Citation
Analytical techniques and field verification method for wind loading analysis of the historic Prowers Bridge

Material Information

Title:
Analytical techniques and field verification method for wind loading analysis of the historic Prowers Bridge
Creator:
Jacobson, Veronica R
Publication Date:
Language:
English
Physical Description:
xvii, 143 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Wind-pressure ( lcsh )
Bridges -- Aerodynamics -- Colorado ( lcsh )
Bridges -- Aerodynamics ( fast )
Wind-pressure ( fast )
Prowers Bridge (Colo.) ( lcsh )
Colorado ( fast )
Colorado -- Prowers Bridge ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 141-143).
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Veronica R. Jacobson.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
123006705 ( OCLC )
ocn123006705
Classification:
LD1193.E53 2006 J32 ( lcc )

Full Text
ANALYTICAL TECHNIQUES AND FIELD VERIFICATION METHOD
FOR WIND LOADING ANALYSIS OF THE
HISTORIC PROWERS BRIDGE
by
Veronica R. Jacobson
B.S., University of Colorado at Denver, 2000
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Civil Engineering
2006


This thesis for the Master of Science
degree by
Veronica Rose Jacobson
has been approved
by
Stephan Durham


Jacobson, Veronica R. (M.S., Civil Engineering)
Structural Response of the Historic Prowers Truss Bridge Due to Lateral Loads
Thesis directed by Associate Professor Kevin L. Rens
ABSTRACT
This thesis examines the analytical techniques and field verification
methods for wind loading analysis on the bottom chord members of the historic
Prowers Bridge. An experimental and analytical study was completed on the 1909
Prowers Bridge over the Arkansas River located near Lamar, Colorado. Field data
from anemometers and clamp-on modular strain transducers was utilized to provide
verification of an analytical deck model of the bridge.
The overall study results indicate that increasing the dead load of the deck
and accounting for the stiffening effect of the deck in the analytical model allows
the windward bottom chord eyebars to satisfy American Association of State
Highway and Transportation Officials (AASHTO) lateral wind loading
requirements. This research provides useful applications to aid rehabilitation and
restoration of historic vehicular steel truss bridges for pedestrian use.
This abstract accurately represents the content of the candidate's thesis. I
recommend its publication.
Signed
/Kevin Lee Rens


DEDICATION
I dedicate this thesis in memory of my mother, Barbara Kaufhold, for her
unconditional support and encouragement over the years to pursue both a bachelor
and a master degree in civil engineering.


ACKNOWLEDGEMENT
I wish to thank:
my advisor, Kevin L. Rens, for his patience and support during these past
years
Steve Durham and Bruce Janson for generously accepting and making the
time to be on my graduate committee
Fred R. Rutz for his generosity in time and patience during these past two
years
my employer for allowing me to rearrange my work schedule and for
providing funding to pursue my master thesis
my family and friends for all their support and encouragement.
I also wish to thank the National Center for Preservation Technology and Training,
under Grant # MT-2210-04-NC-12, for the funding this work and the State
Historical Fund of the Colorado Historical Society under Grant # 2004-M1-019
for the funding of the original equipment. Further, I wish to acknowledge Bent
County, Colorado, the bridge's owner, and the town of Lamar. A special thanks to
Val Moser and Don Brown of Campbell Scientific, Inc. for making remote data
acquisition a reality.
I wish to acknowledge the help and support of Sam Brown, Nick Clough, Kazwan
Elias, Aaron Erftnan, Helen Frey, Shohreh Hamedian, Paul Jacob, and Clint
Krajnik.


TABLE OF CONTENTS
Figures..................................................................x
Tables................................................................xvii
Chapter
1. Overview............................................................1
1.1 Introduction........................................................1
1.2 Previous Research...................................................2
1.3 Goal................................................................3
2. Site Study..........................................................5
2.1 History of Prowers Bridge...........................................5
2.2 Prowers Bridge of Today.............................................7
3. Modeling and Analysis..............................................10
3.1 Modem Analysis.....................................................10
3.2 Prowers Skeleton Model.............................................10
3.3 Prowers Deck Model................................................15
3.3.1 Youngs Modulus (Eeq) of Prowers Deck...........................15
3.3.1.1 Background Information on Youngs Modulus.......................16
3.3.1.2 Youngs Modulus (Eeq baseline) for the Corrugated Steel Deck...18
3.3.2 Modeling Techniques for Plates..................................23
VI


3.3.3 Modeling Results...................................................26
3.4 Skeleton Model Results................................................28
3.5 Deck Model Results....................................................29
4. Field Instrumentation.................................................31
4.1 Introduction..........................................................31
4.2 Instrumentation System of Prowers Bridge..............................32
4.3 Instrumentation Components............................................33
4.3.1 Wind Sensors.......................................................33
4.3.2 Wind Direction Sensor..............................................34
4.3.3 Interval Timer.....................................................35
4.3.4 Strain Transducers.................................................35
4.3.5 Data Logger........................................................38
4.3.6 Wheatstone Bridge..................................................39
4.3.7 Laptop Computer....................................................40
4.3.8 Software...........................................................40
4.3.9 Data Acquisition Program...........................................41
4.3.10 Cables.............................................................42
4.3.11 Digital Cellular Modem with Antenna................................43
4.3.12 Security Box.......................................................45
4.4 Instrumentation Location on Prowers Bridge............................45
5. Field Data Analysis from Prowers Bridge Test..........................53
vii


5.1 General..............................................................53
5.2 Raw Wind Speed Data..................................................55
5.3 Raw Strain Data......................................................55
5.3.1 General Background of Transducer Location..........................55
5.3.2 Axial Forces in Bottom Chord Eyebars...............................55
5.3.3 Flexural Moments in Portal Frames..................................57
5.4 Strain Data Reduction................................................60
5.5 Cable Resistance.....................................................61
5.6 Transducer Strain....................................................62
5.7 Noise Filter ........................................................63
5.8 Delta Strain.........................................................65
5.9 Summary of Reduced Data..............................................66
6. Verification Model Analysis and Comparisons..........................68
6.1 DeckModel............................................................68
6.1.1 Actual Wind Loads..................................................68
6.1.2 Actual Boundary Conditions.........................................72
6.1.3 Internal Member End Restraints.....................................73
6.2 Verification of Youngs Modulus (EEq)................................74
6.3 Comparison of the Verification Deck Model with Field Data............75
6.4 Leeward Bottom Chord Eyebars Consideration...........................77
7. Summary, Conclusions, and Recommendations for Future Research.......78
viii


7.1 Introduction......................................................79
7.2 Summary of Findings...............................................79
7.3 Conclusions.......................................................80
7.4 Recommendations for Future Research...............................81
Appendix
A. Field Instrumentation Set-up for Data Acquisition with PC9000.....83
B. PC9000 Data Retrieval Procedures for Data Acquisition............105
C. LoggerNet Program Set-up and Remote Data Collecting Procedures for
CR5000...........................................................111
D. Data Acquisition Program.........................................125
E. High Spike Raw Wind and Strain Data............................134
F. High Spike Reduced Strain Data.................................137
G. Delta Strain Data Summary .......................................139
Bibliography.........................................................141
IX


FIGURES
Figure
2.1 MAP OF COLORADO................................6
2.2 LOCATION OF PROWERS BRIDGE.....................6
2.3 HISTORIC PHOTOGRAPHY OF PROWERS BRIDGE.........7
2.4 PRESENT STATE OF PROWERS BRIDGE................8
2.5 PRESENT BOTTOM DECK AND SUBSTRUCTURE
OF THE PROWERS BRIDGE..........................9
2.6 PRESENT TOP DECK OF PROWERS BRIDGE.............9
3.1 3D SKELETON MODEL OF THE PROWERS BRIDGE.......11
3.2 SUPERIMPOSED GRAVITY LOADS ON THE PROWERS
SKELETON MODEL................................12
3.3 BOTTOM CHORD RELATIVE AXIAL FORCES DUE TO GRAVITY
LOADS ON THE PROWERS SKELETION MODEL..........12
3.4 AASHTO WIND PRESSURE ON THE PROWERS
SKELETON MODEL................................13
3.5 BOTTOM CHORD AXIAL FORCES DUE TO WIND LOAD ON THE
PROWERS SKELETON MODEL........................14
3.6 PROPPED CANTILEVER CONDITION DIAGRAMS.........14
3.7 DECK MODEL OF THE PROWERS BRIDGE..............15
3.8 CORRUGATED DECK MODEL.........................19
x


3.9 RENDERING OF STEEL STRINGERS ON STEEL FLOOR BEAM
ON THE PROWERS BRIDGE..........................24
3.10 OFFSET MEMBERS AND RELEASE LOCATIONS
ON THE PROWERS BRIDGE IN THE RISA MODEL........25
3.11 RENDERING OF THE STEEL DECK
ON THE PROWERS DECK MODEL.....................25
3.12 GRAPH OF Eeq VERSUS NET AXIAL FORCE DUE
TO AASHTO WIND LOAD PLUS DEAD LOAD............30
4.1 SCHEMATIC DIAGRAM OF PROWERS BRIDGE
INSTRUMENTATION SYSTEM........................33
4.2 ANEMOMETER AND WIND SENTRY VANE (WIND DIRECTION
SENSOR).......................................34
4.3 CAMPBELL SCIENTIFIC MODEL SDM-INT8
INTERVERAL TIMER..............................35
4.4 STRAIN TRANSDUCER WITH POLYETHYLENE COATING...37
4.5 CAMPBELL SCIENTIFIC CR5000 DATA LOGGER........38
4.6 CAMPBELL SCIENTIFIC PS 100 POWER SUPPLY.......39
4.7 FULL BRIDGE WIRING DIAGRAM
FOR TERMINAL INPUT MODULE (TIM)...............40
4.8 REDWING CELLULAR DIGITIAL MODEM FROM AIRLINK
COMMUNICATION.................................44
4.9 YAGI ANTENNA..................................44
4.10 DIAGRAM OF PROWERS BRIDGE ILLUSTRATING
THE LOCATIONS OF THE ANEMOMETERS (WS1-WS7)
AND WIND DIRECTION SENSOR (WD)................46
4.11 ANEMOMETER (WS6) INSTALLED ON PROWERS BRIDGE..46
xi


4.12 ANEMOMETER (WS1) AND WIND DIRECTION SENSOR (WD)
INSTALLED ON PROWERS BRIDGE......................47
4.13 ANEMOMETER (WS2) INSTALLED ON PROWERS BRIDGE....47
4.14 ANEMOMETER (WS4) INSTALLED ON PROWERS BRIDGE....48
4.15 DIAGRAM OF PROWERS BRIDGE ILLUSTRATING
THE LOCATIONS OF THE STRAIN TRANSDUCERS (G1-G16).49
4.16 STRAIN TRANSDUCERS (G4 & G6-G8) INSTALLED ON THE
BOTTOM CHORD EYEBARS ON THE WEST SIDE ON PROWERS
BRIDGE ...................................49
4.17 STRAIN TRANSDUCER INSTALLED ON THE INSIDE FACE
OF THE NORTH END POST (PORTAL) ON PROWERS BRIDGE.50
4.18 STRAIN TRANSDUCER INSTALLED ON THE OUTSIDE FACE
OF THE NORTH END POST (PORTAL) ON PROWERS BRIDGE.50
4.19 INSTALLATION OF THE SECURITY BOX
UNDER THE PROWERS BRIDGE.....................51
4.20 DATA LOGGER AT PROWERS BRIDGE...............51
4.21 YAGI ANTENNA ON PROWERS BRIDGE..............52
5.1 WIND DIRECTION DATA AT PROWERS BRIDGE........54
5.2 WIND SPEED DATA AT PROWERS BRIDGE............54
5.3 STRAIN TRANSDUCER ARRANGEMENT ON BOTTOM CHORD
EYEBARS......................................56
5.4 STRAIN TRANSDUCER LOCATIONS ON THE NORTH END POST
MEMBER (PORTAL)..............................58
5.5 SECTION OF END POST MEMBER (PORTAL) WITH STRAIN
TRANSDUCER...................................59
xii


5.6 STRAIN MEASUREMENTS FOR THE WINDWARD AND
LEEWARD BOTTOM CHORD EYEBARS AT PROWERS BRIDGE ... 63
5.7 STRAIN MEASUREMENTS FOR THE SOUTH END POSTS
(PORTALS) AT PROWERS BRIDGE...................64
5.8 STRAIN MEASUREMENTS FOR THE NORTH END POSTS
(PORTALS) ON PROWERS BRIDGE...................65
5.9 CHANGE IN STRAIN FOR WINDWARD AND LEEWARD
BOTTOM CHORD EYEBARS AT PROWERS BRIDGE........66
6.1 WIND SECTIONS FOR THE FIELD WIND DATA.........70
6.2 WIND PRESSURE PER SECTION.....................72
6.3 ILLUSTRATION OF PROWERS BIRDGE ACTUAL BOUNDARY
CONDITIONS....................................73
6.4 GRAPH OF Eeq VERSUS COMPRESSIVE AXIAL FORCE DUE TO
FIELD WIND LOAD...............................75
A. 1 SERIAL CABLE ATTACHED TO LAPTOP..............83
A.2 SERIAL CABLE ATTACHED TO DATA LOGGER..........84
A.3 SLIP CONNECTION TO PS 100 POWER SUPPLY........84
A.4 POWER SUPPLY WIRE ATTACHED TO DATA LOGGER.....85
A.5 POWER SWITCH FOR DATA LOGGER..................85
A.6 PC9000 DESKTOP ICON...........................86
A.7 PC9000 COMM PORT SCREEN.......................86
A.8 PC9000 TOOLS PULLDOWN MENU....................87
A.9 PC9000 PROGRAM TO DOWNLOAD SCREEN.............88
A. 10 PC9000 DOWNLOAD PROGRAM SCREEN..............88
xiii


A.l 1 PC9000 PROGRAM DOWNLOAD MESSAGE SCREEN..........89
A. 12 PC9000 PROGRAM NOT DOWNLOADED MESSAGE...........89
A. 13 CR5000 DATA LOGGER SCREEN.......................90
A. 14 PC9000 DATA LOGGER CLOCK PULLDOWN MENU..........90
A. 15 PC9000 DATA LOGGER CLOCK SCREEN.................91
A. 16 PC9000 DATA LOGGER FIELD MONITOR PULLDOWN MENU 91
A. 17 PC9000 DATA LOGGER FIELD MONITOR SCREEN.........92
A. 18 PC9000 DATA LOGGER FLAG TOGGLE BUTTON...........93
A. 19 PC9000 DATA LOGGER TABLE PULLDOWN MENU..........94
A.20 PC9000 DATA LOGGER FIELD MONITOR SCREEN
WITH PUBLIC TABLE.............................95
A.21 PC9000 DATA LOGGER PUBLIC TABLE
WITH COMMUNICATION ERRORS.....................96
A.22 PC9000 PROGRAM COLLECT DATA
RETRIEVAL PULLDOWN MENU.......................97
A.23 PC9000 PROGRAM COLLECT DATA SCREEN...........98
A.24 PC9000 DATA LOGGER RESET TABLE TOGGLE........99
A.25 PC9000 DATA LOGGER RESET TABLE MESSAGE BOX...99
A.26 DEEP CELL BATTERY WIRES FROM THE SLIP CONNECTION.100
A.27 SLIP CONNECTION TO PS 100 POWER SUPPLY...........100
A.28 PLUGIN FOR THE DIGITIAL CELLULAR MODEM CABLE
TO DATA LOGGER...............................101
A.29 MODEM CABLE PLUG-IN TO DIGITAL CELLULAR MODEM....101
xiv


A.30 POWER SUPPLY CABLE PLUG-IN
TO DIGITAL CELLULAR MODEM.....................102
A.31 POWER SUPPLY CABLE PLUG-IN TO PS 100 POWER SUPPLY.102
A.32 POWER LIGHT MODEM ON THE DIGITAL CELLULAR MODEM 103
A.33 ANTENNA CABLE PLUG-IN TO DIGITAL CELLULAR MODEM.... 103
A.34 YAGI ANTENNA......................................104
A. 35 REG MODEM LIGHT ON THE DIGITAL CELLULAR MODEM....104
B. 1 SERIAL CABLE ATTACHED TO DATA LOGGER..............105
B.2 SERIAL CABLE ATTACHED TO LAPTOP....................105
B.3 PC9000 DESKTOP ICON................................106
B.4 PC900 COMM PORT SCREEN.............................106
B.5 PC9000 PROGRAM COLLECT DATA RETRIEVAL
PULLDOWN MENU......................................107
B.6 PC9000 PROGRAM COLLECT DATA SCREEN.................107
B.7 PC9000 DATA LOGGER SELECTION PULLDOWN MENU.........108
B.8 PC9000 DATA LOGGER RESET TABLE MESSAGE BOX.........109
B. 9 PC9000 DATA SCREEN FLAG
TOGGLE BUTTON......................................110
C. 1 LOGGERNET 3.1 TOOLS BOX...........................Ill
C.2 LOGGERNET 3.1 SETUP DIALOG BOX.....................112
C.3 LOGGERNET 3.1 ADD DIALOG BOX.......................113
C.4 LOGGERNET 3.1 ADD TAPIREMOTE DIALOG BOX............113
XV


C.5 LOGGERNET 3.1 ADD CR5000 DATA LOGGER DIALOG BOX.114
C.6 LOGGERNET 3.1 SETUP TAPIREMOTE DIALOG BOX.......115
C.7 LOGGERNET 3.1 SETUP DIGITAL NUMBER
PHONE DIALOG BOX...............................116
C.8 LOGGERNET 3.1 SETUP CR5000 DATA LOGGER DIALOG BOX... 117
C.9 LOGGERNET 3.1 CONNECT TOOL BOX.................118
C.10 LOGGERNET 3.1 CONNECT DIALOG BOX..............118
C.l 1 LOGGERNET 3.1 DATA COLLECTION CUSTOM BUTTON..119
C.12 LOGGERNET 3.1 CUSTOM COLLECTION DIALOG BOX....120
C.l3 LOGGERNET 3.1 CUSTOM COLLECTION
RESET TABLE BUTTON.............................121
C.l4 LOGGERNET 3.1 RESET TABLE DIALOG BOX..........122
C.l5 LOGGERNET 3.1 CONFIRM RESET TABLE MESSAGE BOX.122
C.16 LOGGERNET 3.1 CUSTOM COLLECTION CLOSE BUTTON..123
C.l7 LOGGERNET 3.1 FILE PULLDOWN MENU..............124
XVI


TABLES
Table
3.1 Youngs modulus, Eeq for Prowers deck model..........................23
3.2 Summary of maximum axial forces in the windward bottom chord
eyebars on the Prowers Bridge.......................................27
4.1 Transducer factors and strain gage factors...........................37
4.2 26 AWG wire length and 18 AWG cable length...........................43
5.1 Prowers bridge field data (measured) forces summary..................67
6.1 Wind velocities, average section velocities, and average section pressure on
the Prowers Bridge..................................................71
6.2 Comparison of reduced field data to the deck model cases of the
maximum axial compressive forces in windward bottom chord eyebars.... 74
6.3 Prowers Bridge verification summary..................................76
xvii


1.
Overview
1.1 Introduction
In the late 1800's and early 1900's, steel vehicular truss bridges were an
integral part of America's roadway system. Over the years, newer bridges with
more vehicle lanes and higher load capacities have replaced the narrow steel truss
bridges. The history of the truss bridges has slowly disappeared due to demolition
of the bridges or the bridges collapsing from lack of maintenance (Rutz 2004). One
way to preserve the remainder of these historic truss bridges is to convert them to
pedestrian bridges.
Converting the historic truss bridges to pedestrian use allows the general
public to have access to these structures. Conversion also allows public agencies to
continue to maintain historic truss bridges when funds would otherwise not be
applied toward that bridge. Unfortunately engineers attempting to analyze the
historic truss bridge for pedestrian use by the "traditional" skeleton method of
modeling often determine the bridge has inadequate lateral strength to resist design
wind load (Rutz and Rens 2004). The reason for this engineering conclusion is
based on two factors: (1) the current wind load design requirements are higher than
1


the original wind design and (2) using "traditional" skeletal structural analysis to
determine lateral strength. The wind load criteria in these cases are based on the
American Association of State Highway and Transportation Officials (AASHTO)
Guide Specifications for the Design of Pedestrian Bridge. The combination of
these two factors can lead to the incorrect conclusion that the windward bottom
chord members are overstressed and this hinders preservation efforts.
1.2 Previous Research
The University of Colorado at Denver has been studying ways to convert
historic truss bridges to pedestrian use. Previous research by Rutz (2004) focused
on the difference in assumed load paths for a "traditional" skeleton method of
modeling and the actual load paths for an existing historic truss bridge. The
"traditional" skeleton method of modeling only incorporates the truss members,
without the deck or rail elements, and uses either a two-dimensional or a three-
dimensional computer-modeling program.
The design wind pressure of 1.44 kPa to 1.92 kPa (30 psf to 40 psf) was a
common design practice in the era of the Prowers Bridge (Wandell 1916). Today,
the AASHTO Guide Specifications for the Design of Pedestrian Bridge requires
3.59 kPa (75 psf) lateral pressure of loading on all bridge elements for wind design
for pedestrian use (AASHTO 1997). Based on the Carroll (2003) study, using the
2


traditional "skeleton" method of modeling with the current wind design load
requirements, five historic Colorado truss bridges with traditional timber decks
were inadequate for pedestrian conversion. Even though these historic truss
bridges with traditional timber deck did not pass current standard code, the existing
bridges were not showing signs of failure due to wind. With that in mind, Herrero
(2003) developed and tested strain transducers so that existing structures, including
historic truss bridges, could be field tested.
Two historic truss bridges were tested and analyzed by Rutz (2004). The
basic findings of these tests showed a decrease in compressive axial force in the
windward bottom chord eyebars when modeling the timber deck as plate elements.
This was field verified on two existing bridges. To further the wind load research
on historic truss bridges the National Center for Preservation Technology and
Training provided a grant for Prowers Bridge, San Miguel Bridge, Rifle Bridge,
and Blue River Bridge (Rutz, et al 2005). This thesis addresses only the analytical
techniques and field verification method used on the Prowers Bridge.
1.3 Goal
The goal of this study was to examine the stiffening effect of the deck under
lateral (wind) loads, both analytical and by field-testing, of the historic Prowers
Bridge over the Arkansas River near Lamar, Colorado. In order to accomplish this
3


goal, an experimental study utilized data from anemometers and clamp-on modular
strain transducers with remote data retrieval to provide verification of an analytical
deck model of the current Prowers Bridge.
4


2.
Site Study
2.1 History of Prowers Bridge
The Prowers Bridge is a single lane vehicular steel truss bridge that is
shown in Figure 2.1 and Figure 2.2. The bridge consists of one five-panel Pratt
pony truss built in 1921, three nine-panel Camelback Pratt through trusses built
1909, and two six-panel Pratt through trusses built in 1902 and 1906. The Pueblo
Bridge Company (Baughn 2006) built all three of the Camelback Pratt through
trusses. The Prowers Bridge survived a 1921 major flood of the Arkansas River as
shown in Figure 2.3. The original bridge deck was likely timber planks on steel
stringers with steel floor beams. It has steel bottom eyebars and diagonals with
steel rod counterbracing. The railing is a steel lattice with single angle top and
bottom rails.
5


(3>
i
Reefy $
MtvriUm
NjfL'Pj*
CD
Otnvtr
(3> ^
(£3
><<> Spriftfi
(53
I
"f*
S3

'fe*'':: 'i^|
P*M
S3
Figure 2.1 MAP OF COLORADO. The star indicates the City of Lamar (Yahoo
Maps 2006)
Figure 2.2 LOCATION OF PROWERS BRIDGE. It is approximately 16.1
kilometers (10 miles) west of the City of Lamar, CO (Delorme 1997).
6


Figure 2.3 HISTORIC PHOTOGRAPHY OF PROWERS BRIDGE. One the few
steel truss bridges remaining in place after the 1921 disastrous flood.
The Prowers Bridge is shown above with its northern span washed
away (Livak 2004). The circled span was studied in this thesis.
2.2 Prowers Bridge of Today
Over time, the Arkansas River has narrowed so it only passes under one
span of the Prowers Bridge. The 160-foot Camelback Pratt through truss span, as
shown in Figure 2.4 and circled in Figure 2.3, currently crosses the Arkansas River.
It is the middle span of the three 9-panel Camelback Pratt through trusses and the
span studied in this thesis. Until its abandonment in 1994, the Prowers Bridge
served as a highway bridge. It was abandoned when an adjacent replacement bridge
made of steel was constructed. The original bridge is now closed to vehicular
traffic.
7


The original steel railing, steel floor beams, steel stringers, portal members,
top chord and bottom chord bracing, vertical members, diagonals with steel rod
counterbracing, and bottom chord eyebars are still intact. Virtually, all the paint
has weathered away. The north end post (portal) roller bearings have shifted off
their rollers and have been allowed to translate in the transverse direction of the
bridge. The south end post (portal) "pinned" bearings have also shifted and have
been allowed limited translation in the transverse direction of the bridge. The piers
consist of two 1.45 m (4.75 feet) diameter steel pipes filled with concrete with a
steel plate in between the pier columns, as shown in Figure 2.4.
8


Currently the deck support consists of steel floor beams with steel stringers,
as shown in Figure 2.5. The corrugated metal deck has an asphalt overlay, as
shown as shown in Figure 2.6. The original timber deck no longer exists.
Figure 2.5 PRESENT BOTTOM DECK AND SUBSTRUCTURE OF THE
PROWERS BRIDGE. A corrugated steel deck has been filled with
asphalt pavement.
Figure 2.6 PRESENT TOP DECK OF THE PROWERS BRIDGE. A corrugated
steel deck has been filled with asphalt pavement.
9


3. Modeling and Analysis
3.1 Modern Analysis
The "traditional" structural analysis of this bridge was based on a skeleton
frame without deck components. The "traditional" structural analysis and modeling
were completed using matrix algebra and computer programs. RISA-3D software
was used to model and analyze the Prowers Bridge (RISA 2002). Both the Prowers
Bridges skeleton model and deck models were analyzed. In order to study the
lateral stiffness, the program modeled the deck as solid plates. The axial loading
in the bottom chord eyebars were compared with different deck stiffnesses.
3.2 Prowers Skeleton Model
As-built dimensions and sections properties for Prowers Bridge were
obtained from field measurements during the site visits, a previous Colorado State
Department of Transportation inspection report, and through the Carroll (2003)
study. A pin-connected skeleton frame was analyzed under AASHTO load
combinations with RISA-3D software. The boundary conditions of the analysis
consisted of pinned joints at one end and roller joints at other end. This is
illustrated in Figure 3.1. Both the roller and pinned joints were restrained from
10


lateral translation, but the roller joints were free to translate in the longitudinal
direction of the bridge. Both the roller and pinned joints were free to rotate in any
direction, similar to a ball joint. The in-field boundary conditions for the Prowers
Bridge model are discussed in more depth in Chapter 6.
Figure 3.1 3D SKELETON MODEL OF THE PROWERS BRIDGE. The
Prowers Bridge traditional skeleton model based on the steel
members only. On the right side of the bridge, the supports are pinned
as indicated. On the left side, the supports are rollers. The rollers are
free to translate in the longitudinal direction of the bridge.
Figure 3.2 illustrates the superimposed vertical dead loads on the skeleton
model. Gravity loads consist of the skeleton members with an additional ten
percent weight factor to account for lattices, rivets, bolts, pins, and floor plates.
The gravity point loads also include 0.391 kN (88 lbs) for the railing and stringer
dead load plus 12.26 kN/m (0.84 k/ft) for the deck. Figure 3.3 illustrates the axial
forces in the bottom chord eyebars due to the gravity loads on the skeleton model.
11


Figure 3.2 SUPERIMPOSED GRAVITY LOADS ON THE PROWERS
SKELETON MODEL. Graphical representation of superimposed
gravity loads.
Figure 3.3 BOTTOM CHORD RELATIVE AXIAL FORCES DUE TO
GRAVITY LOADS ON THE PROWERS SKELETION MODEL.
Note that the tensile force is the largest in the mid-span.
Figure 3.4 illustrates the wind load on skeleton structure based on the
AASHTO criteria of 3.59 kPa (75 psf)- This is the standard wind pressure for
12


exposed members in truss bridge analysis for pedestrian use. This study was based
on a load case with the wind load orthogonal to the exposed members; no wind
load was added to the deck in an upward or downward direction.
Figure 3.4 AASHTO WIND PRESSURE ON THE PROWERS SKELETON
MODEL.
Figure 3.5 illustrates the wind load effect on the skeleton structure. It
shows the axial forces in the bottom chord eyebars due to wind load. Tensile axial
force was mainly in the leeward side and the compressive axial force was mainly
on the windward side. In the 3D analysis, there was a sign reversal in the member
forces on both sides near the pinned ends. Due to the pinned end that was
restrained in the lateral direction, its behavior was similar to a "propped cantilever"
condition as shown in Figure 3.6 (Carroll 2003).
13


Figure 3.5 BOTTOM CHORD AXIAL FORCES DUE TO WIND LOAD ON
THE PROWERS SKELETON MODEL. Graphical representation of
axial forces in the bottom chord eyebars due to wind on the skeleton
model. The reversal in sign occurs near the pinned ends.
-1 - wi
ltttt I ^
*/
T Tl>r^
Shear

8l . /
u,

Moment

\)J
'"mo*
Figure 3.6 PROPPED CANTILEVER CONDITION DIAGRAMS. The load,
shear, and moment diagrams for a propped cantilever condition from
the AISC Manual of Steel Construction (AISC 2001).
14


3.3 Prowers Deck Model
As show in Figure 3.7, the Prowers deck model was based on a skeleton
model plus 810 plate elements with steel stringers. The corrugated deck was
represented as thin plate elements.
Figure 3.7 DECK MODEL OF THE PROWERS BRIDGE. The skeleton model
plus the stringers and deck.
3.3.1 Youngs Modulus (EEq) of Prowers Deck
The corrugated steel deck had one stiffness in the corrugated direction (the
bridge longitudinal direction) and a different stiffness in the non-corrugated
direction (the bridge transverse direction). RISA 3-D was not capable of defining
multiple stiffnesses in orthogonal directions for plates/shells. An equivalent
Youngs Modulus EEq was used in this study to model a range of deck stiffnesses
for the corrugated steel deck, treating the deck as if it were isotropic. To account
15


for the deck stiffness, Young's modulus E and shear modulus G were calculated
based on methods described in Section 3.3.1.2.
3.3.1.1 Background Information on Young's Modulus
For the uniaxial loading of elastic materials, stress is proportional to strain.
This relationship is known as Hooke's law and is shown in Equation 3.1 (Beer
1992).
<7 = E £ (3.1)
Where:
ct = stress
E = Young's Modulus (modulus of elasticity)
£ = strain
Young's Modulus is the material "stiffness" or the ability to resist
deformation within the linear range.
Solving Equation 3.1 for E:
E = ct / £ (3.2)
For the uniaxial loaded case, ct is defined as:
ct = P / A (3.3)
16


Where:
P = applied load
A = cross-sectional area
Uniaxial strain is defined as:
s = A / L0 (3.4)
Where:
A = change in length of the member
L0 = original length of the member
Equations 3.3 and 3.4 were substituted into Equation 3.2 for the following
equation:
E = (PL0) / (AA) (3.5)
Then Equation 3.5 was rearranged as an element stiffness equation in linear
algebraic form for axial force members:
{P} = [(AE/L0)] {A} (3.6)
The shear modulus G was calculated using Equation 3.7. It is similar to the
modulus of elasticity except it is for shear.
G = E / (2 (1 + v)) (3.7)
17


Where:
v = Poisson's ratio (assumed to be 0.3 for steel)
G = Shear Modulus
3.3.1.2 Youngs Modulus (Eeq baseline)
for the Corrugated Steel Deck
The Prowers current bridge deck consisted of 3.175 mm (.125 in) thick corrugated
steel decking of 50.8 mm x 152.4 mm (2 in. x 6 in.) with the asphalt filled flutes
and an asphalt overlay of 63.5 mm (2.50 in.). For this study, the corrugated deck
was assumed to be connected to steel stringers with puddle welds. The number or
the pattern of the welds was not verified. Due to its current poor condition, the
asphalt pavement was not considered to contribute significantly to stiffness. Thus
the asphalt pavement in the deck model contributed only to dead load.
In all the deck models, the corrugated steel deck was represented as solid
plate elements. Before the deck model was analyzed with solid plate elements, a
modulus of elasticity for the corrugated deck (which is an anisotropic material) was
determined with several simple models described below. The length and the cross
sectional areas of both models were the same.
18


A 25.4 mm (1 in) strip of corrugated deck was modeled using RISA-3D, as
shown in Figure 3.8. The boundary conditions of the model consisted of one fixed
end with no rotation or translation, and the other end was allowed to translate but
restrained from rotation. Based on the global axes, a 4.448 kN (1 kip) load was
applied in the negative x-direction.
Roller End &
Rotation Restraint
A
E, longitudinal of
corrugated deck model
(transverse direction of
Prowers Bridge)
E, transverse of corrugated
deck model (longitudinal
direction of Prowers
Bridge)
Figure 3.8 CORRUGATED DECK MODEL
The RISA model was analyzed to determine a deflection (A) at the roller
end. There was only one force element of 4.448 kN (1 kip) and one displacement
(A) to calculate. Equation 3.6 was simplified to one equation form without the
symbology for force vectors, displacement vectors, and stiffness matrix.
19


P = (AE/Lo) A
(3.8)
Equation 3.8 was rearranged as follows:
k = P/A
(3.9)
Where:
k = (AE/L0), element stiffness coefficient
Then the stiffness (kcon-ugated) of corrugated deck model was evaluated as
follows:
P = 4.448 kN (1 kip) Applied load to a section of the corrugated
steel deck
A = 1.8390 mm (.0724 in) Deflection of the joint in the negative x-
direction from the 3-D RISA model.
kcorrugated = 2.42 kN/mm (13.81 k/in) Stiffness of the corrugated
deck model
A 25.4 mm (1 in) strip of non-corrugated, thin steel plate was also modeled
with the same boundary conditions as the corrugated section. Equation 3.11 was
used to the find the stiffness (kPL) of the plate as shown below.
(3.10)
Where:
kpL P/A
(3.11)


Where:
P = 4.448 kN (1 kip) Applied load to a section of the steel plate
model
A = 0.0432 mm (0.0017 in) Deflection of the joint in the negative
x-direction from the 3-D RISA model
kpL = 103.01 kN/mm (588.23 k/in) Stiffness of the section of the
plate model
A stiffness of ratio between the plate and corrugated deck model was
developed with the normal Youngs Modulus of steel to adjust the stiffness of the
deck. Equivalent Youngs Modulus Eeq Baseline of deck was evaluated to represent a
reduced stiffness for the anisotropic material as shown below.
EEQ Baseline E (kcomjgated / kpL) (3.12)
Where:
E = 200 MPa (29000 ksi) Young's Modulus of steel
kcomigated = 2.42 kN/mm (13.81 kip/in)
kPL = 103.1 kN/mm (588.23 kip/in)
Eeq Baseline = 4.7 MPa (681 ksi) Young's Modulus lower boundary
for the deck
EEQ Baseline represents the lower bound of the Young's Modulus, E in other
words for the deck model in the transverse direction. Eeq Baseline showed that its
reduced stiffness perpendicular to the corrugations significantly reduced the
stiffness of the plate. It was approximately 2.3 percent of the normal Youngs
21


Modulus of steel. The upper bound of E was in the transverse direction of the
bridge which was 200 MPa (29000 ksi) as an isotropic material. The value for the
Prowers Bridge deck model was expected to occur between these bounds.
Equation 3.13 was used to modify EEqBaseline to represent the stiffness of the
deck in both directions. Many E combinations were analyzed to compare to the
field verification in a trial and error process. In Case 3, the case used in the final
results, the Eeq Baseline was modified by a factor 1.26 (26 percent increase) to
accommodate the stiffer direction as shown below.
EEq = EEQ_Baseline X (3.13)
Where:
EEQ_Baseiine= 4.7 MPa (681 ksi) Lower boundary of the equivalent
Young's Modulus for the corrugated deck
X 1.26 Factor of EEq Baseline'
EEq-5.9 MPa (861 ksi) Equivalent Young's Modulus used in the
deck model
The value of EEQ_Baseime was increased by different percentages to
investigate different deck stiffnesses for the deck model. Table 3.1 shows five
different EEq values used for the corrugated steel deck model for the Prowers
Bridge.
22


Table 3.1 Young's modulus, Eeq for Prowers deck model.
Deck Model Young's Modulus (EEq) MPa (ksi) Percent Increase Above EEQ_Baseline
EEQBaseline (Corrugated Deck Model) E = 4.7 MPa (E = 681 ksi) 0%
Case 3: E = 5.9 MPa (E = 861 ksi) 26%
Case 4: E = 60 MPa (E = 8700 ksi) 1176%
Case 5: E- 100 MPa (E = 14500 ksi) 2028 %
Case 6: E= 150 MPa (E = 21750 ksi) 3091 %
Case 7: E = 200 MPa (E = 29000 ksi) 4155%
3.3.2 Modeling Techniques for Plates
The plate elements representing the deck were four-node quadrilateral
elements. They are called "mixed interpolation elements" (RISA 2002) in that they
include interpolation functions for out-of-plane transverse shear (Bathe 1996). This
approach was analogous to incorporating shear deformations with flexural effects
from beam theory and resulted in plate elements that were thin (like the Prowers
Bridge). The plate elements were approximately 0.60 m (1.97 ft) by 0.57 m (1.87
ft) with a thickness of .003 m (.125 in.) except along the edge of the deck, which
was 0.60 m (1.97 ft) by 0.20 m (0.67 ft). The equivalent Young's Modulus (Eeq)
23


was the same for all of the plates. The Eeq cases were analyzed using the deck
model with the same lateral loads.
The RISA rendering in Figure 3.9 shows the steel stringers on the floor
beams. The floor beam and interior stringers were American Standard Shapes and
the exterior stringers were American Standard Channels. The steel stringers and
steel floor beams were pinned together with rivets located at the bottom flange of
the floor beam to the top flange of the steel stringers. The frame elements were
offset from each other with a rigid offset element. The rigid offset elements were
released for bending moment at the location where the members were pinned
together (as shown in Figure 3.10) to represent the assembly of the actual stringers
to the floor beams. The plate elements were offset in a similar way, to represent the
assembly of the actual deck to the stringers as shown in Figure 3.11.
Figure 3.9 RENDERING OF STEEL STRINGERS ON STEEL FLOOR BEAM
ON THE PROWERS BRIDGE.
24


I
Figure 3.10 OFFSET MEMBERS AND RELEASE LOCATIONS ON THE
PROWERS BRIDGE IN THE RISA MODEL.
Figure 3.11 RENDERING OF THE STEEL DECK ON THE PROWERS DECK
MODEL. The rendering shows the corrugated steel deck, modeled as
an equivalent solid plate, with a deck angle on the steel stringers
attached to the steel floor beam.
25


3.3.3 Modeling Results
The mid-span of the bridge was chosen for a location of interest where
wind-induced axial forces were expected to be the maximum. Table 3.2 shows
axial forces of the windward bottom chord eyebars from the skeleton model with
two different deck dead loads. This table also shows axial forces of the windward
bottom chord eyebars from the corrugated steel deck model with six different deck
stiffnesses. The criteria for the wind pressure were based on the AASHTO wind
pressure of 3.59 kPa (75 psf), acting perpendicular to the bridge.
26


Table 3.2 Summary of maximum axial forces in the windward bottom chord
eyebars on the Prowers Bridge. Forces are expressed in kN (kips),
followed by percent of reduction of compression (or increase in
tension) compared to the "traditional" (Case 2) existing deck skeleton
model. (Positive = tension; negative = compression).
Model Axial force due to dead load only Axial force due to wind load only Net axial force due to wind plus dead load
Case 1: Skeleton Model (Timber Deck Dead Load = 160.1 -226.8 -59.60
0.72 kPa (15 psf) Figure 3.5) (36.0) (-51.0) (-13.4)
Case 2: Skeleton Model (Existing Deck Dead Load = 284.2 -226.8 69.8
2.3 kPa (47 psf) Figure 3.5) (63.90) (-51.0) (15.7)
Case 3: Deck Model 278.9 -152.1 129.8
(E = 5.94 MPa (861 ksi) (62.74) (-34.2) (29.2)
Figure 3.7) 1.8% 32.9% 86.0%
Case 4: Deck Model 276.7 -118.3 160.1
(E = 60 MPa (8700 ksi) - (62.20) (-26.6) (36.0)
Figure 3.7) 2.6% 47.8% 129.0%
Case 5: Deck Model 276.7 -118.3 166.6
(E = 100 MPa(14500 ksi)- (62.20) (-26.6) (37.5)
Figure 3.7) 2.6% 47.8% 139.0%
Case 6: Deck Model 276.7 -118.3 172.9
(E = 150 MPa(21750ksi)- (62.20) (-26.6) (38.8)
Figure 3.7) 2.6% 47.8% 148.0%
Case 7: Deck Model 276.7 -118.3 178.9
(E = 200 MPa (29000 ksi) - (62.20) (-26.6) (40.2)
Figure 3.7) 2.6% 47.8% 156.0%
Case 8: Diaphragm 279.0 -27.1 260.7
(Rigid Deck E = Infinite) (62.71) (-6.1) (58.6)
1.9% 88.0% 272.0%
27


The percent change from the skeleton model (existing deck Case 2) was
determined for the deck models (Case 3 through Case 7) from:
% change =100 x
Fskeleton Fdeck
(3-14)
Fskeleton
and for the diaphragm model (Case 8) from:
% change =100 x
(3.15)
Fskeleton
Where:
Fskeleton = calculated force in windward bottom chord from the steel
skeleton model
Fdeck = calculated force in windward bottom chord from the deck
model
Fdiaphragm = calculated force in windward bottom chord from the
diaphragm model
3.4 Skeleton Model Results
The timber deck dead load skeleton model was analyzed to show how the
original windward bottom chord would have behaved with the current AASHTO
wind load requirements. A skeleton model of the original timber deck was
evaluated with an assumed original timber deck dead load of 103 kPa (15 psf). The
original timber deck dead load was approximately three times less the dead load
than the current composite deck. The skeleton model with the original timber deck
(Case 1) indicated that the windward bottom chord eyebars were in compression.
28


Therefore, it failed the current AASHTO (1997) Pedestrian wind criteria. The
windward bottom chord was subjected to buckling as if the original timber deck
were still in place or replaced with similar material to recreate the original (historic)
look. The skeleton model with the existing deck (Case 2) showed that the
windward bottom chord eyebars were in tension; therefore, the model conformed to
the AASHTO Pedestrian wind criteria. This analysis showed that an increase in
deck dead load allowed the windward bottom chord eyebars to conform to current
wind load criteria.
3.5 Deck Model Results
Case 3 thru Case 7 showed higher tensile forces in the windward bottom
chord eyebars as the deck stiffness increased. All cases conformed to the
AASHTO Pedestrian wind criteria. As Young's Modulus for the equivalent deck
plate was increased, it was assumed that the deck was more isotropic with an
increased number of deck welds to the steel stringers. A graph was plotted to
evaluate the relationship between EEq and the net axial force due to AASHTO wind
load plus dead load. A non-linear relationship between deck stiffness and net axial
force due to wind, plus dead load in the deck, was determined as shown in Figure
3.12. Due to wind loads, a stiffer deck had less compressive force in the windward
bottom chord eyebars.
29


0
5
40
NET AXIALFORCE ON THE WINDWARO BOTTOM CHORO EYE BARS (kip*
10 15 20 25 30 36
200
160
160
? 140
&
£ 120
l/J
3
3 ioo
o
9
y> 80
*
| 60
40
20
0
0 20 40 60 80 100 120 140 160 160 200
NET AXIAL FORCE ON THE WINDWARO BOTTOM CHORO EYEBARS (kN)
20000 l
£
i/i
3
15000 3
Figure 3.12 GRAPH OF Eeq VERSUS NET AXIAL FORCE DUE TO AASHTO
WIND LOAD PLUS DEAD LOAD.
Finally in Case 8, the deck model was treated as a rigid diaphragm, the
theoretically stiffest deck. This case provided an upper bound for the wind
resisting effect. This was completed even though it is recognized that a rigid deck
would not be a reasonable design.
30


4.
Field Instrumentation
4.1 Introduction
A field study was completed to verify the analytical methods. Due to the
greatest wind exposure, the 160-foot pin-connected Camelback Pratt through truss
span (shown in Figure 2.3) was the span chosen for instrumentation. The
instrumentation was set up for approximately five weeks, from 03/05 to 04/05. It
was assumed, and later verified, that a high wind event with a wind direction
perpendicular to the bridge would occur during this period. This field
instrumentation set-up was originally designed and tested by Rutz (2004). The
original instrumentation system collected wind speed, wind direction, and
transducer strains at a sampling interval of 0.1 seconds. Some modifications to this
instrumentation were made for the Prowers Bridge. An overview of the Prowers
Bridge instrumentation system, with modifications, is described below.
Seven anemometers were used to obtain wind speed data at multiple
locations over the instrumented span at approximately the same time. A wind
direction sensor was used to obtain the direction of the wind with respect to the
bridge orientation. Sixteen transducers were used to collect the strain at the middle
31


span bottom chord eyebars and end post members (portal) during a wind event.
These members were selected in anticipation of high axial forces or flexural
moments due to lateral loading. A data logger was used to collect and store the
wind direction, wind speeds, and strains. A digital cellular modem with an antenna
was used to access the data logger from a remote location. A deep cell battery was
used to allow the data logger and modem to run until a wind event would trigger
the data collecting program. The battery also allowed remote data downloading.
4.2 Instrumentation System of Prowers Bridge
The field equipment consisted of sixteen strain transducers, seven wind
sensors, one wind directional sensor, a digital cellular modem and antenna, and a
data acquisition logger with a power supply. The schematic diagram shown in
Figure 4.1 represents the instrumentation set-up at the Prowers Bridge.
32


Wim DIRECTION
ANEMOtfTFRS
LEEWARD WINDWARD
ErlBABS EYEBARS
NORTH
PORTAL
Figure 4.1 SCHEMATIC DIAGRAM OF PROWERS BRIDGE
INSTRUMENTATION SYSTEM
4.3 Instrumentation Components
The individual instruments of the Prowers Bridge system are described in
the following sections.
4.3.1 Wind Sensors
Seven R.M. Young model 03101-5 wind sentry anemometers shown in
Figure 4.2 (on the right) were used on the Prowers Bridge.
33


03301-5 WIND SENTRY VANE USS MTG
03310 VANE ASS* M/COUMTCKwCiOiT
6-13 1/8 SET SCREW
03330A POTENTIOMETER ASSY
03101-5 WIND SENTRY ANEMOMETER LESS MIC
6-32 I 1/6 SIT SCREW
03*26 riANoe
^7 ej

'////////////) MOUNT** 3/4* IPS
3/4 PAN mo SCREW
031S OFFSET MOUNTINC 3/4* IPS FOR 03101 ANCUOMCTCR
03133 RING WACHCT
ASSY with 3-*0 I 1/2
PAN HO SCREW
6-32 2/16
PAN HO SCREW
(3 PECO)
woori 03MN-9 low: PRO 04-86
MN0 SENTRY VAT*/ANfMOMfirR loNN *L 0*G 10-90
KfKACtvCHT PARTS 1 Cm A* WOJOOl S
PM fOUHC CO TfiAVtRSC orr M 49664 USA. 6)6 946-J960
Figure 4.2 ANEMOMETER AND WIND SENTRY VANE (WIND DIRECTION
SENSOR). (Campbell Scientific 1996a)
4.3.2 Wind Direction Sensor
One R.M. Young Model 03301-5 Wind Sentry Vane, which was the
anemometer and wind direction sensor, was used as shown in Figure 4.2. To
determine the wind direction of the wind event, the wind direction sensor was
placed over the mid-span and at mid-height. The wind direction was assumed to be
the same over the entire structure for the duration of the wind event.
34


4.3.3 Interval Timer
The seven anemometers were connected to the data logger through a
Campbell Scientific model SDM-INT8 interval timer shown in Figure 4.3. The
interval timer receives and processes signals from the anemometers for data
collecting purposes.
Figure 4.3 CAMPBELL SCIENTIFIC MODEL SDM-INT8 INTERVAL TIMER.
(Campbell Scientific 2005)
4.3.4 Strain Transducers
The strain transducers were designed and verified experimentally by
Herrero (2003). Two main reasons for the production of strain transducers were
ease of field placement and relatively low costs. Strain transducers were tested and
assembled in a lab and had relatively short field time installation compared to the
installation of individual strain gages on field members.
o
WLv^* S3M IN IB
35


Each strain transducer used consisted of a 76 mm (3 in.) diameter steel ring
bolted to steel angles with a strain gage epoxied at 90 degrees from the axis of the
bolts inside the ring surface as shown in Figure 4.4. Except for the back of the
angle clamped against the bridge member, the steel angles were painted with rust-
inhibitive paint. The 76 mm (3 in.) diameter steel ring was protectively coated with
polyethylene to prevent rusting of the ring or gage as shown in Figure 4.4. The
strain transducers were clamped to each member with a C-Clamp on each angle.
The transducers were tested in the lab by Herrero (2003) for slip issues from the
painted or rusted members. There were no significant issues with slippage from the
member. The Model CEA-06-250UW-120 strain gages inside the ring,
manufactured by Vishay Micro Measurements Group, had a gage factor of either
2.065 or 2.095.
The transducer sensed axial strain was amplified by the flexural
deformation of the ring; therefore, the true strain of the member being studied was
obtained by multiplying the transducer strain by a predetermined transducer factor
(Herrero 2003). To determine the proper transducer factor as described by Herrero
(2003), the strain transducers were tested with an upper limit of 89 kN (20 kips).
For the transducers used on the Prowers Bridge, the ranges of the transducers
factors were found to be 4.47 to 4.90 as shown in Table 4.1.
36


Figure 4.4 STRAIN TRANSDUCER WITH POLYETHYLENE COATING
Table 4.1 Transducer factors and strain gage factors.
Transducer Pair (Strain gage factor) Transducer Factors
G1 (2.065) & G2 (2.065) 4.49
G3 (2.095) & G5 (2.095) 4.62
G4 (2.065) & G6 (2.065) 4.47
G7 (2.065) & G8 (2.065) 4.69
G9 (2.065) & G13 (2.065) 4.54
G10 (2.095) & Gil (2.095) 4.90
G12 (2.095) & G14 (2.095) 4.64
G15 (2.095) & G16 (2.095) 4.73
37


4.3.5 Data Logger
A Campbell Scientific model CR5000 data logger as shown in Figure 4.5
was used. The analog inputs used seventeen channels: one channel for wind
direction and wind speed measurements and sixteen channels for a Wheatstone
bridge (full bridge) that incorporated strain transducers and Campbell Scientific
model 4WFB120 Terminal Input Modules (TIM's) (Rutz 2004). A Campbell
Scientific PS 100 power supply as shown in Figure 4.6, in conjunction with a deep
cell battery, was used to run the modem and data logger for the duration of the field
tests. Because the power draw was minimal, the deep cell battery functioned up to
three weeks before needing to be recharged. The level of battery charge and the
data were monitored remotely through the LoggerNet program.
Figure 4.5 CAMPBELL SCIENTIFIC CR5000 DATA LOGGER. (Campbell
Scientific 2001).
38


Figure 4.6 CAMPBELL SCIENTIFIC PS 100 POWER SUPPLY. (Campbell
Scientific 2006).
4.3.6 Wheatstone Bridge
Channel two through channel seventeen were used for the strain transducers
signals with the in-line Wheatstone (3/4 quarter) bridge. The circuitry for the
Wheatstone bridge was provided by Campbell Scientific model 4WFB120
Terminal Input Modules (TIMs), as shown in Figure 4.7. A Terminal Input
Module consisted of three internal resistors connected with a strain transducer (one
resistor) to complete a full bridge. The strain transducer had a resistance of 120
ohms plus or minus a tolerance of 0.01%; the resistor in the TIM circuit matched
the nominal resistance of the strain transducer (Campbell Scientific 1996b).
Because the actual change in resistance of the strain gage was small, a full bridge
configuration was used to give the maximum resolution.
39


Figure 4.7 FULL BRIDGE WIRING DIAGRAM FOR TERMINAL INPUT
MODULE (TIM). The variable resistor represents the strain gage on
the transducer ring. The other three resistors are contained within the
TEM. The TIM connects to the data logger via three pins (labeled H, L,
and AG in the diagram). Excitation voltage is supplied from the data
logger via the lead labeled Vx (Campbell Scientific 1996b).
4.3.7 Laptop Computer
A Dell Inspiron 8500 laptop computer was used in conjunction with the
data logger in the field to verify that the strain transducers and the anemometers
were set-up correctly and running properly. The laptop was also used for remotely
downloading from Denver, Colorado to the bridge site near Lamar, Colorado and
for on-site data downloading from the data logger.
4.3.8 Software
The application software used in on-site field equipment verification,
downloading the data acquisition program to the data logger, and on-site data
40


downloading was PC9000 from Campbell Scientific. (See Appendix A and
Appendix B for on-site field equipment verification procedures and on-site data
downloading procedures.) LoggerNet 3.1 from Campbell Scientific was the
application software used with the laptop computer for remote downloading. (See
Appendix C for detailed remote downloading procedures.)
4.3.9 Data Acquisition Program
The original data acquisition program was developed by Campbell
Scientific and was modified for seven anemometers and remote downloading
capabilities. The data acquisition program had wind conditions criteria that needed
to be met prior to storing strain data, wind speeds, and the wind direction.
Appendix D contains the data acquisition program used on the Prowers Bridge.
The data logger program was triggered to store data when the wind speed
ranged between 32 km/h to 97 km/h (20 mph to 60 mph) or more, with the wind
direction of +/- 45 degrees perpendicular to the bridge. The program was set-up to
store five tables for each different range of wind speed. Each data table contained a
record date and time stamp, the seven wind speeds, the wind direction, and the
output of the sixteen strain transducers. All of this was obtained at a sampling rate
of 0.1 seconds. (See Appendix E for a sample of the field data output from the data
logger program.)
41


4.3.10 Cables
Each strain gage, inside of the transducer ring, had copper lead wires of 26
American Wire Gauge (AWG) soldered to its terminal, as shown in Figure 4.3.
Table 4.2 shows the length of the 26 AWG wire for each transducer.
The cables used for connecting the lead wires to the data logger consisted of
three 18 AWG copper cable with PVC insulation and a foil shield with a grounding
wire. These wires were soldered in the field to the three lead wires from the strain
gages at one end and connected to the data logger by using Terminal Input Modules
at the other end. To provide some shielding from extraneous electrical signals, the
ground wire from the cable was connected to the data logger ground wire. Table
4.2 also shows the 18 AWG cable lengths for each transducer. For the final
computation of the strain reading, the resistance of the wire or cable for each strain
transducer was calculated.
42


Table 4.2 26 AWG wire length and 18 AWG cable length
Transducer # 26 AWG Wire Length m (ft) 18 AWG Cable Length m (ft)
G1 1.83 (6.0) 41.18(135)
G2 2.14(7.0) 41.18(135)
G3 1.83 (6.0) 41.18(135)
G4 1.83 (6.0) 24.4 (80)
G5 1.98 (6.5) 78.69(158)
G6 1.83 (6.0) 29.89 (98)
G7 1.83 (6.0) 29.89 (98)
G8 2.14(7.0) 29.89 (98)
G9 4.58(15.0) 76.25 (250)
G10 1.83 (6.0) 44.84 (147)
Gil 4.27 (14.0) 49.41 (162)
G12 1.83 (6.0) 75.03 (246)
G13 4.58(15.0) 75.03 (246)
G14 1.68 (5.5) 76.25 (250)
G15 2.14(7.0) 47.89(157)
G16 4.27 (14.0) 51.85 (170)
4.3.11 Digital Cellular Modem with Antenna
To remotely download data from Denver, Colorado, a Redwing Cellular
Digital Modem from AirLink Communication (Figure 4.8) was placed on-site and
connected to the data logger. A Yagi Antenna (Figure 4.9) was connected to the
Modem and was attached to the security box under the bridge.
43


Reset
9?9* 9*
/Ifr£lnk* * L
Redwing CDMA
Figure 4.8 REDWING CELLULAR DIGITAL MODEM FROM AIRLINK
COMMUNICATION
A
Figure 4.9 YAGI ANTENNA
44


4.3.12 Security Box
To provide security and shelter to the data logger, battery, and digital
modem over the duration of the field test, a wood security box was rigidly attached
to underside of the deck.
4.4 Instrumentation Location on Prowers Bridge
The locations of the single wind direction sensor and the seven wind
sensors are shown in Figure 4.10 (north is to the left). The wind direction sensor
was oriented on Prowers Bridge such that true north and north on the wind sensor
(0/360 degrees) were aligned with the longitudinal direction of the bridge. WS1
was positioned directly upwind of the centroid of the wind intercept area. WS2 and
WS5 were located 1.15 meters (approximately 3.5 feet) above the top of the end
post members in the portals. WS3 was positioned 3 meters (approximately 10 feet)
below the bridge deck at a mid-height elevation between the bridge deck and the
water surface below. WS4 was positioned 2.5 meters (approximately 8 feet) below
the bridge deck at a mid-height elevation between the bridge deck and the existing
ground below. WS6 and WS7 were positioned approximately at the deck elevation.
The installations of selected anemometers and the wind direction instrument on the
Prowers Bridge are shown in Figures 4.11 through 4.14.
45


WS5
WS2
WS4
W83
Figure 4.10 DIAGRAM OF PROWERS BRIDGE ILLUSTRATING THE
LOCATIONS OF THE ANEMOMETER (WS1 WS7) AND WIND
DIRECTION SENSOR (WD).
Figure 4.11 ANEMOMETER (WS6) INSTALLED ON PROWERS BRIDGE. It
was located on the windward side, approximately at bridge deck
elevation.
46


Figure 4.12 ANEMOMETER (WS1) AND WIND DIRECTION SENSOR (WD)
INSTALLED ON PROWERS BRIDGE. They were located on the
windward side, approximately at the centriod of the bridge truss.
Figure 4.13 ANEMOMETER (WS2) INSTALLED ON PROWERS BRIDGE. It
was located on the windward side in the south end post (portal) above
the bridge members.
47


Figure 4.14 ANEMOMETER (WS4) INSTALLED ON PROWERS BRIDGE. It
was located on the windward side, below the bridge deck on the west
side.
The locations of the sixteen strain transducers are shown in Figure 4.15.
For the bottom chord members, transducers G1-G3 and G5 were clamped to the
leeward bottom chord eyebars, and transducers G4 and G6-G8 were clamped to the
windward bottom chord eyebars. For the portal members, transducers G9 and G12-
G14 were clamped to the end posts member at the south portal, and transducers
G10-G11 and G15-G16 were clamped to the end posts member at the north portal.
The installations of selected instruments on Prowers Bridge are in shown in Figures
4.16 through 4.19.
48


Figure 4.15 DIAGRAM OF PROWERS BRIDGE ILLUSTRATING THE
LOCATIONS OF THE STRAIN TRANSDUCERS (G1-G16)
Figure 4.16 STRAIN TRANSDUCERS (G4 & G6-G8) INSTALLED ON THE
BOTTOM CHORD EYEBARS ON THE WEST SIDE ON PROWER
BRIDGE. Each device had a safety string attached to the deck to
prevent it from falling off the bridge during installation.
49


Figure 4.17 STRAIN TRANSDUCER INSTALLED ON THE INSIDE FACE OF
THE NORTH END POST (PORTAL) ON PROWERS BRIDGE. It
was located just below the end post's knee brace.
Figure 4.18 STRAIN TRANSDUCER INSTALLED ON THE OUTSIDE FACE
OF THE NORTH END POST (PORTAL) ON PROWERS BRIDGE.
It was located approximately 0.6 meters (2 feet) above the roller.
50


Figure 4.19 INSTALLATION OF THE SECURITY BOX UNDER THE
PROWERS BRIDGE. The security box stored the data logger, the
digital cellular modem, and deep cell battery for the five-week testing.
Figure 4.20 DATA LOGGER AT PROWERS BRIDGE. Strain transducer and
wind sensor cable installation to the data logger prior to placing into
the security box under Prowers Bridge deck.
51


Figure 4.21 YAGI ANTENNA AT PROWERS BRIDGE. Yagi Antenna attached
to the security box under Prowers Bridge deck.
52


5. Field Data Analysis from Prowers Bridge Test
5.1 General
The Prowers Bridge was monitored from 03/05 to 04/05. Data was
remotely downloaded daily from the data logger. Model verification data needed
the highest wind event and closest wind direction perpendicular to the Prowers
Bridge. On April 4, 2005, at 7:13 pm a wind event occurred with a maximum
wind speed of 69 km/h (43 mph) at approximately at 290 degrees (20 degrees
skewed from orthogonal direction of the bridge). This wind event was used for
model verification. The wind direction was approximately from the west,
orthogonal to the bridge as shown in Figure 5.1. The local "low" spike was from
time 44 seconds to 47 seconds at the bottom of the event and the local "high"
spike was from 23 seconds to 26 seconds at the top of the event as shown in
Figure 5.2. Appendix E shows a sample of the raw field data table.
53


WIND DIRECTION
Prowers Bridge 04/04/03
0 80 120 180 240
Time (seconds)
| Wind Direction |
Figure 5.1 WIND DIRECTION DATA AT PROWERS BRIDGE. Wind
direction as measured during the test. The wind direction of 270 was
global/local west of the bridge.
WIND SPEED
Prowers Bridge 04/04/05
WS1 WS2 WS3 -WS4 WS5
WS6 WS7 Average WS
Figure 5.2 WIND SPEED DATA AT PROWERS BRIDGE. Wind speed as
measured by the seven anemometers. The bold line shows the average
of all seven anemometers. The circle areas were the "high" spike event
and "low" spike event.
54
Wind Speed (m/s)


5.2 Raw Wind Speed Data
As previously mentioned, the data table with the highest wind speed
event and angle that was the most perpendicular to the bridge was selected for
further evaluation. With the data being logged at intervals of 0.1 second, the
wind speeds from the anemometer showed considerable fluctuation over a very
short time interval as shown in Figure 5.1.
5.3 Raw Strain Data
5.3.1 General Background of Transducer Location
The sections below briefly describe the engineering assessment, the theory
used on the raw strain data to acquire axial forces on bottom chord members, and
the flexural moments on the end post members (portal) of the bridge.
5.3.2 Axial Forces in Bottom Chord Eyebars
Each bottom chord member of Prowers Bridge consisted of two pin-
connected eyebars. These members have high axial forces due to lateral (wind)
loading and thus, were further monitored. At the mid-span of the bridge, a strain
transducer was clamped to each side of each eyebar, as shown in Figure 5.3.
55


C p
<

V_

G1
rl
GAGE
EYEBAR A
I
EYEBAR B

G 5
1
'Pa ^
G2 1 ZZZ 1 1 1 i r 1 uZj 1 ] G3
/ 1 i i ' \ 1
-7&
0.44 mm
*
>
J
Figure 5.3 STRAIN TRANSDUCER ARRANGEMENT ON BOTTOM CHORD
EYEBARS.
The location of the strain gage within the transducer was offset 44 mm
(1.73 in.) from the bottom-chord eyebar centerline as shown in Figure 5.3. The
gage sensed both flexural and axial strain. By using two transducers, one on each
side, and averaging their strain values (Equation 5.1 through 5.4), the flexural
component of strain was compensated for, leaving only axial strain (Rutz 2004).
(5.1)
_ Pa + My_ £, _ Pa My
AE El AE EI
_ fa + £2) _ X £ tq O3 EI EI) _ Pa
2 2 AE
(5.2)
Likewise,
(5.3)
+f5 P
£ =~1? ~-
B
AE
56


The total force in the bottom-chord member (includes both eyebars A and B) was:
f=pa + pb =£aAE + bbAE= (eA + sb)AE
(5.4)
Where:
e'a = strain sensed at eyebar A
sb = strain sensed at eyebar B
y = distance from centerline of eyebar to the gage location
PA = force in eyebar A
Pb = force in eyebar B
F = total force in bottom chord member, comprised of eyebars A
and B
5.3.3 Flexural Moments in Portal Frames
The strain transducers were mounted near the top of the end post members
below the knee brace and 0.45 meters (1.5 feet) above the support of the portal
frame. These members were selected due to high moments with lateral (wind)
loading. The transducer locations are shown in Figure 5.4 for the north end post
members (portal). A typical transducer orientation on the end post section is shown
in Figure 5.5.
In the end post members (portal), the measured strain has both axial and
flexural components as shown below:
s =-----1-
P My
(5.5)
AE El
57


Figure 5.4 STRAIN TRANSDUCER LOCATIONS ON THE NORTH END
POST (PORTAL).
There was a flexural moment at the base of the end post members (portal)
as shown in Figure 5.4. The rotation and translation about the axes of the bottom
chords of the bridge were restrained. They are represented by the fixity symbol at
the supports as shown in Figure 5.4. The dashed line indicates the end post
deformation under a wind load.
58


1 58mm
(6.00in.)
GAGE
Figure 5.5 SECTION OF END POST MEMBER (PORTAL) WITH STRAIN
TRANSDUCER. (Rutz 2004).
While the measured data was a combination of axial plus flexural strain, it
was important to know the end flexural moments in the members for comparison
with 3D deck model. Equation 5.6 through 5.9 used to isolate the flexural strain
and to determine the end flexural moments (Rutz 2004).
CTio E E]0
(5.6)
Where:
a = stress transducer number ten
E = 200 MPa (E = 29000 ksi), Young's Modulus of steel
s = mircostrain, strain data from transducer number ten
Likewise for transducer number fifteen,
59


CT15 E E15
(5.7)
Then the average axial stress for the top portal was found by:
O'total axial upper (^10 + CTl5)/2 (5.8)
With the flexural strain at the top of the member determined, the moment
located at transducer number 10 was found by:
Mio ((cr 10 CTtotal axial upper) EI/y) (5.9)
Where:
I = Moment of Inertia, 0.000116 m4 (279.26 in4)
y = centerline of member to strain transducer, 158 mm (6.00 in)
The same approach was used to find flexural moments for top and bottom
members of both end post members (portal).
5.4 Strain Data Reduction
Data collected from the strain transducers required processing as follows:
Make a correction to account for the resistance of the wire or cable from
the strain transducer to the data logger.
Apply experimentally pre-determined transducer factors to each strain
gage reading to account for the difference between measured strain at the
transducer and true strain in the member.
60


Filter out electrical noise using a 2 second rolling average.
Obtain change in strain ("delta strain") from 3-second time interval of a
"high" spike event and 3-second time interval of a "low" spike event.
The methodology for making these corrections is described in the
following sections and was originally developed by Rutz (2004). (See Appendix
F for a sample of the reduced field data table.)
5.5 Cable Resistance
The wire or cable resistance from the strain transducer to the data logger
reduced the sensitivity of the Wheatstone bridge output. The gage factor was
modified to account for the resistance of the cables and wires with the following
equation:
GF revised = GFjntial* [Rg/Rg+Rl] (5.10)
Where:
GFjnitiai = Gage factor of strain gage inside the transducer
GFrevised = Adjusted gage factor that accounts for the cable or wire
resistance
Rg = Resistance of strain gage (120 ohms for all of the strain
gages used on the Prowers Bridge)
Ri = Resistance of wire or cable used (Vishay 2004)
61


The resistance of the 18 American Wire Gauge (AWG) copper cables from
the data logger to wire lead of the transducer was 0.0209 ohms/meter (0.006385
ohms/foot) (Belove 1986). The resistance of the 26 AWG copper strain gage lead
from the strain gage to the cable was 0.318 ohms/meter (0.041 ohms/foot) (Vishay
2004).
5.6 Transducer Strain
Strain readings from the data logger were converted from millivolt per Volt
to microstrain with the following equations:
s =
4V.
GF{\-2Vr)
TF
(5.11)
With:
Where:
s= Strain (microstrain)
GF = Gage Factor, after the adjustment to account for lead wire
resistance (dimensionless)
TF = Transducer Factor (dimensionless)
V r = Reading from datalogger (mVolt per Volt)
V0ut = Measured bridge output voltage
= Excitation voltage (Campbell Scientific 1996b)
62


Further information and derivation of Equation 5.19 refer to Instruction
Manual for 4WFB120 (Campbell Scientific 1996c). The microstrain reading was
multiplied by the transducer factor (Equation 5.19) to determine the actual strain of
the member.
5.7 Noise Filter
The variations in the strain signals were very evident in the raw strain
data as shown in Figure 5.6. The strain measurements were collected at very
low voltages and could be distorted by any slight disturbance such as wind
blowing the cables.
BOTTOM CHORDS
Prowers Bridge 04J04J06
Figure 5.6 STRAIN MEASUREMENTS FOR THE WINDWARD AND
LEEWARD BOTTOM CHORD EYEBARS AT PROWERS BRIDGE.
The bold trace at the top is the average wind speed to an arbitrary scale.
Measured strain in the leeward eyebar is shown above the measured
strain in the windward eyebar. Both the raw data line and a filtered
line that removes much of the signal noise are shown.
63


By taking a rolling average of several consecutive data points and
advancing the average by the sampling interval of 0.1 second, the data was
filtered and data trends were smoother. The rolling average for the data strain
was based on a study completed by Rutz (2004). In the Rutz (2004) study it was
determined that the 0.5-second to 3-second average closest matched the
amplitude and frequency of the variations in the wind velocity curve. Therefore,
for the bottom chord eyebars, a 2-second rolling average was selected and
applied to the raw data as shown in Figure 5.6. The 2-second rolling average was
also completed on raw data for the end post members (portals), as shown in
Figure 5.7 and Figure 5.8. A 2-second rolling average included 20 consecutive
data points at 0.1-second intervals, adding up to 2 seconds, and advancing the
average at intervals of 0.1 second.
so
40
30
20
510
3
c 0
jg10
-20
-30
-40
-50
Figure 5.7 STRAIN MEASUREMENTS FOR THE SOUTH END POSTS
(PORTALS) AT PROWERS BRIDGE.
SOUTH PORTAL
Prowers Bridge 04/04/09


00
120
Time (seconds)
180
240
WSavg
G9 2-Sec Avq
- G9 G12 G13
-G12 2-Sec Avo G13 2-Sec Ava G14 2-Sec Ava
G14
64


40
NORTH PORTAL
Prowtr* Bridfi* 04*04/06
120
Tim# (seconds)
240
- Average WS/2.5G10 G11 G15 G16
-G102-SecAvg G112-SecAvq G152-SecAvq G162-SecAvg__________
Figure 5.8 STRAIN MEASUREMENTS FOR THE NORTH END POSTS
(PORTALS) AT PROWERS BRIDGE.
5.8 Delta Strain
Every strain gage on the transducer had different residual strain that was
an unknown value. Thus, absolute values of strain data were not used; instead,
changes in strain (delta strain) were used. To eliminate any residual strain
from the strain gage, a local "low" spike of a 3-second interval was averaged at
the bottom of the event and a local "high" spike of a 3-second was averaged at
the top of the event to obtain a strain difference. A delta strain for each of the
transducers was evaluated from the local "low" spike from 44 seconds to 47
seconds at the bottom of the wind event and the local "high" spike from 23
seconds to 26 seconds at the top of the wind event. The final strain value used
65


for comparison to the 3D deck model was the delta strain of the average "high1
spike strains in respect to the average of the "low spike" strains as shown in
Figure 5.9.
BOTTOM CHORDS: Change In strain
Prowers Bridge 04/04/05
Figure 5.9 CHANGE IN STRAIN FOR WINDWARD AND LEEWARD
BOTTOM CHORD EYEBARS AT PROWERS BRIDGE. Both are
baseline traces of the filtered data. Thus, they represent the change in
measured strain, starting from the same point in time as the
corresponding change in wind velocity. For reference, the wind
velocity is shown to an arbitrary scale at the top of the graph.
5.9 Summary of the Reduced Data
After all the strain data was reduced, the delta strains were obtained
shown in Table 5.1. The windward bottom chord eyebars and leeward bottom
chord eyebars were equal and opposite with approximately a +/-3.11 kN (.700
kips) axial force. This shows that the windward and leeward eyebars were
66


experiencing the same amount of loading. This was expected. After the 3D
verification deck models were completed, bending moments for the portal
members were assessed. (See Appendix G for a summary of reduced strain data
for bottom chord eyebars and end post members [portals].)
Table 5.1 Prowers Bridge field data (measured) forces summary.
Member Measured Force
Windward Bottom Chord -3.11 kN (-0.70 kips)
Leeward Bottom Chord 3.11 kN (0.70 kips)
North End Post (Upper Portal) 0.67 m-kN (0.50 ft-k)
North End Post (Lower Portal) -0.48 m-kN (-0.36 ft-k)
South End Post (Upper Portal) 1.33 m-kN (0.98 ft-k)
South End Post (Lower Portal) -1.48 m-kN (-1.09 ft-k)
67


6.
Verification Model Analysis and Comparisons
6.1 Deck Model
The deck models used for verification were the same models developed in
Chapter 3, except for the following:
The wind load was modified to represent the test conditions instead the
uniformly distributed AASHTO wind pressure.
The boundary conditions at the supports were changed to represent actual
in-field boundary conditions.
The conditions of internal member end restraints were changed to represent
actual internal member end restraints.
The methodology for making these changes is described in the following
sections.
6.1.1 Actual Wind Loads
Rather than the uniformly distributed AASHTO wind pressure, the actual
wind pressure from the field test was applied to the Prowers Bridge. During the
field test, the data from the seven anemometers provided the wind velocity. The
wind pressure was really the variable of interest and it varies to the power of 2 to
68


the velocity. Based on Bernoulli's theorem (Weidner and Sells 1966) and
incorporating drag coefficients, the wind velocities from the anemometers were
converted into wind pressures at each location. Equation 6.1 converted the field
wind velocities of the anemometer to wind pressure at each location:
At any elevation:
p= CdCa 0.00256 V2 (6.1)
Where:
p = pressure (psf)
V = velocity (mph)
Cd= 2, which is the Drag Coefficient for most bluff bodies such as
structural shapes (ASCE 1961) as eyebars, channels, and rods.
Cfl = 0.89. Ca is determined from information in ASCE 7-02, Table
C6-1 (ASCE 2002). Ca equals a ratio of average ambient air density
for the actual elevation of the bridge to the average ambient air
density at sea level. Ca is a correction factor for the actual air
density for 1129 m (3704 feet) at Prowers Bridge.
The data from the seven anemometers revealed that the instantaneous wind
speeds were not the same at different sections of the bridge. Therefore, the wind
pressures were not uniform along the bridge. To address this reality and to
account for the different pressure at different areas of the bridge, the bridge model
was divided up into the five sections shown in Figure 6.1. The reason five sections
were used instead of seven was that WS4 wind data showed extremely low wind
speeds due to the adjacent embankment. Thus, WS3 and WS4 wind data was
69


eliminated from the verification model because it would have skewed the results
significantly. WS6 and WS7 anemometers were at the deck level and represented
the lower bridge members sufficiently. The velocities were measured from time 44
seconds to 47 seconds at the low spike and from 23 seconds to 26 seconds at the
high spike as discussed in Section 5.8. The "high" spike values and "low" spike
values for each section were averaged to remove some of the local fluctuation.
Table 6.1 summarizes the reduced averaged wind velocity data for each
anemometer and the average velocity and pressure for each section. The average
wind pressure for each section was applied to the RISA-3D deck model as shown
in Figure 6.2.
Figure 6.1 WIND SECTIONS FOR THE FIELD WIND DATA. Sections
subjected to different uniformly distributed wind pressures. Wind
pressure on Section 1 was determined from a weighted average from
the velocities measured at WS5 and WS1. Wind pressures at the other
sections were similarly determined except for Section 5, which was
determined from WS1 only.
70


Table 6.1 Wind velocities, average section velocities, and average section
pressure on the Prowers Bridge.
Section Anemometer Velocity from anemometer m/s (mph) Section average velocity m/s (mph) Section average pressure Pa (psf)
Section 5-Central WS1 10.0 (22.3) 10.0 (22.3) 110.1 (2.3)
Section 3 -South Upper WS2 14.1 (31.6) 12.1 (27.0) 162.8 (3.4)
WS1 10.0 (22.3)
Not in any Section WS3 7.1 (15.8) Not in any section Not in any section
Not in any Section WS4 1.0 (2.2) Not used in any section Not used in any section
Section 1 -North Upper WS5 5.5 (12.2) 7.7 (17.3) 71.8 (1.5)
WS1 10.0 (22.3)
Section 4 -South Lower WS6 7.5 (16.7) 8.7 (19.5) 86.2 (1.8)
WS1 10.0 (22.3)
Section 2- North Lower WS7 3.2 (7.2) 3.5 (7.9) 62.2 (1.3)
WS1 10.0 (22.3)
71


Figure 6.2 WIND PRESSURE PER SECTION. Wind pressure applied to the five
sections for analysis. North is the direction of the arrow.
6.1.2 Actual Boundary Conditions
In the "traditional" analysis of the deck model, both the roller and pinned
joints were restrained from lateral translation but the roller joints were free to
translate in the longitudinal direction of the bridge. These boundary conditions
were modified to more accurately represent the field conditions. For the actual
boundary condition of the Prowers Bridge, the "pin" boundary condition was
modeled as a joint that was restrained from translation in all three degrees of
freedom. The rotation was restrained except that the rotation about the pin was
released as shown in Figure 6.3. The "roller boundary condition was the same as a
pin, except translation was released in the bridge longitudinal direction, as shown
in Figure 6.3.
72


Figure 6.3 ILLUSTRATION OF PROWERS BRIDGE ACTUAL BOUNDARY
CONDITIONS. This is a photograph of a truss bridge bearing similar
to Prowers Bridge (Rutz 2004). The bottom chord eyebars are at the
right and the physical pin can be seen in the illustration.
6.1.3 Internal Member End Restraints
The traditional analyses of the trusses were based on the assumptions that
all joints at internal members were pinned and free to rotate in any direction
similar to a ball joint. Actual field connections were evaluated and some
modifications were made to the riveted and bolted connections. With relatively
low lateral loading, riveted and bolted connections may not actually slip;
therefore, no rotation would occur. The rivet and bolted connections in the
modified deck model were fixed while connections with physical pins remained
the same.
73


6.2 Verification of Youngs Modulus (EEq)
After the deck model was modified, it was analyzed for all the Eeq cases as
outlined in Section 3. Eeq was compared to the reduced field data discussed in
Chapter 5. The reason for this analysis was to select an EEq so the deck model
would closely represent the response of the real deck. Table 6.2 shows the
windward bottom chord eyebar axial compressive forces for the all cases with the
modified deck model.
Table 6.2 Comparison of reduced field data to the deck model cases of the
maximum axial compressive forces in windward bottom chord eyebars.
Forces are expressed in kN (kips). (Negative = compression).
Deck Model Axial force due to wind load only
Reduced Field Data -3.11 kN
(-0.70 kips)
Case 3: Deck Model -2.73 kN
(E = 5.94 MPa (861 ksi) Figure 3.7) (-0.61 kips)
Case 4: Deck Model -2.14kN
(E = 60 MPa (8700 ksi) Figure 3.7) (-0.48 kips)
Case 5: Deck Model -2.00 kN
(E = 100 MPa (14500 ksi) Figure 3.7) (-0.45 kips)
Case 6: Deck Model -1.88 kN
(E = 150 MPa (21750 ksi)-Figure 3.7) (-0.42 kips)
Case 7: Deck Model -1.78 kN
(E = 200 MPa (29000 ksi) Figure 3.7) (-0.40 kips)
In addition to compiling the table data, a graph was plotted to evaluate the
relationship between Eeq and the net axial force due to field wind load. A trend
was identified, represented by two linear lines with different slopes as shown in
74


Figure 6.4. The first line shows that from zero stiffness to 50 percent of E, axial
force decreased about 75 percent. However, the second line shows that from 50
percent of E to 100 percent of E, axial force decreased about 20 percent. Thus, a
deck had less axial compressive force in the windward bottom chord eyebars due to
wind load based on the stiffness of the deck.
WINDWARD BOTTOM CHORD EYEBARS AXIALFORCE DUE TO WIND LOAD ONLY (MpM
0.614 0.514 0.414 0.314 0.214 0.114 0.014
25000
20000
15000
10000
5000
0
I
£
m
3
>
Figure 6.4 GRAPH OF Eeq VERSUS COMPRESSIVE AXIAL FORCE DUE TO
FIELD WIND LOAD. Note that this graph is different from Figure
3.12 above. In that deck model loads are based on the field wind data
not standard AASHTO wind load requirements.
6.3 Comparison of the Verification Deck Model
with Field Data
The closest relationship between the calculated forces and measured forces
was Case 3, which was the Young's Modulus of 5.94 MPa (861 ksi). Table 6.3
75


shows the comparison of the Case 3 deck model to the measured force from the
field data.
Table 6.3 Prowers Bridge verification summary. Comparison of calculated
forces to measures forces expressed in kN (kips) and m-kN (foot-kips).
Member Calculated Force (Deck Model) Measured Force (Field Data) Correlation: % Difference
Windward Bottom Chord -2.73 kN (-0.61 kips) -3.11 kN (-0.70 kips) 14%
Leeward Bottom Chord 2.74 kN (.62 kips) 3.11 kN (.70 kips) 13%
North Portal Upper 1.52 m-kN (1.12 ft-k) 0.67 m-kN (0.50 ft-k) 56%
North Portal Lower -1.76 m-kN (-1.30 ft-k) -0.48 m-kN (-0.36 ft-k) 73%
South Portal Upper 2.44 m-kN (1.80 ft-k) 1.33 m-kN (0.98 ft-k) 45%
South Portal Lower -2.54 m-kN (-1.87 ft-k) -1.48 m-kN (-1.09 ft-k) 42%
There are good correlations with the measured and calculated data for the
bottom chord members. A fair correlation was observed for south members
(portal) and the north members' (portal) correlations were poor. A possible
reason for a weak north end post member correlation is that the actual bearing
had shifted off its roller. This shift allowed rotation in any direction, similar to a
ball joint. In the RIS A-3D model the roller was released about the pin and
restrained in the other two directions as if the roller had not shifted. The lower
end post of the north portal was also shielded from some wind due the existing
embankment of the newer, adjacent bridge about 15 meters (50 feet) to the west.
76


Thus, due to shielding effects of the embankment on the north end members the
wind pressure in the RISA-3D model was higher than the actual conditions in the
field.
6.4 Leeward Bottom Chord Eyebars Consideration
The main goal of this study was to analyze and verify modeling
techniques for wind loads of the Prowers Bridge. The study mainly focused on
the windward bottom chord eyebars. A load combination that consisted of the
dead load, the wind load, and the pedestrian live load was analyzed because a
stiffer, heavier deck could overstress the leeward bottom chord eyebars.
Based on an allowable stress design per the AASHTO Standard
Specifications of Highway Bridges code, overstress can be 125% for load
combinations that include wind loads. From an historic reference (AISC 1952),
the elastic limit of steel (Fy) was approximately 206.8 MPa (30 ksi). The
allowable stress for tension is 0.55 Fy per AASHTO code (AASHTO 1992).
Thus, the allowable stress for Prowers Bridge is 113.78 MPa (16.5 ksi). A load
combination of the pedestrian live load, 30 percent of the AASHTO wind load,
and the dead load was analyzed on the deck model. The stress on the leeward
bottom chord eyebars was 94.6 MPa (13.72 ksi). In this analysis, the leeward
bottom chord eyebars conformed to AASHTO and was not overstressed. An
77


analysis was not performed on field wind data since it was a lower wind pressure
than AASHTO pressure.
78


7. Summary, Conclusions, and Recommendations
for Future Research
7.1 Introduction
The goal of this study was to examine the stiffening effect of the deck under
lateral (wind) loads, both analytical and by field-testing, of the historic Prowers
Bridge over the Arkansas River near Lamar, Colorado. Analytical studies were
performed to investigate the influence of a deck under wind loading, and a field test
was conducted to verify the analyses.
7.2 Summary of Findings
In this study, "traditional" structural analyses were completed to compare
the effect of wind pressure on the Prowers Bridge skeleton model and on the
same skeleton structure with a deck. The analyses indicated that when a deck
was present, the compressive axial force due to lateral (wind) loads decreased in
the bottom chord eyebars as compared to the traditional skeleton method of
modeling. The analyses also indicated that an increase in dead load of the deck
would cause a decrease in the compressive axial force in the windward bottom
chord members.
79


The field study showed that the measured axial forces in windward
bottom chord members closely represented the calculated axial forces from the
analytical deck model. Therefore, the deck model was a close representation of
the actual in field conditions. The equivalent Young's Modulus was developed
for the corrugated steel deck, treating the deck as if it were isotropic.
7.3 Conclusions
This study shows that with a significant increase in deck dead load results in
a tensile axial force instead of a compressive axial force in the windward bottom-
chord eyebars. The current deck dead load increased by 60 percent than a
traditional timber deck dead load and the traditional skeleton method of modeling
conformed to the AASHTO wind design loading requirements. Therefore, it was
concluded that the compressive axial force in the windward bottom chord eyebars
was not a limiting issue with the current deck at the Prowers Bridge. In addition,
the Prowers Bridge was modeled with the current deck to show the stiffening
effects of the deck when exposed to wind loads. The deck model showed that the
structure was laterally stiffer and that tensile axial force in the bottom chord
eyebars increased by 86% with AASHTO wind requirements. Therefore, the
presence of a deck stiffens a truss bridge laterally.
80


The field study verified that the actual conditions of Prowers Bridge closely
represented the analytical deck model. The lateral strength of the windward bottom
chord members were improved with the deck model. Modeling the plate elements
could be beneficial when analyzing historic truss bridges for pedestrian conversion.
7.4 Recommendations for Future Research
The following constitute areas for further research on preserving historic
truss bridge for pedestrian use.
Analyze different deck configurations on historic truss bridges such that
an optimal deck design can be determined to satisfy the wind loading
requirements of the current code for both the windward and leeward
bottom chords eyebars.
Evaluate the current industry bridge decks and develop a deck that could
replicate a timber deck but have a higher dead load.
Test a sealed bridge model in a wind tunnel to account for the effect of
upward or downward wind pressure on the deck.
Complete a full design check on an historic truss bridge to verify if there
are any other code-related issues for pedestrian conversion.
Complete various lab tests and in-depth model verification on the
corrugated steel decking.
81


Compile a corrugated steel deck table of equivalent Young's Modulus for
computer programs that have only isotropic properties for plate elements.


APPENDIX A
FIELD INSTRUMENTATION SET-UP FOR DATA
ACQUISITION WITH PC9000
1. Turn on the field laptop and place the PC9000 Program CD into the CD drive.
2. Before going out to the field, download the PC9000 Program to the field
laptop. (Once the program is downloaded to the computer, it is not necessary
to do this again as long as the same laptop is used.)
3. Plug in the Serial Cable to the field laptop as shown in Figure A.l.
Figure A. 1 SERIAL CABLE ATTACHED TO LAPTOP
83