ii

Demonstration of Cost-effective Technology to Repair Cracked Girders- A Case Study of I-565 Bridge

By

Dr. Nasim Uddin and Mr. Muhammad Shohel Department of Civil and Environmental Engineering

And

Dr. Uday Vaidya and Mr. J.C Serrano-Perez

Department of Materials Science & Engineering The University of Alabama at Birmingham

Birmingham, AL 35294

Prepared by

UTCA University Transportation Center for Alabama The University of Alabama, The University of Alabama at Birmingham,

and, The University of Alabama in Huntsville

UTCA Report 03405 June 15, 2004

ii

Technical Report Documentation Page 1. Report No.

2. Government Accession No.

3. Recipient's Catalog No. 5. Report Date June 2004

4. Title and Subtitle Demonstration of Cost-effective Technology to Repair Cracked Girders- A Case Study of I-565 Bridge 6. Performing Organization Code

7. Author(s) Dr. Nasim Uddin and Muhammad Shohel Dr. Uday Vaidya and J.C Serrano-Perez

8. Performing Organization Report No. UTCA Final Report 03405

10. Work Unit No. (TRAIS)

9. Performing Organization Name and Address Department of Civil and Environmental Engineering and Department of Materials Science and Engineering The University of Alabama at Birmingham Birmingham, AL 35294

11. Contract or Grant No. ALDOT Research Project 930-549

13. Type of Report and Period Covered Final Report September 2002-June 2004

12. Sponsoring Agency Name and Address Alabama Department of Transportation, 1409 Coliseum Boulevard P.O. Box 303050 Montgomery, Alabama 36130-3050 14. Sponsoring Agency Code

15. Supplementary Notes 16. Abstract High quality and expedient processing repair methods are necessary to enhance the service life of bridge structures. Deterioration of concrete can occur as a result of structural cracks, corrosion of reinforcement, and freeze-thaw cycles. Most infrastructure-related applications of fiber-reinforced plastics use traditional hand lay-up technology. This process is tedious, expensive, labor-intensive, and sensitive to personnel skill level. An alternative to hand lay-up is Vacuum Assisted Resin Transfer Molding (VARTM). This newly introduced technique for civil infrastructure, VARTM, uses single sided molding technology to infuse resin over fabrics while wrapping large structures such as bridge girder spans, wall panels, and columns. There is limited understanding of the problems and the benefits developed when VARTM processing is adapted to wrap fibers such as carbon and/or glass over concrete structures. For example a weak laminate-concrete interface could create problems such as water seepage, fiber wrinkling, premature failure caused by debonding at the interface, and moisture-induced delamination. This project was conducted to improve understanding of the composite-concrete interface, the feasibility of VARTM over a large reinforced concrete structure, and the issues encountered in the practical implementation of VARTM. The laboratory studies performed during this project made possible the implementation of this technology in a full size T-bulb girder of I-565 in Huntsville, Alabama, which is the first of its kind ( to the researcher’s knowledge) to have been retrofitted using VARTM. Unidirectional carbon fiber Sikawrap HEX 103C and low viscosity epoxy resin Sikadur 300 were used in the processing of the concrete specimens. 17. Key Word VARTM, repair, rehabilitation, CFRP, RC bridge girder

18. Distribution Statement

19. Security Classify. (of this report)

20. Security Classify. (of this page)

21. No. of Pages 19

22. Price

iii

Contents

Contents………………………………………………………………………...……………...…iii Figures………………………………………………………………………...……………..…...iv Tables………………………………………………………………………………………...…...vi Executive Summary…………………………………………………………………………..….vii 1.0 Introduction……………………….………………………………………………..……….…1 2.0 Methodology……………………………………………………………………………..…....2 3.0 Project Outline and Background……………………………………………………………....5 4.0 Experimental program………………………………………………………………….……..7 4.1 Pull-off Test………………………………………………………………………………7 4.2 In-Plane Shear Test……………………………………………………………………….8 4.3 Large Beam Test and Instrumentation……………………………………………………9 4.3.1 Experimental results and Analysis………………………………………………….9 5.0 VARTM in I-565 Highway bridge girder………………….…….………………...…….…..11 5.1 Design calculation……………………………………………………………………….11 5.2 Processing Detail………………………………………………………………………..11 6.0 Cost Evaluation for implementation of VARTM……………………………………………14 6.1 Costing of VARTM process……………………………………………………………..14 6.2 Costing of Hand lay-up process………………………………………………………….14 7.0 Conclusions……………………….….…………………………………………………...….16 8.0 References…………………………….……………………………………………………...17

iv

List of Figures

Number Page

2-1 VARTM schematic in rectangular concrete beam 3

3-1 I-565 Highway Bridge in Huntsville, Alabama 5

3-2 Cross section of T-bulb girder showing dimensions in cm. Image of cracked girders treated with polymeric foam injection and external steel reinforcement

6

4-1 Results from pull-off test 8

4-2 Experimental and theoretical results of in- plane shear test 8

4-3 Load vs mid-deflection curve of FVVTM and FV-control beam 9

4-4 Load vs mid-deflection curve of FPVTM beam before and after VARTM strengthening

10

5-1 Release film, distribution mesh placement over placed fabric 12

5-2 Schematic of infusion and vacuum lines placement in T-bulb girder (not in scale)

12

5-3 Infusion progression and flow front throughout flange and web. Straight line indicates flow front. Dark areas indicate wet area; light grey represents dry areas

13

5-4 Final overall view of reinforced section 13

6-1 Initial project cost for repair of 6.4 m2 of prestressed concrete girder using two different processing approaches

15

v

List of Tables

Number Page

4-1 Test matrix small beam (1st phase) 7

4-2 Large beam specimens (2nd phase) 7

4-3 Load, deflection and failure mode of tested beam 9

5-1 Number of CFRP layers calculation 11

6- 1 Variable cost for filed implementation of VARTM process 14 6- 2 Fixed cost for field implementation of VARTM 14 6-3 Tooling cost for hand lay-up process 15

vi

Executive Summary High quality and expedient processing repair methods are necessary to enhance the service life of bridge structures. Deterioration of concrete can occur as a result of structural cracks, corrosion of reinforcement, and freeze-thaw cycles. Most infrastructure-related applications of fiber-reinforced plastics use traditional hand lay-up technology. This process is tedious, expensive, labor-intensive, and sensitive to personnel skill level. An alternative to hand lay-up is Vacuum Assisted Resin Transfer Molding (VARTM). This newly introduced technique for civil infrastructure, VARTM uses single sided molding technology to infuse resin over fabrics while wrapping large structures such as bridge girder spans, wall panels, and columns. There is limited understanding of the problems and the benefits developed when VARTM processing is adapted to wrap fibers such as carbon and/or glass over concrete structures. For example weak laminate-concrete interface may create problems such as water seepage, fiber wrinkling, premature failure caused by debonding at the interface, and moisture-induced delamination. This project was conducted to improve the understanding of the composite-concrete interface, the feasibility of VARTM over a large reinforced concrete structure, and the issues encountered in the practical implementation of the processing route. The laboratory studies performed during this project made possible the implementation of this technology in a full size T-bulb girder of I-565 in Huntsville, Alabama, which is the first of its kind (to the researcher’s knowledge) to have been retrofitted using VARTM. Unidirectional carbon fiber Sikawrap HEX 103C and low viscosity epoxy resin Sikadur 300 were used in the processing of the concrete specimens.

1

1.0 Introduction

The Federal Highway Administration (FHWA) estimates that a significant portion of bridges in the USA, including Alabama, are structurally deficient or functionally obsolete. High-quality and expedient repair methods are necessary for enhancing the service life of bridge structures, and simultaneously making them less vulnerable to vandalism and fire. Reinforced concrete bridge girders are the superstructure elements between the abutments and can suffer from surface deterioration problems. Deterioration of concrete can occur as a result of structural cracks, corrosion of reinforcement, and freeze-thaw degradation. Cost-effective methods with potential for field implementation are required to address the issue of repair and strengthening of bridge structures. Numerous researchers have studied the behavior of FRP- strengthened concrete beams and departments of transportation (DOTs) across the USA have implemented FRP strengthening techniques by the hand lay up process. They include Mayo et al. (1999) who applied bonded FRP-laminates to strengthen and lift load restriction from a simple span, reinforced concrete slab bridge, Bridge G270, in Missouri for the Missouri DOT. Kachlekev et al. (2000) upgraded the capacity of the historic Horsetail Falls Bridge for the Oregon DOT. Shahrooz et al. (2001) examined four 76-year old T-reinforced beams and retrofitted a 45-year old, three span reinforced concrete slab bridge for the Ohio DOT. Besides these, there have been seismic upgrading of bridge columns for I-57 in Illinois, I-580 in Reno, Nevada and I-80 in Salt Lake City, Utah. This project presents a bridge strengthening case that was completed recently for an I-565 bridge girder by newly introduced Vacuum Assited Resin Transfer Technology (VARTM) with a cost effective method. VARTM is an emerging manufacturing technique that holds promise as an affordable alternative to traditional autoclave molding and automated fiber placement for producing large-scale structural parts. It is a process currently widely used in automotive, boat and aerospace applications. It is ideal for large structures and suitable for field implementation. The use of a vacuum makes the resin application more uniform than with traditional hand lay-up, and has also proven to be environmentally friendly. This new processing technique for bridge strengthening uses single-sided tooling, which results in significant savings in tooling costs, and vacuum-bag technology. This process has very attractive cost-advantages in comparison to hand lay-up impregnation methods. Other advantages of VARTM are low process volatile emission (and therefore environmentally friendly/safer), high fiber to resin ratio, good quality and process repeatability. As a result of savings in time, material and labor, a significant cost reduction is achieved as compared to traditional hand lay up processes. In the next section, details of the application of VARTM in concrete have been described.

2

2.0 Methodology VARTM is a composite manufacturing process that involves: (1) the lay-up and vacuum bagging of dry reinforcing fibers in fabric, tape, or bulk form as a prefrom in a one-sided open mold (2) impregnation of the perform with liquid resin using negative pressure (i.e., vacuum), and (3) curing and demolding. Recent advanced technology demonstrators such as the advanced enclosed mast sensor (AEM/S) system and the composite advanced vehicle (CAV) have shown the potential of VARTM technology for the low cost fabrication of large-scale structures requiring thick-section construction and hybrid multifunctional integral armor. The VARTM process is also used extensively in commercial applications such as bridge decks, rail cars and yachts. As an alternative to the labor intensive hand lay-up process, VARTM has the potential to be an attractive process for bridge structures since it can save processing time (especially when many fiber reinforced polymer (FRP) layers are being applied), can make the resin application more uniform than with traditional hand lay-up, and has also proven to be environmentally friendly since it has no direct volatile organic compounds (VOC) emissions. The shortcomings of traditional lay-up techniques are: a) the need of skilled workers to handle the fabric, b) improper wetting, and c) difficult to control in long bridge girders and columns. On the other hand, VARTM is believed to form a strong uniform interface, free of resin-rich areas, without compromising the integrity of the fabric. For the VARTM on concrete application, the surface of the concrete acts as the tool for the infusion process. After the surface of the concrete is prepared, a layer of primer is applied in the area to be infused. A vacuum seal is formed around the beam or column, using vacuum bag sealant tape to ensure an airtight seal. The dry fabric, initially cut to shape, is then placed on the concrete surface and held in place using adhesive tape on the edges. Then, the stretched fabric is covered using a porous media and a distribution mesh, to allow resin flow through the fabric. The infusion and vacuum lines are then placed, and finally covered by a bagging film. The lay-up is de-bulked; the resin/hardener mixture is infused with the aid of a vacuum. VARTM forms a strong uniform interface, since higher consolidation pressure produces lower resin void content and less resin-rich areas. The feasibility of using VARTM was demonstrated on an 1828.8 mm (6 ft) long Reinforced Concrete (RC) beam, as illustrated in Figure 2-1. The resin flowed across the fabric, and filled any surface flaw by the aid of the hydrostatic pressure that was applied by the pressure on the bag. The infusion parameters were optimized so that the infusion was symmetrical and even, and resin rich areas were avoided. Curing took place under vacuum. The main issues in the VARTM process are resin flow and maintaining a vacuum until the resin is cured. Flow of resin is mainly affected by the permeability of the porous media, number of fiber layers and orientation of the fiber, resin viscosity and surface roughness of concrete. The resin flows across the fabric and follows Darcy’s law given in Equation 2-1.

Infusion Line

3

PkJ ∆η

−=

where J is the volume current density, k is the permeability of the porous media, η is the viscosity of the fluid, and P is the pressure that drives the fluid flow (Chawla 1998). The flow fills any surface flaw by pressure applied by the atmosphere on the bag. To fully implement the resin injection technology, one needs to understand the basic principle and the possible variations. When injecting a homogeneous material with a resin, the filling time of the injection of a rectangular strip can be calculated (Brouwer et al. 2003) using the modified Darcy’s law given by

Pk

ltimefill∆

=2

_2ϕη

where ϕ is the porosity of the reinforcement, k is the permeability of the reinforcement, η viscosity of the resin, l flow distance, and ∆P pressure difference. These parameters can be optimized as follows:

• Resin viscosity: It is usually limited to 100-500 mPa s. The filling time is directly proportional to the viscosity; low viscosity resins are more desirable for low injection times, the viscosity of the Sikadur 300 is 300mPa s. • Reinforcement porosity and permeability: The porosity of most common reinforcements is between 0.5 and 0.85. The permeability of the reinforcements can vary greatly and it is a factor that is optimized in this particular case by the use of distribution mesh as explained in this section. • Pressure difference: Large structures can only be manufactured effectively using vacuum pressure; the maximum pressure difference will be 1 bar; and • Length of the part: The maximum length to be infused by an infusion line should be 30 cm.

To have an approximation of the permeability of the reinforcement, the Carman-Kozney equation (Han et al. 2003) is used:

(2-2)

Figure 2-1. VARTM schematic in rectangular concrete beam

(2-1)

4

2

32 )1(16 f

ff

vv

Kd

k−

=

where fd is fiber diameter, fv is volume fraction, and K is the Kozeny constant. Using the Kozeny constant of T-300 carbon of K = 6.9 (Han et al. 2003) and T-300 fiber diameter of d = 7µm, and solving for the experimentally obtained volume fraction, k in the reinforcement is found to be 2.21e-13 m2. This value is very small, would require very long infusion times, and the resin would cure before filling the mold. To improve the permeability, a distribution mesh (made from Polyethylene Terephthalate (PET), weight 50 g/m2) along with a release fabric (porous bleeder release film made of tightly woven silicon coated polyester, weight 74 g/m2) were used in the implementation of Sikadur 300/Sikawrap systems as shown in Figure 2-1. There are also four essential steps involved in the VARTM process to achieve good strengthening: wetting/impregnation, lay-up, consolidation, and solidification. The desired area of the fiber can be wetted to allow the minimum fill time given in Equation 2-2 and to make sure the resin flows entirely around the fiber. To ensure this, the location of the inlet/infusion lines and the number and location of vacuum lines in a designated area are optimized as shown in Figure 2-1.

(2-3)

5

3.0 Project Outline and Background

Figure 3-1 is a typical I-565 bridge in Huntsville, Alabama. In this project we consider this to be a first study demonstration of repairing and simultaneously hardening/strengthening an existing bridge structure. The structure was an I-565 highway bridge, pre-stressed concrete girder, in Huntsville. Two bridge structures (BIN 015820 and BIN 015821) of the I-565 bridge showed significant cracking following construction during 1988-1990. The number of girders with the cracks was approximately 50. Most of them were simply supported. However a diaphragm was used at the support to connect the girders and to provide continuity of the girders. These diaphragms probably restricted the movement of the girder during thermal expansion and created a fixed end action at the support. It is suspected that all the cracking of these 50 girders was due to this end restriction.

The bridge selected for demonstration of the VARTM strengthening technology was on Route/Bin 52 in Madison County, Alabama. It was a 27.13 m (89 ft) long solid pre-stressed concrete girder, cast in situ slab, bridge built in 1991. After visiting the site of some cracked and deteriorated girders, the authors and the Alabama Department of Transportation (ALDOT) selected this bridge girder for evaluation.

The cracks started in the bottom of the flange as flexure cracks (vertical) and continued to propagate into the web in a 45° angle (typical of shear cracks). The typical cracking behavior and geometry of all the girders are illustrated in Figures 3-2. The first approach utilized by ALDOT was to inject the cracks using polymeric foam. This approach helped to control moisture- related damage to exposed steel reinforcement; however, it did not stop crack growth, and was finally discarded. To control cracking, a steel structure was built and put in place as a temporary support.

Figure 3-1. I-565 bridge in Huntsville, Alabama

Figure 3-1: I-565 bridge in Huntsville, Alabama

6

Bottom Flange

Web

2.0 m

Cracks

Overall view of girder and support

10 cm

Web Figure 3-2. Cross section of T-bulb girder showing dimensions in cm. Image of cracked girders treated with polymeric foam injection and external steel reinforcement

160.

02

7

4.0 Experimental program A characterization study was undertaken at the University of Alabama at Birmingham laboratory to investigate the strength of existing girders in comparison to the VARTM repaired/strengthened samples. For screening and identification of potential solution, two types of concrete specimens were fabricated. One was a small prism 101.6 mm x 76.2 mm x 406.4 mm (4″×3″×16″) and the other was a 152.4 mm x 304.8 mm x 1828.8 mm (6″×12″×72″) reinforced concrete beam. Small prisms were used to investigate the bond strength and interface properties between the concrete and carbon fiber reinforced polymers (CFRP). The large specimens were used to investigate the load carrying capacity while reinforced with the same CFRP and epoxy resin system. The work was done according to the test matrix, Table 4-1 and Table 4-2. Detailed results are presented at Uddin et al. (2004).

4.1 Pull off Test

Surface strength of concrete was studied by pull off tests. Sandblasting treatment was found to be an effective means to transfer higher loads through bond. Figure 4-1 shows a summary of the pull off test results. From the graph, it can be seen that the sandblasting improves the result from the pull off test, and that exaggerated roughness (as in rough/epoxy sample) does not necessarily contribute to better bond. Therefore, it was decided that the second phase of the test would be conducted solely with the sand blasting surface treatment.

Treatment No. of Prisms in In-plane

No. of Prisms in Pull-off test Prism size, mm (inch)

As cast 3 3 101.6 × 76.2 × 406.4 (4 ×3×16)

Sand blasting 3 3 101.6 × 76.2 × 406.4 (4 ×3×16) wire brushing rough 3 3 101.6 × 76.2 × 406.4 (4 ×3×16)

Beam No Beam Description Test procedure (Four-point Bending)

and Failure mode Beam Dimensions, mm

(inch)

1 Virgin (FV-Control) Full failure in flexure

Partial failure in flexure

152.4 × 304.8 × 1828.8 (6 × 1 2 × 72)

2 Virgin + Tensile face Reinforced (FVVTM) Full failure in flexure 152.4 × 304.8 × 1828.8

(6 × 12 × 72)

3 Partial failure Repair with CFRP (FPVTM) Full failure in flexure 152.4 × 304.8 × 1828.8

(6 × 12 × 72)

Table 4-1: Test matrix small prism (1st phase)

Table 4-2: Large beam specimens (2nd phase)

8

4.2 In-Plane Shear Test

The interface developed using VARTM between the concrete substrate and the carbon fabric was studied based on the model similar to the test apparatus of in-plane shear used by Bizindavyi and Neale (1999). The assembly is supported in such a way that there is direct shear at the composite-to-concrete interface. This test set up is an adequate way to determine the bond length as well as the general properties and behavior of the formed interface under the chosen conditions of surface preparation and processing. In this test a uni-axial tension was applied on the fabric bonded to the concrete, so that the interface was subjected to direct shear. It is known that the application of a laminate to a concrete beam introduces shear transfer to the concrete/epoxy interface.

The results for ultimate load on both the VARTM specimen and the hand lay-up specimen show that the interface developed using VARTM provides improvement. Moreover, the interface in VARTM is more uniform, and the failure modes encountered in the interface formed by VARTM show better bonding and a cohesive failure in contrast to the failure mode encountered in the hand lay-up (Uddin et al 2004). Experimental results of the hand laid-up and VARTM specimens also agree with experimental derivations of Hiroyuki and Wu (1997) and Tanaka (1996). The analytical models developed by Neuerbauer and Rostasy (1998) and Volnyy and Pantelides (2000) show higher bond strengths than those observed in testing.

Figure 4-1. Results from pull-off test

Figure 4-2. Experimental and theoretical results in-plane shear test

9

4.3 Large Beam Test and Instrumentation

In the second phase of the experiment beams 1, 2 and 3 were called FV-Control, FVVTM and FPVTM beam respectively. The first letter (F) indicates the way the tested beams failed, here flexure, the second letter (V or P) indicates the beam condition, virgin and pre-cracked, and VTM indicates the VARTM strengthening technique. In Table 4-2, reinforcement characteristics and surface preparation of the three beams are shown.

For instrumentation a linear variable displacement transducer (LVDT) was placed under the mid span of the beam to measure the deflection. Two strain gauges were attached beneath the point load and one was attached at the mid span of the beam. All three strain gages were attached at the tension face of the beam. The strain gauges and the LVDT were connected to a data acquisition system (MEGADAC) which recorded data every one second. The load was applied through a load cell which was placed in the middle on the top of the set up. 4.3.1 Experimental Results and Analysis These tests were done to study the behavior of VARTM-strengthened RC beams under flexural conditions. Typical load versus deflection curves for three types of beam have been shown in Figures 4-3 and 4-4. Table 4-3 shows the load, deflection and failure mode of the three tested beams.

Figure 4-3 shows the load deflection curve of the FV-control and FVVTM beams. The FVVTM beam showed higher ultimate load before failure, compared to the non-strengthened beam (Figure 4-3). Figure 4-4 shows the load deflection curve of a pre-cracked beam. From the figure it is seen that the ultimate load was increased significantly after the VARTM strengthening. The tested RC beams were designed for flexure failure. For the FVVTM beam both flexure and shear reinforcement were used. Debonding occurred at the end location of the flexural CFRP plate. For the FPVTM beam no external reinforcement was provided for shear, which is why external shear failure was observed along with concrete crushing for this case.

Beam Failure Load, kN (kip)

Deflection, mm (inch)

Failure Mode

FV-Control 129.4 (29.4) 12.45 (0.49) Flexure Failure FVVTM 187.54 ( 43.32) 11.71 (0.46) Out of plane De bonding FPVTM 185.54 (43.35) 6.15 (0.24) Shear

Table 4-3. Load, deflection and failure mode of Tested beam

10

Figure 4-3. Load vs mid deflection curve of FVVTM and FV-Control beam

Figure 4-4. Load vs mid deflection curve of FPVTM beam before and after VARTM strengthening

11

5.0 VARTM in the I-565 Bridge Girder

5.1 CFRP Design Calculation

For calculation it was assumed that after development of the flexural cracks, the concrete section was fully cracked and the concrete could no longer carry tensile force. To retrofit such a section, it was also assumed that external reinforcement carries all of the nominal capacity of the section. First, the location of the critical section was calculated which is equal to hc+254 mm (hc+10″), where hc is the distance between the centroid of the pre-stressing strands and top of the girder. At this critical section both the pre-stressing force and temperature variation were considered for calculating the required number of CFRP layers. All calculations for required CFRP have been shown Table 5-1. In order to avoid shear failure, shear reinforcement was calculated based on 292.6 kN (65.7 kips) shear force developed at the critical section due to a HS 20 truck. Based on the design guide ACI 440, this gave two layers of CFRP for shear strengthening of that section. The flexural reinforcement were placed along with the shear reinforcement as 0/90/0/90 (0° for flexural and 90° for shear reinforcement) sequence.

5.2 Processing detail

After structural calculations for strengthening, two flexure layers and two shear layers of unidirectional carbon fiber, and epoxy resin were selected for retrofitting the T-bulb girder. A crew of four conducted the material processing on site for a period of two days on a 1.524 m × 4 m area. The first step in the process was surface preparation of the girder using a sandblaster and compressed air. After the surface was cleaned and treated, a coat of resin was applied in the border area to ensure an adequate seal between the concrete and the vacuum sealant tape. After the layer of primer cured, vacuum sealant tape was placed on the surface of the concrete following the pattern of the part to be bagged. The lay up sequence of [0°/90°/0°/90°, i.e. flexural fabric in the longitudinal and shear fabric in the perpendicular direction] was adopted and the flexure plies were laid first so the shear plies (laid up to 0.2 m above the neutral axis to ensure anchoring effects) would help contain the edge peel stresses that usually initiate debonding of

At Critical section Area/ No. of CFRP layer

Nominal Moment Capacity

Moment due to Temperature at

mid span Moment due to

Temperature and Pre-stressing effect

Shear Force for HS 20 Truck

Moment /Shear (5347.5 kip-ft) (741.8 kip-ft) 2024.9 (kip-ft) (65.8 Kips)

Required area of CFRP Af mm2 (in2)

11548.4 (17.9)

1871 (2.9)

4839 (7.5)

890.3 (1.38)

No. of layers of CFRP 3.4 0.47 1.42 1.72

Remarks (CFRP need) 4 1 2 2

Table 5-1. Calculation of number of CFRP layers

12

flexure plies. All plies were held in place against the concrete surface with the aid of 3M-spray adhesive. Once all the plies were put in place, a porous bleeder release film made of tightly woven silicone coated polyester was placed on top of the lay up so that the distribution mesh would not bond to the fabric (Figure 5-1).

Then the distribution mesh, made from PET (Polyethylene Terephthalate), was placed with the infusion and vacuum lines on top of the plies and on the upper part of the bottom flange. Following the distribution mesh and the inlet and outlet channels, the bag was put in place on top of the whole arrangement. Debulking took place for twenty minutes under vacuum, and the bag was straightened and checked for leaks.

The infusion strategy was to place an infusion channel at the middle of the bottom flange, two infusion/vacuum channels at the web/bottom flange interface and two final vacuum channels at the top of the part as seen in Figure 5-2.

Finally infusion was performed and monitored for 90 minutes; the flow front was very homogeneous and symmetrical to each side of the girder showing a linear pattern. The vacuum was continued for 12 hours to ensure proper curing under 635 mm Hg vacuum pressure (see Figure 5-3).

Figure 5-1. Release film, distribution mesh placement over placed fabric

Figure 5-2. Schematic of infusion and vacuum lines placement in T-bulb girder (not in scale)

13

The on-site filling time for the girder was approximately 60 minutes. After debagging, the part was found to have conformed readily to the shape of the girder. A coat of latex paint was applied to the surface of the finished part to act as a seal against any environmental attack such as moisture, dust or particles. A completed view is shown in Figure 5-4.

Figure 5-3. Infusion progression and flow front throughout flange and web. The straight line indicates the flow front. Dark areas indicate wet area, light grey represents dry areas

Figure 5-4. Final overall view of reinforced section

14

6.0 Cost Evaluation for field implementation of VARTM

A generalized analysis was performed for the main factors influencing the economics of the process, and the total price for the initial stage of the project was estimated for both VARTM and hand lay-up techniques. For manufacturing operations, certain costs appear to be independent of the particular production practice or the business. Only the parameters relevant to repair of concrete structures were studied here. 6.1 Cost of VARTM Process Table 6-1 shows the variable cost for the VARTM process and Table 6-2 shows the fixed cost. The total cost for the repair operation of 6.4 m2 using VARTM was $9,797. 6.2 Cost of Hand Lay-up Process Less equipment is needed for hand lay-up when compared to other processes. It is the same as VARTM but without the rental of the vacuum pump. A worker for hand lay-up should be more skilled than a worker for the VARTM process since part of the quality depends on the precision

Description of Variable Cost Quantity/

No. of Workers

$/Day or

$/hour

Days used

Hours used

Cost ($)

Vacuum pump 1 50 3 - 150 Portable generator 1 55 4 - 220

Man lift 1 195 4 - 780 Sand blaster 1 50 1 - 50

Machinery

Portable compressor 1 60 1 - 60 Direct Labor Construction worker 4 38 - 24 3648

Total 4908

Description of Fixed Cost Quantity/Type Unit price ($) Total price ($)

Sikadur 300 Epoxy resin 3 (Drum-5 gal) 194.67 584.01 Sikawrap Hex 103C carbon fabric 3 (Roll-15 m) 55 220 Materials Coating latex based paint 2 (2 gal) 195 780 Beelder lease G Roll-91 m 698.00 69.80 Vacuum bag sealant tape GS-95 dark gray Case-40 rolls 95.02 24.00

Distribution mesh Roll-200m 200.00 20.00 Vacuum bagging film DPT 1000, 0.003in gauge Roll 100m 1,216.29 21.621

Tooling

PVC tubing Roll-30 m $17.19 $34.38

Table 6-1. Variable cost for field implementation of VARTM process

Table 6-2. Fixed cost for field implementation of VARTM

15

of the fabric placement, proper wet-out of the part, and rolling in place. A wage of $45 per hour was used for this estimate based on the work of Daniels and Grogan (2002). So the total direct labor cost would be $4320. The fixed cost and raw materials cost remains the same as for VARTM. Tooling costs are minimum as shown in Table 6-3. The total cost of using hand lay-up for the repair of 6.4 m2 was $10,129. A comparison of the costs of the two processes is shown in Figure 6-1. 6.2 Discussion Hand lay-up and VARTM are very similar in terms of costing since the high tooling cost in VARTM is equivalent in the hand lay-up process by the increased labor rates and longer processing times. However, the final product advantages of VARTM over hand lay-up discussed early make it an attractive and an effective processing route for the repair of large reinforced concrete structures.

Tooling Description Quantity

Unit price ($) Total price ($)

Aluminum Rollers 4 15 60

Rubber squeegees 4 5 20

Table 6-3. Tooling cost for hand lay-up process

Figure 6-1. Initial project cost for repair of 6.4 m2 of prestressed concrete girder using two different processing approaches

$0$2,000$4,000$6,000$8,000

$10,000$12,000

Variablecost

Directlabour

Material Tooling Totalcost

Project Cost

VARTM Hand Lay Up

16

7.0 Conclusions These investigations addressed an emerging liquid molding technique referred to as VARTM. It is a promising low cost innovation to traditional composites manufacturing methods. It uses single-sided tooling (translating to significant savings in tooling costs) and vacuum-bag technology. Laboratory experiments were performed to optimize the processing parameters and material types (fabric and resin) to be used in retrofitting existing structures. An in-plane shear bonding test showed that VARTM can generate a strong interface between composite laminate and the conditioned surface of the concrete, indicating the feasibility of VARTM for bonding composite laminate to concrete surfaces. VARTM was then implemented on a laboratory scale for the repair and upgrade of reinforced concrete rectangular beams, showing a considerable increase in capacity as compared to the unreinforced section, over a wider variety of failure mechanisms. Finally this newly-introduced technique was successfully implemented in the field for the repair of a pre-cracked reinforced concrete T-bulb I-565 bridge girder in Huntsville, Alabama. The field work was completed within two days without any traffic interruption.

From the research and field work conducted in this study, the following conclusions were drawn:

1. Surface preparation studies revealed the need for sandblasting so that VARTM could generate a strong interface between the composite laminate and the concrete surface.

2. Bond testing, mainly the interface shear test, showed that VARTM can generate a strong and similar interface between the composite laminate and the conditioned surface of the concrete. These results confirmed the feasibility of VARTM for bonding composite laminates to concrete surfaces.

3. VARTM was implemented on a laboratory scale for the repair and upgrade of reinforced concrete rectangular beams, showed a considerable increase in capacity as compared to the unreinforced section, for a number of failure mechanisms.

4. VARTM was successfully implemented for the repair of a precracked reinforced concrete T-bulb bridge girder on an I-565 bridge.

5. A simple economical analysis showed that VARTM is a competitive means for reinforcing large structure. The overall cost for this project was estimated and compared with traditional methods. Initial project costs were very similar when compared to hand lay-up techniques. However, life cycle expectancy and maintenance costs were not taken into account and can ultimately give a competitive advantage to VARTM over hand lay-up and other conventional repair methods including steel plate bonding. The completed work was inspected by ALDOT bridge engineers, who were pleased with the quality of the work. They extended the funding to continue research on developing design guidelines and detailed specifications for field implementation of the VARTM process.

17

8.0 References American Concrete Institute (2000), “Guide for the design and construction of externally bonded FRP system for strengthening concrete structures,” ACI committee 440. Bizindavyi, L and and Neale, K. W. (1999), “Transfer lengths and bond strengths for composites bonded to concrete,” Journal of Composites for Construction, ASCE, Vol. 3, No.4. pp. 153- 160. Brouwer W.D., Van Herpet E.C.F.C and Labordus M., (2003), “Vacuum injection moulding for large structural applications,” Composite Part A: applied science and manufacturing 34(2003)551-558. Kachlekev, D., Green, B.K., Barnes, W. (2000), “Behavior of Concrete Specimens Reinforced with Composite materials-Laboratory Study,” Report No. FHWA-OR-RD-00-10, Oregon Department of Transportation. Haffner S.M. (2002), “Cost modeling and design for manufacturing guidelines for advanced composite fabrication,” MIT dissertation. Cambridge, Mass, 120-146. Hiroyuki, Y. and Wu, Z (1997) ,”Analysis of debonding fracture properties of CFS strengthened member subject to tension,” Non-Metallic (FRP) Reinforcement for concrete structures, proceedings of the Third International Symposium, Sapporo, Japan, pp287-294. Mayo, R., Nanni, A., Gold, W., and Baker, M (1999). “Strengthening of Bridge G270 With Externally Bonded Carbon Fiber Reinforced Polymer Reinforcement,” SP-188, American Concrete Institute Proceedings of the Forth International Symposium on FRP for Concrete Structures (FRPRCS4). Baltimore, MD, pp.429-440. Neuerbauer, U and Rostasy, F.S.( 1997) ”Design aspects of concrete structure strengthened with externally bonded CFRP plates,” Proceedings of the seventh International Conference on Structural Faults and Repairs, edited by M.C. Forde, Engineering technics Press, Edinburgh, UK, pp.109-118. Shahrooz, M. Bahram and Boy Serpil (2001), “Retrofitting of Existing Reinforced Concrete Bridges with Fiber Reinforced Polymer Composites,” Ohio Department of Transportation and the U.S. Department transportation, Federal Highway Administration, Report No. UC- CII 01/01. Serrano-Perez, J., C. (2003). “Implementation of Vacuum Assisted Resin Transfer Molding in the Repair of Reinforced Concrete Structures,” Masters of Science Thesis, University of Alabama at Birmingham, USA.

18

Tanaka, T (1996), “Shear Resisting Mechanism of Reinforced Concrete beams with CFS as shear Reinforcement,” Graduation Thesis, Hokkaido University, Japan. Uddin N., Vaidya U.K, Shohel, M. and Serrano-Perez, J.C. (2003), “ALDOT research project # 930-549,” Alabama Department of Transportation. Uddin N., Vaidya U.K, Shohel, M. and Serrano-Perez, J.C (2004), “In Plane Debond Response of Vacuum Assisted Resin Transfer Molded Carbon Fiber Sheets on Concrete Substrates,” Cement & Concrete Composites, (in print), Uddin, N., U. Vaidya, M. Shohel, and S.C. Serrano-Perez, “Cost-effective strengthening of

bridge girder using vacuum assisted resin transfer molding”, J. Advanced Composite Materials, Vol.13, No. 3-4, 2004:pp: 214-235.

Vaidya, U.K., Abraham, A., Mohammed, M., “Manufacturing of Composite Integral Armor for Ballistic Application,” 29th International SAMPE Technical Series, pp. 523-531, Green, Beckwith and Strong, Editors, 1997. Volnyy, Vladimir A. and Pantelides, Chris P. (1999), “Bond Length of CFRP Composites Attached to Precast Concrete Walls,” Journal of Composite for Construction, Vol. 3, No.4, November. pp. 168-176