Concrete Bridge Decks Constructed with Fiber-Reinforced Polymer Stay-in-Place Forms and Grid Reinforcing

Transportation Research Record 1814 ■ 219Paper No. 02-3205

A description is given of laboratory testing, initial test results, and con-tinuing research to develop design procedures and plans for constructionof a two-span highway overpass on US-151 in the state of Wisconsin usingonly a fiber-reinforced polymer (FRP) reinforcement system in the con-crete bridge deck. The use of FRP reinforcing is being pursued to increasethe durability of bridge decks and to reduce unit cost and time. Theunique aspect of the new bridge is that it will use FRP stay-in-place form-work, placed over precast concrete I girders, to serve as formwork and asthe bottom transverse reinforcement of the bridge deck. In addition, aheavy-duty prefabricated FRP grid will be used for the top layer of con-crete reinforcing. Full-scale prototype laboratory testing is being used todevelop design recommendations including effective distribution widths.From the spring to fall of 2003, a construction productivity study com-paring the FRP deck construction sequence with that of identical adjacentsteel reinforced deck will be conducted. Both bridges will be field loadtested before opening to traffic. The FHWA Innovative Bridge Researchand Construction Program is supporting the research and construction.

The primary goal of a good concrete bridge deck design is to safelycarry design loads over the intended design life. Many state depart-ments of transportation find the typical design life of concrete bridgedecks has been reduced or has been exceeded (1, 2). A primary causeof deterioration of concrete bridge decks is the corrosion of steel re-inforcement (3). Corrosion of reinforcement steel adversely affectsthe long-term durability of the concrete through spalling and crack-ing, eventually exposing the deck to excessive deflections and thepossibility of member failure (4). Costs to replace prematurely fail-ing bridge decks (through materials, labor, relocation of traffic, etc.)and the inconvenience to the traveling public are facilitating a searchfor a more durable material to replace steel reinforcement.

The purpose of this research is to evaluate pultruded fiber-reinforcedpolymer (FRP) products as relatively inert materials to satisfy strengthand serviceability concrete deck design requirements. In addition,eliminating steel reinforcement and increasing long-term durability,the pultruded stay-in-place (SIP) FRP deck form and bidirectionalFRP grid panel concrete reinforcement system used in this projectshould provide additional benefits. From an economic point of view,this FRP deck reinforcement system is expected to reduce construc-tion labor cost through ease of assembly and the elimination of tradi-tional cast-in-place concrete deck forming and falsework. Finally, this

reduction of labor cost is related to another key project: decrease intime of bridge deck construction.

From a structural-performance point of view, the main objectiveswere to determine the failure mechanisms and the overall perfor-mance of the FRP deck system through experimentally testing FRPreinforced slabs and beams. It is extremely difficult to accuratelysimulate, to scale in the laboratory, the continuity and end conditionsof a continuous concrete bridge deck. Therefore, it was decided toisolate and measure deck response characteristics, such as momentcapacities and stiffnesses, with simple test setups. In the future, resultsof the tests will be compared with analytical calculations to determinethe effectiveness of design methodologies proposed.

The results from testing conducted in this research project willbe implemented in the construction of one twin of a new twin-bridge structure owned by the Wisconsin Department of Trans-portation (WisDOT). The opposing twin of the new bridge structure,on US-151 near Fond du Lac, Wisconsin, will use conventional steelreinforcement for the bridge deck, thus providing a hands-on oppor-tunity to compare performance and durability of the two materialsin service. This will also provide a unique chance to compare con-structibility issues. To elaborate, the scope of this research project willinclude a constructibility study to quantify the value of reducing theconstruction labor cost and time to construct this FRP reinforcementsystem versus a traditional steel-reinforced concrete deck.

DESCRIPTION OF PROTOTYPE BRIDGE ANDPROJECT PROGRAM

WisDOT proposed, through the FHWA’s Innovative Bridge Re-search and Construction program, to investigate an FRP reinforce-ment system to increase long-term durability of the state’s bridgedecks. This project consists of three main phases: fully investigatethe strength and serviceability behavior through laboratory testing,implement the tested design into a prototype bridge deck, and finally,use a long-term monitoring program. WisDOT is teamed with theCivil and Environmental Engineering Department at the Universityof Wisconsin-Madison as well as with Alfred Benesch and Company,consulting bridge engineers.

The bridge targeted for this FRP reinforcement system is new con-struction on US-151 (average daily traffic of 18,600) over StateHighway 26 near the city of Fond du Lac, Wisconsin. This is a twin,two-lane bridge structure with two continuous 32.7-m (107-ft) spans.The girders are 1.37-m (54-in.) deep prestressed concrete I beams,spaced at 2.65 m (8 ft 8 in.). The deck is 200 mm (8 in.) thick (exclud-ing wearing surface) and 12.75 m (43 ft) wide, with a total deck area

Concrete Bridge Decks Constructed withFiber-Reinforced Polymer Stay-in-PlaceForms and Grid Reinforcing

David A. Dieter, Joshua S. Dietsche, Lawrence C. Bank, Michael G. Oliva, and Jeffrey S. Russell

D. A. Dieter and J. S. Dietsche, Room 2225; L. C. Bank, Room 2206; M. G.Oliva, Room 2212; and J. S. Russell, Room 2304; Engineering Hall, Departmentof Civil and Environmental Engineering, University of Wisconsin–Madison, 1415Engineering Drive, Madison, WI 53706.

of 855.6 m2 (9,200 ft2). Construction is expected to begin in thespring of 2003.

The proposed innovative construction technologies will improveon current construction practices. One proposed technology is sim-plifying the concrete reinforcement assembling process and thereforespeeding up construction. The pultruded SIP FRP deck form and thepultruded grid panels will be prefabricated and presized units anddelivered to the job site. The deck forms will be placed in rapidfashion 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 re-inforcement system, a “field laboratory” will be created to comparein-service performance behavior. The two twin bridges will bemonitored to provide long-term data comparisons in identical harshenvironments.

DECK MATERIALS

The FRP reinforced concrete system is composed of a top and a bot-tom layer (see Figure 1 for picture of an FRP system mock-up). Thebottom tensile reinforcement is composed of a pultruded FRP SIPdeck form (Figure 1) spanning between, but not continuous over, thegirders. The FRP deck form is analogous to the main positive steelreinforcement typically placed perpendicular to the girder system.Each deck form is 457.2 mm (18 in.) wide and overlaps with adja-cent deck forms via a shiplap joint. The SIP forms are stiffened bytwo 76.2-mm (3-in.) square hollow corrugations centered 228.6 mm(9 in.) apart. To ensure composite action through horizontal sheartransfer between the deck form and the concrete, 6.35-mm (1⁄4-in.)aggregate is bonded to most of the horizontal surface area of thedeck with epoxy. Studies such as Harik et al. (5) have experimentallyinvestigated an FRP reinforcement system composed of a top layerof FRP bar reinforcement combined with a similar SIP deck form.This deck form is a nonmetallic pultruded FRP product produced byComposite Deck Solutions Inc., located in Dayton, Ohio.

The bottom reinforcement layer contains no longitudinal re-inforcement per se, which would be analogous to the distribution

220 Paper No. 02-3205 Transportation Research Record 1814

reinforcement typically seen in a traditional steel reinforced deck.However, the actual bridge deck design will contain FRP reinforce-ment bars for negative moment continuity reinforcement over themiddle pier. Beyond the negative moment envelope, a percentage ofthe FRP reinforcement bars will continue to the abutments. HughesBrothers, Inc., located in Seward, Nebraska, will supply the FRP barreinforcement for negative moment continuity for the actual bridgedeck design.

A bidirectional FRP grid panel provides the top transverse and thetop longitudinal (direction with respect to the girders) tensile re-inforcement. The transverse reinforcement for negative moment overthe girder is provided by 50.8-mm (2-in.) deep, I-shaped FRP bars.The longitudinal reinforcement, commonly referred to as temperatureand shrinkage reinforcement, is supplied by oval FRP bars (dark barsin Figure 1). The oval FRP bar stays in the plane of the 50.8-mm I barsby penetrating through and mechanically locking with the I-bar web.The orthogonal members of the grid are spaced at 101.6 mm (4 in.)on center in each direction. Both bars are quite smooth and cannotdevelop adequate bond to the concrete. However, because they liewithin the same plane and are inherently connected, they mechani-cally anchor one another. Pultruded FRP grid reinforced bridge deckstudies were conducted by Bank et al. (6 ). Strongwell, of Chatfield,Minnesota, produces and supplies the FRP reinforcement grid panel.

The concrete mix design used in experimental testing is the samedesign mix the WisDOT specifies for bridge deck construction, aClass D with 19-mm (3⁄4-in.) maximum aggregate size and a nominal28-day compressive strength of 27.6 MPa (4 ksi).

EXPERIMENTAL RESEARCH

To characterize the FRP reinforcement system, two primary areasof experimental research were conducted. First, structural testing ofcomposite FRP reinforced slab and beam specimens was performed.Second, concurrent to the structural testing, an investigation of thephysical and mechanical characteristics of the FRP material wasconducted. The results of selected mechanical properties of the SIPdeck form and grid are shown in Table 1. The latter provided prop-

FRP Deck Form

Bi-Directional Grid

Plastic Chair

FIGURE 1 Mock-up of FRP reinforcement system.

erties used to calculate analytical deflections and strengths of thestructural specimens.

STRUCTURAL TESTING

The laboratory testing was carried out at the Structures and MaterialsTesting Laboratory at the University of Wisconsin-Madison. The FRPreinforced specimens were tested with MTS Systems Corporationclosed-loop servohydraulic actuator test equipment. Simply sup-ported slabs and beams were tested with a single-point load providedby an 890-kN (200-kip) capacity actuator. For two-span continuousbeam testing, a 244.75-kN (55-kip) capacity actuator was added tosupply a second load. The actuators were each mounted on rigid steelframes that were firmly anchored on concrete supports that were pre-stressed into a structural testing floor. Each test specimen was set on203.2-mm (8-in.) wide hydrocal plaster bearings on concrete bents.

The first set of tests conducted was to determine an effective designwheel load distribution width of an FRP reinforced deck panel. Effec-tive wheel load distribution width is directly related to the span lengthof a simply supported section (7, 8). Therefore, three deck panels ofdifferent span lengths were constructed and tested to derive an equa-tion of effective distribution width. A 2.745-m (9-ft) deck panel width[six SIP deck forms at 457.2 mm (18 in.) each] was selected. A 200-mm (8-in.) deck thickness was selected to match the actual struc-tural bridge deck thickness of the proposed bridge on US-151. Wheelload distribution widths, for positive moment, were determined forservice and ultimate loads. Each simply supported deck panel wasloaded over an area that would represent the contact area of a doubletire wheel, 610 × 305 mm (24 × 12 in.) and was centered in the span(see Figure 2 for picture of test setup). In addition to establishingthe distribution width, the deck panels were loaded to find the ulti-mate capacity and to identify the failure mode (punching shear or flex-ure or a combination thereof). From these results, a safety factor of theservice load, 71.2 kN (16 kips), to ultimate capacity could be estab-lished. The transverse flexure cracking, or joint opening (transversewith respect to the main flexural reinforcement), was measured asload was increased above service loading, because this was expectedto be a significant failure mode of the SIP form system. Before thefinal structural testing, each deck panel was conditioned above serviceloading for the purpose of a “break in” to simulate an actual in-servicecondition. To break in the deck panel, it was cyclically loaded from4.45 to 106.8 kN (1 to 24 kips) 10 times. The next step was to moni-tor the strains and deflections under service load of 71.2 kN (16 kips)to determine stiffness behavior, as well as to determine the wheel load

Dieter et al. Paper No. 02-3205 221

distribution width. After the elastic testing was complete, inelastictesting began by loading the deck panel at a uniform rate to ultimatecapacity, again measuring strains and deflections.

Following the slab tests, the positive transverse bending momentcapacity (the transverse tensile reinforcement is provided by the SIPdeck form) was considered. The positive moment capacity neededto be isolated for measurement. This was accomplished by apply-ing a concentrated line load, distributed evenly across the width, ona single-span, simply supported beam. These beams were 200 mm(8 in.) thick and 0.914 m (3 ft) wide (two deck forms wide). Onlyone span length was selected, 2.65 m (8 ft 8 in.), and three beamswere tested to establish repeatability of the results. The load fromthe actuator was applied onto a 0.914-m (3-ft) long, 76.2-mm (3-in.)square steel bar centered in the span. The loads were applied at a uni-form rate until ultimate capacity was reached; strains and deflectionswere measured in both the top and bottom surfaces of the section,concrete, and FRP deck form, respectively.

Subsequently, the transverse negative moment capacity of theFRP reinforcement system, which is provided by the pultruded grid,was studied. To achieve this, two loading actuators were required,one for each span (see Figure 3). Loads were applied evenly andsimultaneously. Each of the two load points, located at the center-line of their respective spans, was applied uniformly across thewidth of the specimen. Three specimens, 0.914 m (3 ft) wide, 5.28 m(17 ft 4 in.) long, and 200 mm (8 in.) thick, were tested.

Deck Material Test Configuration Strength MPa (ksi) Modulus GPa (ksi)

Longitudinal Tension 504.3 ± 111.6 (73.2 ± 16.2) 33.4 ± 4.1 (4841 ± 593.6) Longitudinal Compression 484.4 ± 91.6 (70.3 ± 13.2) 37.5 ± 4.9 (5436 ± 706.6) Transverse Tension 29.6 ± 1.30 (4.3 ± 0.19) Not measured Transverse Compression 82.0 ± 6.5 (11.9 ± 0.94) 7.5 ± 0.58 (1087 ± 84.0) Short Beam Shear 38.6 ± 3.3 (5.6 ± 0.48) Not Applicable Grid Material Longitudinal Tension 571.9 ± 12.3 (83.0 ± 1.79) 30.9 ± 1.07 (4480 ± 155.7) Longitudinal Compression 648.3 ± 5.85 (94.1 ± 0.85) Not measured Transverse Tension 73.7 ± 7.51 (10.7 ± 1.09) Not measured Transverse Compression 170.9 ± 3.24 (24.8 ± 0.47) Not measured Short Beam Shear 51.0 ± 0.90 (7.4 ± 0.13) Not Applicable

NOTE: Values included ± 1 standard deviation. Test data based on a minimum of 5 specimens.

TABLE 1 Selected Mechanical Properties of FRP Material

FIGURE 2 Picture of deck panel test setup.

Finally, to investigate serviceability, a 2,000,000-cycle fatigue testwas conducted. This test was intended to stress the FRP reinforce-ment system through tensile cycles and measure the relative changein stiffness of the reinforced section. The beam was geometrically thesame as the negative moment beams. The same test setup used forthe transverse negative beams was used to test the fatigue beam. Theactuator of one span held a static load of 71.2 kN (16 kips), whereasthe other actuator was loaded from 21.4 kN (4.8 kips) to 92.6 kN(20.8 kips) at 2 Hz. Before the test, the beam was loaded to beyondthe negative moment cracking load to ensure the FRP grid alonewould be subjected to tensile force over the center support. At every200,000 cycles, both spans were simultaneously loaded to 71.2 kN(16 kips), and center deflections were measured.

DESCRIPTION OF SELECTED TEST RESULTS

As mentioned previously, the main purpose of this experimental test-ing was to understand the failure modes and fatigue performance andto determine the ultimate and service load capacity of the system. The

222 Paper No. 02-3205 Transportation Research Record 1814

specimens were tested sequentially: deck panels, positive momentcapacity beams, negative moment capacity beams, and finally thefatigue beam. Only selected conclusions on the primary test resultsare reported; further detailed data will be presented in future papers.

Deck Panels

Loading the deck panels to ultimate capacity revealed that the fail-ure mechanism, based on observation of the cracking pattern, wasclearly punching shear for all three specimens (test data are providedin Table 2). This coincides well with the analytical predictions, inwhich a punching shear [Vc, nominal shear strength provided byconcrete alone, calculated from ACI 318-99 (9, Equation 11-37]failure should have a lower capacity than a flexural failure [Mn, nom-inal flexural capacity based on concrete crushing failure mode, cal-culated from ACI 440.1R-01 (10, Equation 8-4a), using an averagemeasured f ′c of 38.2 MPa (5.54 ksi)]. Other significant conclusionscan be drawn from the data provided in Table 2. First, looking at theultimate load, a large discrepancy can be seen between the analyti-cal and the experimental values. Cutting the deck panel into halvesalong the centerline revealed the reason (see Figure 4). Analyticalcalculation of the punching failure was based on a 200-mm (8-in.)thick concrete section, which is not appropriate as seen from thecracking pattern of the punching failures. The shear crack starts atthe shiplap joint and travels along the interface between the concreteand FRP deck form interface until it reaches the top of the corruga-tion (especially on the outboard side of the deck form that is notcoated with epoxy gravel). This essentially means that only the con-crete from the corrugation top to the load patch provides punchingshear resistance. No capacity can be expected from the shiplap jointto the top of the corrugation because the deck form is smooth andthus has no friction for tensile strength. Therefore, the effectivedepth for analytically calculating the shear capacity must be reducedto 127 mm (5 in.): 200-mm (8-in.) thick slab minus 76.2-mm (3-in.)high corrugation. A final conclusion can be drawn by consideringthe factor of safety (FS) of the service load to ultimate load. Asshown in Table 2, the factor of safety is adequate, between 5 and 6.

The experimental effective live-load distribution widths for each ofthe three deck panels are shown in Table 2. Because of problems with

FIGURE 3 Picture of negative moment test setup.

* Based on an 8� thick section **Based on a 5� thick section

Deflection at Service Wheel Load, 71.2 kN (16 kip), mm (in.) Failure Mechanism

Strength Capacity, kN (kips)

Factor of Safety, Ultimate/Service Load, 71.2 kN(16 kip)

Positive Moment Distribution Width, m (in.)

Deck Panel Span Predicted Exper. Predicted Exper. Predicted Exper. Experimental

AASHTO LFD Predicted

At Service, Exper.

At Ultimate, Exper.

8�-0�

0.508 (0.020)

0.685 (0.027)

Punching Shear

Punching Shear

845.5* (190)*

529.6** (119)**

405.0 (91)

5.68(79)

--(75)

9�-10� 0.787 (0.031)

0.889 (0.035)

Punching Shear

Punching Shear

881.1* (198)*

551.8**

(124)**

387.2 (87)

5.43(84)

--

1.91

1.73(68)

11�-6� 1.78 (0.070)

1.80 (0.071)

Punching Shear

Punching Shear

841.1* (189)*

525.1** (118)**

427.2 (96)

6.00

2.01

2.14

2.14(84)

1.75(69)

1.58(62)

TABLE 2 Analytical Predictions Versus Experimental Results for Deck Panels

Dieter et al. Paper No. 02-3205 223

Edge of Load Patch

Top of Corrugation

Shiplap Joint

0

20

40

60

80

100

120

0 0.5 1 1.5 2 2.5 3Center Deflection, in.

Lo

ad, k

ips

11'-6" Span, Ult. = 96 kips 9'-10" Span, Ult. = 87 kips 8'-0" Span, Ult. = 91 kips

1 in. = 25.4 mm 1 kip = 4.45 kN

FIGURE 4 Postfailure of deck panel showing punching shear failure directly underload patch.

FIGURE 5 Load and center deflection response of three deck panels.

is an outboard side of the FRP deck panel that does not have an epoxygravel coating (Figure 6). Strain gauges indicate a nonuniform strainacross the FRP deck form. In comparing the sides of the deck panelafter testing, a significant difference in cracking patterns is observed(Figure 7). The gravel epoxy outboard side (bonded) of the beamshows many flexural cracks near midspan that transition into flexural-shear cracks as they approach the support. This is indicative of aneffective horizontal shear transfer between the concrete and the FRPdeck form. The smooth outboard side (nonbonded) of the beam showsa few large flexural cracks near midspan, indicative of a lack of bondwith the tensile reinforcement.

One of the three positive moment beams was tested without a topreinforcement grid to examine any differences in structural behav-ior of beams with and without the grid in the positive moment area.Although it is for negative moment reinforcing, there are some indi-cations that the top FRP grid does provide some benefit to postultimatecapacity in the positive moment region. The load-versus-deflectionresults for all three beams (Figure 8) show two beams, D1 and D2,(with the top FRP grid reinforcement) maintaining a sizeable loadthrough large deflections after ultimate, whereas the one positivemoment beam without the FRP grid, D3, does not.

The ultimate strengths of the positive moment beams tested were,on average, 57% [averaged, 320.4 kN (72 kips) versus 565.2 kN(127 kips) of the strength calculated analytically with an averagemeasured f ′c of 43.5 MPa (6.31 ksi) (10, equation 8-4a)]. The differ-ence may be attributed, somewhat, to the lack of complete composite

the data acquisition system, the effective live-load distribution width,at service load, for two of the three specimens could not be deter-mined. Yet, depending on the span, it can be noted with some confi-dence that the experimental effective distribution widths, at ultimate,range from 5% to 24% less than that of AASHTO load factor design(LFD), Section 3.24.3.2 (7), effective width equation [E = 4 + 0.06 �clear span (ft), but not to exceed 2.135 m (7 ft)].

The deck panels also show an interesting load-versus-deflectionresponse (see Figure 5). It can be seen that after failure, the deck panelexhibits an extraordinary capacity to hold a relatively large loadthrough significant deflections. Cracking patterns, ultimate capacitiesfalling short of analytical capacity, and FRP deck form strains indi-cate that the FRP reinforcement system falls short of being a fullycomposite section. The lack of a fully composite section may beattributed to the incomplete coverage of epoxy-coated gravel on thehorizontal surfaces of the FRP deck form, among other factors.

Positive Moment Beams

The testing of the positive moment beams provided additional evi-dence for a not fully composite section between the concrete andthe FRP deck form. This is due to the lack of uniform shear trans-fer between the FRP deck panel and the concrete. The epoxy gravelon the horizontal surfaces of the deck form is intended to provide thebond to transfer the shear between the two materials. However, there

224 Paper No. 02-3205 Transportation Research Record 1814

Epoxy Outboard Side

Smooth Outboard Side

Epoxy/Gravel

No Epoxy/Gravel

01020

30405060

708090

0 0.5 1 1.5 2 2.5 3 3.5 4Center Deflection, in.

Lo

ad, k

ips

D1, Ultimate = 63.5k D2, Ultimate = 77.6k D3, Ultimate =76k, NO TOP GRID

1 in. = 25.4 mm 1 kip = 4.45 kN

FIGURE 6 Epoxy gravel placement on the FRP deck forms.

FIGURE 8 Load versus center deflection plot of the three positive moment beams.

(a) (b)

FIGURE 7 FRP deck form sides of posttest positive moment beam: (a) bonded and (b) nonbonded.

action. Further testing is currently being conducted to investigateimproving composite action by adding epoxy and gravel to cover allhorizontal surfaces of the FRP deck form. The moment–curvatureresults of the positive moment beams are shown in Figure 9.

Negative Moment Beams

Negative moment beams were tested to determine the strength andstiffness behavior of the FRP top grid reinforcement over a support,

such as a girder. Figure 3 shows the test setup of the two-span con-tinuous beam with an actuator to load the center of each 2.47-m(97-in.) span. Ultimate failure occurred at midspan (positive momentregion) under a mechanism of flexure concrete crushing combinedwith shear. Figure 10 shows the load versus center deflection rela-tionship from Span A. Similar to the positive moment beams results,the FRP system permits support of a significant load, relative to theultimate load, through large deflections after the ultimate capacity isreached.

Strain gauges were attached to the FRP grid over the support tomonitor the strains caused by the negative moments. In the same ver-tical cross section, a strain gauge was applied to the concrete in com-pression (bottom of slab centered above the bearing) in the maximumnegative moment region. These two instruments were used to developa moment–curvature relationship over the middle support (Figure 11).Notice the initial slopes are quite steep until the concrete in tensioncracks and the tensile force is transferred to the FRP grids.

Fatigue Beam

The load-versus-deflection data were examined to observe any sig-nificant relative stiffness change to the accelerated flexural fatigueloading. Figure 12 reveals little relative stiffness change, or loss, overthe 2,000,000 cycles. No tensile concrete cracks were discoveredbeyond the initial cracks, from the cracking load, over the support.

PLANS FOR DEVELOPING DESIGN METHODS

With further detailed examination of the data, design methods andguidelines for the FRP reinforcement system will be established.Plans include developing a finite element model (FEM) of the systemusing the experimented structural behavior for comparison. An accu-rate FEM, along with experimental and analytical data analysis, willhelp in developing positive and negative distribution widths.

Dieter et al. Paper No. 02-3205 225

SUMMARY

This research project experimentally examined the structural behav-ior of an FRP SIP deck form and bidirectional grid concrete re-inforcement system. Key experimental results and analysis conductedto date are provided. Accelerated flexural fatigue loading does notsignificantly decrease the stiffness of the reinforcement system. Dis-tribution widths, at ultimate, of the deck panel tests were found torange from 5% to 24% less than AASHTO’s equation for live-loaddistribution. Mode of failure for these deck panels is punching shear,as analytically predicted. The experimental results (strain gaugeinformation) of the positive moment beams indicated that the re-inforcement system does not achieve a fully composite action withthe concrete.

ACKNOWLEDGMENTS

The authors acknowledge the following persons for their contribu-tions: Stan Woods and Gerry Anderson of the Wisconsin Depart-ment of Transportation; Bernie Gallagher of Alfred Benesch andCompany; Tom Strock of FHWA; Bill Lang and John Dreger of theUniversity of Wisconsin, Structures and Material Testing Labora-tory; Brian Beaulieu, Eric Helmueller, Wyatt Henderson, and BeauSanders, University of Wisconsin undergraduate students. The au-thors also acknowledge the financial support from the InnovativeBridge Research and Construction Program.

0

500

1000

1500

2000

2500

0 0.0005 0.001 0.0015 0.002 0.0025 0.003

Curvature

Mo

men

t, in

-k

D1 D2 D3, No top grid

1000 in-k = 113 kN-m

01020

30405060

7080

0 0.5 1 1.5 2Center Deflection, in.

Lo

ad, k

ips

E1 E3 E51 in. = 25.4 mm 1 kip = 4.45 kN

FIGURE 9 Moment–curvature results of the three positive moment beams.

FIGURE 10 Load versus center deflection relationship from Span A of the three negativemoment beams. [NOTE: Because of problems with data acquisition system during test of E5,trend-line is cut short of ultimate load. E5 ultimate load � 316 kN (71 kips).]

REFERENCES

1. Alkhrdaji, T., A. Nanni, and R. Mayo. Upgrading Missouri Transporta-tion Infrastructure: Solid Reinforced-Concrete Decks Strengthened withFiber-Reinforced Polymer Systems. In Transportation Research Record:Journal of the Transportation Research Board, No. 1740, TRB, NationalResearch Council, Washington, D.C., 2000, pp. 157–163.

2. Chajes, M. J., H. W. Shenton, III, and W. W. Finch, Jr. Performance ofGlass Fiber-Reinforced Polymer Deck on Steel Girder Bridge. In Trans-portation Research Record: Journal of the Transportation ResearchBoard, No. 1770, TRB, National Research Council, Washington, D.C.,2001, pp. 105–112.

3. Bradberry, T. E. Concrete Bridge Decks Reinforced with Fiber-ReinforcedPolymer Bars (With Discussion and Closure). In Transportation ResearchRecord: Journal of the Transportation Research Board, No. 1770, TRB,National Research Council, Washington, D.C., 2001, pp. 94–104.

4. Taly, N. Design of Modern Highway Bridges. McGraw-Hill, New York,1998.

5. Harik, I., P. Alagusundaramoorthy, R. Siddiqui, R. Lopez-Anito, S. Morton, P. Dutta, and B. Shahrooz. Testing of Concrete/FRP Com-

226 Paper No. 02-3205 Transportation Research Record 1814

posite Deck Panels. Proc., 5th ASCE Materials and Engineering Con-gress (L. C. Bank, ed.), Cincinnati, Ohio, ASCE, Reston, Va., 1999,pp. 351–358.

6. Bank, L., Z. Xi, and E. Munley. Tests of Full-Sized Pultruded FRPGrating Reinforced Concrete Bridge Decks. In Materials: Performanceand Prevention of Deficiencies and Failures (T. D. White, ed.), ASCEMaterials Engineering Congress, Atlanta, Ga., Aug. 1992, pp. 618–631.

7. Standard Specifications for Highway Bridges, 16th ed. AASHTO,Washington, D.C., 1996.

8. AASHTO LRFD Bridge Design Specifications, 2nd ed. AASHTO,Washington, D.C., 1998.

9. Building Code Requirements for Structural Concrete and Commentary.ACI 318-99 and ACI 318R-99. American Concrete Institute, FarmingtonHills, Mich., 1999.

10. Guide for the Design and Construction of Concrete Reinforced withFRP Bars. ACI 440.1R-01. American Concrete Institute, FarmingtonHills, Mich., 2001.

Publication of this paper sponsored by Committee on Structural Fiber-ReinforcedPlastics.

0.0

200.0

400.0

600.0

800.0

1000.0

1200.0

0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035

Curvature

Mo

men

t, in

-kip

s

E1, Initially Uncracked E3, Initially Uncracked

1000 in-k = 113 kN-m

# of Cycles

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08

Center Deflection, in

Lo

ad, p

ou

nd

s

200,000 600,000

1,000,000 1,600,000

2,000,000

1 in. = 25.4 mm 1 kip = 4.45 kN

FIGURE 11 Negative moment beam test, moment–curvature relationship over middle support.

FIGURE 12 Load–deflection response under 200-kip actuator, 2,000,000-cyclefatigue test.