to the equivalent coefficient as follows. ) ( ) (, t C C t Ci ref p a i p (3 2) w here, Cp ,i is the wind pressure coefficients at pressure tap i referenced to a 3 second gust wind speed at the mean ro of height, at time t for wind angle ; and Ca is an adjustment factor

PAGE 39

39 which is given by the squared ratio of the mean wind speed at reference height refU to the equivalent 3second gust wind speed at mean roof height mrh sec, 3U (Shown in Equation 3 3). 2 mrh sec, 3 2 ref aU U C (3 3) The mean hourly wind speeds at the reference height refU (13.03 m/s) and mean roof height Umrh (6.54 m/s) were determined from the velocity profile for the wind tunnel testing. The ASCE 7 05 provides the Durst curve which relates the wind speed averaged over gust duration, t (in seconds), Ut to hourly mean speed, U3600. However, the curved corresponds to open terrain conditions and an elevation, z = 10 m (Simiu and Scanlan 1996) As already stated, the wind tunnel data was developed for a suburban terrain conditi on and pressure coefficients are intended to be referenced to mean roof height of 14.3 ft (4.2 m). Hence, conversion of the mean hourly speed to 3second gust was done using Equation 3 4 provided by Simiu & Scanlan (1996) o 3600 tz z ln 5 2 ) t ( C 1 ) z ( U ) z ( U (3 4) where, C(t) is the time averaging constant for a given time averaging interval is the squar ed ratio of the friction velocity to the longitudinal turbulence fluctuations; zo is the roughness length; z is the height at which the wind speed is to be evaluated. The calculation of 3second gust wind speed at mean roof height, which was based on C( 3se c ) =2.85; zo = 0.22m, z = 4.2m (14.3ft) and = 5.25, is as follows: o mrh mrh sec, 3z z ln 5 2 sec) 3 ( C 1 U U (3 5)

PAGE 40

40 s / m 33 12 m 22 0 m 2 4 ln 5 2 25 5 85 2 1 ) s / m 54 6 ( Umrh sec, 3 (3 6) The resulting adjustment factor, Ca for re referencing the pressure coefficients is 1.1168. Wind Tunnel Results and Analysis Wind Pressure Coefficients Time Histories The re sulting pressure coefficient time histories were converted to equivalent full scale pressure coefficients using the reduced frequency relationship shown in Equation 3 7. p mV fL V fL (3 7) w here f, L and V are respectively sampling frequency cha racteristic dimension, and wind speed referenced at mean height over a specified duration Subscripts m and p denote model (1:50 scale) and prototype (full scale) buildings respectively. Based on model frequency and 3second gust wind speed at mean roof he ight of 300 Hz and 29 mph respectively, and 3second gust full scale wind speed at mean roof height of prototype for suburban terrain, of 80.27 mph, the prototype frequency is calculated as: mph 27 80 50 f mph 29 1 Hz 300p Hz 46 17 fp (3 8) Using equality of non dimensional time the equivalent full scale duration is given by: m m p pf T f T (3 9) Hz 46 17 s 120 Hz 300 Tp utes min 36 34 Tp (3 10) The equivalent full scale time step for the time histories is:

PAGE 41

41 s 0573 0 Hz 46 17 1 f 1 tp p (3 11) Pressure coefficient time histories in duration of 34 minutes in full scale of the 387 pressure taps were generated for each sample of a wind direction. Calculation for determining the equivalent fullscale duration of the 120 seconds of test period is presented in the appendix. Figure 35 shows time series plots pressure coefficients at pressure taps 1 and 387 for wind direction 0o. Pressure coefficient time series, which are useful for analyzing dynamic responses of low rise b uildings; have been saved in a MATLAB data format, as shown in F igure 33. There are 40 binary files with filenames structured as CSM4 12_Suburban_Cp_data_dir_XXX_Y.mat, where CSM4 12 identifies that the Clemson Standard Model with a roof slope of 412 was used in the wind tunnel study; XXX denotes the wind direction; and Y denotes the sample number. Each file also contains information of full scale geometric properties of the building model tested, sampling frequency and period, time step, locati ons of the pressure taps on model, data sample number, wind azimuth, etc. Observed Statistical V alues of Wind Pressure C oefficients The sample mean, root mean square (RMS) and peak local pressure coefficients were computed for the eight samples of each w ind azimuth The statistical values of pressure coefficients, which are useful for the design of cladding and components such as roof fasteners, purlins and panes, have been evaluated for 34 minute equivalent fullscale a erodynamic pressure coefficient time histories These values were also saved in the forty MATLAB files containing the pressure coefficient time histories. The mean and RMS pressure coefficient values were averaged values of the eight samples:

PAGE 42

42 8 1 n i p i p) n ( C 8 1 ) ( C (3 12) 8 1 n i p i p) n ( C 8 1 ) ( C (3 13) w here, ) ( Ci p and ) ( Ci p are respectively the mean and RMS pressure coefficients for at pressure tap i for wind angle of the entire experiment; and ) n ( Ci p and ) n ( Ci p are respectively the mean and RMS values of time series of the nth sample i for wind angle Contour plots of mean and RMS pressure coefficients measured for each direction are shown in Figures 37 and 3 8. Extreme Value Analysis of Pressure Coefficients Peak values estimated based on a probability distribution function are generally more statistically stable quantities than the observed peaks from individual samples (Ho et al. 2005a) The extreme negative and positive pressure coefficients mea sured from the eight samples of each wind direction were fitted to an Extreme Type 1 Value Distribution. The probability density function (PDF) and the cumulative distributive function (CDF) of the Extreme value Type 1 (also referred to as Gumbel distribut ion) are given by: ) x ( ) x (e e 1 ) x ( f (3 14) ) x (e ) x ( F (3 15) w here, is the location parameter (mode); and is the scale parameter (NIST 2003) The parameters were calculated using the Best Linear Unbiased Estimators (BLUE) (Lieblein 1974) There are three methods proposed by Lieblein (1974) based on sample sizes for the estimation of the location and scale parameters Method one is for an analysis with sample size less than

PAGE 43

43 sixteen. The second method should be used for a study with sample size larger than sixteen but generally smalle r than about fifty. For an analysis with larger same size, method three is to be used. The first method is adopted in this study since the sample size is eight Furthermore, for this analysis the peak negative pressure coefficients were multiplied by nega tive one to make them positive since BLUE analysis was developed for maximum values of the Type I extreme Value distribution. The positive values were then sorted in the ascending order to place them in the following order: 8 2 1x ... x x The loca tion parameter, and the scale parameter were then estimated as follows: 8 1 n i ixa 8 1 n i ixb (3 16) w here, xi is the i th value of the ascending array of maximum values of the eight samples and ai and bi are given by Table 3 2. The best e xpected (mean) peak pressure coefficient mea sured at each pressure tap in a given wind direction is given by: 5772 0 pC (3 17) Figures 39 and 3 10 show the spatial variations of the expected extreme pressure coefficients on the roof of th e building for different wind directions. The roof corners and gable edges experience spatial variations at close distances and higher magnitude of suctions for all wind directions except for wind direction 90o. A nearly even distribution is observed away from roof edges and corners. A similar pattern is observed between the mean, RMS and extreme pressure coefficient distributions of wind azimuth s 45o and 135o are at opposing angles. The same observation is made between the distributions of 0o and 180o. App endix A provides mean,

PAGE 44

44 RMS and extreme pressure coefficients of selected pressure taps for wind azimuths 0o, 45oand 90o. Area Averaged Pressure C oefficients Area averaged pressure coefficients have been derived from pressure coefficient time histories for regions of different size as follows: j jN i N i i i p FA A i t C t j C1 1/ ) ) ( ( ) ( (3 18) w here CF(t) is a area averaged wind pressure coefficient on region j at time t ; Cp(i,t) is the wind pressure coefficient at pressure tap i at time t; Ai is the tributary area of pressur e tap i and Nj is the total number of pressure taps on region j The mean, RMS and extreme values of area averaged pressure coefficients for the eight samples of each direction were determined. The average mean and RMS of were calculated as discussed abov e, while an extreme value analysis was done to determine the peak negative and positive area averaged pressure coefficients. Figures 3 11 to 3 13 display the area averaged pressure coefficients as a function of wind azimuth for corner, ridge corner, eave, ridge, interior, and gable edge, corner, ridge corner, eave, ridge, interior, and gable edge. These pressure coefficient values measured in the different regions on the surface of the building are compared to ASCE 7 05 external wind pressure coefficients f or components and cladding provi ded in Figure 611C of the ASCE 705. Table 3 3 provides a summary of peak local pressure coefficients observed within each of the three zones defined in ASCE 705, the area averaged pressure coefficients and the C&C externa l pressure coefficients corresponding to the zones. It is observed that, the peak local (tap) negative pressures (suctions) are generally higher in magnitude as compared to the ASCE 7 05 provisions for the des ign of components and cladding.

PAGE 45

45 However, the pe ak area averaged pressure coefficients measured for each zone fall within the provisions of ASCE 7 for the various zones. Also, it is observed that, the peak area averaged pressure coefficients for two wind directions (i.e. 0o and 180o; and 45o and 135o) a re identical.

PAGE 46

46 T able 31. Measurement configuration and parameters Model scale 1:50 Sampling frequency 300 Hz Sampling period 120 s Test angles 000 0 045 0 090 0 135 0 and 180 0 Upstream exposure Suburban 3second gust nominal wind tunnel speed at mea n roof height 12.33 m/s Table 3 2. Coefficients of BLUE for Type 1 Extreme Value Distribution (Lieblein 1974) i 1 2 3 4 5 6 7 8 a i 0.274 0.190 0.150 0.121 0.097 0.076 0.056 0.036 bi 0.394 0.06 0.011 0.059 0.087 0.103 0.108 0.102 Table 3 3. Comparison of wind tunnel and ASCE 705 peak pressure coefficients Zones Wind Tunnel ASCE 7 05 C&C Local Peak Area averaged 1 2.73 0.81 0.9 2 4.06 1.84 1.7 3 4.39 2.77 2.6

PAGE 47

47 A B Figure 31. 1:50 Scale house model (CSM 4 12) used in the wind tunnel study Figure 32. Test section arrangement for 1:50 suburban terrain

PAGE 48

48 0 10 20 30 40 50 0 0.3 0.6 0.9 1.2 1.5 U/Uref, Iu Height (m) Wind tunnel, U/Uref log law, U/Uref Wind tunnel, Iu Log law, Iu A B Figure 33. Wind flow characteristics for 1:50 suburban wind tunnel study. A) Mean wind speed and turbulence intensity profiles B ) Longit udinal spectrum of wind speed at full scale height of 32.8 ft Figure 34. Frequency response characteristics of the pressure tubing system 200 mm long -1.37 mm ID 3 00 mm long 0 06 mm ID 1 0 mm long 1.37 mm ID tube connector

PAGE 49

49 A B Figure 35. First 10 minutes time Series of wind pressure coefficient measured in direction 0o a t A ) Tap 1 B ) Tap 2 Figure 36. Format of MATLAB files of pressure coefficient data

PAGE 50

50 -0.4-0.2-0.2-0.2-0.20-0.40.20.4 Wind Azimuth = 000 -0.8-0.6-0.4-0.4-0.4-0.4-0.2-0.2-0.2-0.2-0.2-0.2-0.2-0.2-0.2-0.2-0.200-0.20.2 Wind Azimuth = 045 -0.2-0.2-0.2-0.2-0.2-0.2-0.2-0.2-0.2-0.200.2 Wind Azimuth = 090 -0.7-0.6-0.5-0.5-0.5-0.4-0.4-0.4-0.4-0.3-0.3-0.3-0.3-0.3-0.3-0.2-0.2-0.2-0.2-0.2-0.2-0.2-0.1-0.1-0.1-0.1-0.1-0.1-0.1-0.1-0.1-0.2-0.2-0.4-0.2-0.8-0.6-0.30 Wind Azimuth = 135 -0.4-0.3-0.3-0.3-0.2-0.2-0.2-0.2-0.1-0.1-0.1-0.1-0.1-0.1-0.3-0.1-0.3-0.1 Wind Azimuth = 180 -1 -0.8 -0.6 -0.4 -0.2 0 Figure 37. Spatial distributions of mean wind pressure coefficients

PAGE 51

51 0.20.20.20.40.40.40.20.20.40.20.20.40.20.60.60.20.61 Wind Azimuth = 000 0.20.20.2 020.20.20.20.20.30.30.30.30.30.30.30.40.40.40.40.40.50.5 0.50.50.10.10.20.20.60.60.70.80.20.10.90.40.30.20.40.20.20.5 Wind Azimuth = 045 0.20.20.20.20.10.10.10.20.20.20.10.10.10.30.30.30.30.30.30.40.50.302 Wind Azimuth = 090 0.20.20.20.20.20.20.20.20.20.20.30.30.30.30.30.30.30.3 0.40.40.40.40.40.50.50.50.10.10.60.60.10.10.70.80.10.90.50.411 Wind Azimuth = 135 0.10.10.10.10.10.10.10.20.20.20.30.30.30.30.10.10.10.40.4 0.40.50.20.50.60.3 Wind Azimuth = 180 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Figure 38. Spatial distributions of RMS of pressure coe fficients

PAGE 52

52 -3-2.5-2-2-2-2-1.5-1.5-1.5-1.5-1-1-1-1-1 -0.5-0.5-0.5-0.5-0.5-0.5-1 Wind Azimuth = 000 -3-3-2.5-2.5-2.5-2-2-2-2-2-1.5-1.5-1.5-1.5-1.5-1.5-1.5-1-1-1-1-1-1-1-0.5-0.5-0.5-0.5-1-1-1-1-1-1-0.5-0.5-2-1-0.5-1-3 Wind Azimuth = 045 -1.6-1.6-1.4-1.4-1.4-1.4-1.4-1.2-1.2-1.2-1.2-1.2-1.2-1.2-1.2-1.2-1-1-1-1-1-1-1-0.8-0.8-0.8-0.8-0.8-0.8-0.8-0.8-0.6-0.6-0.6-0.6-0.6 -0.6-0.6-0.6-1-1-1-1-1-1.2-1.2-0.4-1.4-1-1 Wind Azimuth = 090 -4-3.5-3-3-2.5-2.5-2-2-2-2-1.5-1.5-1.5-1.5-1.5-1.5-1.5-1.5-1-1-1-1-1-1-1-0.5-0.5-0.5-1-1-1-1.5-1.5-1-1-0.5-0.5-2-1-1-1.5-1.5-1-1 Wind Azimuth = 135 -3.5-3-2.5-2-2-2-2-1.5-1.5-1.5-1.5-1-1-1-1-1-0.5-0.5-0.5-0.5-0.5-0.5-0.5-2-1-2 Wind Azimuth = 180 -4.5 -4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 Figure 39. Spatial distributions of expected negative peak wind pressure coefficients

PAGE 53

53 0.50.50.5010.50.520.50 Wind Azimuth = 000 0.50.500.510.50.52000 Wind Azimuth = 045 0.50.50.50.50.50.51020 Wind Azimuth = 090 0.20.20.20.20.20.20.20.40.40.40.40.40.20.20.20.20.20.20.20.60.60.60.60.80.20.610.20.21 Wind Azimuth = 135 0.20.20.20.20.20.40.40.40.40.40.40.40.40.40.40.60.20.21 Wind Azimuth = 180 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Figure 310. Spatial distributions of expected positive peak wind pressure coefficients

PAGE 54

54 Figure 311. Area averaged pressure coefficients for reg ions corresponding to zone 1 ( Figure 6 11C in ASCE 7 05).

PAGE 55

55 Ridge R egion Eave Gable End Figure 312. Area averaged pressure coefficients for reg ions correspond ing to zone 2 ( Figure 6 11C in ASCE 7 05).

PAGE 56

56 Ridge Corner Eave Corner Figure 313. Area averaged pressure coefficients for reg ions corresponding to zone 3 ( Figure 6 11C in ASCE 7 05).

PAGE 57

57 CHAPTER 4 APPLICATION OF DAD M ETHODOL GY The database assisted design (DAD) methodology utilizes wind tunnel derived aerodynamic database of wind pressures in combination with structural influence functions to predict the wind load effects on a structural system. In this c hapter, pressure coef ficie nt time histories developed in C hapter 3 are used with experimentally derived structural influences to estimate structural reactions at roof to wall and wall t ofoundation connections. This c hapter opens with an overview of the structural influence fu nctions which were determined by Datin et al. (2010) A detailed explanation of the application of the DAD approach in this study is provided. Finally estimated structural reactions, analysis of results and their comparison with design wind loads based on ASCE 7 05 are presented. Structural Influence Function In structural analysis, the variation of shear, reaction, bending moment or deflection in a st ructure subjected to live load or moving load is generally best described using an influence line (Hibbeler 2006) An influence line therefore represents the variation of any struc tural response at a specific point in a member as a concentrated force moves over the member. Usually, the numerical values of a function for an influence line are determined using a dimensionless unit load. Subsequently the value of a response due to a fo rce applied at any position can be obtained by multiplying the ordinate of the influence line at that position. This concept of influence functions is also adopted in the use of the DAD methodology to predict wind induced structural responses. For steel fr amed structures, influence functions used in DAD based software provided by NIST (2008) are derived analytically from simple 2D model of a single frame. However, due to the high variability in wood properties, this study utilizes

PAGE 58

58 influence functions determined experimenta lly on a 1/3 scale wood house model (shown in Figure 41A ). In the experiment, a point load was applied at grid points (shown in figure 4 1C ) on the roof surface of the model house T he reactions induced by the point loads at 8 roof to wall and 9 wall to foundation connections were measured by load cells (locations shown in figure 4 1B ). Influence coefficients at each load cell were determined by dividing the measur ed reactions by the load applied at the grid points The resulting fraction (percentage) of the applied load provides a relative measure of forces transferred to the connections Structural influence surfaces over of vertical reactions were generated for each load cell. To justify th e experimental procedure used, simple 2 D and 3 D model s o f the building roof system were analyzed in structural analysis software, Visual Analysis 4.0 Truss members were modeled as simply supported beams and were pinned together. For the 3 D model, the roof sheathing was modeled as a plate element. In both cases, a unit load was applied at points on a t russ of the system and an influence line was drawn for reaction at the truss support Figure 43 shows a comparison of the results with the experimentally derived influence coefficients at grid points on a truss for a roof to wall connection of the truss. Influence function determined by 2D analytical model decreased linearly with distance between loading point and the support. However, while the 3D analytical model produced a close agreement with respect to the variat ion of influence function as compared to the experimental work, its influence coefficients differ from experimental results by as much as 30%. Typical experimentally derived influence su rfaces are shown in Figure 4 4.

PAGE 59

59 Evaluating Vertical Reactions Based on DAD methodology Velocity Pressure Estimates of vertical reactions at the roof to wall and wall to foundation connections are based on wind speeds estimates without regard for direction as in the ASCE 705 provision. Velocity pressure, qz ( psf ) was evalua ted at the mean roof height of 14.3 ft as follows: 2 142 1sub ft air zV q (4 1) open ft z sub ftV K V, 33 14 (4 2) w here air is the density of air; Kz is an velocity pressure exposure coefficient whose square root transforms the 3 second wind speed in miles per hour at 33 ft above ground over open terrain, V33ft,open into 3second gust wind speed at mean roof height over a suburban terrain, V14ft ,sub. Kz=0.70 (Table 6 3 of ASCE 705) was used in the DAD approach in order to be consistent with the ASCE 7 procedure Also the design wind speed, V33ft,open=130 mph (figure 6 1 of ASCE 705 (2005) was chosen, for consistency with vertical reactions evaluated on the basis of ASCE 7 05. Hence, mph 8 108 ) mph 130 ( 7 0 Vsub ft 14 (4 3) psf mph qz29 30 ) 8 108 ( 00256 02 (4 4) Pressure Taps and Influence F unctions The roof of the prototype house was d ivided into tributary areas of the pressure taps with the assumption that pressure coefficient measured by a tap was uniform within its geometric tributary area Pressure taps we re identified for the influence function grid points which fell within the tributary of the pressure taps Consequently, two or more influence coefficients were

PAGE 60

60 assigned to a pressure tap and load calculations were based on the tributary area of the influe nce function grid point. Reaction Loads Reaction time histories at eight roof to wall (gable end connection removed) and eleven wall to foundation connections were elevated by UFDAD using the windtunnel pressure coefficient time histories and the struct ural influence functions. The structural loads were determined by: 1 i i i p j i z jA ) t ( C ) N ( q ) t ( R (4 5) w here Rj(t) is structural load estimated at jth load cell at time, t ; qz is the velocity pressure evaluated at the mean roof height; Cpi(t ) is pressure coeffici ent for ith grid point at time t ; ( Ni)j is the influence coefficient at the ith grid point for j t h load cell ; and Ai is the tributary area of ith grid point. DAD R esults and Analysis Observed Statistical Values of Structural Reactions Again, forty files w ere generated in this analysis each containing structural reaction time histories of the 17 load cells. These times histories are useful in determining peak loads and other statistical values of interests. They can furthermore be used effectively in probab ilistic and structural reliability studies. Figures 4 6, 47 and 48 show the time histories of the vertical reaction at connections 5, 10 and 15 for wind directions 0o and 45o. The sample mean RMS and extreme reactions were measured from the time historie s for the forty samples of data. The measured statistical values of reactions for samp le 1 are provided in Table 4 1.The statistical values for the other samples are provided in Appendix B.

PAGE 61

61 Average values of RMS and mean reactions were calculated for each wind direction from the eight sample mean and RMS reactions. These quantities are not directly used in structural design but are useful for reliability analysis Tables 4.2 shows computed RMS and mean reactions. The mean reactions are also shown in Figures 47 to 411. Extreme Value Analysis of Vertical Loads The peak or extreme uplift reactions are important statistical values that are used in structural design. Generally, Lieblein fitted peak values are more statistically stable quantities than the measur ed peaks (Ho et al. 2005a) The measured peak reactions from the eight samples for each wind direction were fitted to Extreme Value Type 1 distributions. The location and scale parameters of these distributions were estimated using Best Linear Unbiased Estimators (BLUE) as proposed by Lieblein (1974) The same procedure explained in Chapter 3 of this report was followed reactions are provided in Table 4 3. These parameters were used to obtain the pro bability density functions (PDF) and the cumulative density functions (CPF) of the peak reactions from which the mean peak reactions were estimated. The best expected (mean) peak reaction load and corresponding standard deviation at the jth connection, in a given wind direction, were respectively determined as follows: j j j5772 0 R 6 / s (4 6) The estimated peak reactions and stand ard deviations from the Type I Extreme V alue an alyses are provided in Table 4 4. Fi gures 47 to 411 show the expected peak and mean reactions at the connections for all the wind directions considered in this study.

PAGE 62

62 It is observed from the plots that, a greater percentage of the uplift loads were transferred through the roof to wall con nections (load cells 15 and 19) at the gable end. The highest uplift reaction (1750 lbs) estimated was transferred through load cell 15 for wind directions 0o and 45o. It is worth nothing that the influence coefficients for this connection are higher at t he gable edge and the ridge corners, where high suctions were observed. Again its influence surface was wide spread than in the case of the other roof to wall connections. Again, though the connections between the gable end roof and wall were discontinued except at the end connections, reasonable uplift loads are transferred through foundation load cells 11, 12 and 13 for the directions. This is explained by the widespread of their influence surface s as a result of the diaphragm action in the wall (similar to deep beam action). However, very small loads are transferred through load cell 14. Relatively high er uplift loads were estimated at all connections for wind direction 0o than the other azimuths. The loads generally decreased as the direction of the inc ident increased. Higher positive reactions were transferred through the load cells for wind azimuth 90o because peak positive pressures enveloped the building for that wind direction. Vertical Reaction B ased on ASCE 7 05 Standard ASCE 7 05 provides separa te provisions for wind design using loads for either the main wind force resisting system (MWFRS) or for components and cladding (C&C) members. Major members of a building, which work together with other members in an assemblage to provide support and stab ility for the overall building, are designed using wind loads provided for MWFRS. Other members, which are directly loaded, are designed for localized wind load effects on relatively small areas using provisions for C&C. Roof to wall connections (truss re actions) which serve as media for trans fer of wind loading on the roof system to walls have in the past (Datin and Prevatt 2007) been designed as

PAGE 63

63 part of the buildings MWFRS H owever, roof to wall connection failures may be likely due to localized wind effects acting on the roof which necessitate estimating wind load reactions at these connections based on C&C pressures. Moreover, ASCE 7 05 lists roof trusses under both C&C and MWFRS which complicates the interpretation of the code for estimating reactions. For this reason, two separate analyses were done using both wind load provisions Velocity Pressure Vertical reactions loads at roof to wall connections of the building were obtained using the Analytical Procedure (Method 2) for low rise buildings (ASCE/SEI. 2005) with the design wind speed of V33ft,open,3sec=130 mph (figure 6 1 of ASCE 705) and a suburban terrain. A velocity pressure exposure coefficient of Kz=0.7 ( obtained from table 6 3 of ASCE 7 05) to modify the design wind speed to account for the suburban terrain condition and a mean roof height of 14.3 ft. This is very consistent with the wind speed used in the DAD es timations. A reduction factor, Kd = 0.85 to account for wind directionality effect was used in the ASCE 7 05 calculations even though no reduction factor was used in the DAD analysis. The importance factor I and the topographic effect factor Kzt were assumed to be unity. The velocity pressure, qz for ASCE 7 05 calculation is as follows : I V K K K 00256 0 q2 sec 3 open ft 33 d zt z z (4 7) 2 2 zft / lb 74 25 ) 0 1 ( ) mph 130 )( 85 0 )( 0 1 )( 7 0 ( 00256 0 q (4 8) ASCE 7 based Design Loads Wind pressure s on the building were determined as the just product of computed velocity pressure and external pressure coefficient s because internal coefficients were not considered in the studies In both MWFRS and C&C analysis, the study external pressure coefficients were determined based on the dimension, a = 3ft (10% of the smallest horizontal dimension).

PAGE 64

64 For ASCE 7 05 MWFRS, e xternal pressure coefficients are defined with respect to roof angles and building surfaces. The building has a roof angle of 18.4o and therefore l i near interpolation s were used to determine ex ternal pressure coefficients. Partitioning of the roof surface into four building surfaces as defined in Figure 610 of ASCE 705 is shown in Figure 412. Partitions were used to determine the pressures on the i ndividual trusses. Table 4 5 shows MWFRS based w ind pressures for the different building surface based on the velocity pressure In the case of C&C, external pressure coefficients were obtained from Figure 6 11C of the ASCE 7 05. Selection of the coefficie nts was based on an effective wind area of 300 ft2, which is a product of 30 ft span length and one third the span length. The building roof surface again, was partitioned into press ures zones as defined in ASCE 7 (see Figure 4 13). Wind Pressures based on C&C provisions for the zones including the roof overhang as given in Table 4 6. Wind load intensity on each truss was calculated by multiplying pressure on the truss tributary area by its tributary width. Each truss was loaded with its pressure intensity and analyzed separately in Visual Analysis 4.0 to determine the structural reaction at its support, for both MWFRS and C&C cases For MWFRS case, the procedure was repeated for all the eight load patterns defined in Figure 610 in ASCE 7 to obtain the ma ximum uplift reaction. Comparing Uplifts Reactions Predicted B ased on DAD vs. ASCE 7 05 For comparison, the DAD based reactions results were multiplied by 0.85 to account for the reduced probability of maximum winds coming from any direction, as it was in the case of the ASCE 7 based estimations. The worst uplift loads estimated at each of the roof to wall connections were selected and compared to the peak estimates based on the ASCE 7 Standard for MWFRS and Components and Cladding (C&C). T he comparison of the results based on ASCE 705 and DAD methodology are presented in Table 4 7. The results are graphically presented i n Figure 413.

PAGE 65

65 It is observed that the DAD based reactions are generally higher than reactions based on ASCE 7 05 MWFRS provisions. At t he gable end supports, ASCE 7 MWFRS underestimate the reaction at that support by as much as 28 % as compared to the DAD estimates. Underestimations of reactions by ASCE 7 MWFRS provisions are high at truss supports closer to the gable end of the house. Ve ry close results are however observed between the results at truss supports which are far away from the gable end. Significantly high reactions (by 6% to 62%) are observed for reactions based on ASCE 7C&C than in the case of the DAD. Using discrepancies o bserved in the two comparisons above, the DAD results have a better agreement with ASCE 7 MWFRS estimates than C&C based estimates.

PAGE 66

66 Table 4 1. Measured peak mean and RMS reactions (lbs) for sample 1 Azimuth 0o 45o 90o 135o 180o Load Cells Neg. Peak M ean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS 1 828 171 86 199 408 96 180 115 460 61 287 90 378 88 235 106 245 23 139 48 2 939 201 7 0 229 434 116 137 134 445 61 276 87 346 83 223 99 242 24 132 47 3 1067 256 63 287 749 140 142 161 414 77 327 102 364 98 173 113 259 30 127 53 4 1187 294 67 328 988 141 184 167 410 82 403 110 386 104 139 120 262 33 134 56 5 1 099 278 40 307 987 177 93 204 442 94 230 112 327 103 50 115 212 29 115 48 6 400 96 22 107 245 62 55 70 178 35 102 44 143 44 49 50 109 12 55 22 7 218 51 13 57 123 33 33 37 103 19 57 24 81 24 29 28 61 7 31 12 8 335 82 18 90 222 50 43 57 131 27 89 34 112 34 38 38 82 10 41 17 9 766 190 37 210 559 116 90 135 275 63 198 78 242 76 75 85 179 22 89 38 10 1149 291 47 320 887 191 110 218 440 102 259 122 354 117 93 131 260 33 132 56 11 1562 396 63 436 1210 260 146 298 602 142 344 168 483 161 127 180 352 45 181 77 12 1391 353 57 389 1080 230 134 264 536 125 310 150 432 142 115 159 312 40 161 68 13 404 103 16 114 316 69 37 79 159 38 86 44 127 42 32 47 90 12 47 20 14 52 14 2 15 42 10 4 11 21 5 11 6 18 6 4 6 11 2 6 3 15 1668 446 67 494 1754 308 158 373 757 157 308 184 509 164 74 182 345 49 182 77 19 1388 315 50 354 1340 328 9 359 660 169 50 185 453 139 86 153 215 36 111 55 20 1343 291 55 322 1 067 310 10 339 629 171 60 187 474 143 101 157 234 32 133 54

PAGE 67

67 Table 4 2. Average d mean and RMS reactions (lbs) Load Cells 0o 45o 90o 135o 180o Mean RMS Mean RMS Mean RMS Mean RMS Mean RMS 1 168 196 103 121 63 93 89 107 32 53 2 197 226 1 23 140 62 89 84 100 32 52 3 250 282 149 168 78 103 99 114 39 58 4 288 321 150 174 83 111 105 121 43 62 5 272 302 187 212 95 113 105 116 38 54 6 94 105 66 73 36 45 45 50 16 24 7 50 56 35 39 19 24 24 28 9 13 8 80 89 53 60 28 35 34 39 13 19 9 186 206 123 140 64 79 77 86 29 42 10 285 315 201 227 103 124 118 132 44 63 11 388 429 275 310 143 171 163 181 60 86 12 346 382 243 275 127 151 144 161 53 77 13 101 112 73 82 38 45 43 47 15 22 14 13 15 10 12 5 6 6 6 2 3 15 437 486 324 385 159 186 167 184 64 88 19 309 349 344 372 169 184 139 154 48 63 20 286 318 326 352 171 186 143 158 45 63

PAGE 68

68 Table 4 3. Parameters (lbs) for Type I Extreme Value Distribution of peak negative reacti ons Load Cells 0o 45o 90o 135o 180o 1 758 61 441 101 380 29 379 28 249 23 2 856 50 463 91 355 35 352 26 241 22 3 1037 40 608 68 391 33 374 21 259 22 4 1168 112 791 82 412 34 399 23 268 22 5 1053 98 875 55 380 38 344 27 225 24 6 369 14 232 21 160 14 148 12 106 7 7 199 11 120 14 88 9 84 6 60 4 8 304 16 206 15 125 9 116 7 82 5 9 698 49 510 32 278 20 256 18 179 11 10 1061 78 824 47 428 30 380 33 261 18 11 1446 105 1123 64 587 42 522 47 355 26 12 1289 96 1000 58 522 38 465 40 315 24 13 377 29 295 17 154 11 137 12 91 7 14 50 4 40 2 21 1 19 1 12 1 15 1608 171 1665 75 630 84 538 46 354 27 19 1257 106 1136 96 598 70 491 39 235 25 20 1125 86 1022 95 563 64 496 27 239 23

PAGE 69

69 Table 4 4. Expected peak negative (uplift) reactions (lbs) and standard deviation (lbs) estimate d from BLUE fitted probability distribution Load Cells 0o 45o 90o 135o 180o Mean Std Mean Std Mean Std Mean Std Mean Std 1 793 78 500 129 397 37 394 35 262 29 2 885 65 516 117 375 45 366 33 254 28 3 1061 52 647 88 410 42 386 27 271 28 4 1232 144 838 105 432 43 413 30 281 29 5 1110 126 907 70 402 49 360 35 239 31 6 377 18 244 28 168 17 155 15 110 9 7 206 14 128 18 93 11 87 8 62 5 8 314 21 215 20 130 12 121 10 84 6 9 726 63 528 41 289 26 267 23 185 14 1 0 1106 100 851 60 446 39 399 42 271 23 11 1506 134 1160 82 611 54 549 60 370 34 12 1344 123 1033 74 543 48 489 52 329 31 13 393 37 304 22 160 14 144 16 96 9 14 52 5 42 3 22 2 20 2 13 1 15 1707 219 1708 96 679 108 564 59 3 69 35 19 1318 135 1192 123 638 89 513 50 249 31 20 1175 110 1077 121 599 81 511 35 252 30 Table 4 5. Pressures based on MWFRS for different building surface Building Surface 2 3 2E 3E GC pf 0.69 0.47 1.07 0.67 Pressure (lb/ft 2 ) 17.76 12.10 27.54 17.25 Table 4 6. Pressures based on C&C for different zones Roof Overhang Zones 1 2 3 2 3 GCp 0.8 1.2 2 2.2 2.5 Pressure (lb/ft 2 ) 20. 6 30. 9 51. 5 56.6 64. 4

PAGE 70

70 Table 4 7. Comparison of uplift reaction estimates based on DAD and ASCE 7 05 Connection Number 1 2 3 4 5 15 19 20 DAD 674 752 902 1047 944 1452 1120 999 ASCE 705 MWFRS 763 763 763 830 830 1043 840 659 C&C 956 956 956 956 1435 1794 1794 1435 Figure 41. 1/ 3 Scale house model for determining influence functions

PAGE 71

71 Figure 42. Locations of load cells and wind direction F igure 43. Grid points for experimental determination of influence coefficients

PAGE 72

72 Figure 44. Influence lines for vertical reactions at a support (roof to wall connection) of an internal truss A B C D -0.2 0 0.2 0.4 0.6 0.8 1 1.2 Figure 45. Typical influence surfaces for vertical reaction determined on 1/3 scale house model A ) load cell 15 B ) load cell 5 C ) load c ell 4 D ) load cell 11

PAGE 73

73 0 100 200 300 400 500 600 -2000 -1500 -1000 -500 0 500 Reaction Loads (lbs)Time (sec) 0 100 200 300 400 500 600 -2000 -1500 -1000 -500 0 500 Reaction Loads (lbs)Time (sec) Figure 46. First 10 minutes load time Series at roof to wall load cell 5 A) wind azimuth 0o B) wind azimuth 45o 1 2 3 4 5 15 19 20 6 7 8 9 10 11 12 13 14 -1800 -1600 -1400 -1200 -1000 -800 -600 -400 -200 0 200 400 600 Load (lbs)Load Cell Negative Extreme Positive Extreme Mean Roof-to-Wall Connections Wall-to-Foundation Connections Figure 47. Estimated mean and peak vertical reactions at load cells for wind azimuth 000o

PAGE 74

74 1 2 3 4 5 15 19 20 6 7 8 9 10 11 12 13 14 -1800 -1600 -1400 -1200 -1000 -800 -600 -400 -200 0 200 400 600 Load (lbs)Load Cell Negative Extreme Positive Extreme Mean Roof-to-Wall Connections Wall-to-Foundation Connections Figure 48. Estimated mean and peak vertical reactions at load cells for wind azi muth 045o 1 2 3 4 5 15 19 20 6 7 8 9 10 11 12 13 14 -1800 -1600 -1400 -1200 -1000 -800 -600 -400 -200 0 200 400 600 Load (lbs) Load Cell Negative Extreme Positive Extreme Mean Wall-to-Foundation Connections Roof-to-Wall Connections Figure 49. Estimated mean and peak vertical reactions at load cells for wind azimuth 090o

PAGE 75

75 1 2 3 4 5 15 19 20 6 7 8 9 10 11 12 13 14 -1800 -1600 -1400 -1200 -1000 -800 -600 -400 -200 0 200 400 600 Load (lbs)Load Cell Negative Extreme Positive Extreme Mean Roof-to-Wall Connections Wall-to-Foundation Connections Figure 410. Estimated mean and peak vertical reactions at load cells for wind azimuth 135o 1 2 3 4 5 15 19 20 6 7 8 9 10 11 12 13 14 -1800 -1600 -1400 -1200 -1000 -800 -600 -400 -200 0 200 400 600 Load (lbs)Load Cell Negative Extreme Positive Extreme Mean Roof-to-Wall Connections Wall-to-Foundation Connections Figure 411. Estimated mean and peak vertical reactio ns at load cells for wind azimuth 180o

PAGE 76

76 Figure 412. Building surfaces for determining wind pressure for each truss based on ASCE 7 provisions for MWFRS A) Transverse direction B) Longitudinal direction Figure 413. Zones for determining wind pressure for each truss based on ASCE 7 provisions for c omponents and c ladding

PAGE 77

77 Figure 414. Uplift reactions at roof to wall load cells based on DAD approach and ASCE 705 provisions

PAGE 78

78 CHAPTER 5 WIND FLOW CHARACTERI ZATION USING TFI COB RA PROBE The flow characteristics of wind largely influence the intensity and fluctuation of wind forces on structures. The experimental component of the study includes a scale house model immersed within a wind flow generated by a UF Wind Generator. A multihole pressure probe cal led the Cobra Probe was used to map the wind flow characteristics to provide the necessary information for analyzing the data. This c hapter discusses the Cobra Probe, its principle of operation and pilot flow measurements conducted prior to using it in the main experimental study. Flow characterization of the wind field produced for the experi ment is also presented in this c hapter. The Cobra Probe The Cobra Probe (shown in Figure 51A) is a robust and compact multi hole pressure probe designed by Turbulent Flow Instrumentation (TFI) Australia, for measuring turbulent wind flows. The Cobra probe was first proposed by Shepherd (1981) for mean flow measurements and furth er developed by Hooper and Musgrove (1997) for resolving turbulence structure. The device has the following advantages: 1. It resolves highly turbulent wind flows into three orthogonal comp onents at high frequency (up to 10000 Hz). 2. It is a robust self contained device. 3. It can withstand significant levels of instrument vibration and maintain accuracy (Watkins et al. 2004) The high sampling frequency of the probe allows for continuous data streaming, with virtually unlimited data to be recorded while avoiding aliasing and reducing measurement of noise. The Probe has a truncated triangular py ramid shaped head with four faces ground flat to 45o (Chen et al. 2000) Each face has a 0.5 mm diameter pressure tap located at its center that is

PAGE 79

79 connected to piezoresistive bridge pressure transducers, and a preamplifier system located within the body of the probe. This arrangement was necessary to minimize the length of tubing so as to achieve a high frequency re sponse as well as mechanica lly protecting the transducers. The probe has an overall length of 160 mm and a body diameter of 14 mm. The design of the probe is shown in Figure 51B. The principle of operation of the probe is to relate the pressure field det ected by four pressure t ransducers to the magnitude of the instantaneous local velocity vector, yaw and pitch angles and the instantaneous static pressure. Pressure signals measured are corrected for transmission effects, using a predetermined transfer fun ction (frequency response function) of pressure tubing (Chen et al. 2000) A Series 100 Cobra Probe (Serial ID 193) was used in the study. It has a velocity range of 3 m/s to 60 m/s and an accuracy of 0.5 m/s and was supplied with an interface unit housing an integrated data acquisition and device control software. Preliminary Experiments Using the Cobra Probe Pilot e xperiments were conducted using the Cobra Probe aimed at (1) comparing flow measurements by the probe and hotwire anemometer (b) understanding the use and principle of the probe and (3) evaluating methods in modeling flow field conditions for the studys m ain experiment. In all the preliminary experiments, flow measurements were usually sampled by the probe at 5000 Hz and down sampled at a rate of 2500 Hz for 120 seconds per each test run. Before each test, the probe was zeroed to remove offset voltages fr om its pressure transducers. Comparing Wind Flow Measurements by the Cobra Probe and Hotwire Anemometer Flow measurements were simultaneously taken by the cobra probe and a hot wire anemometer in a generated wind field Table 5 1 compares the mean axial velocities, U, and

PAGE 80

80 turbulence intensities, Iuu, measured by the Cobra Probe and the hot wire anemometer. Good agreement is observed for the mean velocities. However, significant differences are observed for the turbulence intensities, which may in part be due to errors as a result of relative positions of the instruments. Wind Tunnel Model Several tests were done in a small wind tunnel (shown in Figure 53) to understand the principle and usage of the probe. The wind tunnel consists of a 10 ft long test sec tion, a contraction area, and a plenum and flow straighter section. It is powered by a 2 ft diameter axial fan. The test section is 24 in wide and 8 in high. Results from one test where flow measurements were taken at points (shown in Figure 5 4) within a 2D surface at 0.5 in downstream of the wind tunnel exit are presented. Profiles for all the orthogonal mean velocities measured are shown in Figure 5 5. Corresponding turbulence intensity distributions for longitudinal (Iuu), transverse (Ivv) and vertical (Iww) directions of flow are also displayed in Figure 5 5B. It is seen from the plots that very low transverse and vertical wind velocities are recorded away from the walls as compared to the high axial velocities. Relatively low axial velocities with corr esponding high turbulent intensities are observed near the wall surfaces of the test section. This is due to a high frictional effect near the wall surfaces. Figures 56 to 58 show the 2D variation of the mean velocities and turbulent intensities across t he test section. The spectral contents of wind speeds at the exit of the wind tunnel are displayed in Figure 5 9. Mapping of Wind Field G enerated by UF Wind Generator Velocity and turbulent intensity measurements of the wind flow generated by the UF wind generator were taken prior to this studys experimental work. The main purpose was to investigate if the wind field was uniform within the test area

PAGE 81

81 The probe was mounted on a computer controlled traverse to control the horizontal position. Vertical adju stments of the probe were done manually. The traverse frame was set up in two locations laterally to map the wind flow across the section. Figure 5 10 shows the traverse frame within the test area. The Cobra Probe was used to measure flow characteristics at 90 locations (shown in Figure 511) within a cross section at 16 ft downstream of the wind generator. Each measurement was taken for 60 seconds at a sampling frequency of 10000 Hz and output to file at 5000 Hz. A mean wind velocity of 50 mph (22 m/s) wa s used for all the tests. Contour plots of longitudinal, transverse and vertical wind fields are shown in Figure s 512 and 513. Relatively low longitudinal wind speeds at locations outside the jet field were observed. Longitudinal wind speed was generall y uniform across the cross section but decrease near the wall.

PAGE 82

82 Table 51. Comparison of flow measurements by Cobra Probe and hot wire anemometer Height (in) Cobra Probe Hot wire %Difference U (m/s) Iuu (%) U (m/s) Iuu (%) U Iuu 2 5.47 14.33 5.13 13.38 6.2% 6.6% 6 6.26 13.83 6.36 10.68 1.6% 22.8% 10 5.83 10.15 6.28 11.38 7.7% 12.1% Figure 51. Cobra Probe Figure 52. Cobra Probe and Hot wire anemometer setup for simultaneous measurements

PAGE 83

83 Figure 53. Wind tunnel model used in p ilot studies Figure 54. Measuring points for mapping flow measurements at of the wind tunnel

PAGE 84

84 A B Figure 55. Flow measurements at exit of wind tunnel model A) Mean velocity (m/s) distribution s. B) Turbulence intensity distributions (y is distan ce of grid point relative to the midpoint of test section; D is half the width of test section) 8 10 12 14 16 4 6 8 10 12 Figure 56. Spatial variations of longitudinal velocity (m/s) and turbulence intensity (%) across exit section of wind tunnel model -1 -0.5 0 0.5 -1.5 -1 -0.5 0 0.5 1 1.5 2 Figur e 57. Spatial variations of lateral v elocity (m/s) and turbulence intensity (%) across exit section of wind tunnel model

PAGE 85

85 -1.5 -1 -0.5 0 0.5 1 1.5 2 2 4 6 8 10 Figure 58. Spatial variations of lateral v elocity (m/s) and turbulence intensity (%) across exit section of wind tunnel model Figure 59. Longitudinal, transverse and vertical spectral contents of wind speed at 0.5 in. downstream of the exit of the wind tunnel

PAGE 86

86 Figure 510. Position of traverse frame in test section Figure 511. Location of measurem ents points for flow mapping

PAGE 87

87 Figure 512. Variation of longitudinal wind speed across 2D measurement surface Figure 513. Variation of lateral and vertical wind speed across 2D measurement surface

PAGE 88

88 A B C Figure 514. Spectral co ntents of wind speed A) longitudinal B) Transverse C) Vertical

PAGE 89

89 CHAPTER 6 WIND INDUCED PRESSURE AND STRUCTURAL LOAD MEASUREMENTS The objective of the experimental component of the research is to evaluate the validity of using the DAD methodology to predic t wind induced structural loads as discussed in Chapter 5. This chapter presents the experimental study, results and analysis performed to validate the DAD approach. Materials and Methods Scale House Model The 1/3 scale house (shown in Figure 6 9) was mode led and constructed by Datin et al. (2009) based on non dimensional geometrical scaling laws. The building model has a rectangular floor plan which is 10 ft wide by 13 ft 4 in. long (30 ft by 40 ft at full scale) and a mean roof height of 4 ft 1 in (12 ft 3 in at full scale) with a 6 in. (1 ft 6 in at full scale) overhang around the perimeter It has a gable roof sloping at 18o, with Fink style wood trusses installed at 8 in. spacing (24 in o.c. at full scale) The building model is mounted on a 3 ft high st eel frame clad with oriented standard boards (OSB) sheets. To be representative of typical residential constructions, the gable end trusses were connected only at the truss reaction points and not through bottom chords with the end wall. Pressure and Load Sensors on the B uilding Figure 61 shows the layout of pressure taps and load cells. Twenty nine pressure taps were installed on the building model, 25 on the roof and 4 on the walls (See Figure 2 for the distribution of the taps on the roof). The nearest pressure taps to the gable end edge of the roof are installed 3.5 in. from the edge. Each pressure tap was made using 3 in. long 0.124 in inside diameter (ID) brass tube soldered to a thin metal disc as shown in Figure 62A. Figure 62B shows the distribu tion of pressure taps over the house roof 27 Omega PX 138 pressure

PAGE 90

90 transducers and 2 Setra model 265 transducers were used to capture pressure distributions on the roof and walls of the house. The high port of each transducer was connected to the taps via a 6 in. long 3/16 in. ID vinyl clear tube. The internal pressure was measured using a Dwyer 616 20B transducer. The low ports of the transducers were connected to a single manifold and held at ambient pressure (see Figure 6 3). The specifications of these transducers are given in Table 6 1. Pictures of the three different transducers used are shown in Figure 6 4. Structural reactions on the building were monitored using twenty one 300 single axes compression/tension load cells (Futek model LRF 350). Twelve load cells were installed at roof to wall connections and 9 at the wallto foundation connections (see Figure 6 5). The building was isolated at the instrumented end to capture the total load reactions through this load cell array. All these pressure tra nsducers were factory calibrated However a Fluke 718 1G pressure calibrator with serial number of 9916005(shown in Figure 6 6) was used to confirm the calibration charts. The data acquisition system consisted of a National Instruments (NI) CompacDAQ chassis ( NI cDAQ 9172) with modules NI 9205 and NI 9219. This data acquisition system was controlled by Measurement and Automation Explorer and LabView software version 8.5. Prior to the experiment, the instruments were left overnight to measure room con dition signals to check for possible drift effects in their performances. Test Arrangement Two 16 ft tall by 34 long woo d framed walls were constructed at the exit the UF wind generator to enclose the test building. The windward face of the 1/3 scale model was placed at approximately18 ft away from the exit to provide a flow development length of at least four times the fan diameter (4.5 ft). The house was oriented in three separate directions (0o, 45o and 90o) to the wind flow Figure 68 shows a layout fo r the experimental set up. Figures 59 shows

PAGE 91

91 the different orientations of the house in the experimental setup. The house was anchored to the ground to resist all horizontal and vertical movements as well as overturning during the test. The Cobra Probe was positioned centrally within the test section 6 ft downstream of the contraction of the wind generator to record wind speeds. It was installed at 8 ft above ground to match the mean roof height of the house. A pitot tube, connected t o a D wyer pressure tran sducer ( shown in Figure 67), was installed 2 ft above the roof ridge Wind Generation The UF wind generator (shown in Figure 610) situated at the Powel Family Structures Lab was developed for testing window and door panels. It generates wind forces by u sing an 8fan array of 4.5 ft diameter fans hydraulically powered by four marine diesel engines with a combined 2800 hp. The generator can produce fluctuating winds at speeds up to 120 mph. Further description of its development and operation can be found in Masters et al. (2008) A mean wind velocity of 50 mph (22 mph) was used for all the tests. The wind speed was chosen based on scaling considerations for equi valent 130 mph full scale mean speed. To produce this speed, the wind generator engines were run at 1000 rpm turning the eight fans at 700 rpm. The wind generator produced approximately 6 % turbulence intensity, measured at the mean roof height, for the ex periment. Experimental Procedure and Measurements Tests were conducted with the building in three orientations to incident wind flow: 0o, 90o and 45o. Zer o degree was taken as wind flowing normal to the gable end wall of the building Three repeats were d one at each house orientation. Data was sampled at 200 Hz lasting 10 minutes for each test. The LabView program synchronized measurements by all the pressure and load sensors and saved them into test file output for each test. The Cobra Probe, which was

PAGE 92

92 co ntrolled by TFI device software, was timed using a digital trigger connected to the LabView program. Sampling by all the measuring instruments started at the same time for all the tests. Prior to each test, the Cobra Probe was zeroed and initial readings f rom the transducers and load cells were taken for 120 seconds. These initial records were subtracted from pressure and load measurements before data analysis remove any initial offset of the instruments. The pressure and structural load measurements were l ow passed filtered at 100 Hz and their statistical values were measured using a MATLAB based program Experimental Results, Analysis and Discussion Wind Pressure Measurements Positive peak pressures (7 11 psf) were recorded on the windward walls whereas suctions were recorded on the leeward walls for 0o and 90o azimuths. For 45o degrees, both the instrumented gable end wall and the side wall experienced positive pressures. The internal pressure recorded during all the experiments was zero psf. The wind d istribution over the roof of the building was generally characterized by peak suctions especially for the 0o and 45o azimuths. Wind suctions were highest along the second interior truss (with load cells 4 and 16) for the 0o azimuth. High suctions were reco rded along the gable edge of the roof while relatively low suctions were measured at the interior roof areas of the roof for the 45o azimuth. For the 90o azimuth however, relatively low suctions enveloped the roof of the house except at the eave end area. The wind pressure distributions observed in all these experiments were in good agreements with what is reported in literature and observed from the wind tunnel data. Wind pressures measured were normalized into pressure coefficients by the velocity pressur e at the m ean roof height of the building. The pressure coefficients were evaluated using Equation 5 1.

PAGE 93

93 2 mrh i i mrh pU 2 / 1 ) t ( P ) t ( C ( 51) where Cp,mrh,i(t) pressure coefficient at pressure tap i at time t, referenced to the mean wind speed measured at the mean roof height of the building ; Pi is the measured w ind pressure at tap i the air density; Umrh is mean wind speed (mph) recorded at mean roof height by the Cobra Probe. Pressure coefficients obtained for the nine test repeats are given in Appendix C. Table 6 2 shows the mean, peak and root mean square (RMS) pressure coef ficients averaged from results of the three datasets for each wind direction. The peak negative and mean pressure coefficients contour plot results are also shown in Figure 611. Wind Induced Structural L oads Measured statistical values of structural loads obtained from each test of the experimental study are provided in Appendix D. The three sample data obtained for each direction were augmented. Each augmented time history was then divided into six segments to obtain six 10 minute (full scale) samples. Me an, peak and RMS reactions were measured from these segments and an extreme value analysis (discussed in Chapter 3 and Chapter 4) was used to estimate the expected positive and negative peak reactions. The Lieblein BLUE estimators (Li eblein 1974) used in the analysis of the six samples in each direction are provided in Table 5 3. The mean RMS and expected peak values of structural loads measured by the twenty one load cells are provided in Table 5 4. These statistical values are aver age loads from the three datasets for each wind azimuth. The locations of the load cells on a 3D drawing of the house are shown in Figure 613. Figures 5 14 to 516 show the mean, extreme positive and negative structural reactions. The plots are such that the gable end load cells (15 & 19) are at the center of the roof to wall zone. Load cells on the same truss are mirrored about load cells 15 and 19 on the plots for easy comparison of measurements.

PAGE 94

94 Even though there was no connection between the gable end truss and the gable end wall except at end supports of the truss, there was considerable load transferred from the gable end wall to the foundation. At the 9 0o azimuth, significantly high reactions were recorded at the gable end wall to foundation load ce lls than the other load cells. Another unusual observation made for this direction was that load cell 20 recorded high positive reactions whereas adjacent load cells (19 and 16) recorded relatively high negative reactions. Perhaps, these observations could be explained by fact that the portion of the building where these load cells were located was not completely immersed in the wind flow jet (as shown in Figure 5 8) and therefore subjected to un realistic wind loads. For incident 0o azimuth, it is observed that fairly symmetrical loads were recorded at the roof to wall connections on side walls Also, load cells 4 and 16 recorded highest uplift load (26 psf & 22 psf respectively) on either sidewall suggesting that high suctions experienced along the second interior truss were mainly transferred through its support. Load cells located at the gable end measured equally high uplifts as observed in the DAD based loads of wind tunnel data. Generally, the structural loads measured at the roof to wall connections a re highest at the 0o azimuth and lowest at the 9 0o azimuth This observation was also made in the case of loads computed based on wind tunnel data where the highest loads were estimated for the 0o azimuth and decreased with wind direction increment of 15o. Structural Load Comparison The envelope pressures over the house were used to estimate the structural loads being transferred through roof to wall load cell numbers 4, 5 and 15, and foundation load cell number 11. Predictions were made based on the DAD me thodology using the experimentally derived influence coefficients (described in Chapter 4). For comparison with the directly measured

PAGE 95

95 structural loads, actual pressure time histories (not pressure coefficients) based on the wind flow characteristics of the various tests were utilized in the DAD approach as follows: 1 i i i j i j))] t ( P A ) N [( ) t ( R (5 2) w here Rj(t) is load reaction in pounds at the j th connection at time t ; (Ni)j is i nfluence coefficient at the i th grid point for the j th connection; Ai is tributary a rea of i th grid point; and Pi(t) is pressure measured at the i th grid point at time t Figures 617 to 619 show 1 minute records of measured and evaluated reactions. High correlations are observed between the reaction time histories of the two records. T he highest correlation coefficient of 0.9 was measured between the records of load cells 15 for the 0o wind azimuth. Measured correlation coefficients for the records are given in Table 65. Even though the estimated and measured reactions are not in excel lent agreement as far as magnitude is concerned for some the load cells considered, their records show good trends. It is observed that in almost all the comparisons, t he peaks of time histories of the two records seem to be occurring at the same time. Fur thermore, significantly more fluctuating signals were obtained using the roof pressures relative to the directly measured load cell signals. It is worth noting that a similar observation was made by Zisis & Stathopoulos (2009) in a pa rallel study. Zisis & Stathopoulos (2009) attributed this phenomenon to the attenuation effect due to structural and material damping of the building components thereby lowering the fluctuations in the measured signal. Peak and mean uplift structural loads measured and estimated for the four load cells are provided in Table 66. Good agreements are observed comparing the mean and peak reaction in almost all cases. Small discrepancies (lowest is 2%) are seen between the statistical values for load cell 15 as compared to the others. Generally, the DAD methodology underestimated structural loads expected to be transferred through the other load cells considered. This may be

PAGE 96

96 due to the fact that the effective influence surfaces of these lo ad cells were not completely covered with pressure taps. Hence, pressure loads which were transferred through these load cells were not fully captured. In summary, the DAD methodology adequately predicted the expected loads at roof to wall and wallto foun dation connections based on realistic wind load paths. The ability to estimate structural loads on light frame d wood s tructures using the DAD approach has been demonstrated in this experimental study.

PAGE 97

97 Table 61. Manufactures and specifications of pressur e sensors Transducer Model Manufacturer Number Used Range Output Response Frequency PX 138 001 D5V Omega Engineering 27 1 psi (144 psf) 1 to 6 VDC 1000 Hz 265 Setra Systems, Inc. 2 10 in. H2O (52 psf) 0 to 5 VDC 5 0 Hz 616 -20B Dwyer Instruments, Inc. 1 10 in. H2O (52 psf) 4 to 20 mA 2.5 Hz

PAGE 98

98 Table 62. Statistical values of measured pressure coefficients 0o 45o 90o Press. taps Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Roof press ure taps 1 1.59 0.56 0.06 0.60 0.15 0.04 0.21 0.06 0.93 0.14 0.47 0.24 2 2.70 0.51 1.08 0.56 1.97 0.05 1.56 0.31 2.98 0.33 1.79 0.48 3 2.74 0.48 1.77 0.52 2.70 0.80 1.24 0.88 3.01 0.45 1.94 0.54 4 2.73 0.45 1.35 0.50 2.31 0.56 0.9 8 0.60 2.90 0.51 1.66 0.58 5 2.26 0.52 0.13 0.56 1.64 0.06 1.38 0.20 1.75 0.01 1.46 0.20 6 1.91 0.53 0.25 0.57 0.05 0.11 0.27 0.11 1.68 0.07 1.56 0.24 7 2.95 0.47 0.80 0.50 0.18 0.12 0.41 0.12 1.71 0.02 1.51 0.24 8 2.41 0.42 1.26 0.45 0.69 0.10 0.77 0.13 1.81 0.02 2.00 0.23 9 1.16 0.42 0.05 0.44 3.14 0.56 1.94 0.67 0.86 0.10 1.03 0.16 10 1.30 0.44 0.05 0.46 0.41 0.01 0.40 0.06 0.80 0.10 0.70 0.13 11 2.23 0.47 0.16 0.49 0.17 0.04 0.27 0.06 0.57 0.15 0.70 0.16 12 1.86 0.47 0.54 0.49 0.18 0.02 0.22 0.05 0.86 0.38 1.26 0.60 13 0.89 0.34 0.09 0.35 3.16 0.80 1.41 0.87 1.02 0.10 0.51 0.14 14 1.02 0.39 0.06 0.40 1.25 0.07 1.03 0.17 0.53 0.10 0.55 0.11 15 1.36 0.42 0.05 0.44 0.53 0.10 0.22 0.11 0.44 0.06 0.39 0.08 16 1.60 0.47 0.27 0.49 0.33 0.10 0.12 0.11 0.46 0.08 0.56 0.09 17 0.85 0.34 0.07 0.35 3.28 1.08 0.82 1.13 0.47 0.01 0.64 0.07 18 1.05 0.36 0.07 0.37 2.30 0.42 1.30 0.49 0.40 0.04 0.43 0.07 19 2.08 0.45 0.18 0.47 0.76 0.3 9 0.10 0.40 0.39 0.04 0.33 0.07 20 1.68 0.40 0.33 0.41 1.15 0.35 0.39 0.36 0.37 0.05 0.41 0.07 21 0.88 0.34 0.01 0.35 6.34 1.72 2.53 1.87 1.09 0.05 0.46 0.09 22 1.02 0.36 0.04 0.37 3.48 0.64 1.68 0.77 1.14 0.03 0.29 0.07 23 1.56 0 .40 0.04 0.41 2.55 0.54 1.07 0.59 0.86 0.02 0.20 0.07 24 1.31 0.34 0.03 0.35 3.21 0.92 1.37 0.96 0.91 0.04 0.30 0.09 25 1.59 0.41 0.10 0.43 1.63 0.23 1.26 0.30 0.89 0.05 0.18 0.09 Wall pressure taps 26 0.63 0.19 0.15 0.20 0.60 0.79 1 .04 0.79 0.43 0.45 1.64 0.46 27 1.21 0.31 0.19 0.32 0.39 0.53 0.72 0.54 0.33 0.92 1.36 0.92 28 0.45 0.77 1.01 0.78 0.55 0.73 0.90 0.73 1.02 0.41 0.07 0.42 29 0.86 1.07 1.25 1.07 0.12 0.35 0.56 0.35 0.61 0.05 0.65 0.12 Internal pressure 30 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.03 0.00 0.02 0.01

PAGE 99

99 Table 6 3. Coefficients of BLUE for Type 1 Extreme Value Distribution (Lieblein 1974) i 1 2 3 4 5 6 ai 0.355 0.225 0.166 0.121 0.083 0.048 b i 0.459 0.036 0.073 0.127 0 150 0.1 46 Table 64. Mean, RMS and BLUE estimated peak values of measured structural loads (lbs) Load Cells 0o 45o 90o Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Roof to wall load cells 1 14.3 8.0 0.7 8.3 6.9 3.9 1.1 4.0 0.1 3.7 6.5 3.8 2 10.1 7.1 3.1 7.1 8.6 5.9 3.1 5.9 3.9 0.1 3.8 1.3 3 13.0 8.1 4.6 8.2 7.5 5.1 2.7 5.2 2.0 1.3 4.4 1.5 4 25.6 18.6 11.8 18.7 16.1 12.5 8.9 12.6 2.4 1.4 5.0 1.7 5 19.7 15.7 11.5 15.8 13.0 9.9 7.5 9.9 16.2 9.0 4.9 9.1 15 22.9 15.3 9.1 15.5 17.5 12.1 6.8 12.2 3.4 1.9 6.9 2.5 16 20.7 16.3 12.3 16.3 20.3 15.8 11.5 15.9 7.2 3.0 0.5 3.1 17 9.7 5.6 2.7 5.6 5.7 2.3 1.5 2.4 3.4 1.1 0.1 1.3 18 10.7 7.1 4.3 7.2 13.5 9.2 5.2 9.3 6.4 1.0 1.9 1.6 19 20.1 13.9 8.8 14.0 19.0 13.0 7.2 13.1 16.3 8.5 2.5 8.8 20 19.9 14.2 9.9 14.3 19.8 14.2 7.7 14.3 5.0 9.3 13.2 9.3 21 11.0 6.5 2.7 6.6 7.3 3.7 1.1 3.9 6.0 0.7 1.6 1.3 Wallto foundation load cell s 6 7.6 3.4 0.2 3.6 14.3 9.9 5.8 10.0 6.7 0.3 6.3 2.3 7 17.5 11.8 5.7 11.9 5.1 1.1 3.2 1.8 6.4 12.7 19.4 12.9 8 23.3 17.7 11.1 17.8 4.8 0.5 3.8 1.5 7.3 14.7 21.4 14.9 9 22.3 18.1 13.3 18.1 10.7 7.3 3.9 7.4 5.3 2.3 8.8 3.2 10 2 1.8 18.1 13.7 18.1 16.7 13.7 10.8 13.7 26.9 17.4 10.5 17.5 11 18.5 13.7 8.8 13.8 19.4 14.5 10.2 14.6 32.3 21.0 11.8 21.2 12 18.8 14.6 10.4 14.7 17.6 13.5 9.5 13.6 44.7 29.8 19.0 30.1 13 10.7 8.1 5.5 8.2 9.7 7.4 4.9 7.4 1 8.2 11.7 6.7 11.8 14 2.7 4.7 6.3 4.7 2.0 0.2 1.8 0.7 5.7 1.8 1.5 2.1

PAGE 100

100 Table 65. Correlation coefficients measured between time histories of measured and estimated reactions Azimuth 0 o 45 o 90 o Load Cell 4 5 11 15 4 5 11 15 4 5 11 15 Coefficie nt 0.83 0.85 0.56 0.90 0.48 0.30 0.30 0.41 0.47 0.43 0.43 0.45 Table 66. Comparison of directly measured and DAD based estimated reactions Wind Direction Peak Uplift (lbs) Mean (lbs) Load Cell 4 5 11 15 4 5 11 15 0o DAD 19.6 14.7 16.4 22.0 9.7 8.0 9.6 12.3 Measured 25.9 19.7 18.5 22.9 18.6 15.7 13.7 15.3 `45o DAD 11.3 11.4 11.4 25.3 5.0 5. 6 7.2 10.3 Measured 15.9 13. 0 19.4 17.5 12.5 9.9 14.5 12.1 90o DAD 7.8 5.7 5. 3 8.8 0.5 0.2 0.5 0.4 Measured 3.3 16.0 32.3 3.36 1.4 9.0 21.0 1.9 Figure 61. Locations of pressure taps and load cells on the scale house model.

PAGE 101

101 A B Figure 62. A) A pressure tap sample B) Layout of pressure tap over the building roof Figure 63. Interior of the house mod el

PAGE 102

102 A B C Figure 64. Pressure Sensors A) Omega PX 138 B) Setra 265 C) Dwyer 616 A B Figure 65. Futek load cells A) Roof to wall connection B) Wall to foundation connection Figure 66. F luke pressure calibrator

PAGE 103

103 Figure 67. Dwy er Transducer and pitot tube for wind velocity measurements

PAGE 104

104 F igure 68. Layout of experimental setup

PAGE 105

105 A B C D Figure 69. A) Completed setup B) House at 0o orientation C) House at 45o orientation D) House at 90o orientation

PAGE 106

106 F igure 610. UF Wind Generator -2.6-2.4-2.2-2.2-2-2-1.8-1.6-1.6-1.6-1.4-1.4-1.4-1.2-1.2-1-1-1.8-2-2.8-1.8-2-2.2 Azimuth = 000 -6-5-4-4-3-3-3-2-2-2-2-1-1-100-1 Azimuth = 045 -3 -2.5-2-1.5-1.5-1-1-1-0.5-0.5-1 Azimuth = 090 6422866 -6 -5 -4 -3 -2 -1 Figure 611. Spatial distribution s of p eak pressure coefficient

PAGE 107

107 -0.5-0.45-0.4-0.4-0.35-0.35-0.45-0.4 Azimuth = 000 -2-1.5-1-1-0.5-0.5-0.50000 Azimuth = 045 -0.6 -0.5 -0.4-0.3 -0.2-0.1000 0.10.10.20.20.30.4 Azimuth = 090 54 -2 -1.8 -1.6 -1.4 -1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 Figure 612. Spatial distribution s of m ean pressure coefficient Figure 613. Location of load cells on the house

PAGE 108

108 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 -50 -40 -30 -20 -10 0 10 20 30 Load (lbs)Load CellsWind Direction 000 Negative Extreme Positive Extreme Mean Roof-to-Wall Connections Wall-to-Foundation Connections Figure 614. Mean and expected p eak v alues of measured structural loads for wind direction 0o 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 -50 -40 -30 -20 -10 0 10 20 30 Load (lbs)Load CellsWind Direction 045 Negative Extreme Positive Extreme Mean Roof-to-Wall Connections Wall-to-Foundation Connections Figure 615. Mean and expected peak values of measured structural loads for wind direction 45o

PAGE 109

109 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 -50 -40 -30 -20 -10 0 10 20 30 Load (lbs)Load CellsWind Direction 090 Negative Extreme Positive Extreme Mean Roof-to-Wall Connections Wall-to-Foundation Connections Figure 616. Mean and expected peak values of measured structural loads for wind direction 90o 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 4 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 5 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 11 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) DAD estimated Measured Load Cell 15 Figure 617. Comparison of measured and estimated reactions for wind direction 0o

PAGE 110

110 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 4 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 5 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 11 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) DAD estimated Measured Load Cell 15 Figure 618. Comparison of measured and estimated reactions for wind direction 45o 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 4 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 5 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) Load Cell 11 0 10 20 30 40 50 60 -30 -20 -10 0 10 Time (Sec)Load (lbs) DAD estimated Measured Load Cell 15 Figure 619. Comparison of measured and estimated reacti ons for wind direction 9 0o

PAGE 111

111 CHAPTER 7 CONCLUSION AND RECOMMENDATIONS This c h apter provides a summary of work undertaken and key findings made. Significant conclusions which can be drawn from the study as well as recommendation s for fu ture research in this study area are discussed. Summary Light frame d wood structures (LFWS) are by far the largest contributor to monetary losses associated hurricane damages. Inadequate load transfer mechanism provided in wood constructions account significantly to structural failures of such buildings. The need for studies to provide better understanding of wind load paths on LFWS buildings cannot be overemphasize This research was aimed at estimating wind induced structural loads transferred through roof to wall and wall to foundation connections on LFWS The estimations were based on D atabase A ssisted D esign (DAD) methodology, developed by the National Institute of Standards and Technology (NIST). This methodology, which utilizes large databases of aerodynamic pressures and climatological information, has been used by other researchers to predict structural responses due to wind forces on steel portal frame buildings. This current study, apart from extending the application of the DAD approach to LFWS also demonstrates the validity of this methodology to adequately evaluate structural reactions on LFWS The project was accomplished in two phases: a hybridized analytical approach and an experimental approach. In the former, spatially distributed pressure coefficients were der ived from wind tunnel data for a 1/50 scale model house. These pressure coefficients (time histories) were combined with structural influence coefficients (developed on a 1/3 scale model wood house) through computer based analysis to generate structural re actions. Peak uplift reactions were estimated from a Lieblein BLUE fitted distribution of the measured peak reactions. These

PAGE 112

112 were then compared to wind design loads based on ASCE 7 provisions for both Main Wind Force Resisting Systems (MWFRS) and Component s and Cladding (C&C). It was realized that using MWFRS pressures underestimated the reactions by up to 32% while C&C provisions resulted in highly conservative estimates (up to 60% overestimation). In the experimental study, the 1/3 scale wood house model was instrumented with pressure and load sensors. The house was subjected to wind forces while load and pressure measurements were simultaneously taken. Based on the DAD approach, structural reactions were estimated using measured roof pressures and results were compared to directly measured structural loads. Even though significant ly fluctuating reaction records were obtained using the DAD in comparison to the directly measured reaction records, the two quantities were highly correlated. Again, good agreeme nts were found comparing mean and peak values of the estimated and measured reactions. Conclusions The conclusions of this study can be summarized as follows: 1. Local peak pressure coefficients derived from wind tunnel analysis are considerably higher in mo st cases than ASCE 7 05 component and cladding external pressure coefficients However, an excellent agreement is observed between the wind tunnel area averaged pressure coefficients and the ASCE 7 standard provisions. 2. ASCE 7 05 MWFRS provisions produced lower peak reactions ( average of 21% lower) than predictions based on DAD methodology while predictions based on ASCE 705 components and cladding were generally higher .(average of 33 % higher) than DAD based reactions. 3. Despite limited match of the wind g enerated in t he experiment to realistic wind flows, overall the roof pressure distributions on the 1/3 scale house were reasonably matched to the wind tunnel pressure spatial distributions in the literature. 4. The structural reaction time history obtained us ing the DAD method had greater dynamic content than the time history of directly measured structural reactions. Despite this fact there was still good agreement between the peak values and the mean values of DAD results and the measured reactions.

PAGE 113

113 5. DADbase d reactions were highly correlated with directly measured structural reactions 6. Both the analytical and experimental components of this study confirmed that maximum uplifts loads are transferred through the gable end supports 7. Lastly, a wider spread of loa d sharing through wall to foundation connection was observed in both analytical and experimental studies. This may be due to diaphragm (deep beam) actions of the building walls. Finally, the study has not just extended the application of DAD methodology to LFWS but has evaluated its validity to adequately evaluate structural reactions on LFWS through an experimental study. Recommendation Notwithstanding the significant findings reported in this study and its possible influence on design practices for LFWS, further research is needed in order to fully comprehend wind load paths in LFWS. Recommendations for future research in this discipline are discussed below. 1. The study house is a simple rectangular gable roof structure. Most typical residential structures have complex shapes which may result in different pressure distributions on houses and subsequently different wind load paths. It is recommended that studies are done on LFWS with complex building shapes and different configurations. 2. In future experimental studies, load cells should be installed around the whole perimeter of the 1/3 scale building at roof to wall and wallto foundation interfaces. This will ensure that wind load paths are not unduly affected by the stiffness of the load cells. 3. Also, the en tire roof of the house should be equipped with pressure sensors so that the DAD methodology could be validated at all critical connections. 4. A better boundary layer condition should be created for future experiments so that comparison can be made between fi eld and wind tunnel measurements. 5. Reliability studies should be carried out to quantify the uncertainty parameters for load estimations on LFWS based on DAD approach.

PAGE 114

114 APPENDIX A MEAN, RMS AND EXTREM E VALUES OF WIND TUN NEL PRESSURE COEFFICIENTS Table A 1. Peak mean and RMS pressure coefficients of selected pressure taps 0 o 45 o 90 o Press. Taps Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS 1 2.74 0.37 0.16 0.43 4.14 0.81 0.17 0.92 1.74 0.25 0.12 0.28 2 2. 43 0.37 0.21 0.43 4.13 0.83 0.15 0.94 1.67 0.22 0.11 0.25 3 2.22 0.37 0.35 0.43 4.39 0.78 0.22 0.90 1.60 0.23 0.10 0.26 4 2.16 0.36 0.34 0.42 3.80 0.79 0.19 0.89 2.02 0.22 0.21 0.25 5 2.10 0.32 0.43 0.39 3.56 0.71 0.23 0.80 1.19 0 .20 0.12 0.23 6 1.84 0.22 0.49 0.28 2.65 0.60 0.23 0.67 1.02 0.20 0.10 0.23 7 1.79 0.15 0.48 0.21 2.17 0.52 0.16 0.58 0.88 0.20 0.09 0.22 8 1.42 0.13 0.46 0.18 1.86 0.49 0.06 0.53 0.93 0.22 0.06 0.24 9 1.20 0.18 0.33 0.21 1.72 0.5 0 0.03 0.54 1.06 0.30 0.00 0.31 10 1.13 0.11 0.39 0.15 1.60 0.45 0.01 0.48 1.01 0.23 0.04 0.25 11 0.86 0.09 0.28 0.13 1.46 0.42 0.01 0.44 1.11 0.25 0.03 0.27 12 0.86 0.07 0.29 0.11 1.40 0.33 0.09 0.36 0.96 0.20 0.08 0.22 13 0.72 0.08 0.31 0.11 1.27 0.31 0.08 0.34 0.94 0.21 0.08 0.23 14 0.70 0.08 0.24 0.11 1.19 0.28 0.09 0.31 0.94 0.21 0.06 0.23 15 0.65 0.06 0.27 0.09 1.22 0.24 0.09 0.27 0.97 0.20 0.08 0.22 16 0.60 0.05 0.25 0.09 1.16 0.22 0.10 0.25 1.04 0. 20 0.09 0.22 17 0.55 0.05 0.24 0.08 1.20 0.18 0.16 0.21 1.09 0.20 0.10 0.23 18 0.56 0.08 0.21 0.10 1.12 0.22 0.14 0.24 1.19 0.23 0.08 0.25 19 0.58 0.08 0.20 0.10 1.22 0.21 0.12 0.23 1.35 0.23 0.09 0.25 20 0.56 0.09 0.18 0.11 1.36 0.21 0.12 0.24 1.41 0.24 0.08 0.26 21 0.53 0.07 0.19 0.09 1.50 0.20 0.14 0.24 1.43 0.23 0.10 0.26 22 0.78 0.84 2.34 1.04 1.22 0.51 2.06 0.65 1.06 0.62 2.10 0.76 262 0.40 0.10 0.13 0.11 0.52 0.13 0.12 0.14 0.74 0.15 0.36 0.18 263 0.44 0.09 0.16 0.11 0.50 0.11 0.21 0.13 0.64 0.11 0.43 0.14 264 0.45 0.07 0.18 0.09 0.46 0.10 0.23 0.12 0.57 0.08 0.45 0.12 265 0.44 0.07 0.20 0.09 0.48 0.10 0.24 0.12 0.55 0.08 0.44 0.12 266 0.45 0.07 0.22 0.09 0.48 0.08 0.27 0.10 0.45 0.05 0.44 0.10 267 0.50 0.06 0.23 0.08 0.42 0.08 0.29 0.10 0.43 0.04 0.46 0.09 268 0.57 0.05 0.25 0.08 0.48 0.08 0.28 0.11 0.46 0.05 0.47 0.10 269 0.63 0.06 0.26 0.09 0.45 0.09 0.29 0.11 0.46 0.06 0.45 0.10

PAGE 115

115 Table A 1. (Contd) 0o 45o 90o Press. Taps Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS 270 0.56 0.07 0.24 0.09 0.45 0.08 0.29 0.11 0.43 0.05 0.46 0.10 273 0.77 0.08 0.33 0.11 0.47 0.09 0.28 0.11 0.48 0.07 0.47 0.11 274 0.86 0.13 0.34 0.16 0.52 0.13 0.26 0.15 0.53 0.12 0.41 0.14 275 1.02 0.14 0.33 0.17 0.50 0.12 0.27 0.14 0.49 0.10 0.43 0.13 276 1.09 0.12 0.44 0.17 1.15 0.05 0.47 0.08 0.39 0.01 0.50 0.08 277 1.30 0.20 0.46 0.25 1.34 0.04 0.50 0.0 8 0.40 0.02 0.52 0.09 278 2.15 0.32 0.36 0.38 1.90 0.07 0.40 0.12 0.58 0.05 0.47 0.10 279 2.15 0.37 0.40 0.42 2.28 0.09 0.53 0.23 0.48 0.04 0.53 0.10 280 2.29 0.37 0.28 0.43 2.29 0.16 0.58 0.33 0.54 0.04 0.56 0.10 281 2.22 0.36 0. 26 0.42 2.08 0.20 0.66 0.35 0.64 0.06 0.54 0.12 282 2.13 0.37 0.21 0.42 1.90 0.21 0.73 0.34 0.81 0.12 0.51 0.17 367 0.57 0.09 0.21 0.11 1.26 0.16 0.28 0.20 1.59 0.25 0.60 0.31 368 0.63 0.08 0.20 0.10 1.28 0.17 0.33 0.20 1.61 0.22 0 .67 0.29 369 0.63 0.07 0.20 0.10 1.18 0.16 0.38 0.20 1.58 0.20 0.69 0.28 370 0.63 0.07 0.21 0.09 1.31 0.16 0.44 0.21 1.59 0.20 0.76 0.28 371 0.69 0.07 0.23 0.09 1.37 0.17 0.43 0.21 1.50 0.20 0.80 0.27 372 0.66 0.07 0.25 0.09 1.31 0.17 0.41 0.22 1.37 0.20 0.73 0.27 373 0.65 0.07 0.24 0.10 1.47 0.17 0.47 0.22 1.51 0.20 0.75 0.27 374 0.69 0.06 0.26 0.09 1.43 0.17 0.47 0.22 1.43 0.20 0.78 0.27 375 0.83 0.07 0.32 0.10 1.35 0.18 0.46 0.23 1.43 0.21 0.72 0.27 376 0.92 0.07 0.34 0.10 1.52 0.18 0.50 0.23 1.48 0.20 0.73 0.27 377 0.97 0.15 0.26 0.17 1.40 0.21 0.46 0.26 1.56 0.31 0.67 0.36 378 0.96 0.12 0.30 0.15 1.50 0.21 0.45 0.25 1.58 0.25 0.71 0.30 379 1.02 0.18 0.24 0.21 1.43 0.25 0.37 0.29 1.60 0.31 0.71 0.35 380 1.17 0.12 0.41 0.17 1.29 0.14 0.50 0.20 1.32 0.18 0.86 0.25 381 1.26 0.15 0.39 0.20 1.38 0.13 0.53 0.19 1.45 0.18 0.86 0.25 382 1.45 0.18 0.36 0.24 1.32 0.11 0.56 0.18 1.38 0.16 0.90 0.25 383 1.60 0.24 0.41 0.29 1.20 0.11 0.63 0.19 1.44 0.18 0.91 0.26 384 2.01 0.28 0.42 0.34 1.23 0.11 0.70 0.20 1.47 0.19 0.85 0.27 385 2.25 0.32 0.43 0.39 1.73 0.08 0.83 0.21 1.62 0.18 0.90 0.27 386 2.52 0.39 0.44 0.47 1.86 0.02 0.85 0.19 1.75 0.18 0.8 5 0.27 387 3.71 0.47 0.39 0.59 3.91 0.04 0.89 0.33 1.45 0.18 0.73 0.26

PAGE 116

116 APPENDIX B MEASURED STATISTICAL VALUES OF VERTICAL REACTIONS DERIVED FROM WIND TUNNEL DATA

PAGE 117

117 Table B 1. Measured peak mean and RMS reactions (lbs) for sample 2 Azimuth 0o 45o 90o 135o 180o Load C ells Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS 1 812 178 101 204 375 99 170 117 411 69 352 99 439 10 7 151 124 272 28 177 51 2 878 207 72 234 400 119 120 136 376 69 318 95 403 100 140 116 254 28 159 49 3 1031 262 58 292 577 146 99 164 408 87 300 112 440 117 145 131 256 35 134 56 4 1220 301 50 333 730 146 155 170 423 95 310 1 21 468 124 157 139 266 38 134 59 5 1008 285 19 313 869 183 47 207 382 106 176 122 365 123 104 134 225 35 104 51 6 374 99 25 109 212 64 31 71 163 40 115 48 176 53 50 58 111 14 56 23 7 204 52 16 58 107 34 20 38 92 21 68 26 1 00 29 29 32 62 8 33 13 8 300 84 15 92 192 51 25 58 120 31 91 38 137 40 39 44 85 11 39 18 9 688 195 24 214 484 120 50 136 268 71 187 85 296 90 86 99 185 26 87 40 10 1044 299 30 327 785 197 60 221 414 115 245 134 433 139 121 152 273 39 127 60 11 1423 407 41 446 1071 268 80 302 570 159 331 185 596 191 165 209 374 54 173 82 12 1270 363 36 397 952 238 72 268 508 141 298 164 531 170 147 186 332 48 154 73 13 372 106 10 116 280 71 19 80 151 42 84 49 155 50 42 55 97 14 44 21 14 49 14 2 15 38 10 2 11 21 6 12 7 20 7 5 7 13 2 6 3 15 1476 458 61 504 1644 315 119 374 669 175 220 201 558 195 127 212 343 58 162 83 19 1286 323 98 360 1024 344 58 369 698 183 14 198 507 163 64 176 243 44 102 60 20 1303 300 100 331 910 325 31 348 647 186 19 200 507 168 74 182 229 40 129 59

PAGE 118

118 Table B 2. Measured peak mean and RMS reactions (lbs) for sample 3 Azimuth 0o 45o 90o 135o 180o Load C ells Neg. Peak Mean Pos. Peak RMS Neg. Peak load cells Neg. Peak Mean Pos. Peak RMS Neg. Peak load cells Neg. Peak Mean Pos. Peak RMS Neg. Peak load cells Neg. Peak Mean 1 771 179 104 205 399 100 184 119 384 72 495 101 356 77 184 95 317 33 153 55 2 858 2 06 96 233 494 123 164 140 405 72 440 97 342 73 163 89 287 33 147 53 3 1061 259 85 289 601 151 156 171 478 90 437 114 357 86 148 101 315 41 142 60 4 1327 297 54 328 779 153 170 179 519 98 477 123 385 91 157 107 320 44 144 64 5 1227 281 19 308 1041 192 95 219 454 109 243 125 336 91 66 103 262 40 135 56 6 361 98 26 109 249 66 51 74 180 41 159 49 146 39 51 45 133 17 64 25 7 200 52 17 58 128 35 31 39 98 22 95 27 82 21 31 25 75 9 35 14 8 293 83 17 91 219 53 44 61 144 32 121 38 114 30 40 34 100 13 49 19 9 696 193 27 212 553 125 100 143 328 73 244 87 249 67 79 76 216 30 107 44 10 1087 295 20 324 903 205 134 233 498 118 313 136 371 103 101 117 320 45 159 65 11 1479 402 22 441 1237 280 178 318 681 164 421 188 511 142 138 160 438 62 218 89 12 1321 358 20 393 1104 248 161 282 606 145 378 167 457 125 125 142 390 55 193 79 13 389 105 4 115 324 75 44 84 177 43 106 50 135 37 35 42 114 16 56 23 14 53 14 1 15 43 11 5 12 24 6 14 7 19 5 5 6 15 2 7 3 15 1723 450 31 497 1980 334 231 402 720 180 349 204 507 145 79 163 408 66 214 90 19 1400 321 49 359 1139 352 19 379 667 186 34 200 542 122 36 136 275 49 113 65 20 1160 29 9 97 330 1104 335 9 361 657 189 42 203 505 125 44 139 278 46 140 64

PAGE 119

119 Table B 3. Measured peak mean and RMS reactions (lbs) for sample 4 Azimuth 0o 45o 90o 135o 180o Load Cells Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS 1 787 174 86 201 873 98 257 117 353 74 349 101 358 86 200 103 291 34 120 53 2 912 202 73 229 979 120 228 137 324 73 307 97 331 81 16 8 97 289 34 108 52 3 1162 256 68 287 718 147 237 167 361 90 309 113 390 95 143 110 295 41 112 59 4 1471 295 67 328 926 150 264 174 418 97 324 122 431 101 139 116 317 45 122 63 5 1291 279 47 308 874 186 148 211 365 107 151 123 333 101 98 112 317 41 91 55 6 409 97 21 107 267 64 85 72 148 41 115 49 138 43 52 49 117 17 45 24 7 212 51 13 57 142 34 48 38 81 22 69 27 78 23 33 27 66 9 26 13 8 354 82 15 91 220 52 70 59 117 32 89 38 112 33 39 37 90 1 3 34 19 9 837 191 35 211 518 121 154 138 264 73 179 87 249 74 74 83 198 31 73 43 10 1266 293 48 322 847 199 203 224 408 117 226 135 367 114 107 127 292 46 107 64 11 1720 399 61 438 1159 271 273 306 557 162 304 187 501 157 145 175 399 63 146 87 12 1536 355 54 391 1028 240 245 271 496 144 273 166 449 139 128 155 354 56 130 78 13 448 104 14 114 305 72 67 81 146 43 76 49 131 41 37 46 103 16 38 23 14 59 14 2 15 44 10 7 12 20 6 10 7 19 6 5 6 14 2 5 3 15 2008 449 94 497 1757 322 293 382 616 177 205 201 535 161 134 178 458 67 135 89 19 1378 318 100 357 1132 340 5 366 590 181 11 194 463 135 80 149 309 51 75 65 20 1073 295 71 326 1012 323 21 348 560 185 10 198 470 138 91 1 53 301 49 89 64

PAGE 120

120 Table B 4. Measured peak mean and RMS reactions (lbs) for sample 5 Azimuth 0o 45o 90o 135o 180o Load Cells Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Pe ak Mean Pos. Peak RMS 1 745 158 121 185 499 101 151 120 396 45 335 80 383 88 213 106 242 33 142 54 2 851 188 102 216 516 124 134 141 373 44 286 76 345 84 171 99 241 33 128 53 3 1016 240 97 270 628 152 156 17 1 429 56 239 86 372 98 136 113 271 41 132 59 4 1100 276 102 310 802 153 176 177 449 59 255 92 400 105 118 120 270 44 136 63 5 970 261 65 291 850 190 117 215 380 73 147 93 327 104 51 116 220 40 118 55 6 360 89 34 100 266 66 62 74 175 27 95 37 151 44 53 50 102 17 59 25 7 199 47 20 53 139 35 35 39 97 14 59 21 86 24 34 27 57 9 34 14 8 298 76 27 85 226 53 49 60 134 21 69 29 117 34 38 38 80 13 44 19 9 665 178 59 198 542 125 109 141 292 48 132 65 254 76 66 85 173 30 97 43 10 1006 273 78 302 877 204 152 229 441 79 168 102 370 117 76 131 253 46 142 64 11 1373 371 103 412 1191 279 205 313 605 109 228 141 510 161 105 180 346 62 195 88 12 1227 331 92 367 1057 247 183 277 537 97 204 125 453 143 94 160 308 55 173 78 13 360 97 26 107 314 74 51 83 159 29 58 37 133 42 26 47 90 16 50 23 14 48 13 4 14 44 11 6 12 22 4 8 5 19 6 4 6 12 2 6 3 15 1489 420 111 470 1607 327 220 386 611 125 259 154 526 165 74 183 349 66 182 89 19 1255 296 82 335 1298 354 12 382 597 141 61 158 549 138 51 153 255 50 105 65 20 1053 272 71 304 1213 335 38 361 545 141 67 158 558 141 56 157 249 47 140 64

PAGE 121

121 Table B 5. Measured peak mean and RMS rea ctions (lbs) for sample 6 Azimuth 0o 45o 90o 135o 180o Load Cells Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS 1 683 159 95 189 6 39 106 168 124 366 56 281 89 427 88 294 107 237 34 127 54 2 795 189 88 220 479 125 146 142 345 55 285 85 405 84 265 100 223 35 127 53 3 1055 242 81 275 715 148 140 168 379 68 311 96 382 98 233 114 239 42 120 60 4 1301 279 90 315 928 147 173 172 376 72 348 102 390 104 214 121 255 46 114 64 5 1180 264 73 295 962 185 71 210 343 85 209 102 387 105 90 117 213 42 88 56 6 362 90 36 102 263 66 49 74 156 32 103 42 156 44 88 51 101 17 50 25 7 186 48 20 54 152 35 29 39 87 17 59 23 88 24 54 28 57 9 29 14 8 316 77 26 87 236 53 39 60 120 25 86 32 122 34 65 39 79 14 36 19 9 753 180 56 201 576 123 82 140 265 57 184 73 274 76 128 86 174 31 77 44 10 1169 275 80 307 906 201 105 227 406 92 244 113 423 118 159 132 255 47 111 65 11 1596 375 111 418 1229 275 140 310 555 128 330 156 582 163 216 182 347 65 153 89 12 1424 334 96 373 1099 243 127 274 493 113 297 139 516 144 194 161 309 58 135 79 13 419 98 2 7 109 322 73 34 82 146 34 84 41 153 43 54 48 89 17 39 23 14 56 13 4 14 43 10 3 12 20 5 11 6 21 6 6 6 11 2 5 3 15 1883 425 119 475 1649 325 160 385 547 142 310 169 590 168 133 186 335 69 136 91 19 1124 300 139 342 1192 342 1 370 531 157 44 172 536 140 61 155 226 52 82 66 20 1162 276 133 310 1072 325 4 351 494 157 58 172 531 144 72 160 220 49 105 65

PAGE 122

122 Table B 6. Measured peak mean and RMS reactions (lbs) for sample 7 Azimuth 0o 45o 90o 135o 180o Load Cells Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS 1 826 159 142 190 453 108 176 125 416 62 285 92 416 91 221 109 232 33 13 6 54 2 899 188 147 219 481 127 142 144 395 61 217 88 416 86 164 102 229 34 125 52 3 1003 240 152 273 544 151 122 170 434 76 222 102 373 100 118 116 251 41 132 59 4 1068 276 172 312 729 150 139 174 446 80 266 109 393 107 125 1 22 251 44 134 63 5 1020 261 95 292 835 187 77 212 409 93 186 111 380 106 57 117 214 40 106 55 6 369 90 54 102 219 67 44 75 187 35 79 44 154 45 48 51 108 17 55 25 7 205 47 31 54 122 36 27 40 104 19 47 24 86 25 31 28 60 9 3 1 14 8 295 76 42 86 193 54 36 60 144 27 62 35 118 35 32 39 79 13 42 19 9 659 178 92 200 484 125 75 141 318 63 142 78 265 78 57 87 171 30 90 43 10 998 273 121 305 791 204 99 229 482 102 203 122 410 120 75 134 248 46 130 64 11 1361 371 161 416 1075 278 135 312 664 141 273 169 564 165 101 184 333 62 177 88 12 1208 331 143 370 958 246 122 277 590 125 246 150 500 146 90 163 294 56 157 78 13 351 97 40 108 284 74 34 83 174 37 71 44 148 43 25 48 85 1 6 46 23 14 45 13 5 14 40 11 4 12 24 5 10 6 21 6 3 6 12 2 6 3 15 1549 420 141 471 1672 327 166 387 650 157 290 185 613 169 64 186 351 67 171 89 19 1259 293 73 336 1199 343 5 371 576 166 72 181 540 140 39 155 252 50 107 64 2 0 1118 272 114 306 1033 326 13 353 570 168 69 183 522 144 49 159 276 47 134 64

PAGE 123

123 Table B 7. Measured peak mean and RMS reactions (lbs) for sample 8 Azimuth 0o 45o 90o 135o 180o Load Cells Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Pea k RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS 1 846 165 109 194 432 115 164 131 386 62 318 93 398 88 210 105 265 33 137 55 2 917 195 100 224 431 134 130 149 339 61 299 8 9 349 83 176 99 266 33 126 54 3 1096 248 98 279 635 157 131 175 377 76 301 103 413 98 169 113 284 41 121 61 4 1185 285 108 318 832 156 150 179 411 82 337 110 461 104 202 120 309 45 120 66 5 1085 269 61 297 859 193 77 216 433 95 173 113 440 104 162 116 272 41 104 57 6 389 93 35 104 225 70 38 77 155 36 97 45 179 44 63 50 105 17 50 25 7 214 49 20 55 112 37 24 41 83 19 59 24 98 24 38 27 60 9 29 14 8 327 79 29 88 203 56 34 62 132 28 79 35 139 34 50 38 83 13 36 20 9 750 184 63 204 504 130 67 145 309 64 164 79 309 76 111 85 188 31 80 44 10 1130 282 83 311 803 211 91 234 483 103 216 123 471 117 162 131 278 46 121 66 11 1538 384 107 423 1093 288 123 320 656 142 294 170 647 161 222 180 379 63 167 91 12 1371 342 97 377 975 255 111 283 584 126 264 151 576 143 197 160 339 56 148 81 13 400 100 27 110 286 76 31 84 172 38 75 45 170 42 57 47 98 16 43 23 14 53 13 4 14 39 11 4 12 24 5 10 6 24 6 7 6 13 2 6 3 15 1863 432 114 478 1647 336 132 393 853 159 240 187 702 166 233 184 383 68 160 92 19 1372 308 115 349 1172 351 16 376 779 168 49 184 497 138 120 153 224 52 117 67 20 1189 283 92 314 1158 334 5 357 662 170 56 186 514 142 136 158 235 49 144 67

PAGE 124

124 APPENDIX C PRESSURE COEFFICIENT S MEASURED IN THE 1/ 3SCALE HOUSE TEST Table C 1. Measured peak mean and RMS pressure coefficients for test repeat 1 0 o 45 o 90 o Press. Taps Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Roof Pressure Taps 1 1.47 0.52 0.03 0.55 0.15 0.05 0.22 0.07 0.90 0.16 0.44 0.26 2 2.42 0.50 1.05 0.54 1.90 0.06 1.55 0.34 3.50 0.36 1.66 0.50 3 2.90 0.47 1.85 0.51 2.77 0.82 1.29 0.90 3.26 0.49 1.61 0.57 4 2.62 0.44 0.92 0.49 1.94 0.56 1.73 0.60 3.00 0.54 1.31 0.60 5 2.05 0.49 0.08 0.52 1.54 0.07 1.24 0.20 1.86 0.01 1.51 0.21 6 1.85 0.50 0.22 0.53 0.04 0.12 0.26 0.12 1.63 0.06 1.47 0.25 7 2.48 0.45 0.71 0.48 0.15 0.13 0. 53 0.14 1.92 0.01 1.51 0.25 8 1.86 0.41 1.48 0.44 0.63 0.11 0.81 0.15 2.05 0.05 2.17 0.24 9 0.98 0.40 0.01 0.41 3.30 0.57 1.91 0.68 0.81 0.10 1.09 0.16 10 1.31 0.42 0.31 0.44 0.42 0.03 0.39 0.06 0.76 0.10 0.71 0.12 11 2.32 0.45 0.43 0 .47 0.14 0.05 0.24 0.06 0.59 0.14 0.70 0.16 12 2.05 0.46 0.47 0.48 0.15 0.03 0.24 0.05 0.08 1.15 1.98 1.15 13 0.79 0.33 0.20 0.34 3.62 0.82 1.00 0.88 1.01 0.11 0.46 0.15 14 1.02 0.37 0.06 0.38 1.38 0.06 1.04 0.16 0.59 0.09 0.47 0.11 1 5 1.39 0.41 0.11 0.42 0.47 0.08 0.18 0.10 0.44 0.05 0.42 0.07 16 1.52 0.46 0.21 0.48 0.32 0.09 0.11 0.10 0.56 0.07 0.71 0.09 17 1.03 0.32 0.01 0.33 3.18 1.12 1.08 1.17 0.41 0.01 0.77 0.07 18 1.27 0.36 0.05 0.37 2.53 0.43 1.64 0.50 0.32 0.04 0.47 0.07 19 2.21 0.45 0.03 0.47 0.75 0.38 0.04 0.39 0.39 0.04 0.36 0.07 20 1.54 0.40 0.09 0.41 1.28 0.34 0.42 0.36 0.29 0.04 0.40 0.06 21 0.92 0.34 0.02 0.35 6.61 1.71 2.91 1.86 0.96 0.05 0.36 0.08 22 1.20 0.36 0.04 0.37 3.70 0.67 1.57 0.80 0.65 0.03 0.21 0.06 23 1.86 0.40 0.08 0.42 2.46 0.54 1.04 0.59 0.55 0.02 0.21 0.05 24 2.02 0.34 0.01 0.36 3.23 0.89 2.05 0.93 0.82 0.04 0.24 0.08 25 1.53 0.43 0.27 0.44 1.58 0.26 1.19 0.33 0.66 0.05 0.15 0.08 Wall Pressure Taps 26 0.58 0.19 0.17 0.20 0.64 0.83 1.07 0.83 0.41 0.46 1.73 0.47 27 1.14 0.30 0.00 0.32 0.39 0.56 0.73 0.56 0.23 0.92 1.26 0.92 28 0.46 0.76 0.98 0.76 0.54 0.72 0.88 0.72 0.93 0.43 0.14 0.44 29 0.93 1.08 1.26 1.08 0.11 0.36 0. 60 0.36 0.64 0.04 0.57 0.11 Interior Pressure Taps 30 0.01 0.00 0.01 0.00 0.00 0.01 0.02 0.01 0.00 0.01 0.02 0.01

PAGE 125

125 Table C 2. Measured peak mean and RMS pressure coefficients for test repeat 2 0o 45o 90o Press. Taps Neg. Peak Mean Pos. Peak RMS Neg Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Roof Pressure Taps 1 1.66 0.53 0.06 0.57 0.16 0.03 0.20 0.05 0.99 0.12 0.56 0.23 2 2.43 0.49 1.08 0.53 1.98 0.05 1.58 0.30 2.35 0.29 1.69 0.44 3 2.92 0.46 1.82 0.51 2.64 0.80 1.11 0 .87 2.55 0.40 1.96 0.49 4 2.63 0.44 1.40 0.48 2.30 0.56 0.56 0.60 2.65 0.48 1.60 0.55 5 2.46 0.50 0.34 0.53 1.66 0.05 1.56 0.20 1.51 0.02 1.55 0.18 6 1.96 0.50 0.18 0.54 0.06 0.10 0.26 0.10 1.54 0.07 1.67 0.22 7 4.54 0.45 0.80 0.48 0 .29 0.10 0.35 0.11 1.35 0.01 1.39 0.23 8 2.80 0.41 1.08 0.44 0.67 0.09 0.81 0.12 1.85 0.02 1.85 0.21 9 1.13 0.40 0.02 0.42 3.27 0.55 2.02 0.67 0.75 0.10 0.94 0.15 10 1.19 0.43 0.11 0.45 0.36 0.01 0.36 0.06 0.74 0.10 0.81 0.12 11 1.84 0.45 0.05 0.47 0.20 0.03 0.28 0.05 0.56 0.14 0.69 0.15 12 1.51 0.46 0.34 0.48 0.20 0.01 0.21 0.04 1.49 0.20 0.73 0.47 13 0.76 0.34 0.00 0.35 2.97 0.80 1.88 0.87 0.92 0.10 0.50 0.13 14 0.98 0.38 0.07 0.39 1.16 0.08 1.16 0.17 0.37 0.0 9 0.62 0.11 15 1.36 0.41 0.08 0.43 0.47 0.10 0.20 0.12 0.34 0.06 0.34 0.08 16 1.52 0.46 0.38 0.48 0.33 0.10 0.13 0.12 0.43 0.08 0.57 0.09 17 0.75 0.35 0.04 0.36 3.47 1.06 0.67 1.11 0.50 0.02 0.43 0.07 18 1.05 0.35 0.06 0.36 2.10 0 .42 1.12 0.49 0.36 0.03 0.35 0.06 19 1.97 0.44 0.42 0.46 0.75 0.39 0.19 0.40 0.35 0.04 0.31 0.06 20 1.89 0.39 1.14 0.41 1.04 0.35 0.25 0.36 0.39 0.05 0.49 0.07 21 0.95 0.33 0.04 0.34 5.94 1.69 2.75 1.85 0.73 0.04 0.23 0.08 22 0.91 0 .36 0.03 0.36 3.67 0.64 1.71 0.76 1.54 0.03 0.36 0.06 23 1.59 0.39 0.02 0.41 2.40 0.54 1.10 0.59 0.67 0.01 0.20 0.06 24 0.78 0.33 0.06 0.34 3.00 0.91 0.98 0.95 0.99 0.05 0.24 0.09 25 1.60 0.41 0.08 0.42 1.76 0.23 1.44 0.30 0.97 0 .04 0.18 0.08 Wall Pressure Taps 26 0.53 0.18 0.18 0.20 0.59 0.77 1.10 0.77 0.41 0.40 1.89 0.42 27 1.08 0.30 0.18 0.32 0.40 0.52 0.74 0.52 0.27 0.89 1.30 0.90 28 0.41 0.74 0.99 0.75 0.55 0.72 0.91 0.72 1.01 0.38 0.00 0.39 29 0.71 1.04 1.23 1.05 0.09 0.33 0.56 0.34 0.53 0.06 0.67 0.12 Interior Pressure Taps 30 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.08 0.00 0.02 0.01

PAGE 126

126 Table C 3. Measured peak mean and RMS pressure coefficients for test repeat 3 0o 45o 90o Press. Taps Neg. Peak Mean P os. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Roof Pressure Taps 1 1.63 0.62 0.09 0.66 0.14 0.05 0.21 0.06 0.90 0.13 0.40 0.25 2 3.24 0.55 1.12 0.60 2.03 0.03 1.55 0.28 3.10 0.33 2.03 0.51 3 2.41 0.49 1.62 0.54 2 .68 0.78 1.32 0.86 3.23 0.46 2.26 0.56 4 2.95 0.46 1.74 0.52 2.70 0.55 0.65 0.60 3.06 0.52 2.07 0.59 5 2.28 0.59 0.03 0.62 1.73 0.05 1.36 0.20 1.87 0.02 1.32 0.21 6 1.93 0.59 0.36 0.63 0.05 0.11 0.27 0.11 1.87 0.09 1.54 0.26 7 1.84 0.51 0.89 0.54 0.10 0.12 0.36 0.12 1.87 0.05 1.64 0.26 8 2.58 0.43 1.24 0.47 0.78 0.10 0.71 0.13 1.54 0.00 1.97 0.23 9 1.35 0.45 0.18 0.47 2.86 0.54 1.91 0.66 1.01 0.11 1.05 0.17 10 1.41 0.48 0.05 0.50 0.44 0.01 0.46 0.06 0.89 0.11 0.59 0.13 11 2.53 0.50 0.10 0.52 0.18 0.04 0.30 0.06 0.55 0.16 0.70 0.17 12 2.03 0.50 0.81 0.52 0.20 0.01 0.21 0.04 1.02 0.18 1.08 0.20 13 1.13 0.36 0.06 0.38 2.89 0.79 1.34 0.86 1.13 0.11 0.56 0.15 14 1.05 0.41 0.04 0.42 1.21 0.07 0.89 0 .17 0.61 0.10 0.58 0.12 15 1.33 0.45 0.03 0.46 0.64 0.10 0.28 0.12 0.54 0.06 0.42 0.09 16 1.75 0.50 0.21 0.52 0.33 0.10 0.12 0.11 0.38 0.08 0.41 0.10 17 0.77 0.35 0.24 0.36 3.18 1.06 0.71 1.12 0.50 0.01 0.72 0.08 18 0.82 0.38 0.11 0.38 2.28 0.41 1.14 0.48 0.53 0.04 0.48 0.08 19 2.05 0.47 0.08 0.49 0.78 0.39 0.08 0.40 0.42 0.04 0.33 0.08 20 1.63 0.41 0.04 0.43 1.15 0.35 0.50 0.37 0.42 0.05 0.34 0.08 21 0.77 0.35 0.02 0.36 6.48 1.75 1.92 1.90 1.57 0.05 0.79 0.1 1 22 0.94 0.37 0.05 0.38 3.06 0.62 1.77 0.74 1.22 0.04 0.30 0.10 23 1.23 0.41 0.06 0.42 2.79 0.54 1.06 0.60 1.38 0.02 0.19 0.09 24 1.11 0.34 0.04 0.35 3.41 0.95 1.07 0.99 0.92 0.05 0.42 0.10 25 1.65 0.41 0.10 0.42 1.57 0.21 1.17 0.27 1.04 0.05 0.21 0.10 Wall Pressure Taps 26 0.77 0.19 0.11 0.20 0.57 0.77 0.96 0.77 0.48 0.48 1.31 0.50 27 1.41 0.31 0.39 0.33 0.38 0.52 0.68 0.52 0.50 0.95 1.52 0.95 28 0.49 0.82 1.06 0.82 0.56 0.74 0.90 0.74 1.11 0.42 0.07 0.43 29 0.93 1.08 1.27 1.08 0.14 0.34 0.53 0.35 0.64 0.06 0.72 0.13 Interior Pressure Taps 30 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00

PAGE 127

127 APPENDIX D STRUCTURAL REACTIONS MEASURED IN THE 1/3 SCALE HOUSE TEST Table D 1. Measured peak mean and RMS reactions for test repeat 1 0 o 45 o 90 o Load Cell Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Roof to wall load cells 1 13.5 7.7 0.9 7.9 6.2 3.7 1.3 3.8 1.4 5.0 9.1 5.1 2 10.9 6.9 3.1 7.0 6.6 4. 5 1.7 4.6 5.1 1.0 3.3 1.4 3 13.5 8.0 4.1 8.1 6.8 4.8 2.6 4.8 2.7 0.8 4.2 1.2 4 26.4 19.5 11.1 19.6 13.9 11.1 7.9 11.1 2.9 1.1 4.8 1.4 5 22.3 18.2 14.6 18.3 11.4 8.7 6.6 8.7 17.8 9.8 4.5 9.9 15 20.5 13.4 7.4 13.6 15.3 10.7 6.1 10.7 4.2 0.8 6.6 1.7 16 22.8 18.3 14.5 18.3 19.4 15.1 10.9 15.2 8.6 4.6 2.3 4.7 17 11.4 5.8 2.8 5.9 5.1 2.0 1.5 2.1 3.6 1.4 0.0 1.4 18 11.8 7.1 4.1 7.2 12.2 8.1 4.2 8.2 7.0 1.4 1.9 1.8 19 21.2 13.6 8.7 13.8 17.2 12. 0 5.8 12.1 16.0 7.9 1.0 8.2 20 20.8 14.3 10.1 14.4 17.4 13.1 6.7 13.2 3.9 8.3 13.4 8.4 21 10.3 6.3 2.8 6.4 5.9 2.8 0.7 2.9 7.4 1.0 1.7 1.4 Wallto foundation load cells 6 7.3 2.6 1.5 2.9 12.5 8.0 4.1 8.1 9.5 2.2 5.4 2.8 7 17. 3 11.7 5.6 11.8 4.1 0.3 3.8 1.2 4.5 12.5 20.9 12.7 8 23.3 18.3 11.9 18.3 3.8 0.3 4.3 1.2 5.1 14.1 21.6 14.3 9 24.2 19.7 15.1 19.8 9.6 6.8 4.0 6.9 6.8 1.4 9.1 2.6 10 24.7 20.8 15.2 20.9 15.1 12.4 9.7 12.4 29.3 18.6 10.0 18.8 11 20.6 15.1 8.8 15.3 16.5 12.2 7.6 12.3 37.3 24.6 13.7 24.8 12 21.7 16.6 11.2 16.6 16.0 12.6 8.8 12.7 50.3 31.8 19.1 32.0 13 11.2 7.2 3.7 7.3 8.3 6.5 4.6 6.6 20.1 12.6 7.1 12.7 14 3.1 5.8 8.1 5.8 1.2 0.4 2.0 0.6 6.7 2.6 1.5 2. 8

PAGE 128

128 Table D 2. Measured peak mean and RMS reactions for test repeat 2 0o 45o 90o Load Cell Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Roof to wall load cells 1 15.0 8.3 1.8 8.6 7.1 3.7 0.4 3.8 0. 4 3.1 6.2 3.2 2 10.5 7.3 2.0 7.4 9.3 6.5 3.1 6.5 3.3 0.6 4.3 1.1 3 12.9 8.1 4.6 8.2 7.7 5.3 2.1 5.3 1.5 1.6 4.7 1.8 4 25.5 18.4 10.1 18.6 17.0 13.3 9.3 13.4 2.3 1.6 6.1 1.9 5 19.6 14.9 9.6 15.0 14.5 10.5 7.5 10.5 13.5 8.1 4.2 8.2 15 23.2 14.7 8.5 15.0 18.9 13.1 7.1 13.2 4.0 2.5 7.9 2.9 16 23.3 16.1 8.2 16.2 21.3 16.2 10.9 16.3 6.2 1.8 0.6 1.9 17 11.7 5.7 2.2 5.8 6.3 2.4 1.4 2.6 3.6 0.9 0.4 1.0 18 11.2 7.4 3.5 7.4 13.9 9.7 4.8 9.8 6.9 0.6 2 .3 1.3 19 19.6 13.2 8.2 13.3 19.8 13.6 6.9 13.7 15.4 8.2 2.6 8.4 20 21.2 15.0 9.0 15.1 21.9 15.3 7.5 15.4 4.2 9.6 14.7 9.7 21 11.6 6.4 1.7 6.5 8.0 4.2 0.8 4.3 6.7 0.5 2.0 1.0 Wallto foundation load cells 6 8.0 3.5 1.4 3.7 15 .3 10.7 5.3 10.7 5.9 0.7 7.4 2.0 7 17.8 11.9 4.4 12.0 6.0 1.3 4.7 2.0 6.0 13.0 20.3 13.1 8 23.9 17.7 9.3 17.8 5.6 0.7 4.8 1.6 6.8 14.9 22.7 15.1 9 22.6 17.8 11.4 17.8 11.8 7.7 2.8 7.8 3.4 2.7 10.4 3.3 10 21.5 17.8 10.8 17.9 17.9 14.4 11.3 14.4 23.9 16.0 9.5 16.1 11 19.0 14.0 7.8 14.1 20.9 15.7 11.0 15.8 28.7 18.8 9.2 19.0 12 16.6 12.6 9.0 12.7 19.5 13.8 9.0 13.9 43.2 28.3 13.9 28.6 13 10.0 7.2 4.9 7.2 10.2 7.5 2.4 7.5 18.9 11.2 4.9 11.4 14 2.8 4.6 6.1 4.7 2.5 0.4 3.9 0.8 6.4 1.6 2.2 2.0

PAGE 129

129 Table D 3. Measured peak mean and RMS reactions for test repeat 3 0o 45o 90o Load Cell Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Neg. Peak Mean Pos. Peak RMS Roof to wall load cel ls 1 14.6 8.0 0.2 8.3 7.9 4.2 1.2 4.3 2.4 2.9 5.4 3.1 2 10.0 7.0 1.9 7.1 9.4 6.6 3.6 6.6 5.1 0.7 4.6 1.4 3 13.9 8.1 3.3 8.2 8.0 5.3 2.7 5.4 2.9 1.4 4.6 1.7 4 25.8 17.9 9.0 18.0 16.7 13.2 9.0 13.2 4.7 1.4 5.6 1.9 5 18.6 14 .1 10.0 14.2 13.8 10.4 7.7 10.5 22.4 9.1 4.6 9.3 15 26.5 17.8 10.9 17.9 18.2 12.5 6.3 12.6 4.2 2.3 7.3 2.8 16 18.6 14.4 10.1 14.5 21.2 16.0 11.4 16.1 10.7 2.6 0.2 2.8 17 9.3 5.1 2.0 5.2 6.6 2.4 1.5 2.5 4.7 1.1 0.5 1.3 18 10.8 6.9 4.0 7.0 15.1 9.9 5.4 10.0 8.5 1.0 2.6 1.8 19 21.9 14.9 8.2 15.0 19.6 13.5 6.8 13.6 23.1 9.5 3.2 9.8 20 18.7 13.4 8.0 13.5 20.6 14.3 7.0 14.5 6.2 9.9 13.2 9.9 21 12.1 6.7 3.0 6.8 8.8 4.2 1.1 4.3 9.2 0.5 2.2 1.5 Wal l to foundation load cells 6 8.4 4.1 0.6 4.2 16.1 11.1 6.8 11.2 8.9 0.7 7.8 2.2 7 18.3 11.8 5.8 11.9 6.3 1.8 2.7 2.2 3.5 12.8 19.8 12.9 8 23.5 17.1 10.9 17.2 6.0 1.0 3.4 1.7 5.3 15.1 21.8 15.2 9 20.8 16.8 11.8 16.8 11.4 7.4 3.3 7 .5 8.6 2.8 8.9 3.5 10 19.3 15.6 11.4 15.6 17.4 14.3 10.7 14.4 30.8 17.6 10.4 17.7 11 17.1 12.0 7.1 12.1 20.7 15.6 10.3 15.7 34.2 19.8 11.3 20.0 12 19.0 14.6 10.2 14.7 18.0 14.1 10.3 14.2 52.3 29.5 18.1 29.7 13 12.5 10.1 7 .7 10.1 10.7 8.2 5.9 8.2 21.4 11.3 6.1 11.4 14 1.3 3.5 5.8 3.6 2.3 0.5 1.0 0.7 6.0 1.3 2.5 1.8

PAGE 130

130 LIST OF REFERENCE S ASCE/SEI. (2005). "Minimum design loads for buildings and other structures." 7 05, American Society of Civil Engineers, Reston, VA. Chen, J., Haynes, B. S., and Fletcher, D. F. (2000). "Cobra probe measurements of mean velocities, Reynolds stresses and higher order velocity correlations in pipe flow." Experimental Thermal and Fluid Science 21(4), 206217. Datin, P. L., and Prevatt, D. O. (2007). "Wind Uplift Reactions at Roof to Wall Connections of WoodFramed Gable Roof Assembly." 12th International Conference on Wind Engineering, Australasian Wind Engineering Society, Cairns, Australia. Datin, P. L., Prevatt, D O., and Mensah, A. "Performance Based Wind Engineering: Interaction of Hurricanes with Residential Structures." Engineering Research and Innovation Conference Honolulu, Hawaii. Davenport, A. G., Surry, D., and Stathopoulos, T. (1978). "Wind loads on low rise buildings." Universtiy of Ontario, London, Ontario, Canada. Doudak, G., McClure, G., Smith, I., Hu, L., and Stathopoulos, T. (2005). "Monitoring Structural Response of a Wooden Light Frame Industrial Shed Building to Environmental Loads." Journal of Structural Engineering, 131(5), 794805. FEMA. (2005). "Mitigation Assessment Team Report: Hurricane Ivan in Alabama and Florida Observations, Recommendations, and Technical Guidance." FEMA 489, Federal Emergency Management Agency. Ginger, J. D., and Let chford, C. W. (1993). "Characteristics of large pressures in regions of flow separation." Journal of Wind Engineering and Industrial Aerodynamics 49, 301310. Ginger, J. D., Reardon, G. F., and Whitbread, B. J. (2000). "Wind load effects and equivalent pr essures on low rise house roofs." Engineering Structures 22(6), 638646. Hibbeler, R. C. (2006). Structural Analysis Pearson Prentice Hall, New Jersey. Ho, T. C. E., Surry, D., Morrish, D., and Kopp, G. A. (2005a). "The UWO contribution to the NIST aerod ynamic database for wind loads on low buildings: Part 1. Archiving format and basic aerodynamic data." Journal of Wind Engineering and Industrial Aerodynamics 93(1), 130. Ho, T. C. E., Surry, D., Morrish, D., and Kopp, G. A. (2005b). "The UWO contributio n to the NIST aerodynamic database for wind loads on low buildings: Part 1. Archiving format and basic aerodynamic data." 93(1), 130. Hooper, J. D., and Musgrove, A. R. (1997). "Reynolds stress, mean velocity, and dynamic static pressure measurement by a four hole pressure probe." Experimental Thermal and Fluid Science 15(4), 375383.

PAGE 131

131 Jang, S., Lu, L.W., Sadek, F., and Simiu, E. (2002). "Database assisted wind load capacity estimates for low rise steel frames." Journal of Structural Engineering, 128(12), 15941603. Kopp, G. A., and Chen, Y. (2006). "Database assisted design of low rise buildings: Aerodynamic considerations for a practical interpolation scheme." Journal of Structural Engineering, 132(6), 909917. Lieblein, J. (1974). "Efficient Methods fo Extreme Value Methodology." National Bureau of Standards, Washington, D.C. Liu, Z., Prevatt, D. O., Gurley, K. R., Aponte Bermudez, L., and Reinhold, T. A. (2009). "Field Measurement and Wind Tunnel Simulation of Hurricane Wind Loads on a Single Family Dwe lling." Engineering Structures (in press). Main, J. A., and Fritz, W. P. (2006). "Database Assisted Design for Wind: Concepts, Software, and Examples for Rigid and Flexible Buildings." National Institute of Science and Technology, Gaithersburg, MD. Martin K. G., Gupta, R., Prevatt, D. O., Datin, P. L., and van de Lindt, J. W. (2010). "Evaluation of System Effects and Structural Load Paths in a WoodFramed Structure." Journal of Architectural Engineering, Submitted for review 25 Feb. 2010. Masters, F., Gur ley, K., and Prevatt, D. O. (2008). "Full Scale Simulation of Turbulent Wind Driven Rain Effects on Fenestration and Wall Systems." 3rd International Symposium on Wind Effects on Buildings and Urban Environment, Tokyo, Japan. National Institute of Standard s and Technology, N. (2008). "windPressure DAD software for rigid, gable roof buildings." http://www.itl.nist.gov/div898/winds/wind_pressure/wind_pressure.htm, Date acces sed: 12/20/09. NIST. (2003). "NIST/SEMATECH e Handbook of Statistical Methods." National Institute of Standards and Technology. Retrieved March 8, 2007, from http://www.itl.nist.gov/div898/handbook/ Rigato, A., Chang, P., and Simiu, E. (2001). "Database assisted design, standardization, and wind direction effects." Journal of Structural Engineering, 127(8), 855860. Rosowsky, D., and Schiff, S. (2003). "What are our expectations, objectives, and per formance requirements for wood structures in high wind regions?" Natural Hazards Review 4(3), 144148. Rosowsky, D. V., Walsh, T. G., and Crandell, J. H. (2003). "Reliability of residential woodframe construction from 1900 to present." Forest Products Journal 53(4), 1928.

PAGE 132

132 Sadek, F., Diniz, S., Kasperski, M., Gioffre, M., and Simiu, E. (2004). "Sampling Errors in the Estimation of Peak Wind Induced Internal Forces in Low Rise Structures." Journal of Engineering Mechanics 130(2), 235239. Sadek, F., and S imiu, E. (2002). "Peak Non Gaussian Wind Effects for Database Assisted Low Rise Building Design." Journal of Engineering Mechanics 128(5), 530539. Shepherd, I. C. (1981). Four Hole Pressure Probe for Fluid Flow Measurements in Three Dimensions ." Journal of Fluids Engineering, Transactions of the ASME 103(4), 590594. Simiu, E., and Miyata, T. (2006). Design of Buildings and Bridges for Wind, A Practical Guide for ASCE 7 Standard Users and Designers of Special Structures John Wiley & Sons, Inc., New Jer sey. Simiu, E., Sadek, F., Whalen, T. M., Jang, S., Lu, L.W., Diniz, S. M. C., Grazini, A., and Riley, M. A. (2003). "Achieving safer and more economical buildings through database assisted, reliability based design for wind." Journal of Wind Engineering and Industrial Aerodynamics 91(1215), 15871611. Simiu, E., and Scanlan, R. H. (1996). Wind Effects on Structures Fundamentals and Applications to Design, John Wiley & Sons, Inc., New York. Simiu, E., and Stathopoulos, T. (1997). "Codification of wind loads on buildings using bluff body aerodynamics and climatological data bases." Journal of Wind Engineering and Industrial Aerodynamics 6971, 497506. St. Pierre, L. M., Kopp, G. A., Surry, D., and Ho, T. C. E. (2005). "The UWO contribution to the NIST aerodynamic database for wind loads on low buildings: Part 2. Comparison of data with wind load provisions." Journal of Wind Engineering and Industrial Aerodynamics 93(1), 3159. Stathopoulos, T. (1979). "Turbulent Wind Action on Low rise Buildings," Univ erstiy of Ontario, London, Ontario, London. Suresh Kumar, K., and Stathopoulos, T. (2000). "Wind loads on low building roofs: a stochastic perspective." Journal of Structural Engineering, 126(8), 944956. van de Lindt, J. W., Graettinger, A., Gupta, R., Skaggs, T., Pryor, S., and Fridley, K. J. (2007). "Performance of Wood Frame Structures during Hurricane Katrina." Journal of Performance of Constructed Facilities 21(2), 108116. Watkins, S., Mousley, P., and Vino, G. "The Development and Use of Dynamic Pr essure Probes With Extended Cones of Acceptance (ECA)." 15th Australasian Fluid Mechanics Conference The University of Sydney, Sydney, Australia. Whalen, T., Simiu, E., Harris, G., Lin, J., and David, S. (1998). "The use of aerodynamic databases for the e ffective estimation of wind effects in main wind force resisting

PAGE 133

133 systems:: application to low buildings." Journal of Wind Engineering and Industrial Aerodynamics 7778, 685693. Whalen, T. M., Sadek, F., and Simiu, E. (2002). "Database assisted design for wind: Basic concepts and software development." Journal of Wind Engineering and Industrial Aerodynamics 90(11), 13491368. Whalen, T. M., Shah, V., and Yang, J.S. (2000). "A Pilot Project For Computer Based Design of Low Rise Buildings for Wind Loads The WiLDE LRS User's Manual." Purdue University, West Lafayette, IN. Zisis, I., and Stathopoulos, T. (2009). "WindInduced Cladding and Structural Loads on Low Wood Building." Journal of Structural Engineering, 135(4), 437447.

PAGE 134

134 BIOGR APHICAL SKETCH Akwasi Frimpong Mensah was born and raised in Takoradi, Ghana. He graduated with a Bachelor of Science degree in c ivil e ngineering from the Kwame Nkrumah University of Science and Technology in May 2006. He served as a teaching assistant for a year with the same institution after graduating and later on worked as a Civil/ Structural Engineer with Comptran Engineering and Planning Associates, Accra for another year. He joined University of Florida in pursuit of a master degree in August 2008. He anticipates rec eiving a degree of Master of Science in Civil Engineering in August 2010. The author hopes to practice as an engineer and also lecture in the discipline