EXPERIMENTAL AND ANALYTICAL STUDY OF CONCRETE BRIDGE DECKS CONSTRUCTED WITH FRP STAY-IN-PLACE FORMS AND FRP GRID REINFORCING by David A. Dieter A thesis submitted in partial fulfillment of the requirements for the degree of Masters of Science (Civil Engineering) at the UNIVERSITY OF WISCONSIN – MADISON 2002
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
EXPERIMENTAL AND ANALYTICAL STUDY OF
CONCRETE BRIDGE DECKS CONSTRUCTED WITH FRP
STAY-IN-PLACE FORMS AND FRP GRID REINFORCING
by
David A. Dieter
A thesis submitted in partial fulfillment of
the requirements for the degree of
Masters of Science
(Civil Engineering)
at the
UNIVERSITY OF WISCONSIN – MADISON
2002
i
ABSTRACT
This thesis describes laboratory testing, test results, and research to develop selected design
recommendations for construction of concrete bridge decks using only a Fiber Reinforced
Polymers (FRP) reinforcement system. The use of FRP reinforcing is being pursued to
increase the durability of bridge decks. The unique aspect of the new reinforcement system
is that it will use FRP stay-in-place formwork to serve as formwork and as the bottom
transverse reinforcement of the bridge deck. In addition, a bi-directional prefabricated FRP
grid will be used for the top layer of concrete reinforcing. Full-scale prototype laboratory
testing on FRP reinforced concrete slabs and beams was conducted to compare results of
experimental load distribution widths to American Association State Highway and
Transportation Officials (AASHTO) design equations used for steel reinforcement. In
addition, testing provided performance criteria for the reinforcement system on strength
factors of safety, failure modes and fatigue. Using experimental results for verification, an
analytical model of a prototype slab/girder bridge was built to determine the negative
moment distribution width of the reinforcement system. Test results indicate AASHTO’s
effective distribution width equation for steel reinforcement is applicable. In addition, design
recommendations are given for punching shear capacity based on American Concrete
Institute (ACI) 318 Building Code Requirements. Analytical modeling indicates that the
FRP reinforcement system is adequate for use in the prototype bridge, especially from a
strength perspective.
ii
ACKNOWLEDGEMENTS
I would like to thank Stan Woods and Gerry Anderson from the Wisconsin Department of
Transportation for their cooperation. Without their help and proactive contributions this
project would of never materialized. I would also like to recognize Tom Strock from the
Federal Highway Administration for his constructive ideas. Also a key to this project was
Bernie Gallagher of Alfred Benesch and Company. Bernie’s exhibited innovative ideas and
endless enthusiasm during our brainstorming research and design meetings. On behalf of the
University of Wisconsin research team, I would like to thank Innovative Bridge Research and
Construction (IBRC) for the funding for which was the fuel that drove this research.
I like to extend my gratitude to Bill Lang and John Dreger at UW Structures and Materials
Testing Laboratory. Their expertise and willingness to help ensured this project was carried
out expeditiously. UW undergraduates Brian Beaulieu, Eric Helmueller, Wyatt Henderson
and Beau Sanders provided essential help during the construction, setup and testing of the
research specimens.
Finally, I would like to extend my deepest appreciation and admiration to Dr. Lawrence
Bank, Professor Michael G. Oliva, and Professor Jeffrey Russell for their effort, guidance
and understanding. I hold your wisdom, support and timely encouragement as a key for my
successful completion of this thesis. I cannot express enough about the importance of the
solid education provided by these three gentlemen.
Figure 1.1 Mock-Up of FRP Reinforcement System Figure 1.2 Typical cross-section in the primary direction of the proposed FRP
reinforcement Figure 2.1 A typical cross-section and assembly of a steel SIP deck form used in
conventional bridge deck construction Figure 3.1 I.F. Corporation casting bed Figure 3.2 FRP deck forms Figure 3.3 Grid with 2.5” chair and tie Figure 3.4 Longitudinal FRP bi-directional grid splice, Panel C Figure 3.5 A Beam E specimens after forms were removed Figure 3.6 Placement of concrete with chute and consolidation with vibrator Figure 3.7 Crane lifting a test specimen of the flat bed truck at SMTL Figure 3.8 Reynolds Movers move Panel B into SMTL Figure 3.9 Cross-section view of the primary direction of modified FRP reinforced
Panel C2 Figure 3.10 During the construction of Panel C2 at UW-SMTL Figure 4.1 Load contact area of Concrete Deck Panels (A, B, C, C2) Figure 4.2 Reaction frame for Concrete Deck Panels (A, B, C) Figure 4.3 A view of the concrete block supporting a concrete deck panel Figure 4.4 Series of ±0.1” LVDTs to the right of the actuator Figure 4.5 An example of a concrete strain gauge used Figure 4.6 MTS 458.20 controller, instrument power supply, Validyne signal
amplifier, multiplexer and A/C DaqBook converter Figure 4.7 Elastic test pattern for LVDT displacements of a concrete deck panel
quadrant Figure 4.8 Instrument type and location on the top surface of Panel A for the
inelastic test Figure 4.9 Instrument type and location on the bottom surface of Panel A for the
inelastic test Figure 4.10 Plan view of a positive moment beam, Beam D Figure 4.11 Positive moment beam test set-up Figure 4.12 Three-inch square bar line load assembly Figure 4.13 Instrument type and location for top surface of Beam D1 and D2 Figure 4.14 Instrument type and location for bottom surface of Beam D1 and D2 Figure 4.15 Plan view of negative moment beam (Beam E) Figure 4.16 Negative moment beam test set-up Figure 4.17 Instrument type and location of the top surface of negative moment beam Figure 4.18 Instrument type and location of the bottom surface of negative moment
Figure 4.20 Instrument type and location for top surface of concrete deck Panel C2 (inelastic test) Figure 4.21 Instrument type and location for bottom surface of concrete deck Panel
C2 Figure 4.22 Profile strain gauge location of section A-A of Figure 4.20 Figure 5.1 Preload concrete surface cracks at one supported end of Panel A Figure 5.2 Load versus actuator stroke for the inelastic test of Panel A Figure 5.3 Difference between the stroke and potentiometer readings, Panel A Figure 5.4 Oval concrete cracking pattern, post-inelastic test of Panel A Figure 5.5 Deflection profile of Panel A at the center span Figure 5.6 Longitudinal strains recorded along the centerline of span, Panel A Figure 5.7 Differential deflection of the middle two FRP deck forms Figure 5.8 Load versus adjusted actuator stroke for the inelastic test of Panel B Figure 5.9 View across one support showing uplift of one of the four corners and a
large vertical crack during Panel B test Figure 5.10 Difference between the stroke and potentiometer readings for Panel B Figure 5.11 Plan view of the oval concrete crack pattern on the top surface of Panel B Figure 5.12 Deflection profile of Panel B along the span centerline Figure 5.13 Material strains recorded along the centerline of span, Panel B Figure 5.14 Load versus lateral displacement opening of the center shiplap joint of
Panel B Figure 5.15 Lateral opening of the center shiplap joint for Panel B Figure 5.16 Load versus actuator stroke for the inelastic test of Panel C Figure 5.17 Difference between the stroke and potentiometer readings for Panel C Figure 5.18 View of the oval concrete pattern on the top surface of Panel C Figure 5.19 Elevation view of the bottom surface of FRP deck forms deflecting apart Figure 5.20 Deflection profile of Panel C across the center of span Figure 5.21 Material strains recorded along the center of span, Panel C Figure 5.22 Lateral opening of center shiplap joint for the inelastic test of Panel C Figure 5.23 Post-Failure Panel C Showing Punching Shear Failure Under Load Patch Figure 5.24 Load versus actuator stroke for the inelastic test of Beam D1 Figure 5.25 Depth of section versus material strain as a function of load, Beam D1 Figure 5.26 The concrete crushing near the load for both sides of Beam D1 Figure 5.27 Load versus actuator stroke for the inelastic test of Beam D2 Figure 5.28 Depth of section versus material strain as a function of load, Beam D2 Figure 5.29 Flexure-shear cracking on left and flexure cracks on the right, for Beam
D2 Figure 5.30 Two chunks of concrete dislodged at the end of Beam D2 Figure 5.31 Load versus actuator stroke for the inelastic test of Beam D3 Figure 5.32 Depth of section versus material strain as a function of load, Beam D3 Figure 5.33 Relationship of 3 concrete strain gauges at the same distance from the
centerline Figure 5.34 Material strains through the depth of section for 12” away from the
centerline Figure 5.35 Horizontal shear and flexural-shear cracking of Beam D3
ix
Figure 5.36 Load versus deflection curves for Span A and Span B for Beam E1 Figure 5.37 Strains from the section centered over the middle support Beam E1 Figure 5.38 Top surface strain gauge between the negative moment cracks, Beam E1 Figure 5.39 Strains through the depth of section as a function of load for Beam E1 Figure 5.40 Failure of Beam E1 for each side of beam in Span A. Figure 5.41 Load versus deflection curves for Span A and Span B for Beam E3 Figure 5.42 The first two concrete cracks over the middle support, Beam E3 Figure 5.43 Strains from the section centered over the middle support, Beam E3 Figure 5.44 Strains through the depth of section as a function of load, Beam E3 Figure 5.45 Each side of Beam E3 showing a shear failure occurring at 67 kips,
Beam E3 Figure 5.46 Concrete cracks on each side of Span A of Beam E5 Figure 5.47 Cracking at failure on side of beam without gravel epoxied to FRP form Figure 5.48 Additional cracks of Beam E5 Figure 5.49 Load versus actuator stroke for the inelastic test of Panel C2 Figure 5.50 Difference between the stroke and potentiometer readings for Panel C2 Figure 5.51 Top surface cracking pattern of Panel C2 Figure 5.52 Differential deflection of the center FRP deck forms in Panel C2 Figure 5.53 Deflection profile of Panel C2 along the center of span Figure 5.54 Strain profile of the five strain gauges, through the depth of section,
Panel C2 Figure 5.55 Other two Strain gauges 9” from center of load, along center of span,
Panel C2 Figure 5.56 Lateral displacement of center shiplap joint of Panel C2 Figure 5.57 Post mortem cut along the center of span for Panel C2 revealed the start
of a punching shear failure Figure 5.58 Preload concrete cracks for the fatigue beam Figure 5.59 Relative stiffness of the 200-kip span during the fatigue test after
various number of cycles Figure 5.60 55 kip actuator relative stiffness of 2 million-cycle fatigue test Figure 5.61 Load versus deflection curves for Span A and Span B for Fatigue Beam Figure 5.62 Cracking on each side from the inelastic test of fatigue beam Figure 6.1 Moment-curvature relationship for the positive moment beams Figure 6.2 Positive moment distribution for the elastic load of 20.8 kips for a 9’
wide panel Figure 6.3 Positive moment distribution at panel failure Figure 6.4 Load perimeter area to define bo Figure 6.5 Load versus deflection curves for all of the positive moment beams Figure 6.6 Negative moment curvature relationship centered over the middle
support, Beams E Figure 6.7 Load versus deflection curves for Span A of the negative moment beams Figure 7.1 Comparison of experimental (inelastic test) and analytical deflections at
24” from center of load along center span Figure 7.2 Deflection profile for experimental (elastic test) and analytical
measurements at center span
x
Figure 7.3 Deflection profile for experimental (elastic test) and analytical measurements along A24
Figure 7.4 Moment distribution profiles over a 16’ width Figure 7.5 Negative moment profile of prototype deck for negative moment
xi
LIST OF TABLES
Table 3.1 The list of all 8” thick specimens constructed Table 3.2 Summary of concrete pours Table 4.1 Selected Mechanical Properties of FRP Material Table 4.2 Data record table for Panel A Ultimate Test (2/26/01) Table 4.3 Data record table for Beam D1 (4/10/01) Beam D2 (4/6/01) Table 4.4 Data record table for Beam E1 (5/9/01), Beam E3 (5/16/01), Beam E5
(5/2/01) Table 4.5 Data record table for the inelastic test of Beam E2 (fatigue beam) Table 4.6 Data record table for Panel C2 Ultimate Test (6/27/01) Table 5.1 Measured widths of the first concrete crack to develop on the top surface,
Beam E1 Table 5.2 Measured widths of the first concrete crack to develop on the top surface,
Beam E3 Table 5.3 Concrete strengths (± one standard deviation) tested according to
ASTM C-39 Table 6.1 Table of normalized curvature values to establish an approximate linear
relationship Table 6.2 Live load distribution widths for the 9’ wide concrete deck panels Table 6.3 Punching shear values and calculated capacities for the concrete
deck panels Table 6.4 Strength and effective stiffness results of the concrete deck panel tests Table 6.5 Strength and effective stiffness results of the positive moment beams Table 6.6 Strength and effective stiffness results of Span A for the negative
moment beams Table 6.7 Span-to-deflection ratios for the tested FRP specimens Table 6.8 Comparison of measured and calculated stiffness for the tested FRP
specimens Table 9.1 Design Methodologies for Steel and FRP Reinforcement for Concrete
Bridge Deck Based on 8’-8” Span Table 9.2 Performance of Positive Moment Beams to ODOT criteria for 3’ wide, 8”
thick simply supported beams
1
1 INTRODUCTION
1.1 Introduction
As it can be noted from driving throughout the United States, the use of concrete
bridge decks is ubiquitous. In fact, nearly 70% of the existing bridge decks in the United
States are cast-in-place reinforced concrete (Bettigole, 1997). Many state Departments of
Transportation find the typical design life of a number of these concrete bridge decks have
been reduced, or have been exceeded (Alkhrdaji et al., 2000 and Chajes et al., 2000). Today,
as it has been for many years, the conventional method to reinforce a concrete bridge deck is
to use steel reinforcing bars. A primary cause of deterioration of a number of these concrete
bridge decks is due to the corrosion of these steel reinforcement bars (Bradberry, 2001).
Corrosion of steel reinforcement adversely affects the long-term durability of the concrete
because it leads to spalling and cracking, eventually softening the deck and leading to the
possibility of member failure (Taly, 1998).
With excessive deflections and/or the onset of member failure, these concrete bridge
decks become defined as structurally deficient. A bridge that falls into this category is either
rated for a lower load capacity, or if the damage is severe enough, taken completely out of
service (AASHTO, 2000). This has caused economic costs in rehabilitating or replacing
these bridges, as well as imposing an economic penalty to the traveling public. This also
inherently drives up the anticipated life-cycle cost. This life-cycle cost directly measures the
initial costs (like materials and installation) as well as the long-term operational costs (which
depends on such variables as expected maintenance and repair). The corrosion of steel
2
reinforcement has a negative impact upon the life-cycle costs by necessitating unplanned
costs to mitigate premature deterioration. These unintentional costs to replace prematurely
failing bridge decks (through materials, labor, relocation of traffic, etc.) and the
inconvenience to the traveling public are leading a search for a more durable material to
replace steel reinforcement.
Wisconsin Department of Transportation (WisDOT) teamed up with the University of
Wisconsin-Madison Civil and Environmental Engineering Department, Federal Highway
Administration and industry to test and implement a new technology to increase long-term
durability as well as to reduce construction time to assemble the deck reinforcement. The
new technology proposed by the UW is the use of a non-metallic glass Fiber Reinforced
Polymer (FRP) reinforcement system, to replace the use of traditional steel reinforcement.
The main advantage of using FRP is that it is a naturally inert material, and therefore,
resistant to most corrosive agents anticipated for highway application. WisDOT and
University of Wisconsin applied and were awarded funding through the Federal Highway
Administration’s Innovative Bridge Research and Construction program (IBRC) to
investigate the use of a the Stay-In-Place (SIP) FRP deck form in combination with a bi-
directional FRP top grid concrete reinforcement system.
In addition to increasing long-term durability, this FRP reinforcement system should
provide additional benefits. From an economic point of view, this FRP deck reinforcement
system is anticipated to reduce construction labor cost through ease of assembly and the
elimination of traditional cast-in-place concrete deck forms and falsework. It is anticipated
that reduced labor costs may offset the higher material cost, compared to steel reinforcing.
3
The traditional methods to design steel reinforced concrete bridge decks are not
entirely applicable with concrete bridge decks reinforced with FRP. In general, a lower
modulus of elasticity, a lack of ductility, variation in possible geometrical shapes, the
reliability of composite action, as well as other differences from design with steel, which
means that this FRP system has required experimental study. In order for the proposed FRP
reinforcement system to provide the presumed long-term durability for concrete bridge
decks, this reinforcement must satisfy the strength and serviceability requirements imposed
by the Load Factor Design (LFD) AASHTO Standard Specifications for Bridge Design, 16th
edition. Because of the relatively short time FRP materials have been on the market and
available for use in structural applications, there has not been enough time for the evolution
of a prescribed design methodology for the use of FRP in reinforced concrete bridge decks.
Therefore, in general, the purpose of this research thesis is to evaluate this proposed FRP
reinforcement system’s structural performance and determine if it can satisfy some critical
general strength and serviceability requirements for concrete bridge deck design through
laboratory testing and the use of some analytical tools.
1.2 FRP Reinforcement System and Design Methodology
The FRP reinforced concrete system is made up of two layers, a top and bottom layer
(see Figure 1.1 for picture of a FRP system mock-up). Figure 1.2 illustrates a typical cross-
section of the proposed FRP reinforcement system (cross-section cut is parallel to traffic).
The bottom tensile reinforcement is comprised of a pultruded glass FRP Stay-in-Place (SIP)
deck form spanning between, but not continuous over, the girders, which are parallel to the
traffic direction. A glass FRP bi-directional grid provides the top layer of concrete
4
reinforcement. These two glass FRP products to reinforce concrete bridge decks have been
previously studied (Bank et al, 1992 and Harik et al, 1999), but they have never been studied
together.
The FRP deck form is analogous to the main positive steel reinforcement typically
placed at the bottom of the deck perpendicular to the girder system. Each deck form is 18”
(457.2 mm) wide and overlaps with adjacent deck forms via a shiplap joint. Each of the SIP
forms are stiffened by two 3” (76.2 mm) square hollow corrugations centered 9” (228.6 mm)
apart. To ensure composite action through horizontal shear transfer between the deck form
and the concrete, ¼” (6.35 mm) aggregate is bonded to most of the horizontal surface area of
the form with Concresive 1090 epoxy. This corrugated FRP deck form has been
previously used in bridge deck construction in the Ohio Department of Transportation’s
Salem Avenue Bridge project. The performance of the FRP SIP deck form in the Salem
Avenue Bridge has been documented by Reising et al (2001). Composite Deck Systems of
Dayton, Ohio supplies the FRP deck form.
Figure 1.1 Mock-Up of FRP Reinforcement System
FRP Deck Form
Bi-Directional Grid
Plastic Chair
5
A pultruded bi-directional glass FRP grid panel (typically 4’ wide and lengths can
vary) provides the top transverse and the top longitudinal (direction with respect to the
girders) tensile reinforcement. The transverse reinforcement for the deck’s negative moment
over the girder is provided by 2” (50.8 mm) deep, “I” shaped, FRP bars. In the longitudinal
direction, the temperature and shrinkage reinforcement is supplied by oval FRP bars (dark
bars in Figure 1.1). The oval FRP bar stays in the plane of the 2” “I” bars, by penetrating
through and mechanically locking with the “I” bar web. The orthogonal members of the grid
are spaced at 4” (101.6 mm) on-center in each direction. Both bars are quite smooth and
cannot develop adequate bond to the concrete. Since they lie within the same plane and are
inherently connected, they mechanically anchor one another. The use of the pultruded grid
system to reinforce concrete slabs and bridge decks has been studied previously by Bank et
al. (1992,1993, and 1995). For the experimental phase, a smaller bi-directional grid was used
as a splice between adjacent main reinforcement bi-directional grid panels to provide
continuity of reinforcement. However, in the prototype bridge FRP reinforcement bars will
be used as a substitute to provide the reinforcement continuity. The bi-directional FRP grid
panel and splice grid are proprietary products of Strongwell Incorporated located in
Chatfield, Minnesota.
6
Epoxied 1/4" GravelShiplap Joint
FRP Bi-directional Grid
FRP Bi-DirectionSplice Grid
0'-1 1/2"
8"
FRP Deck Form Figure 1.2 Typical cross-section in the primary direction of the proposed FRP reinforcement
The design methodology presented in American Concrete Institute (ACI) Guide
440.1R-01, “Guide for the Design and Construction of Concrete Reinforced with FRP Bars”
was useful to calculating reinforcing requirements (ACI, 2001). Like most ACI documents
on the design of reinforced concrete members, a strength design philosophy was adopted.
The section was designed as over-reinforced to force a concrete crushing failure, since from a
ductility point of view, this is a marginally more desirable failure than a FRP rupture.
Presumed analytical strength and service requirements were calculated assuming an effective
distribution width as provided by the LFD AASHTO Highway Bridge Design Specification,
16th edition for steel reinforced decks (AASHTO, 1996). The FRP deck form reinforcement
was assumed to be adequate for serving as tensile reinforcement to the composite section
based on previous experience. With the areas of the grid bars set for the top grid, the only
design variable left to determine was the spacing, 4” on center, of the grids for negative
moment capacity. The designed cross sectional characteristics of the proposed FRP
reinforced system is depicted in an illustration of a typical 18” wide cross section shown in
Appendix A.
7
1.3 Objectives
From a structural performance point of view, the main quantitative and qualitative
objectives are as follows:
1. To examine the FRP reinforcement’s ability to laterally distribute the load in the
positive moment region and compare the experimental results to an existing
AASHTO distribution width equation that is intended for use with conventional steel
reinforced decks;
2. To determine the mode and strength at failure and evaluate the factors of safety, and
to compare the governing equation for mode of failure to the capacity measured;
3. To develop a finite element model of a FRP reinforced concrete bridge deck of the
prototype bridge and to determine negative moment distribution width for the deck
over support girders when at a load causing the concrete to start to crack over the
girder, also, to determine these loads required to initiate longitudinal concrete
cracking over the girder;
4. To investigate the effects of accelerated fatigue loading on the FRP reinforcement
system and the mechanical bond between the FRP deck form and the concrete;
5. To qualitatively determine the post-ultimate ductility of the FRP reinforcement
system;
6. To evaluate qualitative and quantitatively the difference between partial and full
coverage of the epoxy coated aggregate on the FRP SIP deck form on the composite
strength in bending; and
7. To evaluate the constructability of the FRP reinforcement components for use in
bridge construction (i.e., how well they fit together).
8
1.4 Scope
The IBRC project consists of three main phases: to investigate the strength and
serviceability behavior of the composite concrete reinforcement system though laboratory
testing, to implement the tested design into a prototype bridge deck, and finally, to complete
a long-term monitoring program. The scope of this thesis, however, is limited to
investigating the strength and some serviceability behavior through laboratory testing and
analytical modeling. The results from the testing conclude that this FRP reinforcement
system is suitable for use in the prototype bridge and for bridges in general. Therefore, it
will be implemented in the construction of one part of a new twin bridge structure owned by
the Wisconsin Department of Transportation. The other twin of the new bridge structure, on
USH 151 near Waupun, Wisconsin, will use conventional steel reinforcement for the bridge
deck, thus providing a hands-on opportunity to compare performance and durability of the
two materials in service. A detailed description of the prototype bridge structure is provided
in Appendix A.
The scope needed to achieve the objectives of this project is outlined as follows:
1. Constructed and experimentally tested, through 3-point bending with a load footprint
of approximately equal to the contact area of a double wheel tire, four FRP reinforced
concrete deck panels to determine the distribution width, mode of failure and strength
factors of safety.
2. Constructed and experimentally tested, through 3-point bending, three FRP reinforced
concrete beams to isolate and determine the positive moment bending characteristics
of the concrete/FRP deck form composite section.
9
3. Constructed and experimentally tested three 2-span continuous FRP reinforce
concrete beams to isolate and determine the negative moment bending characteristics
of the concrete/FRP bi-directional grid panel composite section.
4. Constructed and experimental tested one 2-span continuous FRP reinforce concrete
beam to isolate and determine the fatigue resistance of the FRP materials and the
mechanical bond between the FRP deck form and the concrete.
5. Used the experimental values from the above tests to verify and construct a finite
element model for a section of the prototype bridge that better replicated the scale,
continuity, and end conditions expected in the actual prototype bridge deck. This was
done to determine the negative moment distribution width and the load that produced
concrete cracking over the girder.
Not apart of this scope is the formal comparison of service load deflections of the
FRP reinforced section to AASHTO serviceability criteria. Harik et al (1999) studied the
AASHTO LRFD service limit state for the service load deflection performance of a concrete
section reinforced with the same SIP FRP deck form for the tensile reinforcement, in
combination with, Glass Fiber Reinforced Polymer (GFRP) rebar for the compression
reinforcement.
10
2 LITERATURE REVIEW
2.1 Introduction
Fiber Reinforced Polymer high-performance materials are being diversely used for
structural applications in highway bridges. This chapter presents the literature review of
some previous studies for the use of FRP materials, specifically for highway bridge deck
applications. Most papers reviewed in this chapter address, in some part, the design
performance of their specimens according to AASHTO bridge design specifications. First
part of the chapter will review a few studies where the entire superstructure system is
composed of an FRP modular core slab supported on steel girders, and even in one case, a
study where a FRP core system alone acts as the highway superstructure. The next portion of
the literature review discusses selected studies where FRP bars are implemented as the
concrete composite reinforcement for slabs. Another area of the literature review will
examine studies of the same, or similar, concrete FRP reinforcement grid studied in this
thesis. Also included in this literature review is a look into the conventional use of non-
composite steel SIP forms with reinforced concrete as a bridge deck. The final part of the
literature review examines a couple of studies where the same FRP SIP deck form studied in
thesis has been employed as the tensile reinforcement for composite concrete slabs.
2.2 Bridge Decks Constructed from FRP Composites
Recently, a number of states have begun constructing bridges where the entire
structural deck system is comprised of FRP composites. A fully FRP bridge deck has many
beneficial traits over a conventional steel reinforced concrete deck, such as being lighter,
11
corrosion resistant and a higher strength-to-weight ratio. Delaware Department of
Transportation (DelDOT) has replaced an existing concrete slab, on steel girders, with a
lightweight glass FRP deck. Chajes et al (2001) detailed the rehabilitation of this existing,
low-volume, 35 ft clear span bridge and then documented the performance of a field load
test. The design requirements called for a 10-inch thick GFRP slab, which weighs 30 lb/ft2,
to continuously span over six longitudinal girders spaced at 33.8 inches. For a wearing
surface, the slab had a shop applied polymer concrete finish. Although the paper did not
address the design details, it did hint that the polymer composite slab was governed by
deflection limits.
The field test was conducted with a fully loaded 10-wheel dump truck and both the
performance of the GFRP deck and the girders were of interest. Three different load
applications were completed: static, slow roll and dynamic. The test results for
displacements, peak strains, bending of the girders, transverse load distribution, longitudinal
and transverse bending of the deck were recorded. According to the test results, the GFRP
composite bridge slab performed well within the serviceability and strength requirements.
Of specific interest, the superstructure achieved a deflection ration of L/2100 (limit
recommended by code is L/800) without composite action between the deck and the girders.
One design aspect the Chajes et al (2001) study did not address, because of low-
volume traffic, was the long-term fatigue performance of the FRP composite slab. In order
to jump from low to high volume traffic use, the fatigue performance of these FRP composite
slabs needed to be experimentally documented. A study by Lopez-Anido (1998) investigated
the long-term performance of a modular FRP composite bridge deck, similar to the Chajes et
al system, under cyclic loading. The Lopez-Anido study examined modular FRP deck panels
12
supported on longitudinal girders. The FRP composite deck system performed well, without
major degradation in stiffness or strength, when it was subjected to 2 million load cycles of
an HS20-44 design truck adopted from the LRFD design philosophy for highway steel
structures. These results help provide some confidence for FRP as a competent material to
provide fatigue resistance.
A step beyond the use of FRP composites to act as the deck supported on a steel or
concrete stringer superstructure system is to employ FRP composites as the superstructure
itself. The New York State Department of Transportation replaced a significantly
deteriorated 25-foot span, steel reinforced concrete slab bridge (ADT of 300 vehicles per
day) with a two-piece superstructure cell core system that provides stiffness in two
directions. The two-foot thick cell core system was designed according to AASHTO strength
and serviceability criteria. After the original superstructure was removed, the two-piece
system was installed within six hours and the bridge was reopened to traffic a few weeks
later. Before being opened to traffic, the bridge was proof tested with HS25 truck loadings to
check the structure’s integrity, to establish base line condition and to compare actual
performance with theoretical calculations. Proof testing concluded that the load capacity was
greater than what was determined by analytical analysis and the bridge’s span to deflection
ratio was L/2010. This study also determined that FRP composites could be a cost-effective
alternative for short span bridge superstructures.
2.3 Concrete Bridge Decks Reinforced with FRP Bars
Probably the most natural use of FRP materials to reinforce concrete bridge decks is
to replace conventional steel reinforcing bars with FRP bars. Bradberry (2001) of the Texas
13
Department of Transportation (TexDOT) documented the structural design of a concrete
bridge deck with the top mat of concrete reinforcement made of GFRP bars. The designed
nominal deck thickness was 8 inches and was supported by five prestressed concrete girders
spaced at 7.84 feet, center-to-center. The main design issues discussed, relative to this thesis,
were design ultimate strength, calculated crack width and shear strength of the concrete
bridge deck designed with GFRP bars. Two main conclusions from this paper were: the
author found that the bar spacing in the top GFRP mat was controlled by the serviceability
issue of crack width and in the author’s judgment, the significant lack of ductility displayed
by FRP bars was not a barrier for their use in concrete bridge decks. The conclusions exhibit
the main weaknesses for use of FRP bars, the lack of stiffness and ductility. Although, the
FRP bars do provide a way to prevent deterioration of concrete bridge decks where steel has
performed inadequately.
Bradberry (2001) did state concerns about unanswered question about long-term
durability issues with GFRP. Kumar et al (1998) did conduct a study on GFRP reinforced
concrete deck-stringer systems under fatigue loading. They measured the rates of
degradation of strength and stiffness and compare the performance of GFRP concrete decks
to previous research that used steel reinforcement. The main conclusion is that FRP
reinforcement bars performed well, when compared to steel, in the crack propagation zone.
This study also found that 2 million fatigue cycles, with a load at about 20% of ultimate
strength, could be conservatively assumed to be 80% of the design life of the FRP bars.
These results provide some confidence that FRP bars can safely satisfy fatigue serviceability
requirements for concrete reinforcement of bridge decks.
14
2.4 Concrete Bridge Decks and Concrete Slabs Reinforced with FRP Grids
Survey of the literature addressing FRP composites to reinforce concrete bridge decks
reveal that the main alternative to FRP bars is the use of FRP grids or gratings. Bank et al
(1992, 1993, and 1995) conducted some of the earliest testing that investigated the
performance of FRP gratings. The Bank et al study (1992) investigated full size bride deck
panels reinforced with two different commercially provided orthogonal gratings (Safe-T-
Grate and Duradek) along with control slabs reinforced with steel bars. Each full size
slab was designed according to AASHTO strength and serviceability recommendations using
a nominal HS-25 wheel loading. The slabs were only reinforced in the bottom layer to test
positive moment behavior and failure. As designed, the FRP slabs failed due to concrete
compression failure and the controls specimens failed due to the yielding of steel bars (and
then concrete crushing). This study helped demonstrate the behavior of the different failure
modes between the FRP and steel reinforcement. Experimental service load deflections of
the FRP reinforced slabs were close to the analytical values calculated. Also, the safety
factors of ultimate strength to service load were greater than three. In addition, the ultimate
loads of the FRP slabs were greater than the load that caused yielding in the steel reinforced
control slab. This study provided early evidence that FRP gratings could be a workable
reinforcement alternative for bridge decks.
In 1993, Bank and Xi investigated analytical and experimental agreement of concrete
slabs reinforced with FRP gratings using the same gratings from the 1992 study. Dependent
upon the modulus of elasticity of the FRP, they concluded that the analytical models
illustrated acceptable predictions. The accuracy of the slab stiffness, however, is dependent
15
upon the accuracy of the value for the FRP modulus of elasticity. Another conclusion, at that
time, is that addition study is required to predict the shear strength of these slabs.
Bank and Xi, in 1995, evolved their research to investigate the shear, or punching
shear, capacity of the concrete slabs reinforced with FRP gratings. Punching shear capacity
of a reinforced concrete slab-stringer bridge system is important because the generally
accepted failure of a concrete deck is due to punching shear (exception to overhangs) (Bank
and Xi, 1995 and Bradberry, 2001). In this study, punching shear capacity was tested with
variables of: grating bar spacing, grating panel layout, and panel splice detail. Although
there were no steel reinforced slabs tested for comparison, the main conclusion is that the
punching failure behavior was similar to what is typically seen for a steel reinforced slab. In
addition, since concrete alone provides the punching shear resistance, the variables play little
role to capacity, although they did influence the post-ultimate behavior.
A more recent study by Matthys and Taerwe (2000) examined flexural and punching
shear behavior of concrete slabs reinforced with two types of NEFMAC grids (carbon and a
mixture of carbon and glass fibers). With the FRP slabs designed to be comparable to a
reference steel reinforced slab, this study provided a direct comparison of performance
between FRP and steel. From a bending perspective, the authors conclude that the benefit of
the added strength of the FRP grids is only partially used since the controlling design criteria
was for serviceability deflections. They also conclude that the safety factors for strength are
higher, however, the ductility is lower than steel. From a punching shear capacity
perspective, two main conclusions stand out. First, the FRP slabs, with comparable
reinforcement strength as the steel, showed lower punching shear resistance and stiffness
(although with an increased reinforcement ratio and or depth of section this can be
16
overcome). Second, introducing an equivalent reinforcement ratio (to take into account the
axial rigidity of the reinforcement), ρE/Es, into the analytical predictions from a variety of
international code equations provided satisfactory factors of safety for the experimental
results of the punching shear capacities.
2.5 Use of Steel SIP Forms with Reinforced Concrete Bridge Decks
The conventional use of steel stay-in-place deck forms for bridge applications is
commonplace. They are used, however, only for support of wet concrete between girder
spans and are not applicable to replace or reduce the bottom transverse steel reinforcement.
They are not intended nor are they permitted to act compositely with the concrete to provide
strength after the concrete has set. The main reason they are not permitted to be used as
reinforcement is because they are subject to long-term corrosion problems. Their main
purpose is to facilitate construction by reducing labor through the elimination falsework and
formwork.
The steel stay-in-place forms can be used for steel or concrete girders. Figure 2.1
shows a typical cross-section and assembly of a steel SIP form for a steel stringer or girder
bridge deck system. They are not composite for bridge applications, however, they can be
used as composite reinforcement in building applications. In concrete floor design a steel
SIP deck form alone can provide enough reinforcement in positive moment regions,
however, in building applications these forms are generally isolated from exposure to
moisture.
17
A few of the larger manufacturers of steel SIP deck forms for building and bridge
applications are United Steel Deck, Inc. located in South Plainfield, NJ and Epic Metals
Corporation from Rankin, PA.
Depth
Pitch(varies)
Coverage(varies)
Top Flange(varies)
Stringeror Grider
Stringeror Grider
Deck
Supportangle leg
1"min. bearing
1/2" flange widthSpan 1"
A
A
Section A-A
Figure 2.1 A typical cross-section and assembly of a steel SIP deck form used in conventional bridge deck construction
2.6 Concrete Bridge Decks and Concrete Deck Panels Reinforced with FRP SIP Forms
The use of the SIP FRP deck form to provide a form for wet concrete as well as the
tensile reinforcement for a composite concrete deck is a relatively new concept. Harik et al
(1999) have pioneered the experimental investigation of a similar GFRP deck form (18” wide
each, stiffened with two corrugations) that was studied in this thesis. The Harik et al study
investigated three-point bending of the SIP FRP form/concrete composite deck panels, with
FRP bars acting as compression reinforcement, designed according to ASSHTO LFRD
18
(1998) and LFD (1996) specifications. Their goal was to load three composite concrete deck
panels, of varying depths and spans, with an AASHTO standard HS25 truck wheel load and
observe the service load deflections, the performance to cyclic loading, the strength capacity
and the mode of failure.
The failure mechanism observed from these three-point bending tests was the same as
the one observed in the testing conducted for this thesis. The positive moment test revealed a
flexure-shear cracking and a de-bonding of the FRP deck form from the concrete at failure.
The service load deflections of all specimens were well below the allowable. In addition, the
panel stiffness did not diminish after being subjected to cyclic loading. Finally, the factors of
safety were greater than three for all specimens. Therefore, as designed according to
AASHTO criteria, the selected strength and serviceability requirements were satisfied.
These results breed confidence that the SIP FRP deck form can easily satisfy the
serviceability deflection criteria.
The Ohio Department of Transportation was confident enough with the FRP concrete
reinforcement system used in the Harik et al study that they implemented it into an actual
bridge and Reising et al (2001) documented its performance. The ODOT Salem Avenue
Bridge Project retrofitted a 5-span bridge with four different FRP composite panel systems
and one of the four systems was the SIP FRP deck form/bar concrete reinforcement. From a
global performance perspective, Reising et al found that this system did not reduce the
overall stiffness of the bridge. Additionally, such values for girder distribution and impact
factor were within the recommendations according to AASHTO specifications.
19
3 PREPERATION OF TEST SPECIMENS
3.1 Introduction
This IBRC project required the construction of concrete bridge deck panels and
beams reinforced with FRP for a series of serviceability and strength testing. I.F.
Corporation, on east side of Madison, Wisconsin, was contracted by the UW to prepare,
assemble, pour and deliver pre-cast deck test specimens. This chapter documents the test
specimens, the casting facility, assembly methodology, concrete specifications, transportation
and an overall evaluation of the deck test specimen preparation at I.F. Corporation. Also
documented, is the construction of an additional concrete deck panel at the Structures and
Materials Testing Laboratory (SMTL) on the University of Wisconsin–Madison Engineering
Campus.
3.2 Test Specimens
It was decided that 11 deck specimens were required to determine the stated
objectives in the Introduction. Table 3.1 details the specimens that were constructed for
laboratory testing. The laboratory testing did not require the inclusion of a wearing surface;
therefore, UW Professor Michael G. Oliva designed all specimens with 1.5” of concrete
cover above the bi-directional FRP grid panel. The chapters to follow will address in more
detail the purpose of the test specimens. Appendix B illustrates the location of strain gauges
applied during the assembly process, for some of the test specimens. Panels A, B and C were
fitted with 6 strain gauges each, located on the bottom of the FRP deck forms, for one-half of
the symmetric concrete deck specimen (2 per panel) along the intended center-line of the
20
span. Three of the negative moment deck specimens, Beams E, were fitted with one strain
gauge each, centered on the web of the “I” beam of the bi-directional grid, centered over the
intended support (See Appendix B for gauge locations).
Table 3.1 The list of all 8” thick specimens constructed
Specimen Category Name Span Purpose Panel A 11’-6” Panel B 9’-10” Panel C 8’-0”
Concrete Deck Panels (all 9’-0” wide)
Panel C2 8’-0”
To determine distribution widths, mode of failure and strength
safety factors
Beam D1 Beam D2
Positive Moment Beams
Beam D3
All 8’-8” Test the positive moment properties of the corrugated deck form
Beam E1 Beam E3 Beam E4
Two Span Negative Moment Beams
Beam E5
Each span 8’-0”
Test the negative moment properties of the bi-directional grid over a support
Two Span Fatigue Beam Beam E2 Each span 8’-0”
Test the fatigue strength of the FRP system and the mechanical bond
Figure 3.1 I.F. Corporation casting bed
21
3.3 Precast Facility
At I.F. Corporation, a 9’ x 80’ concrete pre-cast forming bed was utilized to prepare
the deck specimens. The casting bed seen in Figure 3.1 had the ability to form concrete
panels up to 8” thick. The pre-casting bed geometry would not allow a single pour for all
specimens; therefore two separate pours were completed. For curing, the casting bed was
enclosed and heated at 60 degrees for up to 3 days after casting.
Figure 3.2 FRP deck forms
3.4 Assembly Methodology
After assembling the FRP reinforcement for the first concrete deck panel, Panel A, it
was quickly learned how to expedite the process. A sequential order was developed through
experience. First, the pour bed was swept clear of all debris to ensure the FRP deck forms
could lay flat. Next, as seen in Figure 3.2, duct taped was used to enclose the FRP deck form
corrugation end openings to prevent concrete from migrating inside the voids during concrete
22
placement. At this point, lap joints between the adjacent forms were connected together with
screws at approximately 5’ intervals. This was a safeguard procedure to ensure stability of
the forms for the concrete pour only and was not intended to provide any structural
continuity. Preparing ahead for tie down support of the FRP bi-directional grids, wire ties
were anchored to the top face of the FRP deck form corrugations with screws at
approximately 4’ intervals. Next, the FRP grids were placed, along with transverse splices,
into position and propped up with 2.5” plastic chairs. As seen in Figure 3.3, the plastic chairs
were placed upon the top surface of the corrugations.
Figure 3.3 Grid with 2.5” chair and tie
From experience it was found that the FRP grids were sufficiently stiff to walk upon
with chair spacing intervals of 3’ to 4’. Once the grids were into position, they were tied
down with the previously anchored wire ties. Then the transverse splice grids seen in Figure
3.4, used to supply temperature and shrinkage reinforcement continuity between the main bi-
23
directional grids pieces, were tied tightly below the main grids. The assembly process
provided a stable system to withstand walking on and for concrete placement.
Three beams, D1 thru D3, of the same length were assembled to test the positive
moment properties. Beams D1 and D2 had top FRP bi-directional grids, while D3 was cast
without this FRP grid to compare results. These beams did not contain any grid splices;
therefore they were constructed quite simply with one FRP grid and two, side-by-side deck
forms. The order of assembly of the FRP reinforcement components mimic that which was
derived for Panels A, B and C.
Figure 3.4 Longitudinal FRP bi-directional grid splice, Panel C
The negative moment properties of the FRP reinforced system were tested with four
E-type deck beams (see Appendix B for geometry and layout). Three of the four negative
24
moment beams had gauges (E1, E3 and E5) to measure top layer grid strains of the main
reinforcement over the intended support. The remaining negative moment beam, E4, had no
bi-directional grid reinforcement over the intended support, to provide results for negative
moment capacity of an un-reinforced section. The remaining beam, E2, had the FRP grid
reinforcement over the support, but was intended for fatigue testing.
The Beam E specimens were spliced longitudinally with FRP 8” splices at each
longitudinal, bar made-to-fit over the top flange. Theses splices did not provide any strength
capacity, but rather they were to provide proper alignment between grids. A special aspect of
the positive reinforcement scheme for Beams E was an 8” discontinuity or gap of the FRP
deck forms over the support, since the forms simply span between girders and aren’t
continuous over the girders. Photo 5 shows a Beam E right after the forms were removed.
Figure 3.5 A Beam E specimen after forms were removed
25
3.5 Concrete Placement and Specifications
The concrete mixed used was a Wisconsin DOT Class D mix (see Appendix B for
mix design) typically used by the WisDOT for bridge decks (WisDOT, 1996). The concrete
design strength was dropped, however, from the typical 4500 psi to 4000 psi. A 7” slump
was specified to increase the concrete workability and to minimize honeycomb pockets. It
was noticed that one drawback of the FRP grids was that the main reinforcement “I” beams
could easily cause void pockets without proper consolidation. These pockets typically
formed between the outside “I” beam and the concrete forms that run parallel to the main
reinforcement. An additional measure taken to help prevent this problem was to hold the
maximum aggregate size to ¾”.
Figure 3.6 Placement of concrete with chute and consolidation with vibrator
The ready-mix concrete was supplied by M and M Concrete and was delivered from
the truck by chute and spread with a shovel (Figure 3.6). During the placement, a concrete
pencil vibrator facilitated the concrete movement between tight places and ensured good
26
consolidation. After striking off extra concrete, a float finish was applied to the surface of
each specimen (The smoother the finish, the easier it is to see cracks!). Three cylinders were
cast for each test specimen with an additional three for each pour, of which there were two
separate pours.
On January 12th, 2001, Panels A, B and C were poured. About two weeks later, on
January 22nd, the outstanding beams were poured. The original plan was to have 48 hours at
60 degrees for curing. After testing a 2-day cylinder from the first pour, however, the
compressive strength was around 1640 psi, which was below the 2000-psi strength required
before the beams were transported to SMTL. Therefore, for the second and last pour, the
cure time was increased to 60 degrees for three days. After three days, a cylinder tested at
3380 psi, a significant increase for an additional day of curing. Even though the design
mixes are the same, they were from different batches of concrete, which may also have had a
slight impact on the additional strength.
Figure 3.7 Crane lifting a test specimen of the flat bed truck at SMTL
27
3.6 Transportation
I.F. Corporation was also contracted to load and deliver the test specimens, via a flat
bed semi tractor-trailer, to SMTL. As seen in Figures 3.7 and 3.8, a crane provided by
Reynolds Movers, whom also moved the test specimens into SMTL, unloaded the specimens.
Figure 3.8 Reynolds Movers move Panel B into SMTL
3.7 Evaluation
The casting process at I.F. Corporation and the delivery of the test specimens was a
success. Beam E4 (negative moment beam without top FRP grid reinforcement) cracked in
half at the centerline, transversely, on the trip to SMTL. This test specimen had absolutely
no reinforcement within the center 8” of the beam. Since the beam lifted onto the truck at
I.F. Corporation without difficulty, the beam most likely cracked due to the vibration caused
by the ride on the back of the truck.
28
3.8 Construction of an Additional Concrete Deck Panel
After the first phase of testing, the original FRP reinforcement system displayed two
behavioral concerns, partial composite action and a susceptibility to longitudinal concrete
cracking above the shiplap joint. Therefore it was decided, based on these initial test
observations, that an additional concrete panel should be constructed and tested with a
modified FRP reinforcement system to address these concerns. The two modifications to the
FRP reinforcement system were: additional aggregate coverage to all horizontal surfaces of
the FRP deck panel to improve the composite action and the addition of a 4” wide FRP
Fibergrate Molded Grating, 2” square pattern, to reinforce the concrete directly above the
FRP deck panel shiplap joint. Figure 3.9 depicts a typical cross-sectional view of the
modified FRP reinforced section of concrete Panel C2 (transverse to the corrugations). This
additional concrete panel was constructed at SMTL.
Shiplap Joint
FRP Bi-directional Grid
FRP Bi-DirectionSlice Grid 2" Square
FRP Grating
Epoxied Gravel
FRP Deck Form Figure 3.9 Cross-section view of the primary direction of modified FRP reinforced Panel C2
This methodology of assembly of the reinforcement parts was very similar to the
process outlined earlier in this chapter. The only difference was the panel was formed with
plywood buttressed with 2” x 6” whalers. Figure 3.10 shows Panel C2 during the
29
construction process; it also shows the additional coverage of epoxy-aggregate on the FRP
deck forms and the inclusion of 4” wide x 1” deep FRP Fibergrate Molded Grating placed
above the corrugation valleys with the shiplap joints.
Panel C2 was equipped with internal strain gauges applied directly to the FRP
components before the concrete was poured. These were included to allow the measurement
of strains through the depth of the reinforced cross section. The exact same concrete mix
used for the original test specimens was ordered from the same company, M&M Concrete.
Figure 3.10 During the construction of Panel C2 at UW- SMTL
3.9 Concrete Quality Control
For concrete quality control, three 28-day concrete test cylinders were poured for
each specimen cast. In addition, from each batch of concrete (i.e., each concrete truck) an
additional three concrete cylinders were made. Totally, there were three concrete batches:
30
one for Panels A, B and C; another for all Beam D, Beam E and Fatigue Beam specimens,
and one more for Panel C2. Table 3.2 summarizes the specifics of each of the concrete
pours. The mix design for the WisDOT Class D concrete is shown in Appendix B.
Table 3.2 Summary of concrete pours
Specimens Date Poured
Mix design #208, M&M Concrete, WDOT Class D
Cylinders Made
Panels A, B, C 01/12/01 4000 psi/7” Slump/ ¾” Agg. 12 Beam D, E1 and
SG17, Concrete 8 -3.91 Top of SG7 SG2, Concrete 9 -3.455 Top of 5 SG3, FRP 10 -4.04 Top flange of I bar SG4, FRP 11 -3.97 Bottom flange of I SG5, FRP 12 -3.92 Top of corrugation SG6, FRP 13 -4.02 Bottom of section SG7, FRP 14 -3.97 Bottom of SG17
Note: all SG calibration constants are x1000
4.6.4 Description of Specimen Loading
The loading steps and the loading sequence described in Section 4.2.4 were used for
the modified deck panel test. The conditioning step and the elastic step went according to the
plan spelled out for the other concrete deck panels. The ultimate capacity test, however, was
cut short of achieving an ultimate failure. The reaction frame of the 200 kip actuator started
to lift off the structural floor when the load approached an ultimate value of around 120 kips.
As a result, the ultimate or inelastic test was cut short of the goal of monitoring a post-
ultimate behavior of the modified concrete deck panel.
70
5 TEST RESULTS
5.1 Introduction
The description that follows was organized by tests that were conducted for a similar
purpose. In part, the main topics for each section will describe the following: how the test
was conducted, visual information obtained during the test, data measured during the test,
when the test was halted and judgment of failure. The broad aspects of how each test was
conducted will be recalled, but the details of this testing can be found in Chapter 4. This
chapter will focus on the inelastic test results for each specimen, except for the fatigue test.
Not all data the from all the tests conducted will be presented, however, when test data from
either the basic conditioning or elastic steps can help address the objectives, then this
information was included. Finally, it should be noted that all test data does not include
selfweight effects.
5.2 Concrete Deck Panel Test
In general, the concrete deck panel test was designed to answer three key unknowns
of the proposed FRP reinforcement system: how does the lack of bottom distribution steel
and presence of shiplap joints between the FRP deck forms affect the distribution of the load,
the mode of failure and the ultimate capacity. Each concrete deck panel was subjected to
sequential testing steps of conditioning, elastic and inelastic testing. The methodologies of
conditioning, elastic and inelastic testing steps were described in Section 4.2. For the
concrete deck panel, the results obtained from the elastic tests were recorded, however, at
this point this information was not directly essential to address the research objectives. In
71
situations where these results can help meet the objectives, they will be incorporated into the
test results. On the other hand, inelastic testing, which includes testing through the elastic
region, naturally produced informative data requisite for developing load distribution
characteristics, determining failure mechanism and establishing strength.
Inelastic testing involved loading each specimen over a centered surface area
approximately equivalent to the contact area of a double tire wheel, see Figure 4.1. The 16th
edition of AASHTO LFD design specification defines the contact area as a function of the
wheel load. Since the wheel loading in the inelastic testing phase was not static and the
contact area remained constant, the area shown in Figure 4.1 was judged sufficient. The
monotonic load was applied under manual load control at a relatively slow rate. While
loading, deflections and material strains were recorded. In addition to the digital data, visual
and audio observations were also recorded. For example, all preloading concrete cracks were
traced in marker and noted. Near the ultimate capacity point, the actuator was switch from
load to stroke control and loading, indirectly through displacement control, continued. Each
concrete deck panel was loaded past ultimate capacity, except for Panel C2, until a
satisfactory level of displacement was reached. The test was terminated when a significant
deflection beyond the ultimate capacity was safely achieved.
5.2.1 Panel A, 11’-6” Span
Being the first test conducted of the entire project, audio and visual details were
somewhat neglected. As can be seen in Figure 5.1, all concrete surface cracks that existed
before loads were applied were traced and noted with a pen marker. There were a handful of
these preload size surface concrete cracks for Panel A that had little consequence on, or
72
contribution to, the structural performance. As the test progressed, all new concrete cracks
and continued growth of existing concrete cracks were traced and the load at a specific crack
length was noted. In the end, the marked concrete cracks helped to illustrate the values of
tension forces and how they progressed as load was applied.
Figure 5.1 Preload concrete surface cracks at one supported end of Panel A
LabView software was used to record the data from all of the instruments used. This
software allowed a live, real-time, view of a load versus mid-span deflection relationship, via
a graph, while the load was applied to the specimen. The deflection was monitored at a point
12” offset from the center of the load patch, along the span centerline of the concrete deck
panel (4” Potentiometer). Hard data recorded during the inelastic test pertained to vertical
deflections and material strains at locations, on Panel A, depicted in Chapter 4.
Not surprisingly, the initial run experienced a few errors. The data acquisition system
experienced an operator mistake and data was not recorded for any instrument input data for
the first 36 kips of load. The loss of this instrument data inhibited any data analysis within
the initial elastic load versus deflection region. Although small, this undoubtedly introduced
an unknown error into the data, because the initial offsets at no load are not known and the
73
data cannot be properly zeroed. Another error was neglecting to document the audio
information, such as peculiar cracking noises, as the test was conducted. Although it can be
easily overlooked, some audio information can provide explanations to structural behavior.
Adjustments were made and both problems were resolved before the next concrete deck
panel test was performed.
0
10
20
30
40
50
60
70
80
90
100
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Stroke, inch
Lo
ad, k
ips
P yield = 80k
K effective= 88.1 k/inch
Figure 5.2 Load versus actuator stroke for the inelastic test of Panel A
The profile of the load versus actuator stroke response of the inelastic test is
illustrated in Figure 5.2. As will be discussed, since the potentiometer original offsets were
unknown and the uplift of the reference frame near the ultimate load, the stroke data was
used rather than the potentiometer deflection data. There seemed to be a region of softening
just before failure that maintained a high load resistance through increasing deflection.
Although service deflections are not apart of the scope of this thesis, it should be noted that
the interpolated stroke deflection at 20 kip (approximately HS20-44 with impact) was 0.198”,
which equals L/700. This stroke data includes some small deflection contributed by
74
compression of the elastomeric pad under the actuator load. The actual deck deflections,
which did not include these extraneous deformations, would show stiffer behavior. The
stroke versus deflection relationship can be defined as follows: an initial elastic region of
high stiffness, followed by a slight decrease in stiffness, then ultimate capacity, and finally
followed up by a post-ultimate softening/degrading region.
At a stroke defection of 1.70”, the ultimate capacity of 96 kips was achieved. Soon
after reaching the ultimate capacity, each of the four corners of the deck significantly lifted
(easily observable) off their respective support blocks, which is classical behavior for a
simply supported slab. The lifting of the corners can introduce a disagreement between
stroke data and relative displacement from instruments mounted on the reference frame,
because the reference frame was directly attached to each corner. So, the difference between
the stroke (at the center) and the 4” potentiometer (12” offset from center) readings was
graphed in Figure 5.3 to determine if the lifting of the four corners introduced error into the
potentiometer and the LVDTs. There was a constant difference, relatively, until 1.61” or just
before ultimate. This constant difference indicates that lifting of the corners did not adversely
affect the displacement measurements until just before ultimate capacity was reached. After
1.61” the difference between the stroke and the potentiometer sharply increases at a constant
rate because of the onset of punching shear. Therefore, only the stroke values were valid.
The post-ultimate test was terminated at a stroke deflection of 4.10” and still carrying 63
kips. Although difficult to visually define, a yield point of 80 kip and an initial stiffness of
88 kips per inch were characterized by a standard technique proposed by Preistley (1992),
which will be used for all specimens to follow for relative comparisons. The Preistley
method is explained in detail in Appendix D.
75
0
20
40
60
80
100
120
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Difference between stroke and potentiometer, in.
Load
, kip
Figure 5.3 Difference between the stroke and potentiometer readings, Panel A
There was no concrete cracking along the unsupported sides of the panel, leading to a
conclusion that there was little load distributed all the way out to the unsupported edges. The
ends of the panel that rested on the reaction blocks did see some vertical concrete cracking,
especially in the middle region of each end. Some of these vertical cracks along the
supported ends were connected to the oval surface cracks seen in Figure 5.4.
76
SUPPORT
SUPPORT
Figure 5.4 Oval concrete cracking pattern, post-inelastic test of Panel A
Figure 5.5, shows the top surface deflections across the centerline section at various
load levels. The locations of the potentiometer and the LVDTs used to measured these
In general, the test results of the accelerated beam, Beam E2, test lead to the
conclusion that there is no concern for fatigue failure of either the material or the mechanical
bond between the concrete and the FRP deck forms. The testing results documented in
Section 5.6 and Figures 5.59 and 5.60 reveal no significant stiffness loss, for service wheel
load, over 2 million cycles. The maximum load of the cycle was arbitrarily set, 20.8 kips and
the 16 kip range was chosen to represent AASHTO HS-20 truck without impact. The
maximum and amplitude loads were large enough to be considered to induce a significant
non-reversal stress for elastic load levels. Assuming a linear elastic relationship in the early
elastic range with a cracked moment of inertia of 135.2 in^4, yt = 4.50” and negative moment
of 360 kip inches (including selfweight) at the worst loading of each cycle (20.8 kips in one
span and 16 kip in the other) the stress in the FRP grid can be determined by,
fgrid = M20.8*yt / Icr, (6.8)
The maximum stress experienced in the grid per cycle is about 12 ksi, which is about 14.4%
of the tensile strength for the bi-directional grid. One conclusion is that the fatigue life of the
FRP grid must be higher than 12 ksi for 2 million cycles. Looking back, this stress level
162
could have been increased. However, for a bridge application with loading in the early
elastic range the grid should not experience any problems with fatigue strength.
The other fatigue issue dealt with the bond fatigue strength of the mechanical shear
transfer between the FRP deck form and the concrete, which is germane for composite
action. In the absence of possible impacts from environmental issues, there is no evidence
from the fatigue testing to suggest that there will be any fatigue strength issues with this
bond.
In addition to performing adequately for fatigue, Table 6.6 shows the fatigue beam
provided equal strength capacities from the inelastic test, as the other 3 negative moment
beams, even after the fatigue beam was subjected to 2 million cycles.
6.6 Performance of Test Specimens Based on Stiffness
This section goes beyond the scope of this thesis, however, it was conducted to
provide some comparison between the measured and calculated stiffnesses of the specimens.
Table 6.7 summarizes the performance of the FRP reinforced test specimens by comparing
the span-to-deflection ratios (i.e., L/800) at a service wheel load, of 16 kips, experienced
during each respective inelastic test (selfweight not included). Typically in bridge design,
the maximum allowable deflection includes impact loading, however, these specimens do not
directly represent an actual bridge member. For example, the distribution widths of the
panels and the end conditions of all specimens were not representative of an actual
slab/girder bridge deck. Therefore, comparing the values in Table 6.7 to AASHTO criteria
was not applicable. Instead, these values are for relative comparison of performance and the
HS44-20 service wheel load (without impact) was selected somewhat arbitrarily.
163
Table 6.7 Span-to-deflection ratios for the tested FRP specimens
Specimen Span, Center-to-Center of Bearing
Center Span Deflection at 16
kips, in
L/ (Span/Deflection)
Panel A 11’-6” 0.087* L / 1586 Panel B 9’-10” 0.019 L / 6210 Panel C 8’-0” 0.027 L / 3556 Panel C2 8’-0” 0.054 L / 1778 Beam D1 8’-8” 0.068 L / 1529 Beam D2 8’-8” 0.112 L / 929 Beam D3 8’-8” 0.096 L / 1083 Beam E1 8’-0” 0.038 L / 2526 Beam E3 8’-0” 0.042 L / 2286 Beam E5 8’-0” not available --- Beam E2 (Fatigue Beam)
8’-0” 0.039 L / 2462
* Data was interpolated because data was not recorded until 36 kips into inelastic test Table 6.8 compares calculated stiffness based on gross transformed and cracked
moments of inertia to the initial (based on secant line to 16 kips) and effective (based on the
Priestley Method) stiffnesses measured during the inelastic tests, respectively. The
calculated stiffnesses for the panel specimens were based on the AASHTO positive moment
distribution widths, equation 6.3. The calculated stiffness for the beams was based on the
full width, 3’, for each specimen. The measured initial stiffness and calculated stiffness,
based on gross transformed, were compared at a 16 kip load. The measured effective
stiffness and calculated cracked stiffness were compared at the same load used to determine
the measured effective stiffness using the Priestley Method, 75% of the peak load. The
values for the longitudinal gross transformed and the cracked moments of inertia based on
ACI 440H are located in Appendix D.
Basic deflection equation, PL3/48EI, was used to determine the deflections for the
positive moment simply supported panels and beams with concentrated load at center of
164
span, P, and the moment of inertia, I (gross transformed or cracked). To find the deflection
of the 2-span continuous negative moment beams, a computer analysis was conducted for 50
kip concentrated load in each span, which was close to the 75% of the ultimate load seen in
each of Beams E. Again the gross transformed and cracked moments of inertia were
accounted for in the computer analysis. The modulus of elasticity, Ec, for each specimen was
determined from the compression strengths found from the concrete cylinder tests conduced
on the day of the test and using ACI 318 equation, 57,000*(f’c)1/2. It should be noted that
the calculated cracked stiffness, Kcracked, assumes the whole beam was of a cracked section,
which is not true. To improve accuracy of a true effective stiffness of the beam for research
purposes, instead of using cracked moment of inertia, the cracked and uncracked regions of
the beam should be delineated and the moment area integrated with the respective stiffness to
find the true deflection. This, however, was not done for this comparison.
Table 6.8 Comparison of measured and calculated stiffness for the tested FRP specimens
Note: * Kinitial is based on secant line to 16 kips Γ Stiffness based on LVDT data, other values based on actuator stroke
165
Comparing the values in Table 6.8 of the simply supported panels and beams (A, B,
C, C2, D1, D2, and D3), it can be generally stated (expect for Panels A and B) the measured
stiffnesses were below the calculated values. With many variables involved it is difficult to
definitively determine a cause for these discrepancies, although there does not seem to be a
systematic pattern of error. On the other hand, the negative moment beams performed very
well when measured stiffnesses are compared to the calculated stiffnesses. The performance
of the negative moment beams gives some confidence that a continuous FRP concrete deck
on a slab/girder system should easily meet deflection requirements.
166
7 ANALYTICAL MODEL TO DETERMINE NEGATIVE MOMENT DISTRIBUTION WIDTH
7.1 Introduction
One of the stated objectives of the research program was determining a negative
distribution width over the girders for the load that will cause concrete tension cracking on
the top surface of the concrete deck. Cracking of the concrete over the girders can allow
water ingression into the deck. Water in the section, exposed to freeze-thaw cycles, will
affect the long-term durability of the bridge deck by forcing the concrete to spall and crack.
In design, to prevent concrete cracking above the girder, a distribution width of the deck
must be assumed as resisting the wheel load and the moment in that width must remain
below the cracking value. Determining the negative moment distribution width employs the
same procedure as used for the positive moment distribution, except it is defined over the
girder instead of in the middle of the span.
Another difference between the positive and negative moment distribution widths is
that the negative moment distribution width was determined by analytical modeling with
shell finite elements modeling the 2-D deck. The experimental results were used, however,
as a benchmark of verification for the analytical model. The instantaneous elastic defections
measured for Panel C2 were used as a basis, because it’s reinforcement system (added
aggregate coverage on the surface of the FRP deck form) is closest to representing the
reinforcement system to be used in the prototype bridge.
The purpose of this Chapter was to layout the methodology taken to derive the
negative moment distribution width. First part of this chapter will provide an overall
167
description of the process. Another part of the methodology will define the section
properties of the orthrotropic plate model, as well as, to explain the analytical computer
program and boundary conditions used. Then the modeling methodology for the prototype
bridge will be described and finally, the distribution width at the girder will be determined
using the same analytical procedures used for the positive moment distribution width in
Chapter 6. In addition to the negative moment distribution width, the positive moment
distribution width was re-evaluated using the same model. The positive distribution width
was investigated with the wheel loads in adjacent spans that should cause the cracking
moment of the concrete on the bottom surface of the deck form.
7.2 Overall Description of the Methodology
First, the cross-sectional properties of the orthogonal FRP reinforcement system were
calculated in each direction for the positive moment region. These values were used to
define shell element properties in the structural analysis computer program SAP 2000 for the
concrete deck panel test of Panel C2. The experimental Panel C2 test was duplicated
analytically in SAP 2000. The instantaneous deflection values and profiles in the elastic
region of the analytical model were compared to the experimental values. Once the quality
of the analytical model was verified experimentally (by comparing deflections) for the
positive moment regions, the same approach was now used for a model of the prototype
bridge deck with wheel load centered in adjacent spans to determine the negative moments
developed over the girder. These moments were compared to the cracking moment derived
in Section 6.4.1 from the negative moment beam tests. Then the wheel load values in the
model were adjusted until the loads created a negative moment over the girder that would
168
just cause the concrete to crack in tension. At this load, the moment profile of the section
over the girder was used to determine the distribution width. This is the distribution width
that can be used in design to examine to see if certain wheel loading on this FRP
reinforcement system will cause cracking.
The objective was to determine the distribution width of the negative moment region
just as the concrete cracking occurs. With that stated, and reasoning that a continuous
section of the model will be stiffer than a simply supported concrete deck panel, an initial
assumption was made that the entire section should remain with gross section of inertia (and
full composite action) up to the first concrete crack over the girder. All section properties
were defined, however, so that each individual shell element’s cross sectional properties
could be easily changed in the model to it’s cracked section if the positive cracking moment
in that shell element was reached before the negative cracking moment developed. The
model was run to prove the initial assumption on stiffness of a continuous section and the
model was validated by the experimental results.
7.3 Defining the Section Properties
The first step was to define positive moment transformed gross and cracked section
properties for the cross section of the orthrotropic FRP reinforced deck in both directions, see
Appendix D for the illustrated parameters used for a typical 18” cross section. The
“longitudinal” direction is orientated perpendicular to the girder supports. The transverse
direction was the direction of the temperature and shrinkage reinforcement of the top bi-
directional grid, and was parallel to the supports. The cracked section properties of the FRP
169
reinforced system were calculated according to Equation 7.1 which was provided by ACI
Guide 440H.
Icr = (bd3/3)k3 + nfAfd2 * (1-k)2 (7.1)
Where,
k = (2ρfnf + (ρfnf)2)1/2 - ρfnf (7.2)
The variables are defined as,
b = width of retangular cross section, in; d = effective reinforcement depth, in; nf = modular ratio of FRP to concrete; Af = area of FRP reinforcement, in2; and ρf = FRP reinforcement ratio. The gross section property was transformed to include the FRP deck form, which is slightly
less than a true gross sectional value of a true rectangular section (see Appendix D). The
difference between the gross and the gross-transformed section is due to the voids in the
concrete created by the corrugations of the FRP deck form and the fact that the modular ratio
between the two materials is slightly above unity.
To accurately model instantaneous deflection in the experimental test of Panel C2,
four possible moment of inertias for each shell element were defined:
1. Longitudinal transformed gross moment of inertia; 2. Longitudinal cracked moment of inertia; 3. Transverse transformed gross moment of inertia; and 4. Transverse cracked moment of inertia.
Incidentally, using Equation 7.1 to calculate the transverse cracked moment of inertia,
(number 4 in the above list) the area FRP material in the two corrugations was conservatively
ignored and the area, Af, of bottom material of the FRP deck form alone was used, however,
this moment of inertia was never used in this model analysis. After establishing the moments
170
of inertia, the respective cracking moments were calculated according to Equations 6.6 and
6.7. All of the aforementioned moment of inertia and cracking moment values are shown in
Appendix D.
7.4 Finite Element Modeling
To model the orthogonal properties for Panel C2, and to model the continuous bridge
deck, the structural analysis computer program SAP2000 Nonlinear Version 7.44 was used.
The model of Panel C2, as well as the bridge deck model, was defined with shell sections.
For the bridge model, it was assumed that the elastic deflections were small enough not to
The concrete deck panel was modeled with 6” x 6” square shell elements. Each
element was initially set to the gross-transformed section properties in each direction of the
orthogonal material. The supports were model as simply supported with pin supports along
bearing line and rollers on the other bearing line. A load was applied as a pressure to an area
representing the wheel load as illustrated in Figure 4.1. Three load cases were created:
deadload only, liveload (wheel load) only, and a combination thereof. The combination load
case was analyzed to examine for any uplift forces at the supports. Otherwise all
experimental data comparisons from testing excluded the deadload, since the instruments
were applied after the panels were in the testing setup (preventing the ability to measure
deadload deflections and strains).
The first objective was to verify the analytic model accuracy by comparison with the
experimental liveload deflections in the elastic range. Therefore, the first step for
171
verification was to simulate liveload application and to compare elastic analytical deflections
at a common point with the experimental deflection measurements. So, at 8 kip intervals the
analytical deflection at 24” from the centroid of the panel along the center of span was
compare to the experimental results at the same location and load. At the onset, all cross-
sectional properties had a gross-transformed moment of inertia. In addition, the reactions at
each load increment were monitored for any indication of tension forces. If tension was
encountered in the analytical model, the support at a node was released to prevent the
introduction of hold down forces due to an improper support condition model.
Figure 7.1 illustrates the comparison of liveload deflections, at 8 kip intervals
between the model for Panel C2 and the respective inelastic experimental test results. The
graph clearly displays a close relationship between the measured elastic deflections from the
test of Panel C2 and the analytical model. The analytical model did not require any reduction
of section properties from the initial gross-transformed moment of inertia. The comparison
was conducted up to 48 kips because the loading for the investigation of the negative
moment distribution was limited to the elastic region just before cracking over the girder
occurs, which was below 48 kips. According to the model results, the longitudinal positive
cracking moment was not reached, directly under the wheel load, at a liveload of 16 kips
(M16kip = 4.80 and Mcr = 5.50 kip-inch per inch width) Incidentally, examining the
combination load case revealed that each of the four corner nodes did experience net uplift
forces at the 24 kip load interval and thus were released, but no other nodes showed uplift
forces up to 48 kip.
172
0
10
20
30
40
50
60
0 0.01 0.02 0.03 0.04 0.05Deflection, in
Load
, kip
s
Panel C2 Model Uncracked Figure 7.1 Comparison of experimental (inelastic test) and analytical deflections at 24” from
center of load along center span
With the encouraging sign that experimental, data from inelastic test, and the
analytical measurements were very close at one location, the next step was to compare
deflection profiles. Comparing deflection profiles give an indication of lateral stiffness,
which was critical for building a valid analytical model to investigate load distribution.
Therefore, the deflection profiles were compared using the measurements from the elastic
test of Panel C2 along two lines, at center span and A24 or 24” toward the support from the
center span (see Figure 4.8). The elastic test deflections were used because more ±1.0”
LVDTs were set up along the center span, thus providing a better deflection profile. The
profile measurements from the elastic test and the model, at a 16 kip load, are compared in
Figures 7.2 and 7.3. The results in Figure 7.1 were based on a span of 8’-0”. With the
magnified deflection scale of Figure 7.2 it was clear that deflections of the test panel were
not exactly matched by the analytic model. A second analysis was undertaken with a revised
173
effective span length. Analytical results are sown in Figures 7.2 and 7.3 with an effective
span of 8’-0” center-to-center of bearing, and with 7’-4” clear span between front of
bearings.
-0.020
-0.015
-0.010
-0.005
0.000-54 -36 -18 0 18 36 54
Distance from the Center, in
Def
lect
ion
from
16
kip,
in.
C2 Elastic Model 7 ft -4 in span Model, 8 ft span
Figure 7.2 Deflection profile for experimental (elastic test) and analytical measurements at
center span
-0.012
-0.010
-0.008
-0.006
-0.004
-0.002
0.000
-54 -36 -18 0 18 36 54
Distance from the Center, in.
Def
lect
ion
fro
m 1
6 ki
p, i
n.
C2 Elastic Model 7 ft -4 in span Model, 8' span
Figure 7.3 Deflection profile for experimental (elastic test) and analytical measurements along A24
The deflection profiles are quite similar in shape and it was encouraging that the
analytical gross-transformed section properties defined for the model were validated by the
experimental results. It is interesting to note that the experimental deflection values were
174
approximately halfway between the analytical values obtained using center-to-center and
clear spans, which makes sense of a true span being from front quarter bearing to front
quarter bearing, or 7’-8”. At this point the conclusion was made that the positive moment
region can be accurately modeled. In Section 6.4.1, the negative moment region section
behavior was validated by good agreement between the loads that caused concrete cracking
over the support (Beams E1 and E3) and the analytical calculations using the gross-
transformed section properties. With analytical validation for modeling of both the negative
and positive moment regions the next step was to model the prototype bridge deck.
7.4.2 Modeling the Prototype Bridge Deck
The bridge deck model had the same geometric measurements that will be seen in the
Wisconsin DOT prototype bridge described in Appendix A. The center-to-center girder
spacing is 8’-8” and the bearings on top of each girder are 16” wide. Because of the presence
of shear stirrups from the top of the girders cast into the concrete deck, there will be some
flexural continuity between the deck and the girder. In this model that flexural restraint was
not directly modeled and pinned connections were placed along each edge of the bearings.
Thus the model span was taken as the clear span and the flexural restrain of the girders was
indirectly accounted for by having more than one node, in the span direction, pin connected
to the girder/support. Although a deck length should theoretically be considered equal to the
bridge length, a conservative length of about twice the center-center span length, 16’, was
chosen for the model. Wheel loads were placed in the center of two adjacent spans forcing
an assumed maximum moment over the top of the center girder. Spacing of wheel loads in
accordance with AASHTO could have be considered here, but it was conservatively assumed
175
that at some point in the life of the concrete deck the undesirable condition of two loads
being placed centered in adjacent spans would occur. Two additional spans were added to
the model, one to each end of the loaded spans. A 28-day concrete strength of 4000 psi,
specified by the WisDOT for bridge use, was assumed.
Learning from the process of modeling Panel C2 elastic deflections, the section
properties remained equal to the gross-transformed moment of inertia. The moment of
inertia properties would not change if the negative moment section cracks first. With the
positive and negative cracking moment essentially the same, Mcr (Mcr = 4.79 kip-in / in
width, f’c = 4000 psi), the negative moment over the girder should consistently be larger than
the positive moment in the span as the load increases in the elastic region and, thus, crack
first.
Essentially the process of modeling involved meshing the concrete deck into 6”
square shell elements, placing the pinned connections and loading the center of the adjacent
spans until the maximum moment over the girder reached the cracking moment. At this load,
the magnitude was noted as the load to cause longitudinal flexural cracking (in direction of
traffic and girder centerline) of the concrete over the girder. Then the moment profile along
the cross-section, parallel to the girder centerline, was subjected to Equation 7.3.
= (2 * m(x)dx) / mDistribution Width0
96"
maximum (7.3)
Equation 7.3 was another approach to determine the distribution width. It is
fundamentally the same as Equation 6.2, and described in Section 6.2.1, but the method to
determine the total moment (the numerator) is different. Since there was a 16” bearing span
176
acting as the top of girder, it is more accurate to use the analytical model moment profile to
determine the total moment than using a 2-D beam formula (Equation 6.2 determined the
total moment with a simple 2-d beam formula, P*L/4). Equation 7.3 determined the negative
moment distribution width by summing the area under the moment distribution profile, the
total moment, and divided it by the cracking moment, Mcr, or the maximum moment per inch
of width (as will be explained later, this same method was used to analytically determine the
positive moment distribution width in the middle of the span). Figure 7.4 illustrates the
negative moment distribution profile from the analytical model for the section above the
front edge of the bearing on the center girder where the maximum negative moment
developed for liveload of 35 kip wheel load (selfweight neglected) centered in adjacent
spans.
-7
-5
-3
-1
1
3
5
-96 -80 -64 -48 -32 -16 0 16 32 48 64 80 96
Distance from center of load, inches
Mom
ent,
kip-
in/ i
nch
wid
th
Negative Moment Positive Moment
+Mcr or Mmax =4.79 k/in per inch width
-Mcr or Mmax =4.79 k/in per inch width
Figure 7.4 Moment distribution profiles over a 16’ width
The area under (actually above) negative moment distribution curve and the total
moment was 347 inch-kips at a Mcr or Mmax of 4.79 inch-kip per inch of width, which
177
occurred at a liveload of 35 kips centered in adjacent spans. Using Equation 7.3, the negative
moment distribution width at the onset of longitudinal flexural concrete cracking over the
girder was 73”. Figure 7.5 is an illustration of a three dimensional moment profile in the
bridge model.
Figure 7.5 Negative moment profile of prototype deck for negative moment
The positive moment distribution was also determined from the same model. Wheel
loads in the adjacent spans were adjusted until the longitudinal positive cracking moment
(cracking in the directions of traffic) was reached in the center of the adjacent spans. The
magnitude of the load was noted and moment distribution profile along the center span (in
the direction of traffic) was also graphed in Figure 7.4. The area under positive moment
distribution curve (the total moment) was 271 inch-kips at a Mcr or Mmax of 4.79 inch-kip per
inch of width, which occurred at a liveload of 28 kips centered in adjacent spans. Again,
178
using equation 7.3, the moment distribution width at the onset of the concrete cracking at the
bottom of the concrete deck at mid-span was 57.5”. This brings about one final assumption
for the negative moment distribution width, that the cracking in the positive moment region,
which occurs at 28 kip wheel loads centered in adjacent spans, does not affect the negative
distribution width at 35 kip. Another words, with wheel load centered in adjacent spans,
longitudinal concrete cracking in the positive moment region (bottom of the slab) will occur
before longitudinal concrete cracking in the negative moment (top of slab and over the
girder)
The positive and negative moment distribution widths derived from the model of the
prototype bridge cannot be directly compared to AASHTO distribution width, equation 6.3,
used for the experimental concrete deck panels, because each represents a different support
condition. According to AASHTO-LFD Bridge Design Specification, 16th edition, the
experimental concrete deck panels represent support conditions of a slab supported on
abutment or piers, AASHTO 3.24.3.2, Case B. The prototype bridge deck model represents a
slab supported on beams or stringers, AASTHO 3.24.3.1, Case A. The design liveload
moment for Case A, given by Equation 7.4, and the total moments for the negative and
positive distribution curves derived by the analytical model, however, can be relatively
compared on the basis of moment per foot width. It should be noted that Case A is
presumably for strength design and the positive and negative moment distribution widths for
a slab-stringer support condition are serviceability issues, so strictly speaking, this was not
comparing “apples to apples”.
LLM = P(S+2)/32 ft-lb (7.4)
179
For an HS20-44 truck wheel load without impact on the prototype bridge, P is 16,000
pounds and the span, S, is 8.67 feet. For the given wheel load and span the design liveload
moment for the prototype bridge deck, LLM, would be 64 kip-inches per foot width (5.34
kip-inch per inch width). This design moment can be reduced by 20% if the slab is
continuous over three or more stringers, which is the case for the prototype bridge.
Therefore, the design moment is 51.2 kip-inches per foot width (4.26 kip-in per inch width).
The LLM to cause longitudinal cracking for the negative and positive moment distribution
curves were derived by determining the area ±6” each side of Mmax under the curves (see
Figure 7.4). The LLM per foot was calculated as 56.5 and 54 kip-inches per foot width (4.71
and 4.50 kip-inch per inch width) for negative and positive moment to cause longitudinal
cracking. Conservatively, the LLMs to cause longitudinal cracking over the girder and in the
slab are greater than the LLM for the service design moment prescribed by AASHTO.
It should be noted that selfweight of the deck was not included. For the sake of the
real prototype bridge, the selfweight of the deck will contribute 11% and 6% of the liveload
moments produced by the wheel loads used for the negative and positive moment distribution
widths, respectively.
180
8 CONCLUSIONS
8.1 Introduction
In the introduction of this thesis solid reasons were presented for the use of FRP
products to reinforce concrete bridge decks to increase long-term durability. FRP as a
material cannot replace the conventional steel reinforcement, however, without proper
verification of its performance. Condensing the objectives, the purpose of this experimental
study of the FRP reinforced bridge deck was to evaluate its performance under bridge
liveloads. In the end, conclusions from this investigation will impact decisions on whether or
not the proposed FRP reinforcement scheme is adequate for bridge use.
In a bit more detail, two general objectives helped drive the experimental study of the
FRP reinforced concrete bridge decks. First, the proposed FRP reinforcement system uses an
unconventional shape and a stay-in-place deck form for the top and bottom reinforcement
layers, respectively. The unconventional bottom reinforcement scheme introduces a shiplap
joint, a structural discontinuity, and neglects the use of distribution reinforcement in the
bottom layer. These progressive issues needed to be investigated for their impact on the
reinforced deck’s ability to carry bridge wheel loads. Second, the absence of a specific
design guide for this type of FRP reinforcement in concrete bridge decks leaves an engineer
to use an “ad-hoc” collection of design guides. The engineer is without the typical design
tools and equations used for steel reinforced bridge decks. Therefore, determining the mode
of failure and a distribution width are a couple of empirical objectives that will help simplify
the design process and build confidence that the studied AASHTO equations used in this
181
thesis to design a steel reinforced concrete bridge deck can be used for this specific FRP
reinforcement system.
8.2 Observations and Conclusions
Examining the results, many conclusions and observations could be and have been
drawn. In this section, only the major observations and conclusions, however, will be
presented. In addition, the qualitative and quantitative performance observations and
conclusions below will, in general, relate back to the stated objectives presented in the
introduction and are presented in bulleted form. Finally, the following observations and
conclusions will be separated into categories related to serviceability, strength, detailing and
construction issues.
8.2.1 Serviceability
1. From the test results from this thesis and according to the conclusions of Sanders
(2001), the added aggregate coverage alone increased the resistance to longitudinal
concrete cracking above the shiplap joint without distribution reinforcement, and
thus significantly increased the load required, from 5 kip (Panel C) to 47 kip (Panel
C2), to initiate the opening of the shiplap joint directly under the wheel load.
2. An analytical model for the prototype bridge, verified by experimental results in the
elastic region of Panel C2, indicated that the negative and positive moment caused
by wheel loads centered in adjacent spans will be distributed laterally 73” over the
girder and 57.5” at midspan, respectively. The analytical model also indicates the
loads centered in adjacent spans to cause a negative moment crack over the girder is
about 31 kip, including selfweight of the deck.
182
3. The amount of aggregate coverage impacts the performance of the FRP deck form
as the positive moment reinforcement. Positive moment beam tests indicated that
the partial coverage of aggregate on the top surface of the FRP deck form causes
inefficient partial composite behavior in bending.
4. Results from experimental testing of the 9’ wide concrete deck panels
conservatively indicate that the AASHTO distribution equation (AASHTO 3.24.3.2,
Case B) for a steel reinforced concrete slab supported on an abutment or piers is
applicable for use in design of the proposed FRP reinforcement system, with or
without total coverage of aggregate on the top surface of the FRP deck form. In
addition, from analytical model, the AASHTO service design equation for an HS20-
44 truck, Equation 7.4, for a steel reinforced concrete deck supported on three or
more beams or stringers is applicable for use in the prototype bridge to ensure the
prevention of longitudinal flexural cracking in the positive and negative moment
regions.
5. After 2 million cycles of load, the accelerated fatigue beam test indicated no
concern for fatigue failure of the FRP material, the mechanical bond between the
aggregate on the surface of the FRP deck form and concrete and, finally, there
seems to be some sign of stiffness loss, but not significant, for loads above the
service wheel loading. Therefore, similarly to most steel reinforced concrete bridge
decks, fatigue is not a crucial design issue.
183
8.2.2 Strength
1. Because no panels or beams failed by FRP rupture, there was a significant post-
ultimate ductility. This is a behavioral trait that can be expected from a traditional
steel reinforced concrete bridge deck.
2. The factors of safety for the ultimate strength (due to punching shear) to service
load for an HS20-44 load, without impact, are between 5 and 6 for deck forms with
partial aggregate coverage and 8.1+ for total aggregate coverage. Strength results
of Deck Panels C and C2, normalized for concrete strength, indicate that complete
aggregate coverage increased the ultimate strength and stiffness by 49% each. The
above conclusions are based on the assumption that the 4” wide FRP Fibergrate
Molded Grating as FRP reinforcement over the shiplap joints doesn’t contribute to
longitudinal strength or stiffness. Therefore, full coverage of aggregate on the FRP
deck form surfaces is recommended.
3. Concrete deck panel tests indicated that the mode of failure is punching shear and
the ACI 318 guide, Equation 6.4, can be used with a modification. If the
conventional definition of the effective depth parameter, d, is used, then a reduction
coefficient, Ag, should be used. This value of this coefficient depends upon the
amount of aggregate coverage on the top surface of the FRP deck form. With
partial aggregate coverage (original product from CDS, Inc.), Ag = 0.60, and with
aggregate to all horizontal surfaces, Ag is unity. Therefore, equation 6.4 should be
used as,
Vc = Ag * 4 * (f’c)1/2 * bo * d (8.1)
184
8.2.3 Detailing and Construction
1. From the negative moment beam tests, beam failures were based on shear followed
by a de-bonding of the form at the edge of the plaster bearing on the middle
support. Thus, the capacity of the negative moment reinforcement, the top bi-
directional grid, was never reached. It is surmised that if the FRP deck form had
been supported on the plaster bearing, the capacity of the beam would have
increased. Therefore, for the sake of providing capacity against pullout,
embedment length for positive moment reinforcement and also to properly transfer
shear to the support, the FRP deck form end must be supported upon a bearing, such
as on top of a girder or on grout.
2. From practical experience of assembling the FRP reinforcement to construct the
specimens, the system could be easily and quickly build. It was also noted that
concrete did consolidate well, using a pencil vibrator, except for areas where the top
flange bi-directional grid was positioned close to the form. For practical use,
however, this consolidation problem in a slab/girder bridge system will not be a
common experience.
Neglecting possible environmental degradation issues, this FRP reinforcement system
is applicable for use in the WisDOT prototype bridge. This conclusion is only based upon
the results from the experimental and analytical tools used in this thesis.
185
9 SUMMARY AND RECOMMENDATIONS
9.1 Design Criteria
Most Load Factor Design AASHTO criteria governing the design of a reinforced
concrete deck for a slab-on-stringer system are given in AASHTO section 3.24. A typical
continuous steel reinforced concrete slab over girders has the main steel reinforcement
oriented perpendicular to traffic. So, how does the results of the research relate to the
traditional design methodology used by AASHTO? Table 9.1 was constructed to summarize
and compare the design methodology for a traditional steel reinforced concrete deck and one
based on the results of the research on the prototype bridge conducted for this thesis. This
table is intended as a design aid for an engineer to design a deck reinforced with the proposed
FRP system. Other criteria considered in a concrete deck design, but not in the table, are
deck reinforcement details, such as minimum cover and spacing of bars, which are checked
against governing rules and must be applied with engineering judgment for the FRP
reinforcement.
By and large, designing a steel reinforced deck is a relatively easy and quick process,
however, this design methodology is based upon a long and well-documented history of
performance. When the material to reinforce the concrete is changed from steel to FRP, with
unconventional issues pointed out earlier, this design process was not certain and hence there
was a need to conduct the research for this thesis.
186
Table 9.1 Design Methodologies for Steel and FRP Reinforcement for Concrete Bridge Deck Based on 8’-8” Span
Criteria With Steel Reinforcement Based on AASHTO LFD 3.24.3.1, Case A
With FRP Reinforcement Based
on Thesis Results
FRP Reinforcement Research Conclusion
Distribution Width, in
32/(S+2) = 36” +M = 57.5” -M = 73”
AASHTO Equation is applicable for FRP, see Chapter 8.2.1, #2
Positive Moment
Design bottom layer steel reinforcement perpendicular to traffic based on MLL = P*(S+2)/32, plus impact and deadload,
Same moment as steel but check capacity using ACI 440H, where reinforcing area provided by form
FRP deck form provides over designed capacity for typical effective spans
Negative Moment
Design top layer steel reinforcement perpendicular to traffic based on P*(S+2)/32
Size “I” bar area of bi-directional grid and/or spacing based on ACI 440H
Finite Element Model shows deck remains in gross section under normal wheel loads
Distribution Reinforcement
Compute steel in bottom slab based on a percentage of main reinforcement = 67% ≥ 220/S1/2
None required Deck performed acceptably without distribution reinforcing. See 8.2.1, #1
Temperature and Shrinkage Steel
Perpendicular to main reinforcement AASHTO 8.20 minimum of 0.125 in2 per foot in each direction
Same as steel, select size and spacing of bars in bi-directional grid panel based on modular ratio
Shear Considered safe for shear
Strength capacity controlled by punching shear
Considered safe for shear
Fatigue Considered safe Performed well under 2 million cycles of load
No concern for FRP fatigue resistance, see 8.2.1, #5
9.2 Strength of FRP-SIP Reinforced Deck
From the experimental testing, the FRP system (designed according to ACI 440H for
flexural capacity) has provided satisfactory punching shear capacity under a HS44-20 truck
load, which is the generally accepted mode of failure for normal spans. Even though the
187
flexural capacity was not reached, the punching shear capacities from the concrete panel test
provided strength factors of safety of 5 to 8 times service load. The punching shear behavior
under closely spaced wheel loads, where critical shear perimeters overlap has not been
investigated but the design using ACI punching shear criteria should still apply.
9.3 Design for Serviceability
The FRP reinforcement system can provide adequate fatigue resistance when exposed
to 2 million cycles, where each load cycle induced 14% of the ultimate stress in the bi-
directional grid. Fatigue resistance under higher stress levels and exposed to reversal
stresses, however, are not known. Experimental testing also determined that the lateral load
distribution widths in the positive moment region for various span lengths were higher than
what current AASHTO criteria for steel reinforced deck allows. Use of current AASHTO
methods is acceptable. Deflection criteria were not formally addressed and still remain
uncertain for the FRP system. However, Harik et al. (1999) have determined that the SIP
FRP deck form studied in this thesis can satisfy the deflection criteria. Ohio Department of
Transportation (ODOT) was confident enough to implement this FRP deck form into an
actual bridge.
Since deflection criteria was not completely examined, it would be instructive to
compare the deflection and shear results of the positive moment beams to an actual
performance specification used for the Ohio Department of Transportation’s Salem Avenue
Bridge, where ODOT implemented the SIP FRP deck form. Table 9.2 lists the performance
of the three simply supported positive moment beams to deflection performance criteria set
188
by ODOT for the FRP deck based on reinforced concrete deck analysis. It can be seen that
the simply supported positive moment beams performed much better than the criteria.
Table 9.2 Performance of Positive Moment Beams to ODOT criteria for 3’ wide, 8” thick simply supported beams
Positive Moment Beams ODOT Criteria Span, c/c beams (ft) 8’-8” 8’-0” 9’-0” Clear Span (ft) 8’-0” 7’-6” 8’-6” Service Load (kip) 16 12 Simple Span ∆ (in) 0.068 0.112 0.96 0.125 0.18 From the analytical testing, it was shown that for an 8” concrete deck with 8’-8”
center-to-center span with section properties of the FRP reinforced concrete, that the wheel
loads to initiate longitudinal flexure cracking (in the direction of traffic) in the positive and
negative moment regions are much larger than the 16 kip service wheel load. In fact, the live
load moment derived from Equation 7.4 is less than the moment required to initiate cracking
in either region. Other span lengths and deck thickness, however, would need to be looked at
on a case-by-case basis, since no history of performance has been established. This research
is applicable for normal girder spacing between 8’ and 12’ and any spacing outside the
ranges of the concrete deck panel tests may deviate from the results described here.
9.4 Further Investigation
Even though this FRP system has performed well relative to serviceability and
strength criteria, there are some issues that are still unknown and should be investigated
further. The minimum edge bearing of the FRP deck form to ensure that the form does not
pull out and delaminate before the punching or flexural shear capacity is reached is not
certain. The test specimens performed adequately with little or no bearing, but improved
behavior could be achieved. Another outstanding issue deals with reversal of and higher
189
levels of stress effects upon the fatigue capacity. Long-term environmental effects upon the
FRP deck form could pose a durability problem and should be considered for further study.
Other tests (Helmueller, 2001) have shown that moisture collection between the FRP does
not lead to serious de-bonding problems when subjected to freeze thaw action and the FRP is
not susceptible to corrosion. It may be beneficial to further investigate the possibility of
longitudinal concrete cracking above the shiplap joints between the FRP deck forms. As a
final point, other issues related to detailing and construction, such as tying down the FRP
deck forms prior to concrete placement, need to be considered, but these construction
solutions may be best developed through ingenuity and creativity in the field.
9.5 Improving the FRP Reinforcing
To conclude the summary of this thesis, improvements to the FRP reinforcement
system are suggested based on test results, experienced gained from the construction of
specimens, and some insight of future construction issues. Most of the improvements
reflected in the test results are related to the SIP FRP deck form. First, the deck form should
incorporate complete coverage of aggregate to all horizontal surfaces to ensure efficient bond
and thus increased performance in strength and stiffness. In addition, it has been shown that
through the lack of structural continuity, the shiplap joints were a point of weakness for
strength. This could be improved in two ways: 1.) the shiplap joint could be redesigned to
provide a shear connection in both directions between adjacent deck forms, and 2.) the deck
forms could fabricated at larger width increments, e.g., 3’, and thus reduce the number of
joints (however this depends on fabrication, shipping and handling limitations). It also may
be economically beneficial to investigate the redesign of the FRP deck form using less
190
material. Adding aggregate and reducing the cross-sectional area of the FRP material could
structurally economize the flexural capacity of the system and reduce cost of the product.
9.6 Constructability
This new system should substantially reduce deck construction time because of the
use of lightweight prefabrication components that can be placed quickly without heavy
equipment. From a constructability perspective, there might be an easier way to provide
temperature and shrinkage reinforcement continuity between adjacent bi-directional grid
panels. In the research, a smaller 12” wide splice grid was tied below the main
reinforcement, but from experience this could require too much time to construct properly. A
natural solution for a composite girder/slab superstructure system that requires negative
moment reinforcement over a pier or bent is to run a percentage of this reinforcement to the
abutments. In the process, the continuation of reinforcement will also provide capacity
against possible concrete cracking above the shiplap joints and act, in some respect, as
distribution reinforcement.
191
REFERENCES
Alampalli, S., O’Connor, J., and Yannotti, A. P. (2000), “Fiber Reinforced Polymer Composites for Superstructure of a Short-Span Rural Bridge.” Albany, N.Y.: Transportation Research and Development Bureau, New York State Department of Transportation, (Special report; 134) Alkhrdaji, T., Nanni, A., and Mayo, R. (2000). “Upgrading Missouri Transportation Infrastructure: Solid RC Decks Strengthened with FRP Systems.” 79th Annual TRB Meeting (CD-ROM), National Academy of Science, Washington, D.C. Bank, L.C., Xi, Z. and Munley, E. (1992). “Tests of Full-sized Pultruded FRP Grating Reinforced Concrete Bridge Decks.” In Materials: Performance and Prevention of Deficiencies and Failures, (ed. T.D. White), ASCE Materials Engineering Congress, Atlanta, GA., August 10-12, pp. 618-631. Bank, L.C., and Xi, Z. (1993). “Pultruded FRP Grating Reinforced Concrete Slabs.” In Fiber-Reinforced-Plastic for Concrete Structure – International Symposium, (eds. A. Nanni and C.W. Dolan), SP-138 American Concrete Institute (ACI) pp. 561-583. Bank, L.C., and Xi, Z. (1995). “Punching Shear Behavior of Pultruded FRP Grating Reinforced Concrete Slabs.” Proc. 2nd Int. Symp. On Non-Metallic (FRP) Reinforced for Concrete Structures, (ed. L. Taerwe), E & FN Spon, London pp. 360-367. Bettigole, Neal H. and Robinson, R. (1997). Bridge Decks: Design, Construction, Rehabilitation, Replacement. American Society of Civil Engineers (ASCE) Press, New York, 1997.
Bradberry, T.E. (2001). “FRP-Bar-Reinforced Concrete Bridge Decks.” 80th Annual TRB Meeting (CD-ROM), National Academy of Science, Washington, D.C. Chajes, M., Shenton, H., and Finch, W. (2001). “Performance of a GFRP Deck on Steel Girder Bridge [Paper 01-3290].” 80th Annual TRB Meeting (CD-ROM), National Academy of Science, Washington, D.C. Dietsche, J. S. (2002). “Development of Material Specification for FRP Structural Elements for the Reinforcing of a Concrete Bridge Deck.” MSCE Thesis, Department of Civil and Environmental Engineering, University of Wisconsin-Madison. Harik, I., Alagusundaramoorthy, P., Siddiqui, R., Lopez-Anito, R., Morton, S., Dutta, P. and Shahrooz, B. (1999). “Testing of Concrete/FRP Composite Deck Panels.” Proceedings of the Fifth ASCE Materials and Engineering Congress (ed. L. C. Bank), Cincinnati, May 1999, ASCE, Reston, VA. pp. 351-358.
192
Helmueller, E.J. (2001). “The Effect of Freeze-Thaw and Aggregate Coating on the Bond Between FRP Stay-in-Place Deck Forms and Concrete.” Independent Study, Department of Civil and Environmental Engineering, University of Wisconsin-Madison. Kumar, S.V., and GangaRao, H.V.S. (1998). “Fatigue Response of Concrete Decks Reinforced with FRP Rebars.” Journal of Structural Engineering, 124(1), ASCE, 11-16. Lopez-Anido, R., Howdyshell, and P., Stevenson, L. D. (1998). “Durability of Moduar FRP Composite Bridge Decks Under Cyclic Loading.” 1998 Durability of Fibre Reinforced Polymer (FRP) Composites for Construction (CDCC). pp. 611-622. University of Sherbrooke, Quebec, Canada. Matthys, S., and Taerwe, L. (2000). “Concrete Slabs Reinforced with FRP Grids. I: One-Way Bending.” Journal of Composites for Construction, ASCE, 4(3), 145-153. Matthys, S., and Taerwe, L. (2000). “Concrete Slabs Reinforced with FRP Grids. II: Punching Resistance.” Journal of Composites for Construction, ASCE, 4(3), 154-161. Priestley, M.J.N., “The US-PRESSS Program Progress Report.” 3rd Meeting of the U.S.-Japan Joint Technical Coordinating Committee on Precast Seismic Structural Systems (JTCC_PRESSS), San Diego, CA, November 18-20, 1992 Reising, R.M.W., Shahrooz, B.M., Hunt, V.J., Lenett, M.S., Christopher, S., Neumann, A.R., Helmicki, A.J., Miller, R.A., Kondury, S., Morton, S. (2001). “Performance of a Five-Span Steel Bridge with Fiber Reinforced Polymer Composite Deck Panels [Paper 01-0337].” 80th Annual TRB Meeting. Sanders, B.M. (2001). “Structural and Physical Tests of CDS-FRP Reinforced Deck Form and Concrete.” Independent Study, Department of Civil and Environmental Engineering, University of Wisconsin-Madison. Taly, N. Design of Modern Highway Bridges. McGraw-Hill, New York, 1998. AASHTO (1996). Standard Specifications for Highway Bridges. 16th Ed., Washington, D.C. AASHTO (1998). LRFD Highway Bridge Design Specifications. 2nd Ed., Washington, D.C. AASHTO (2000). Manual for Condition Evaluation of Bridges. 2nd Ed., Washington, D.C ACI. (1999). Building Code Requirements for Structural Concrete and Commentary. ACI 318-99 and ACI 318R-99. American Concrete Institute, Farmington Hills, Michigan. ACI. (2001). Guide for the Design and Construction of Concrete Reinforced with FRP Bars. ACI 440.1R-01. American Concrete Institute, Farmington Hills, Michigan.
193
ASTM. Standard Method of Testing Compression Strength of Cylindrical Concrete Specimens (C39-93). Philadelphia: American Society for Testing and Materials, 1993. Wisconsin Department of Transportation (1996). Standard Specifications for Highway and Structure Construction. 1996 edition.
194
APPENDIX A
195
DESCRIPTION OF PROTOTYPE BRIDGE
The bridge targeted for this FRP reinforcement system is new construction on US Highway
151 (A.D.T. of 18,600) over State Highway 26 near the city of Waupun, Wisconsin. This is
a twin, 2-lane bridge structure with two continuous 32.7 m (107’) spans. The girders are 1.37
m (54”) deep pre-stressed concrete “I” beams, spaced at 2.65 m (8’- 8”). The deck is 200
mm (8”) thick (excludes wearing surface), 12.75 m (43’) wide with a total deck area of 855.6
m2 (9,200 ft2). Construction is expected to begin in the spring of 2003.
The proposed innovative construction technologies should improve upon current
construction practices. One proposed technology is simplifying the concrete reinforcement
assembling process and, therefore speeding up construction. The pultruded SIP FRP deck
form and the pultruded grid panels will be pre-fabricated and pre-sized units and delivered to
the job site. The deck forms will be placed in rapid fashion without the need for time-
intensive falsework or formwork, which is usually needed in conventional concrete deck
pours. With the companion bridge decked with a traditional steel reinforcement system, a
“field laboratory” will be created to compare in-service performance behavior. The two twin
bridges will be monitored to provide long-term data comparisons in identical harsh
environments.
196
DESIGNED FRP REINFORCEMENT FOR A TYPICAL 18” CROSS SECTION
= d
FRP Bi-directional Grid
1 1/2"
1/4"
8"
2"
1 1/2"
6.93
"
1'-6"
3"
FRP Deck Form
Positive moment section
Area of FRP deck form = 11.89 in2 Area of 2” “I” bars = 2.77 in2
1/4"
1 1/2"
8"
1 1/2"
2"
1'-6"
d = 5.69"
3"FRP Bi-directional Grid
FRP Deck Form
Negative moment section
197
APPENDIX B
198
ILLUSTRATION OF STRAIN GAUGE LOCATIONS FOR THE TEST SPECIMENS
ILLUSTRATION OF STRAIN GAUGE LOCATIONS FOR THE TEST SPECIMENS
(Specimen Assembly)
Panels E
Panel E1
Panel E1
Panel E1
E1
Location of Panel Letter onConcrete Deck Surface
E2
E3
2" Ø PVC
Round Bars = 4" o.c."I" Beams = 4" o.c.
For All Panel E's, Bi-Directional Grid Bar Spacing:
For All Panel E's, Reinforcement Clearence = 1-1/2" @ CL
Gauge Locations
202
WISCONSIN DOT CLASS D CONCRETE DESINGN MIX
(provided by M&M Concrete) DATE: 01/11/2001 ATTN: Dave PROJECT: U.W. Madison CONTRRACTOR: IF Corp. PART OF STRUCTURE: Panels CLASS OF CONCRETE: WDOT Class D MIX DESIGN NUMBER: 208
MATERIAL TYPE SUPPLIER PORTLAND CEMENT ASTM-C150 HOLNAM INC. GRANCEM CEMENT ASTM-C989 HOLNAM INC. FLYASH ASTM-C618 C Mineral Solutions COURSE AGGREGATE #1 ASTM-C33 Prairie Avenue COURSE AGGREGATE #2 ASTM-C33 Prairie Avenue FINE AGGREGATE ASTM-C33 Mann Brothers Air Mix 250 ASTM-C260 Euclid Chemical Company WR ASTM-C494 Euclid Chemical Company MR ASTM-C494 Euclid Chemical Company Eucon 37 ASTM-C494 Euclid Chemical Company Accelgaurd NCA ASTM-C494 Euclid Chemical Company CaCl ASTM-C494 Euclid Chemical Company Fiberstrand F ASTM-C1116 Euclid Chemical Company
Prepared By: Jamie M. Kuchnickl
203
APPENDIX C
204
EXPERIMENTAL INSTRUMENT LOCATION FOR ULTIMATE TEST
PANEL B, TOP CONCRETE SURFACE
CL
Panel B Span
9' Width
(9'-10", nts)
4" Potentiometer
SG7, SG4 denotes strain gauge
LVDT14, LVDT15,... denotes LVDT
SG17
12"9"
LVDT15
12"
LVDT16
12"
LVDT17
Control
SG6
15"
6"
7"
SG4
LC
6"
205
EXPERIMENTAL INSTRUMENT LOCATION FOR ULTIMATE TEST
PANEL B, BOTTOM FRP SURFACE
SG2 9' Width
SG7
SG was applied to FRP,but not used in test
CL
13"
Control
SG5
21"
9"
SG3
LVDT14, LVDT15,... denotes LVDT
LVDT13 to monitor horiz.joint opening (transverse)
1-1/2"
SG7, SG4 denotes strain gauge
(9'-10", nts) Panel B Span
6"
Shiplap joint between the FRP deck forms
206
EXPERIMENTAL INSTRUMENT DATA RECORD FOR THE ULTIMATE TEST
DESCRIPTION OF PRIESTLEY STANDARD METHOD TO CALCULATE A
STIFFNESS AND A YIELD STRENGTH FOR A SPECIMEN TEST RESULT
0
10
20
30
40
50
60
70
80
90
100
0 0.5 1 1.5 2 2.5 3 3.5
Displacement, in
Lo
ad, k
ips
Pmax
.75*Pmax
Disp. @ .75 Pmax Yield Disp.
Pyield
Priestley (1992) algorithm is used as a standard method to determine initial stiffness
and yield strengths for relative comparisons between specimen testing results. In general, the
sequential method to determine the relative stiffness and yield strength is defined below and
exhibited in the above illustration.
1. Determine the maximum load, Pmax, and identify the displacement result for 0.75*Pmax, “Disp. @ .75 Pmax”
2. Determine the relative stiffness, keff, by (0.75*Pmax)/(Disp. @ .75 Pmax), kip/inch
3. Determine the yield strength, Pyield, by identifying the test load which corresponds to the displacement value calculated by Pmax/keff
__________________________________________________________________________ Priestley, M.J.N., “The US-PRESSS Program Progress Report”, 3rd Meeting of the U.S.-Japan Joint Technical Coordinating Committee on Precast Seismic Structural Systems (JTCC_PRESSS), San Diego, CA, November 18-20, 1992
214
DESIGNED FRP REINFORCEMENT FOR A TYPICAL 18” CROSS SECTION
= d
FRP Bi-directional Grid
1 1/2"
1/4"
8"
2"
1 1/2"
6.93
"
1'-6"
3"
FRP Deck Form
Positive moment section
Area of FRP deck form = 11.89 in2 Area of 2” “I” bars = 2.77 in2