Design and Construction of a Bridge Deck using Mild and Post-Tensioned FRP Bars

Transcription

1 SP Design and Construction o a Bridge Deck using Mild and Post-Tensioned FRP Bars by R. Fico, N. Galati, A. Prota, and A. Nanni Synopsis: The type o structure selected or this research project is a our-span continuous concrete slab having GFRP bars or top and bottom mats and CFRP reinorcement or internal post-tensioning. The bridge is located in Rolla, Missouri (Southview Drive on Carter Creek). One lane o the bridge was already built using a conventional our-cell steel reinorced concrete box culvert. One lane and sidewalk needed to be added and were constructed using FRP bars. This study includes design o the FRP portion o the bridge using existing codes, validation o the FRP technology through a pre-construction investigation conducted on two specimens representing a deck strip, and construction o the bridge; the results showed how FRP, in the orm o GFRP as passive and CFRP bars as active internal reinorcement, could be a easible solution replacing the steel reinorcement o concrete slab bridges, and speciically the enhanced shear capacity o the slab due to the CFRP prestressing. Keywords: anchoring systems; bridge deck; FRP tendons; posttension 1121

2 1122 Fico et al. Raaello Fico is a Ph.D student at the University o Naples Federico II, Italy. He perormed his undergraduate thesis in collaboration with the University o Missouri-Rolla on the same topic o this paper. He was the supervisor o the installation o Southview bridge deck. Nestore Galati is a research engineer at the University o Missouri Rolla. He obtained is PhD in Civil Engineering at the University o Lecce-Italy where he also received his B.Sc. in Materials Engineering. He obtained his M.Sc. degree in Engineering Mechanics at the University o Missouri-Rolla. His research has concentrated on the area o retroitting o masonry and upgrade and in-situ load testing o reinorced concrete structures. He is an EIT in the United States. Andrea Prota is Assistant Proessor o Structural Engineering at University o Naples Federico II, Italy. He is member o ib WG 9.3 and Associate Member o ACI 440 Committee. His research interests include seismic behavior o RC and masonry structures, use o advanced materials or new construction and or retroitting o existing structures, and use o innovative techniques or structural health monitoring. Antonio Nanni FACI, is the V & M Jones Proessor o Civil Engineering at the University o Missouri Rolla, Rolla, MO and Proessor o Structural Engineering at the University o Naples, Italy. He was the ounding Chair o ACI Committee 440, Fiber Reinorced Polymer Reinorcement, and is the Chair o ACI Committee 437, Strength Evaluation o Existing Concrete Structures. INTRODUCTION Reinorced and prestressed concrete (RC and PC) structures are acing a worldwide problem, which is the corrosion o the steel as a result o aging and aggressive environments. Steel corrosion leads to member degradation, endangers structural integrity and may even cause catastrophic ailures. Research has been carried out in an eort to ind the solution or this problem. Fiber Reinorced Polymers (FRP), have been proposed or use in lieu o steel or reinorcement and prestressing tendons in concrete structures. The promise o FRP materials lies in their high-strength, lightweight and noncorrosive, non-conducting, and non-magnetic properties. The interest in the use o FRP composites in prestressed concrete is mainly based on durability issues. Corrosion o prestressing steel tendons can cause serious deterioration o inrastructure. Properties like high tensile strength and high resistance to corrosion make FRP composites good candidates or prestressed and post-tensioned tendons. Moreover, unlike steel reinorced concrete sections, members reinorced with FRP bars have relatively small stiness ater cracking. Thereore, serviceability requirements have also been examined.

3 RESEARCH OBJECTIVES FRPRCS The objectives o this research project were as ollows: 1. Evaluate the easibility, behavior and eectiveness o an innovative deck system, showing how glass and carbon FRP (GFRP and CFRP), in the orm o passive and active internal reinorcement, could be a solution replacing steel reinorcement. 2. Provide analytical data in support o the enhanced shear capacity o the concrete slab due to the CFRP post-tensioning. DESCRIPTION OF THE PROJECT The City o Rolla, MO has made available a bridge (Southview Drive on Carter Creek) to demonstrate the use o FRP bars and tendons in new constructions. One lane o the bridge was already constructed using conventional our-cell steel RC box culvert. It consists o a steel RC slab about 0.25 m thick, as depicted in Figure 1. The slab deck is continuous over three intermediate RC vertical walls, and the overall length o the bridge is roughly 12 m. The new deck was built on three conventional RC walls as or the existing structure. The construction o the FRP reinorced slab, plus a 2 m wide conventional RC sidewalk on the opposite side, allowed extending the overall width o the bridge rom 3.9 m to 11.9 m, as Figure 2 shows. The construction o the bridge started on July 2004 and inished on October THEORETICAL ANALYSIS The analysis and design o the FRP concrete bridge deck was based on the ollowing assumptions: Nominal properties or FRP reinorcing material taken rom the manuacturer published data and considered as initial guaranteed values to be urther reduced to take into account the environmental reduction actors as given in ACI 440.1R- 03; Load conigurations consistent with AASHTO Speciications; Design carried out according to ACI 440.1R-03; and Eects due to the skew neglected. The design properties o concrete and FRP bars used are summarized in Table 1. Summary o the structural analysis Bending moments and shear orces are summarized in Table 2. Columns (1) and (3) represent the actors to be applied to dead and live load, respectively. Columns (2) and (4) show unactored moment and shear, due to dead and live load, respectively. Flexural design The lexural design o a FRP RC member is similar to the design o a steel RC member. The main dierence is that both concrete crushing and FRP rupture are potential mechanisms o ailure.

4 1124 Fico et al. The ultimate lexural capacity o the bridge deck was calculated according to ACI 440.1R-03 using the ollowing equation: β1c β1c Mn = A ε E d + Apε pep dp 2 2 (1) where A and A p is the area o GFRP and CFRP reinorcement, respectively; ε and ε p their respective strains; E and E p the respective moduli o elasticity; d and d p represents the distance rom extreme compression iber to the centroid o the tension GFRP and CFRP reinorcement, respectively. β 1 c is the depth o the equivalent rectangular stress distribution in concrete. Assuming the post-tensioned tendons strained at 65% o their ultimate capacity and a 35% o total losses, the theoretical moment capacity is ound to be 64 kn-m. Since the ailure mode is controlled by the rupture o the GFRP bars, the post-tensioning does not inluence the moment capacity. Shear design The concrete contribution to the shear capacity, V c,, o the bridge deck was calculated according to ACI 440.1R-03 as ollows: A E Vc, = Vc (2) AE s s where V c represents the nominal shear strength provided by the concrete and the ratio (A E /A s E s ) takes into account the axial stiness o the FRP reinorcement as compared to that o steel reinorcement. A s reported in Eq. (2) can be ound by determining the area o steel reinorcement required to match the actored FRP lexural capacity, φm n. The concrete contribution to the shear capacity, V c, was determined according to ACI318-99, sec. 11.4, which prescribes or post-tensioned members the ollowing equation: Vd ' u p Vc = (0.05 c ) bwdp (3) M u Ultimate shear and moment laid out in eq.(3) are taken at a cross-section on the central support where both negative bending moment and shear orce reach their maximum values, as summarized in Table 2. The actored shear capacity is φ Vn =φ Vc, = 162kN / m which is higher than the shear demand V u =160.5 kn/m. I the post-tensioning action was not considered, the shear capacity would have been φ Vn =φ Vc, = 95kN / m which is much smaller than the demand (the shear capacities have been computed rom an analysis perormed on the cross section lush with the vertical wall representing the support). The post-tensioning allowed an increase o shear capacity o more than 70%.

5 FRPRCS As an alternative approach to the shear capacity o the bridge deck, the ollowing equation (Tureyen and Frosh 2003), now under consideration or adoption by ACI Committee 440, could be used: ` 2 c bc V = (4) c, 5 where c is the position o the neutral axis at service. Using such approach, the shear capacity would be kn/m. Serviceability Crack width, long-term delection, creep rupture and atigue have been also computed according to ACI 440. The slab was partially post-tensioned, thus permitting some tensile stresses at ull service load. Crack width o FRP reinorced lexural members can be expressed as suggested by ACI 440 as ollows: 2200 w = k 3 b dca E β (5) where β is the ratio o the distance rom the neutral axis to extreme tension iber to the distance rom the neutral axis to the center o tensile reinorcement, k b is a bond dependant coeicient equal to 1.2, represents the stress at service in the FRP, d c is the thickness o the concrete cover measured rom extreme tension iber to the center o the bar, A is the eective tension area o concrete deined as the area o concrete having the same centroid as that o tensile reinorcement, divided by the number o bars. The crack width o the slab, calculated with eq.(5), yields w= mm, much smaller than the allowed value suggested by ACI 440 and equal to 0.5 mm. The long-term delection can be calculated as suggested by ACI 440 as ollows: = LL + λ( DL LL ) (6) where λ = 1.2 represents the multiplier or additional long-term delection. Eq. (6) yields to = 1.3mm, smaller than the suggested AASHTO value o 3.8 mm. To avoid creep rupture o the FRP reinorcement under sustained loads, the stress level in the FRP bar should be limited to the value suggested in ACI 440.1R.03. Speciically, when GFRP reinorcement is used, the stress limit has been set to be equal to 0.20 = 87 MPa. The stress at service in the FRP is: = Eε = 16MPa (7) This value is much smaller than the allowable stress and thereore creep rupture and atigue are not an issue or the long term behavior o the bridge.

6 1126 Fico et al. PRE-CONSTRUCTION INVESTIGATION Two identical specimens representing a deck strip 457 mm wide, 254 mm deep and 7 m long, were abricated and tested as continuous slabs over three supports (Galati et al., 2004). One specimen was used to investigate the lexural behavior ( lexural specimen), while the other one the shear behavior ( shear specimen ). The testing allowed validating the design calculations both in terms o lexure and shear capacities. The specimens had the same reinorcement percentage o the bridge deck: they were reinorced using 3 φ19 GFPR bars as top and bottom mat and 2 φ9 CFRP bars as posttensioning tendons. The position o the tendons was varied along the specimen in order to match the moment demand. In addition, in order to reproduce the actual ield conditions, also φ13 GFRP bars spaced 305 mm on center were placed in the transverse direction Figure 3 shows a detailed layout o the specimens reinorcement, while Figure 4 shows the position o the CFRP tendons. Test Setup Each slab was tested as a continuous member on three supports, comprising o a 3.6 m and 1.8 m long span. The positions o the two loading points were chosen such to orce lexural and shear ailure or the lexural and shear specimens respectively. They were placed at the mid-spans or the lexural specimen (See Figure 5-a)), and 0.9 m away rom the central support in the case o the shear specimen (See Figure 5-b)). The lexural specimen was instrumented using three load cells: two o them were placed under the loading points while the third one was placed under one o the supports in order to determine the end reaction and thereore the real distribution o moments in the slab. Loads were applied to 102 x 457 x 25 mm steel plates resisting on the slab. The loads were generated by means o 30 ton hydraulic jacks reacting against a steel rame. The loading rate was the same or the two spans until reaching 85 kn. Ater that, the load in the shorter span was kept constant while the one in the longer one was increased until ailure. This solution was adopted in order to avoid shear ailure at the central support. Linear Variable Displacement Transducers (LVDTs) were positioned at the loading points (two or each loading point) and at the supports in order to record maximum displacements and support settlements. A total o 21 Electrical Strain Gages (ESG) were used to monitor the strain at the most critical sections. They were placed on each GFRP bar in correspondence with loading points and the central support. In addition, at the same locations, an additional strain gage was attached on the compressive ace o the slab in order to have an additional backup point while determining the experimental momentcurvature response o the slab. For the shear specimen, the number o load cells was reduced to two. Loads were applied to 102 x 457 x 25 mm steel plates resisting on the slab by means o a 100 ton hydraulic jack. Two additional LVDTs were inserted in order to also measure the maximum displacement which, in this case, was not at the loading points.

7 FRPRCS The load was applied in cycles o loading and unloading. An initial cycle or a low load was perormed on every specimen to veriy that both the mechanical and electronic equipment was working properly. Test results and discussion For the lexure specimen, the irst lexural crack was observed on the longer span when the load was approximately 66 kn on both spans. As the loads were increased, some o the cracks started to extend diagonally to orm shear cracks. At maximum loads, 163 kn and 100 kn or long and short spans respectively, concrete crushing at the top was observed. This was immediately ollowed by a sudden shear ailure which also caused the rupture o the CFRP bars due to kinking (see Figure 6-a)). The maximum bending moment was equal to 102 kn-m (when considering C E =1, φ=1), which compared to the theoretical value o 98 kn-m corresponds to an increase o about 4%. For the shear specimen, the irst crack was observed at a load approximately o 89 kn at the central support where the moment was maximum. As the load was increased, the newly ormed cracks between the central support and the loading point started to extend diagonally to orm a shear crack. The ailure o the specimen occurred or an applied load equal to 273 kn due to diagonal tension shear (see Figure 6-b)). The shear specimen presented a shear capacity equal to 142 kn, larger than the theoretical capacity ound using the ACI 440.1R-03 (70 kn, with post-tensioning), with a percentage increase equal to 103%, and to 46% when considering the Tureyen and Frosh approach, conirming the conservative approach o ACI 440 or the shear. SOUTHVIEW BRIDGE DECK CONSTRUCTION The slab is 0.25 m thick, 12 m long and 6 m wide. It is supported by three intermediate reinorced concrete vertical walls. The our span lengths are, rom North to South, 2.89 m 3.38 m, 3.30 m and 2.35, respectively, on centers. Substructure construction The erection o the substructure and the extension o the existing abutments and walls were perormed irst. GFRP bars were inserted into the central wall in order to create a connection between the slab and the substructure (see Figure 7). Slab construction Ater laying the slab ormwork, 13 mm thick neoprene pads were placed on the two abutments and on the two walls that did not have the GFRP anchoring rebars, in order to avoid restraining horizontally the slab and thereore to eectively post-tension it.hence the bottom and the top layers o GFRP mild reinorcement were placed (see Figure 8).

8 1128 Fico et al. A wooden board was built to make the slab end perpendicular to the tendons. This operation was perormed in order to ease the post-tensioning phase (Figure 9). This was ollowed by the placing o plastic ducts to house the CFRP tendons. Such ducts were tied to the GFRP rebars; thirteen additional ducts were also placed (in case o imperect posttensioning), straight on the top side o the slab, as shown in Figure 10. T connectors were used on each end o ducts in order to allow the injection o the grout ater the pulling o the tendons. Strain gages were also attached on the longitudinal GFRP bars in order to monitor the slab during the service lie. Figure 11 shows the bridge deck ater pouring. A week ater the curing o the slab, the CFRP tendons were post-tensioned. The pulling was achieved by means o the jack already used or the specimens. Figure 12 shows the pulling device. It comprises an open steel box having enough room to push the wedges inside the chuck ater pulling the tendon, a hydraulic jack to apply the pulling orce, a round steel plate, a load cell to measure the load, a second plate and an outer chuck. The steel wedge anchorage system used to lock in the CFRP bar and to react against the hydraulic jack was a three part system developed at the University o Waterloo, Canada. It included an outer steel cylinder, a our-piece wedge and an inner sleeve (see Figure 13). The inner sleeve was made out o copper to be deormable. The our-piece wedge was placed evenly around the inner sleeve and inserted into the outer steel cylinder. The anchorage system was later secured by tapping the inner sleeve and ourpiece wedge into the outer steel cylinder with a special hammer. The tendons were pulled at both terminations by means o two hydraulic jacks, connected to two pumps using load steps o kn per side. The use o a single pump with two jacks connected in series was also considered, but it was not selected because it did not allow to ensure uniorm load applications on the tendons. The applied load was measured using a data acquisition system connected to a computer allowing to monitor the applied load in real time. Ater reaching the desired load the wedges were pushed inside the inner chuck, so each jack was released, engaging in such way the inner chucks. At this point FRP tendon was cut with an electric saw (see Figure 14). The grout injection ollowed, ater sealing the chucks to avoid grout leaking. Ater 4 days o curing the inner chucks were removed. Figure 15 shows a picture o the bridge ater completion. Finally, a barrier with GFRP reinorcement will be built on the new FRP side, in order to have a comparison during time with the steel reinorced concrete barrier on the opposite side, which is also going to be built. It will contribute to show the increased durability o a bridge deck using FRP materials as reinorcement, mainly due to the absence o corrosion.

9 CONCLUSIONS FRPRCS The ollowing conclusions can be drawn: The post-tensioning allowed increasing the shear capacity o the slab by more than 70%. Moreover serviceability requirements are ully satisied. The shear specimen presented a shear capacity much larger than the theoretical computed with the ACI 440.1R-03, conirming the very conservative approach o the guideline. The Tureyen and Frosh approach on shear appears to be less conservative than ACI 440.1R-03. Utilizing FRP in the orm o reinorcing bars allows or the use o many steel- RC concrete practices. The abrication and installation details were nearly identical to the methods regularly utilized or steel-reinorced slabs. The installation o the bridge highlighted the act that having an eicient system is as important as having the adequate components. As well, or a new technology its learning curve must be overcome beore its applications can be conducted proiciently. ACKNOWLEDGMENTS The authors would like to acknowledge: the City o Rolla providing this opportunity and Dave Brown and Bill Cochran or their assistance in the dierent parts o the project; the inancial support o the Missouri Department o Transportation, the University Transportation Center (UTC) at the University o Missouri Rolla (UMR) or the research component o the project; Hughes Brothers or providing the FRP reinorcement; Jason Cox o Master Contractors LLC or reinorcement placement and post-tensioning. REFERENCES Reerenced standards and reports The standards and reports listed below were the editions used or this project. AASHTO 1994 Manual or Condition Evaluation o Bridges 1996 Design Code or Highway Bridges 2002 Standard Speciications or Highway Bridges, 17th Edition American Concrete Institute Building Code Requirement or Structural Concrete and Commentary 440.1R-03 Guide or the Design and Construction o Concrete Reinorced with FRP Bars

10 1130 Fico et al. Cited reerences Galati N, G. Boschetto, A. Rizzo, R. Parretti and A. Nanni, 2004: Pre-Construction Investigation or the Rehabilitation o a Bridge Using Internal FRP Technologies, 4th International Conerence on Advanced Composite Materials in Bridges and Structures - ACMBS-IV, July 20-23, 2004, Calgary, Alberta, Canada, CD-ROM version, 12 Hughes Brothers Inc, October 2003 Hughes Brothers Reinorcements Product Guide Speciication, Web site: Sika Corporation, July 2003, SikaGrout 300 PT Guide Speciication, Web site: Tureyen A. K. and Frosh R. J., 2003: Concrete Shear Strength: Another Perspective, ACI Structural Journal, VOL. 100, NO.5 pp LIST OF NOTATIONS A = the eective tension area o concrete, deined as the area o concrete having the same centroid as that o tensile reinorcement, divided by the number o bars A = area o FRP reinorcement A p = area o post-tensioned FRP reinorcement A s = area o steel reinorcement b = width o rectangular cross section b w = width o the web c = distance rom extreme compression iber to the neutral axis C = environmental reduction actor or various iber type and exposure conditions E d = distance rom extreme compression iber to centroid o tension reinorcement d c = thickness o the concrete cover measured rom extreme tension iber to center o bar or wire location closest thereto d p = distance rom extreme compression iber to centroid o post-tensioned reinorcement deined as the mean modulus o a sample o test specimens E = guaranteed modulus o elasticity o FRP deined as the mean modulus o a sample o test specimens E = guaranteed modulus o elasticity o post-tensioned FRP p E s = modulus o elasticity o steel ' c = speciied compressive strength o concrete = stress in the FRP reinorcement in tension k b = bond-dependent coeicient M = nominal moment capacity n

11 FRPRCS M u = actored moment at section, V = nominal shear strength provided by concrete with steel lexural reinorcement c V = nominal shear strength provided by concrete with FRP lexural reinorcement c, V n = nominal shear strength at section V u = actored shear orce at section w = crack width β = actor depending on the concrete strength 1 = long term delection = delection due to the dead load DL = delection due to the live load LL ε = strain in FRP reinorcement ε p = strain in post-tensioned FRP reinorcement λ = multiplier or additional long-term delection φ = strength reduction actor

12 1132 Fico et al. Figure 1 Views o the Former Bridge Figure 2 Cross Section o the Bridge Ater the Expansion Figure 3 Specimen Reinorcement Layout Figure 4 Longitudinal Section o the Slab or One o the Central Spans

13 Figure 5 Test Setups FRPRCS

14 1134 Fico et al. Figure 6 Flexure and Shear Specimens Failures Figure 7 Placement o the GFRP Rebars into the Central Wall

15 FRPRCS Figure 8 Neoprene Pads and GFRP Mild Reinorcement Details Figure 9 Wooden Board Figure 10 Slab Section B-B

16 1136 Fico et al. Figure 11 Poured FRP Bridge Deck Figure 12 New Pulling Machine Ready or the Use Figure 13 Steel Wedge Anchorage System

17 FRPRCS Figure 14 Wedges Insertion and Cutting o the Tendon Figure 15 Southview Bridge-Deck Completed

18 1138 Fico et al.