L

DEVELOPMENT OF FRP BRIDGE DECKS AN ANALYTICAL INVESTIGATION -

Dr. N ahla K. Hassan Ain Shams University Faculty of Engineering

Structural Engineering Department Cairo, Egypt

Dr. Emile Shehata Wardrop Engineering Inc.

Winnipeg, Manitoba Canada, R3C 4M8.

Brea Williams The University of Manitoba

ISIS Canada, Winnipeg, Manitoba CanadaR3T 5V6.

Prof. Sami Rizkalla North Carolina State University

Dept. of Civil Eng. & Construction, Campus Box 7533, Raleigh, NC27695,USA

KEY WORDS: Bridges, Decks, FRP Composites, Finite Element, Tsai-Wu, Failure Criterion

ABSTRACT Building a functional transportation infrastructure is a high priority for any nation. Equally important is maintaining and upgrading its condition to keep pace with increasing usage, higher traffic loads and new technologies. In recent years, highway bridge decks constructed from modular fiber reinforced polymeric (FRP) materials have been the subject of a great deal of research interest, providing not only alternatives to traditional materials and structural systems, but opportunities for more rapid rehabilitation and even functional enhancement over the original design. Experimental studies on the modular all-fiberglass reinforced plastic FRP bridge deck produced encouraging results. Finite Element Analysis (FEA) was used to get a detailed behavior of the these bridge decks, as well as to predict the maximum load capacity, deflections, strains and failure modes. The proposed GFRP bridge decks consist of a series of triangular tubes produced by filament winding process, which are combined with pultruded glass fiber bars and plates to create a single deck module. The ANSYS Finite Element program is employed to analyze the tested specimens at service loading conditions and Tsai-Wu failure criteria is utilized to examine the failure regions of the deck and the different failure modes. The predicted results from the Finite Element Analysis are then compared to the experimental results. This paper summarizes the F.E. model used, analytical results, and the comparison with the experimental data.

INTRODUCTION One area of technology that contributed to the space race was the usage of new "high-tech" materials, including fiber reinforced polymer (FRP) composites. Composites offer inherent advantages over traditional materials (like steel, concrete, and aluminum) including high strength

-to-weight ratio, design flexibility, corrOSIOn and fatigue resistance, low maintenance and extended service life. The United States ignited the "technological evolution" of FRP Composites. Today, however, other countries as Canada, Japan and Western Europe understood the potential of composites in construction applications, including repair and rehabilitation of civil infrastructure. It is a well­known fact that more than 40% of USA's highway bridges are classified as either structurally or functionally deficient. Specially, the carrying capacity of a total of 22,064 reinforced and pre­stressed concrete bridges has been significantly reduced and represents the most serious type of deterioration. These bridges must be replaced with a new system that adequately serves future traffic needs and offers superior advantages compared to conventional bridge systems. Analytical and experimental study programs on modular all-Fiberglass Reinforced Plastic (FRP) Bridge deck produced encouraging results. The studies determined that composite bridge decks present several potential advantages over traditional systems. First, their relative lightweight may translate into increased live load capacity of an existing bridge or additional traffic lanes with the same existing girder system. Second, modular units may be fabricated off-site and quickly assembled at the bridge site, with saving in both construction time and cost. Third, when properly designed, composites will resist a wide range of corrosive environments. In the preliminary analytical work reported by Henry, Ahmed and Plecnik, and Plecnik and Azar, the performance of several glass reinforced polymer bridge deck configurations were investigated using SAP IV. The results revealed that the design was always controlled by the deflection limit states rather than the strength limit states and that the X -shaped deck has the lowest deflection. McGhee and McGhee et aI, presented results of the mathematical optimization such that the objective function was represented by the weight of the deck, while the behavioral constraints were: ultimate strength, local buckling and span/800 deflection limit states. The study concluded that the triangular deck cross-section could efficiently provide fabrication savings. Bakeri and Bakeri and Sunder concluded that using the hybrid concept composed of glass-fiber reinforced polymer, carbon fiber reinforced polymer and light weight concrete resulted in a bridge system having a deflection less than the s/800. Also West Virginia University together with Creative Pultrusions produced a honeycomb shaped core bridge deck of 9m long and 6.6m wide modular GFRP bridges in West Virginia in 1997. The bridge is constructed to resist the AASHTO HS-25 loads. Finally, Martin Marietta Inc. joined with Glassforms Inc. in 1997 to install two GFRP bridges. These demonstration projects are becoming more common and numerous each year. In light of the above, this paper represents a study of the detailed behavior of GFRP bridge decks consisting of three triangular modules bonded together and adhered to top and bottom pultruded plates as well as GFRP bars. The ANSYS finite element program is utilized to get a deep insight of the stress and strain distribution in the deck, and the maximum deflections reached as well as the mode of failure. The Tsai-Wu failure criterion is applied in the analytical investigation. Also a brief summary of the GFRP bridge deck fabrication process, the testing procedure and experimental results conducted by Brea Williams at University of Manitoba is reported. A comparison between theoretical and experimental results regarding strains and deflections is provided. Finally a parameteric study is presented to indicate the importance and contribution of each of the top and bottom plates, tubes and GFRP bars to strength of the deck at service load according to AASHTO 1996 specifications.

MATERIALS & FABRICATION The deck consisted of three filament wound tubes, approximately 200mm in height. An eight layer laminate design was used for the tubes with a layering sequence of [90/ ± 45/ ± 10/ ± 45/90], with a thickness of 0.67mm for each layer. Due to the rounded comers of the triangular sections, pultruded GFRP bars were placed in the section. GFRP puUruded plates of thickness of 15mm were adhered to the top and bottom of the tubes to create one modular unit. The plates consisted of five laminates bonded together with a layering sequence of [ 0/90/0/90/0], each layer thickness is 2.6mm. A schematic of the deck cross-section is shown in Fig.(1). The material properties of the fibers and the epoxy are given in table (1) and table (2). The fabrication process is summarized as follows; (B. Williams ): 1. Triangular shaped styrofoam manderls were custom-made. A thin layer of epoxy­

prepregnated chopped fibers was applied to strengthen the mandrel. 2. Three 3.5m long mandrels were filament wound with Owens Coming glass fiber rovings to

produce the eight layer laminate designed for the tubes. A custom-made epoxy resin with a 24-hour pot life was used.

3. The three triangular tubes were placed on the GFRP plate and bonded with added resin. The GFRP bars were placed between the tubes.

4. The top plate was positioned and bonded with added resin. Upon completion of assembly, the deck was wrapped and sealed in a plastic bag.

5. The deck was cured at 180"F for 8-10 hours, while a vacuum pump worked to remove excess resin from the deck and minimize voids. When curing was complete, the deck module was cut to the desired length.

TESTING AND RESULTS The deck was tested under simply supported conditions with a single 250x250mm point- load applied at the center of its 3m span. This load simulated the wheel load of the HS-30 truck according to the AASHTO specifications. The load was applied using a 1000 KN machine using stroke control at 0.75mm1min, as shown in Fig.(2). The experimental results (B. Williams) showed that all decks demonstrated similar linear behavior under the applied load. This included uniform stiffness throughout loading, interrupted by buckling of plates. For example, for specimen F2-TB-2 plate buckling occurred at a load level of300 KN and the deck failed due to delamination ofthe bottom plate at a load level of 414 KN. The three modes of failure observed, for all the seven specimens, were plate buckling, plate delamination and buckling of the tubes. Table (3) summarizes the important failure parameters for the tested decks.

THEORETICAL MODELING OF BRIDGE DECKS Geometric Modeling The geometric modeling of the GFRP bridge deck cross-section is represented by 20 key-points in terms of coordinates in x, y, and z direction, then extruded to form a full scale dimension of the bridge deck of width 500mm and height of 200mm and a length of 3m. Due to symmetry only half of the deck is analyzed using F.E program. To create symmetry at the middle, the translation in X-direction, the rotations in Y and Z-directions are restrained as shown in Fig.(3). The radius of the fillets at the ends of the triangles is of radius 20mm to meet with the dimension requirements of the tested deck so that these triangles (filament wound tubes) are within the top

and the bottom plates. The GFRP bars are modeled as semi-circles of area equal to the area of the original bars in the experimentally tested FRP deck. The load is applied in the middle on an area of 250mmx250mm to simulate exactly the HS-30 truck according to the AASHTO specifications (1996) as a uniformly distributed load. The load is converted into point loads at nodes of the elements of the top plate mesh as shown in Fig.(3). The size of the meshed elements ranged from 2.5cm at the region of load application to 5cm at the boundaries of the deck. The boundary conditions are chosen to simulate exactly the testing procedure. At one end the three translations in X, Y, and Z directions are restrained and the rotations are free. At the other end only two translations in X, and Y directions are restrained and the rest of translations and rotations are free as shown in Fig.(3).

Material Modeling As for the material modeling, the ply strength and stiffuess properties of each of the top or bottom plates and the tubes are first computed using spread sheet data by the Micro-soft Excel program based on the micro-mechanical analysis. The micro-mechanical analysis is based on Chamis to estimate the following engineering strength and stiffuess properties for each ply: i) longitudinal modulus E1, ii) transverse modulus Ez, iii) shear modulus G12, iv) longitudinal and transverse tensile strength SL +, ST +, v) longitudinal and transverse compressive strength SL- , ST-, vi) shear strength S. These equations are given by Mallick as follows:

1- Longitudinal modulus El : Er =EfVf+EmVm

2- Transverse modulus Ez : lIE2 = Vf l Ef + Vml Em

3- Shear modulus G12 : lIG12 = Vfl Ef + Vm I Gm

4- Major Poisson's ratio V12 :

V12 = VfVf + Vm Vm

5- Longitudinal Tensile Strength SL + :

st = Sf V f + Sm V m

6- Longitudinal Compressive Strength SL - : SL- = 2 Vcr Vf Em Ef 13 (I-Vf) ]1/2

7- Transverse Tensile Strength ST + : ST+ = (E2Sm+)/(EmF)

F = lI(d/s[(Em/Er)-I] +1)

8- Transverse Compressive Strength ST- : ST- = (E2 Sm-) I (Em F )

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

9- Shear Strength S : S = (Ez Sm ) / ( Em Fs)

Fs = 11 (dls)[ (Gm /Gr)-l] +1

Where dis = (4Vd 'IT) 'h

Where Ef : fiber modulus Em : matrix modulus V f : volume of fiber V m : volume of matrix Gf : shear modulus of fiber Gm : shear modulus of matrix Vf : Poisson's ratio of fiber Vm : Poisson's ratio of matrix Sm + : tensile strength of matrix Sm- : compressive strength of matrix Sm : shear strength of matrix

(10)

(11)

(12)

These computed ply strength and stiffuess properties are incorporated in the F.E. program for each component of the FRP deck as given in Table (4). For the top and bottom plates, each is composed of five layers of sequence : [ 0/90/0/90/0] and of thickness 2.6mm. For the tubes, each side is composed of eight layers of sequence: [90/ ± 45/ ± 10/ ± 45/90] and of thickness of 0.67mm. As for the epoxy joining the three triangular tubes from inside is considered as a single layer ofthickness of 0.5mm. In the F.E.A., the regions "A, B, C", shown in Fig. (1), comprising the top plate, epoxy and the tube are modeled as one unit consisting of 14 plies. These plies are 5 layers for the plates, one layer for the epoxy as a 0° lamina, and 8 layers for the tubes of the following sequence: [0/90/0/90/0/0/90/±45/± 10/±45/90]. For the fillets and parts "D, E", they are modeled as 8 layers of [90/ ± 45/ ± 10/ ± 45/90]. For the inside part of the tubes (regions F, G) they are modeled as 17 plies of the following sequence: [90/ ± 45/ ± 10/ ± 45/90/0/90/ ± 45/ ± 10/ ± 45/90]. They could be defined as 8 layers of the tube, one layer for the epoxy as a unidirectional ply, and 8 layers for the other tube. The top or bottom GFRP bars, in regions H, I, J, K, are modeled with the top or bottom plates and a layer of epoxy in between as 7 layers. Defined as 5 layers for the top plate, one layer of epoxy of 0.5mm, and a layer of the bar of 9.5mm thickness). They have the following sequence: [0/90/0/90/0/0/0]. The strength properties for the plates, tubes, and bars are incorporated in the Finite Element ANSYS program in the failure criterion data as explained in the next section. The same method is applied for defining the layer properties of other FRP deck configurations, either with or without top and bottom plates and with or without top and bottom bars.

FINITE ELEMENT PROGRAM The ANSYS Program models composite materials by using specialized elements called layered elements. Shell 99 is used for layered applications of structural shell model. The element is composed of 100 layer, and has six degrees of freedom at each node, three translations in the

nodal x, y, and z directions and three rotations about the nodal x, y, and z axes. The geometry, node locations, and the coordinate system for this element are shown in Fig. (4). The element is defined by: eight nodes, average layer thickness, layer material direction angles and orthotropic material properties. The input of data is achieved in a layer form. The force-strain and moment­curvature relationships defining the matrices for linear variation of strain through the thickness is given by the ANSYS theoretical manual volume IV.

Failure Analysis The design analysis of any structure is performed by comparing stresses and strains, due to applied load, with allowable strength or strain capacity of the material. In case of biaxial or multi-axial stress field, a suitable failure theory is used for this comparison. Therefore when a loading condition exceeds those defined by the failure criterion the material is considered to have failed. Under plane stress conditions, Tsai-Wu theory predicts failure in an orthotropic lamina when the following equality is satisfied;

J31 all + J32 0'22 + J36't12 + J311 0'2 11 + J322 0'2 22+ J366 't2

12 + 2J312 alla22 =1 (13) Where, J31 ,J32 , J311 ,J322 ,J366 are calculated using tensile, compression, and shear strength properties, and J312 is the strength interaction term between all ,0'22 and can be approximated to; -1I2[J311 J322 ]112 ~ F12 ~ 0, where J312 is taken as 1 in this investigation.

THEORETICAL RESULTS Cases Studied Eleven different cases of FRP bridge decks were studied utilizing the ANSYS program. Three of them were similar to the experimentally tested specimens to verify the theoretical modeling of the FRP bridge deck. The other eight cases were investigated to indicate the importance and the effect of each of the following: i) the top and bottom plates, ii) the GFRP bars and their contribution to strength and stiffness of the FRP bridge decks. Also to get a deep insight of the behavior and failure modes of the bridge decks under both ultimate and service loads.

Verification of Results Table (5) shows the three models studied. i) Case "A" with top and bottom plates and bars similar to specimen "F2-TB-a" according to the experimental testing. ii) Case "B" with top plate and bars and corresponds to specimen "F2-T-a". iii) Case "C" with top and bottom bars only and corresponds to specimen "FR-a" tested experimentally. Maximum deflections at the middle of the bridge deck, and maximum strains as predicted by the F.E.A. were compared to the experimental results as shown in Fig. (5) and Fig. (6a) ,(6b) for the three previous cases. A very close agreement is obvious for both the theoretical and experimental results with a maximum difference of 9% for cases "A" and "B". As for case "C" a great difference in strain and deflection, reaching 34%, is mainly because that the tested specimen was originally tested before until cracking and premature failure occurred. The top and bottom plates were removed except for a very thin layer of O.5mm thickness and the specimen was tested again as specimen "FR-a". Of course that was not the case for the theoretical analysis, which led to the increase of the experimental strain than the theoretical one. The failure mode of case "A" that failed by top plate buckling is shown in Fig.(7). As for case "C" the failure mode is shown in Fig.(8), which shows clearly the tube buckling. These two figures show the maximum

deflections reached for the FRP bridge decks of values 70.456mm and 96.9mm respectively at the center of the deck and of values reaching zero deflections at the ends.

Application Cases After the verification of the theoretical F.E. model used, eight other cases were solved utilizing the ANSYS program at service load ofa value of I40kN according to the AASHTO HS-30 truck load. These cases clearly showed the effect, and the contribution to strength and stiffness of the top and bottom plates and the GFRP bars as shown in Table (6) . By comparing cases "1, 3, 7 and 8", the contribution of the top and bottom plates to stiffness is about 60% of the total stiffness where as, by comparing cases "1, 4, 5 and 6", the GFRP bars (mainly used to fill the space between the filament wound tubes and the plates) comprise only about 10% of the stiffness. From the above, it is clearly shown that the contribution of the filament wound tubes to strength and stiffness is about 30%. The maximum deflections of some cases of the FRP decks are shown in Figs.(9) and (10) which indicate that the maximum deflection is about 62.9cm for case 3, and 22.5cm for case 1. These values (Ll52, Ll160) are way beyond the service deflection limits ofLl400 (L is the span of the deck) which is recommended for wood by the AASHTO Standards. This deflection limitation is used since there does not exist any code limits for the FRP bridge decks, and also wood has a laminated structure, which is some how similar to FRP composites. However, based on the results above, the GFRP decks as a structural system appear to be sufficiently ductile. This implies that in actual bridge loading conditions there would not be any danger of catastrophic brittle failure. Therefore, it is recommended to increase the number of modules of the bridge deck, or to increase the top and bottom plate thickness, or to use filament wound top and bottom plates and finally to control the angle of fiber orientation for the filament wound tubes to achieve maximum stiffness. As for the strains, the regions of maximum value of measured compressive strain at the top of the plate is at 300mm from the center of the deck (beyond the loading area) ranged from 2.1 ml. strain to 3.8 ml. strain. For the bottom plate the maximum tensile strain, that could be measured, at the center of the deck ranges from 2.8 ml. strain to 8.85 ml. strain as shown in Figs.(ll) and (12) for cases 3 and 5 respectively. The common failure mode predicted by the ANSYS Program is the top plate buckling for all the studied cases except for the bridge decks without top plate, which failed by tube buckling. The used Tsai-Wu criterion in the F.E.A. predicted the regions of failure and the layer at which failure occurred efficiently as shown in Fig.(13).

CONCLUSIONS 1- The theoretical analysis of FRP bridge decks, in terms of geometric, material and layer

modeling, proved to be very efficient to predict the ultimate and service load deflections and strains. Good agreement was achieved between the theoretical and experimental results with a difference of 15% maximum.

2- The contribution of top and bottom plates, GFRP bars, and tubes to the strength and the stiffness is about 60%, 10% and 30% respectively.

3- The deflection limits at service load ranges from Ll52 to Ll160. Enhancement of stiffness to FRP bridge decks to meet the service limits of deflections of Ll400 can be achieved through the increase of stiffness of the top and bottom plates. This could be done by increasing the plate thickness or using filament wound plates and changing the angle of fiber orientation of the different plies (needs further investigation).

4- The Tsai-Wu criterion used in the F.E.A predicted the failure regions ofthe deck efficiently. 5- The common mode of failure is top plate buckling except for the decks without top plate,

which failed by tube buckling. 6- GFRP is commonly perceived as being a brittle material, providing minimal advance warning

of failure. However, based on the results of this investigation, the GFRP decks as a structural system appear to be sufficiently ductile. This implies that in actual bridge loading conditions there would not be any danger of catastrophic brittle failure.

7- Using the ANSYS Program to analyze bridge decks behavior gave promising results that leads to more and further investigation needed in that field especially under fatigue loading.

ACKNOWLEDGEMENT The authors would like to express their deepest gratitude to all staff who shared effort in producing this paper especially ISIS Canada, Structure's lab. and Computer center at the University of Manitoba, Winnipeg, CANADA.

REFERENCES S.H. Ahmad and I.M. Plecnik (1989) "Transfer of composite technology to design and

construction of bridges," U.S. DOT Report, Sept.. American Association of State Highway and Transportation Officials, LRFD bridge design

specifications, (1998), (AASHTO Manual), pp.2-12. P.A. Bakeri (1989) "Analysis and design of polymer composite bridge decks", Thesis presented

to the department of civil engineering in partial fulfillment of the requirements for the Degree Master of Science at the Massachusetts Institute of Technology, September.

P.A. Bakeri and S.S. Sunder,(1990) "Concepts for hybrid FRP bridge deck system," Serviceability and Durability of Construction Materials, B.A. Suprenant (ed.), Proc. First Materials, Engng. Congress, SACE, Denver, Colorado, Vol. 2, pp. 1006-1014.

C.C. Chamis, (1984) "Simplified composite micro-mechanics equations for hygral, thermal and mechanical properties", SAMPE Quarterly, 15: 14.

1.A. Henry, ( 1985) "Deck girders system for highway bridges using fiber reinforced plastics," M.S. Thesis, North Carolina State University, USA.

K.K. McGhee, (1990) "Optimum design of bridge deck panels using composite materials," Thesis presented to the faculty of Eng. and Applied Science, in partial fulfillment of the requirements for the Degree Master of Science, Univ. of Virginia,.

K.K. McGhee, F.W. Barton and W.T. McKeel, (1991) "Optimum design of composite bridge deck panels," Advanced Composite Materials in Civil Engineering Structures, Proc. Specialty Conference, ASCE, Las Vegas, Nevada, pp. 360-370 ..

Martin Marietta Materials.(1999) "Composite Technology.". http://www.martinmarietta.comlcorpsite/ aboutmmmltechnologiescomptechnology.asp

P.K. Mallick (1993) "Fiber-reinforced composites, materials, manufacturing, and design ", Marcel Dekker.

Owens Coming, (1999) "Bridges". http://www.owenscorning.com/owens/composites/applications/infra/bridge.html.

I.M. Plecnik and W.A. Azar, (1991) "Structural components, highway bridge, deck applications," International Encyclopedia of Composites, I. Lee and M. Stuart (eds), Vol. 6, pp. 430-445,. Swanson Analysis Systems, (1992), Inc., ANSYS manual.

Williams Brea, (2000) "The development of GFRP bridge deck modules" Thesis presented to the dept. of civil engineering in Partial Fulfillment of the Requirements for the Degree Masters of Science in Civil Eng., University of Manitoba. Williams B., S.Rizkalla, E.Shehata, D.Stewart and K. Church, (2000) "Development of modular GFRP bridge decks" Advanced Composite Materials in Bridges and Structures 3rd International Conference, Ottawa, Ontario, Canada.

Table (1): Glass Fiber Roving Properties

Proper~ Value Specific Gravity 2.624 Tensile Strength [MPal 1700 Tensile Modulus [GPa] 72.4 Strain at Failure 4.6% Poisson's Ratio 0.22 Thermal Expansion [1 0-6/oC] 5.8

Table (2): Epoxy Resin Properties

Property Value Specific Gravity 1.163 Tensile Strength [MPa] 64.8 Tensile Modulus fGPal 3.15 Poisson's Ratio 0.27 Percent Elongation 9.9 Heat Deflection Temperature rOC] 103

Table (3) : Failure Parameters ofFRP Bridge Deck

Deck Failure Load Failure Mode Maximum (kN) Deflection (mm)

FI-TB 190 Top plate buckled 33 F1-TB-a 365 Bottom plate delaminated 63 F2-TB-a 414 Bottom plate delaminated 69 F2-T-a 285 Top plate buckled and tube buckled 89

FR-TB-a 387 Top plate buckled 57 FR-T 162 Top plate buckled 34 FR-a 212 Tube buckled 100

Table (4): Strength & Stiffness Properties of Plies

Property Plate Layers Tube Layers Longitudinal Modulus (GPa) 31.8195 32.6505 Transverse Modulus (GPa) 8.191 8.3841 Shear Modulus (GPa) 3.244 3.3207 Poisson's Ratio 0.25344 0.2529 Longitudinal Tensile Strength (MPa) 703.0 724.2 Longitudinal Compressive Strength (MPa) 639.54 639.1 Transverse Tensile Strength (MPa) 50.06 49.56 Transverse Compressive Strength (MPa) 158.9 157.34 Shear Strength (MPa) 32.55 32.22

Table (5)' Verification Models CASE MAX. MAX. AXIAL STRAIN (ml.s)

STUDY LOAD(kN) DEFLECTION (mm) TOP BOTTOM Predicted Experim. Predicted Experim. Predicted Experim.

A ~~/ 414 70.45 68.7 -4.9 -4.7 9.9 9.7

B vlY 285 86.46 95.0 -3.8 -4.2 6.4 7.0

C '\,ff~/ 212 96.9 100 -5.2 -4.4 10.5 16.0

Table (6): Effect of Plates and Bars CASE STUDY MAX. DEFLECTION (mm) MAX. AXIAL STRAIN (ml.s)

Predicted Experimental Top Bottom

1 ~y 22.50 18.8 -2.4 2.80

2 Vlv 41.30 38.7 -3.4 6.80

3 \l~/ 53.23 58.0 -3.8 5.85

4 VDv 25.38 ----- -2.5 3.00

5 ~ 23.63 ---- -2.1 2.80

6 ~

22.70 ---- -2.3 2.70

7 ~~/ 30.23 ---- -2.8 3.20

8 ~~/ 51.18 ---- -3.5 3.50

GFRP Plate

GFRP

Triangular GFRPTube

GFRP Bar for Filler

l3.0

192.5

l3.0

300

Fig. (1) : Cross Section Of Deck (Dimensions In Mm)

Fig.(2): Test Set-Up

J\N

FRP BRIDGE DECK (TOP & BOTTOM PLATES INTACT)

Fig.(3) Load & Boundary Conditions For The Bridge Deck

K

t r,z \f.-s

L

Y,v

J-x~ J

z,w

Fig.(4) Shell Element 99

100

96.9 95.0

86.4

(mm)

--------------------~---------~----------

-------------------- ---

8.8%

70.4 • _____ _ 3% 68.7

CASE A CASEB

Fig. (5) Predicted & Experimental Deflections

AXIAL STRAIN (ml.s)

5.2

4.9 4.7

4.4 4.2

3.8

""'"'F"'---------------- ---~-------- -----

- ---~-- -------

CASE A CASEB

Fig. (6a) Predicted & Experimental Top Axial Strain

• EXPERIMENTAL

• PREDICTED F.E.A

• EXPERIMENTAL

• PREDICTED F.E.A

AXIAL STRAIN (ml.s)

16.0

10.5 9.9

9.7

7.0

6.4

CASE A

• EXPERIMENTAL

• PREDICTED F.E.A

CASEB

Fig. (6b) Predicted & Experimental Bottom Axial Strain

FRP BRIDGE DECK (TOP & BOTTOM PLATES INTACT)

ANSYS 5.5.3 OCT 3 2000 16:55:07 NODAL SOLUTION STEP-1 SUB -1 TIME-1 UY (AVG) RSYS-O Power-Gr-aphics EFACET-1 AYRES-Mat DMX -70.32 SMN --70.32 _ -70.32 _ -62.507 _ -54.693 liliiii -46.88 _ -39.067 o:;;;;:J -31.253 c:::J -23.44 _ -15.627 _ -7.813

o

Fig.(7) Deflection (Uy) Of Case "A" At Ultimate Load

FRP BRIDGE DECK (TOP & BOTTOM PLATES REMOVED

ANSYS 5.5.3 OCT 18 2000 15:26:23 DISPLACEMENT STEP-1 SUB -1 TIME=1 PowerGraphics EFACET-1 AYRES-Mat DMX -96.969

"'DSCA-1 XV -1 YV -1 ZV -1

"'DIST-232.959 "XF --270.426 "'YF --142.213 "'ZF =1227

Z-BUFFER

Fig.(8) Deflection (Uy) For Case "c" At Ultimate Load

FRP BRIDGE DECK TOP & BOT. PLATES SERVICE

ANSYS 5.5.3 OCT 17 2000 14:49:46 NODAL SOLUTION STEP-1 SUB -1 TIME-1 UY (AVG) RSYS-O PowerGraphics EFACET-1 AVRES-Mat DMX -62.962 SMI'! --62.962 SMX -.223589 _ -62.962 _ -55.941 lIB! -48.921 lIB! -41.9 _ -34.88 [:::::!I -27.859 C=:J -20.838

&II =~~7~~8 - .223589

Fig.(9) Deflection (Uy) For Case "3" At Service Load

FRP BRIDGE DECK ( AT SERVICE LOAD )

ANSYS 5.5.3 OCT 17 2000 13:16:27 NODAL SOLUTION STEP-1 SUB -1 TIME-l UY (AVG) RSYS-O PowerGraphics EFACET-l AVRES-Mat DM}{ -22.502 SMN --22.502 ----liliiii F':!VI CJ IiilIIll -

-22.502 -20.002 -17.502 -15.002 -12.501 -10.001 -7.501 -5.001 -2.5 o

Fig.(10) Deflection (Uy) For Case "I" At Service Load.

FRP BRIDGE DECK (TOP & BOT. PLATES

ANSYS 5.5.3 OCT 17 2000 14:41:04 NODAL SOLUTION STEP-1 SUB -1 TIME-1 EPTOZ (AVG) RSYS-O PowerGraphics EFACET-1 AVRES-Mat OM){ -62.962 SMN --.018288 SM){ -.010498 -----I, 'FAR] c::::J --

-.018288 -.01509 -.011891 -.008693 -.005494 -.002296 .903E-03 .004101 .0013 .010498

Fig.(11) Predicted Strain (Ez) For Case "3" At Service Load

FRP BRIDGE DECK TOP BARS & TOP &

ANSYS 5.5.3 OCT 17 2000 14:28:06 NODAL SOLUTION STEP-1 SUB -1 TIME-1 EPTOZ (AVG) RSYS-O PowerGraphics EFACET-1 AVRES=Mat DMK -23.63 SMN --.00966 SMK =.00536 _ -.00966 _ -.007992 _ -.006323 _ -.004654 _ -.002985 I "(it! -.001316 c:::J .353E-03 ..:1 .002022 _ .003691

.00536

Fig.(12) Predicted Strain (Ez) For Case "5" At Service Load

FRP BRIDGE DECK & B OTl'OM PLATES

ANSYS 5.5.3 OCT 5 2000 11:00:43 AVG ELEMENT STEP=l

SOLUT

SUB =5 TIME=2.5 FCMX TOP

(AVG)

DMX SMX -----1""'W"i'l. ~

r:::J --

=76.019 =152.326 o 16.925 33.85 50.775 67.7 84.626 101.551 118.476 135.401 152.326

Fig.(13) Failed Regions Predicted By Tsai-Wu Failure Criteria