flexural behavior of corroded pretensioned girders

129PCI Journal | Spr ing 2014

Bridge infrastructure in North America is aging, with more than 40% of the bridges in Canada and the United States over 50 years old1 and in need

of significant maintenance, rehabilitation, or replacement. Many prestressed concrete bridges are subjected to cor-rosive environments. One of the major factors for this is the limited corrosion resistance in old structures and the continually increasing use of deicing salts in cold regions.2 When considering rehabilitation options, it is important to account for the differences between prestressed concrete structures and reinforced concrete structures. The conven-tional approach of chipping and patching is not always viable for pretensioned concrete structures because of the risk involved in exposing a stressed, heavily corroded strand. Prestressed strands are often designed as the main flexural reinforcement, making their corrosion an extreme-ly critical deterioration of the structure that would require immediate intervention.

Prestressing steels are highly susceptible to corrosion, and this can significantly affect the service life of prestressed concrete structures. Although corrosion-induced deteriora-tion of reinforced concrete structures has been observed throughout concrete’s history, it was not until the 1980s that such deterioration was observed in prestressed con-crete structures.3 Two corrosion-induced failures were the collapse of the Saint Stefano Bridge in Italy in 19904 and

■ This paper assesses the effect of corrosion of seven-wire steel strands on the residual capacity of pretensioned concrete T beams subjected to static flexural loading and the effective-ness of carbon-fiber-reinforced polymer (CFRP) repair.

■ The experimental variables were mass loss due to corrosion (0%, 2.5%, 5%, and 10%) and repair condition (unrepaired and repaired beams using adhesively bonded CFRP sheets).

■ Test results showed reductions in flexure capacity of up to 76% and, at midspan, in deflection up to 26% at 10% mass loss.

■ CFRP repair was able to restore the load capacity to that of the control beams; however, the reduction in ductility due to corro-sion was not reversible.

Flexural behavior of corroded pretensioned girders repaired with CFRP sheets

Adham El Menoufy and Khaled Soudki

Spr ing 2014 | PCI Journal130

localized strand rupture. The load test results showed a reduction in the static capacity of 55% to 65%.

Numerous researchers have investigated the use of fiber-reinforced polymers (FRPs) in repairing corroded steel reinforced concrete beams and have shown good results. Shahawy et al.8 tested 16 full-scale T beams with varied repair schemes to investigate the static and fatigue per-formance of reinforced concrete beams strengthened with carbon-fiber-reinforced polymer (CFRP) laminates. Beams were underreinforced 6 m (20 ft) long T beams; 10 beams were tested under static loading, while the remaining 6 were tested under fatigue loading. Results showed that CFRP strengthening increased the static load capacity by up to 70% with four layers of CFRP laminate, while ductility decreased with the increased number of lay-ers. In addition, results indicated that fatigue-critical beams could be effectively rehabilitated using CFRP laminates. Soudki et al.9 investigated the behavior of CFRP-strength-ened reinforced concrete beams in a corrosive environ-ment. The investigation comprised constructing and testing 11 reinforced concrete beams. Beams were subjected to 300 cycles of wetting and drying in a 3% NaCl solution and repaired using CFRP sheets. Test results showed that CFRP strengthening significantly enhanced the perfor-mance of reinforced concrete beams, nearly doubling the load capacity of unstrengthened beams.

Other researchers have investigated rehabilitation tech-niques such as externally prestressed CFRP sheets and rods to restore strength loss due to impact.10,11 These inves-tigations have yielded promising results. To the authors’ knowledge, limited research is documented in the literature on the behavior of corroded pretensioned concrete mem-bers. The study presented in this paper will address this gap in the literature by investigating the flexural behavior of corroded pretensioned girders strengthened with CFRP sheets.

Experimental program

The experimental program consists of ancillary and full-scale beam testing. The ancillary tests were conducted to investigate the rate of accelerated corrosion on the pre-stressing strands. This was achieved by embedding nonpre-stressed seven-wire strand in prisms (100 × 150 × 300 mm [4 × 6 × 12 in.]) constructed from 30 MPa (4400 psi) concrete mixed with a NaCl solution at a 2.1% chloride concentration by mass of cement. The strands had a 50 mm (2 in.) cover. Nine prisms were constructed and corroded to 2.5%, 5%, and 10% mass loss using a galvanostatic accelerated corrosion approach. A stainless steel hollow tube with 8 mm (0.3 in.) diameter, acting as the cathode, was embedded in the concrete parallel to the prestress-ing strand, which acted as the anode. A current density of 2 µA/mm2 (1300 µA/in2) was impressed to form an artificial corrosion cell as recommended in the literature.12

the collapse of the pedestrian bridge over Lowe’s Motor Speedway in North Carolina in 2000.5

The consequence of corrosion in prestressed concrete ele-ments is far more serious than in reinforced concrete due to the high mechanical stresses applied to the strands.6 Typi-cally, prestressing strands are stressed to between 60% and 75% of their ultimate strength, so any loss of cross section due to corrosion creates stress raisers in the strand. Critical stress raisers are a function of mass loss, geometry (depth, width, and radius), and location of the pits. The prestress-ing steel becomes highly susceptible to local yield and/or fracture with the reduction in cross section.3

Darmawan and Stewart3 conducted a study to quantify the effects of pitting corrosion on the capacity of prestressing steel strands and found that the corroded wire failed at the location of the deepest pit, reducing the tensile capacity of the strand by approximately 40%. The authors used the data to formulate a probabilistic model to predict the maxi-mum pit depth along any length of wire at any time during its service life. They used the model to simulate a real-life corrosion scenario at a current density of 0.01 µA/mm2 (6.5 µA/in.2) for 20 years. They found that the mean loss of cross section can exceed 30%, and that there is a 5% chance of reduction in cross section up to 45%. After 10 years of corrosion there is a 15% chance of at least a 20% reduction in cross-sectional area. These results highlight the effect of pitting corrosion on the struc-tural integrity of prestressing strand.

Ngoc et al.7 conducted a study on 8 mm (0.31 in.) pre-stressing wires exposed to accelerated corrosion. The main variable considered in the study was the stress (0%, 70%, 80%, and 100% of the 1500 MPa elastic limit). Their findings showed that highly stressed wires can exhibit an additional 15% mass loss versus nonprestressed wires. However, stress appears to have no effect on the composi-tion of the rust products.

Rinaldi et al.6 conducted an experimental investigation to study the flexural behavior of prestressed concrete beams with prestressing strand corroded to 7%, 14%, and 20% mass loss. The beams were partially prestressed with three seven-wire stands (12.7 mm [0.5 in.] in diameter) stressed to 66% of their ultimate stress. The concrete compressive strengths were 34, 42, and 47.4 MPa (4900, 6100, and 6900 psi). An aggressive induced current of 400 mA was used to accelerate corrosion while ponding the beams with a 5% NaCl solution. Results revealed that corrosion of the prestressing strand significantly affected the behavior of the prestressed member in terms of load-bearing capacity, ductility, and failure mode. The control beams failed by concrete crushing in the overly reinforced section, while beams that were corroded to 7% mass loss exhibited a combined failure of localized rupture and concrete crush-ing. Beams corroded to 14% and 20% mass loss failed by


The beams were designed in accordance with the Cana-dian Precast/Prestressed Concrete Institute (CPCI) Design Manual13 and satisfied all stress requirements at transfer. Beams were constructed using 40 MPa (5800 psi) concrete and had the following dimensions: flange width of 400 mm (16 in.) with 50 mm (2 in.) thickness and a web height of 250 mm (9.8 in.) and thickness of 100 mm (4 in.). The beams were prestressed using a single seven-wire low-relaxation strand with a nominal diameter of 12.7 mm (0.5 in.) and nominal tensile capacity of 1860 MPa (270 ksi). The prestressing strands had a 60 mm (2.4 in.) concrete cover and were prestressed to 70% of the ultimate

The beam tests comprised testing six pretensioned full-scale T beams to quantify the effect of corrosion of the seven-wire steel strands on their static capacity. One beam was kept as control (that is, not corroded). Six beams were subjected to accelerated corrosion to achieve low 2.5% (2 beams), medium 5% (2 beams), and high 10% (2 beams) mass loss. Two of the six corroded beams were repaired using CFRP sheets following corrosion to examine the feasibility of CFRP repair to restore the capacity of cor-roded pretensioned beams. All beams were tested statically in four-point bending to failure. Table 1 summarizes the experimental program.

Figure 1. Beam cross-section details. Note: All measurements are in millimeters. 10M = no. 3; 1 mm = 0.394 in.; 1 kN = 0.225 kip; 1 MPa = 145 psi.

300

Elevation

50A

1000 Salted concrete in flexure zone

One 12.7mm strand

Section A

250

150

One 12.7 mm diameter 1860 MPa low relaxation standJacking force = 128.9 kN/strand

Four D8-02

300

5025

0

Five D8 bars

10M stirup

10060

150

400

60

3600

Table 1. Testing program

Phase Specimen typeNumber of specimens

Mass loss, % Testing type

Ancillary tests Prism (100 × 150 × 300 mm)

3 2.5

Mass loss analysis3 5.0

3 10.0

Full-scale T beams

Unstrengthened

1 0.0

Static (flexural)

1 2.5

1 5.0

1 10.0

Strengthened1 5.0

1 10.0

Note: 1 mm = 0.0394 in.; 1 MPa = 145 psi.


current density of 2 μA/mm2 [1290 μA/in.2]. This was achieved by connecting the steel strands as anodes and the stainless steel tubes as cathodes to the power supplies. The target corrosion values were 2.5%, 5%, and 10% mass loss.

Upon completion of accelerated corrosion, all beams were taken out of the corrosion chamber. Two corroded beams were repaired with CFRP sheets. Prior to static testing, concrete cylinders were tested for compressive strength. The CFRP repair configuration was designed in accor-dance with ACI 440.2R (2008)14 to restore any reduction in capacity (Fig. 3). According to the manufacturer, the cured carbon-fiber-reinforced sheets and epoxy had tensile strength of 894 MPa (130 ksi), modulus of elasticity of 65,087 MPa (9440 ksi), tensile elongation of 1.33%, and thickness of 0.382 mm (0.015 in.). A flexural CFRP sheet 100 mm (4 in.) wide was applied to the soffit of the beams and ran along the full clear span. In addition, two CFRP U-wraps were applied at the location of the loading points to provide confinement to the concrete along the corroded zone, while two additional CFRP U wraps were placed to-ward the beam ends to prevent delamination of the flexure sheet. The fiber orientation for the flexure sheet was in the longitudinal direction (parallel to the prestressing strand), while for the U wraps it was in the transverse direction. The CFRP sheets were placed in accordance with the manufacturer’s specifications and left at room temperature for seven days to allow the epoxy to fully cure prior to load testing.

The beams were tested in four-point bending in a four-post testing frame. The load was applied by a servo-hydraulic actuator of 280 kN (63 kip) capacity with a stroke of 250 mm (10 in.). The static loading was applied in dis-placement control at a rate of 1 mm/min (0.04 in/min). The distance between loading points was 600 mm (24 in.), and the clear beam span was 3400 mm (134 in.). This gives a shear span/depth ratio of 4.6, which is typical for pre-stressed concrete bridge girders. Failure during static beam tests was defined by strand rupture or concrete crushing accompanied by a 20% drop in load.

Linear variable differential transducers (LVDTs) were used to measure the midspan deflection. Crack opening mea-surements within the constant moment zone were moni-tored using LVDTs mounted across three cracks (Fig. 4). The beams were loaded to initiate cracking; the first three cracks within the constant moment zone were marked. The beams were unloaded and three ½ in. (13 mm) LVDTs were placed, one across each crack. The beams were then reloaded to failure at a rate of 1 mm/min (0.039 in./min) while the LVDTs captured the crack opening measure-ment. The concrete compressive strain was measured by a 60 mm (2.4 in.) concrete strain gauge mounted onto the top concrete surface at midspan. The average tensile strain at the concrete surface was also measured using a 500 mm

tensile strength of the strands (129 kN [29 kip] or 1302 MPa [189 ksi]). Five 8 mm (0.3 in.) plain bars were placed longitudinally in the compression zone, and 8 mm bars at 150 mm (6 in.) on center were placed transversely as shrink-age and temperature reinforcement. The shear reinforce-ment consisted of single-legged, 10M (no. 3), epoxy-coated stirrups at 75 mm (3 in.) spacing placed in the shear zones to ensure flexural failure. No stirrups were placed in the con-stant moment zone. Figure 1 shows the beam cross-section details. To ensure that only the central tension region of the beams corroded, this region (1000 × 150 mm [39 × 6 in.]) was cast using salt-bearing concrete with 2.1% chloride concentration by mass of cement. In addition, a stainless steel tube (8 mm diameter) that acted as the cathode during accelerated corrosion was placed parallel to the prestressing strand within the central corrosion zone.

The prestressed beams were constructed at a precast con-crete plant to ensure consistent quality in fabrication (Fig. 2). The beams were steam cured for 24 hours, after which standard concrete cylinders (100 × 200 mm [4 × 8 in.]) were tested for compressive strength and the prestressing force was transferred at a nominal concrete strength of 26 MPa (3800 psi).

The beams were corroded in a corrosion chamber using an accelerated corrosion technique with an impressed

Figure 2. Beam along the prestressing bed during fabrication.


Figure 3. CFRP sheets repair configuration. Note: All dimensions are in millimeters. CFRP = carbon-fiber-reinforced polymers; 1 mm = 0.394 in.

3600

600

225 1501500150

A

A

CFRP U-wrap

Flexural CFRP sheet

CFRP U-wrap

Elevation

Soffit View

Figure 4. Linear variable differential transformers for crack-opening measurements at the elevation of the strand.


(20 in.) long fiber-optic sensor placed on the web at the location of the embedded prestressed strand. For beams wrapped with CFRP sheets, two 5 mm (0.2 in.) strain gaug-es were installed on the flexural CFRP sheet at midpoint in the longitudinal direction. A data acquisition system was used to acquire all measurements.

Following load testing, gravimetric mass loss measure-ments were conducted in accordance with ASTM G1 (2003)15 to determine the extent of corrosion in the strand. The corroded steel strands were extracted from the tested beams, and the corroded portion was cut into smaller samples. The samples were weighed and cleaned and the mass loss determined.

Experimental results

This section discusses the test results in terms of mass loss, corrosion cracking, and effect of corrosion and CFRP repair on the structural behavior of prestressed corroded beams. Table 2 summarizes the test results.

Mass loss results

Upon completion of the accelerated corrosion process, the seven-wire strands were extracted from the prisms, and gravimetric mass loss analysis was conducted to determine the extent of corrosion. The results of the mass loss analy-sis showed that Faraday’s law overestimates the duration required for 2.5% mass loss, reasonably predicts the dura-tion for 5% mass loss, and underestimates the duration for 10% mass loss. Based on these findings, it was decided to decrease the duration for the 2.5% beam specimens and in-crease the duration of the 10% beam specimens by 25% of the required corrosion time predicted by Faraday’s law. The 25% change in corrosion duration correlates to the differ-ence between the predicted and actual extent of corrosion observed in the ancillary tests. Figure 5 shows a corroded seven-wire prestressing strand before and after cleaning as

part of the gravimetric mass loss analysis. Table 2 gives the mass loss results for the strands in the prestressed beams.

Corrosion cracking

The width of the corrosion cracks at midspan was moni-tored and recorded during the corrosion period; Fig. 6 shows the corrosion-induced crack width measurements at midspan. Cracks were stable (no significant increase in width) up to 2.5% mass loss, exhibited a sharp increase between 2.5% and 5% mass loss, and then increased at a lesser rate between 5% and 10% mass loss.

Effect of corrosion on structural behavior

Figure 7 presents the load–versus–midspan deflection for the prestressed beams at 0%, 2.5%, 5%, and 10% mass loss, as well as calculated cracking and ultimate loads. The load deflection response of the control (uncorroded) beam was typical of a prestressed beam. The load deflec-tion exhibited a trilinear response with a distinct change in stiffness at cracking, followed by a gradual transition into a nonlinear response. The beam did not exhibit a distinct yield plateau; this is expected because the prestressing strands do not have a yield plateau. The load deflection of the control beam corresponded well to the elastic response calculated based on the effective moment inertia.13,16

The corroded beams had a response similar to that of the uncorroded beams, with an initial linear response that transitioned into a nonlinear one but without a distinct change in stiffness at the cracking load. The initial stiffness of the beams was not affected by corrosion. However, the cracking load, ultimate load, and midspan deflection of the beams decreased as the extent of corrosion increased. The ultimate load capacity and deflection of the corroded beams were compared with those of the control beam. The beam at 2.5% mass loss exhibited a 6.5% reduction in

Table 2. Beam test result summary

Mass loss due to corrosionConcrete strength,

MPa

Cracking Ultimate

Theoretical ActualLoad,

kNDeflec-

tion, mmLoad,

kNChange in

load, %Deflection,

mmReduction in deflection, %

0% n/a 48.69 ± 3.49 33.6 3.0 65.28 n/a 141.40 n/a

2.5% 2.66 ± 0.23 48.69 ± 3.49 31.4 2.9 61.00 -6.56 104.09 26.39

5% 5.87 ± 1.2 49.89 ± 4.66 24.0 2.0 58.80 -9.93 78.57 44.43

10% 9.33 ± 1.68 49.89 ± 4.66 23.0 1.8 48.30 -26.01 33.34 76.42

5% repaired 5.87 ± 1.2 48.69 ± 3.49 38.7 3.8 70.30 +7.63 42.76 69.76

10% repaired 9.33 ± 1.68 49.89 ± 4.66 31.6 2.6 63.20 -3.13 43.45 69.27

Note: n/a = not applicable. 1 mm = 0.0394 in.; 1 kN = 0.225 kip; 1 MPa = 145 psi.


600 mm [24 in.]). Using this approach, the average tensile strain in the strand at failure was calculated as 34,600 µε, 25,000 µε, 19,400 µε, and 12,700 µε for beams with 0%, 2.5%, 5%, and 10% mass loss, respectively. In addition, the average tensile strain within the constant moment zone (center 600 mm between loading points) was measured by a fiber-optic sensor and plotted versus the applied load (Fig. 10). The fiber-optic sensor was damaged during test-ing of the control beam, and no average tensile strain was recorded. The average tensile strain in the corroded beams decreased with increased corrosion. Because failure was by strand rupture, this observation correlates well with the observed decrease of the cumulative crack opening and midspan deflection at failure as corrosion progresses. The maximum average tensile strain from the fiber-optic sensor correlated well with the calculated average strain from the crack width measurements. This indicates that the fiber-optic sensor captured the response accurately.

Beam failure was defined by complete strand rupture or by a 20% drop in load. The control uncorroded beam failed by strand yielding followed by a multiple-wire rupture at one location (Fig. 11). The failure pattern of the corroded beams differed, as failure was initiated by one or two wires rupturing at one location accompanied by a drop in load.

ultimate load and a 26% reduction in midspan deflection, while the beam at 5% mass loss had 9.93% and 44% reduc-tions in ultimate load and midspan deflection, respectively. The beam at 10% mass loss had a reduction of 26% in ultimate load and a 76% reduction in ultimate midspan deflection.

The load–versus–concrete compressive strain behavior (Fig. 8) correlates well with the load–versus–midspan deflection behavior (Fig. 2). It is observed that, similar to the ultimate midspan deflection, the concrete compressive strain decreased with mass loss. Beams at 2.5% and 10% mass loss exhibited 27% and 54% reductions in concrete compressive strain, which correlates well with the reduc-tion in midspan deflection. However, the beam at 5% mass loss suffered a reduction of 20% in concrete compressive strain versus 44% in midspan deflection.

The crack opening widths of the first three flexure cracks were measured. Figure 9 presents the load–versus–cumu-lative crack opening curves. The average tensile strain in the strand can be indirectly calculated using measured cu-mulative crack opening. This can be done by dividing the cumulative crack opening by the gauge length (the outer distance between LVDTs used to measure crack width was

Figure 5. Seven-wire strand before (top) and after cleaning (bottom) showing pitting at 5% mass loss.


Figure 6. Corrosion crack width versus predicted mass loss. Note: 1 mm = 0.394 in.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

0% 2% 4% 6% 8% 10% 12%

Cor

rosi

on c

rack

wid

th, m

m

Theoratical mass loss, %

Figure 7. Load versus midspan deflection curves. Note: 1 mm = 0.394 in.; 1 kN = 0.225 kip.

0

10

20

30

40

50

60

70

80

0 20 40 60 80 100 120 140 160

Load

, kN

Midspan deflection, mm

Control 0%

2.5% not repaired

5% not repaired

10% not repaired

Theoretical elastic response

Design cracking and ultimate loads


Figure 8. Load versus midspan concrete compressive strain having different mass losses. Note: 1 kN = 0.225 kip.

0

10

20

30

40

50

60

70

80

0 500 1000 1500 2000 2500

Load

, kN

Compressive strain,

0%-not repaired

2.5%-not repaired

5%-not repaired

10%-not repaired

Figure 9. Load versus cumulative crack opening within the constant moment zone. Note: 1 mm = 0.394 in.; 1 kN = 0.225 kip.

0

10

20

30

40

50

60

70

80

0 2 4 6 8 10 12 14 16 18 20 22

Load

, kN

Total crack opening, mm

0%

2.5%

5%

10%


Figure 10. Load versus average tensile strand strain at different corrosion levels. Note: 1 kN = 0.225 kip.

0

10

20

30

40

50

60

70

80

0 10,000 20,000 30,000 40,000 50,000

Load

, kN

Strain,

2.5%-not repaired

5%-not repaired

10%-not repaired

Figure 11. Multiwire rupture at the same location for the uncorroded beam.


was restored to 3% below that of the control (uncorroded) beam, while the ductility remained roughly the same as that of the unrepaired corroded beam.

The concrete compressive strain and the CFRP tensile strain data at midspan were monitored. Figure 14 shows load–versus–concrete compressive strain curves. The concrete compressive strain decreased while the CFRP tensile strain increased with the increase in mass loss. This indicates that as corrosion progressed, the resistance of the corroded strand to the applied load decreased, and as a result the CFRP sheet contributed more in resisting the applied load. The repaired beams both failed in flexure by strand rupture followed shortly by rupture of the CFRP sheet (Fig. 15). These results illustrate that repair with CFRP sheets is a viable method to restore the static capac-ity of corroded prestressed concrete beams.

Conclusion

Based on the test results, the effects of corrosion and CFRP repair on the behavior of pretensioned concrete beams are evident. The following conclusions can be made:

• Faraday’s law overpredicts the duration of accelerated corro-sion exposure required for low mass loss values and under-estimates the duration for high mass loss for steel strands.

Shortly after, another wire ruptured at an adjacent location (Fig. 12). This is attributed to the stress concentrations at the corrosion pits at different locations along the strand.

Effect of CFRP repair

Two beams corroded to 5% and 10% mass loss were repaired with externally bonded unidirectional carbon-fiber fabric in conjunction with epoxy resin. A 2.5% mass loss did not have a significant effect on the structural behavior of the beam; therefore it was not repaired.

Figure 13 shows the load-deflection response of the CFRP repaired–versus–unrepaired beams. The load-deflection response was bilinear, with a distinct change in stiff-ness beyond the cracking load, and gradually changed to a nonlinear curve. The CFRP repaired corroded beams had higher stiffness and, consequently, higher cracking load than the unrepaired beams. CFRP repair was able to restore the reduction in load capacity due to corrosion. For the beam corroded to 5% mass loss and repaired with CFRP sheets, the load capacity exceeded that of the control (uncorroded) beam by 7.64%, while the midspan deflection was decreased by 25% compared with the unrepaired beam corroded to the same mass loss, leading to a total reduc-tion of 70% in ductility. For the beam corroded to 10% mass loss and repaired with CFRP sheets, the load capacity

Figure 12. Wires ruptured at adjacent locations in the corroded beam.


Figure 13. Load versus midspan deflection comparison of the repaired and unrepaired beams. Note: 1 mm = 0.394 in.; 1 kN = 0.225 kip.

0

10

20

30

40

50

60

70

80

0 20 40 60 80 100 120 140 160

Load

, kN

Midspan deflection, mm

Control 5% unrepaired 5% repaired 5% repair design capacity 10% not repaired 10% repaired 10% repair design capacity

Repair design capcity for 5% Repair design capcity for 10%

Figure 14. Load versus concrete and carbon-fiber-reinforced polymer midspan strains for corroded and repaired beams. Note: 1 kN = 0.225 kip.

0

10

20

30

40

50

60

70

80

-2000 -1000 0 1000 2000 3000 4000 5000 6000 7000

Load

, kN

Strain,

5% repaired-concrete strain

5% repaired-CFRP strain

10% repaired-concrete strain

10% repaired-CFRP strain


• CFRP repair was effective in restoring the load capacity of corroded pretensioned beams. Beams can be effectively restored using CFRP sheets, and the stiffness of the strengthened member increases in comparison with its counterpart. However, the reduction in ductility was not reversible.

References

1. Pakniat, P., and A. Hammad. 2008. “Optimiz-ing Bridge Decks Maintenance Strategy Based on Probabilistic Performance Prediction Using Genetic Algorithm.” In Annual Conference of the Canadian Society for Civil Engineering 2008: Partnership for Innovation, 2057–2067. Quebec City, QC, Canada: Canadian Society for Civil Engineering.

2. PCI Bridge Design Manual Steering Committee. 2011. Precast Prestressed Concrete Bridge Design Manual. MNL-133. 3rd ed. Chicago, IL: PCI.

3. Darmawan, M. S., and M. G. Stewart. 2007. “Ef-fect of Pitting Corrosion on Capacity of Prestress-ing Wires.” Magazine of Concrete Research 59 (2): 131–139.

• Corrosion significantly reduced the load capacities and midspan deflections of pretensioned T beams.

– The beam at 2.5% mass loss exhibited a 6.5% reduction in ultimate load and 26.4% reduction in midspan deflection.

– At 5% mass loss the beam suffered reductions of 44% and 10% in load capacity and midspan deflection, respectively.

– The beam with 10% mass loss had 76% and 26% reductions in load capacity and midspan deflec-tion, respectively.

• The cumulative flexure crack width and average strand tensile strains increased as corrosion progressed, while the concrete midspan compressive strain decreased.

• Failure was characterized by rupture of the strand in all beams, but the rupture pattern differed between corroded and uncorroded beams. For the uncorroded beams, rupture of multiple wires occurred at the same location, while for the corroded beams, wire failure occurred at adjacent locations depending on the pit location.

Figure 15. Typical failure of a prestressed beam repaired by carbon-fiber-reinforced polymer sheets.


4. Proverbio, E., and G. Ricciardi. 2000. “Failure of a 40 Year Old Post-tensioned Bridge Near Seaside.” In EUROCORR International Conference Proceedings 2000: Past Successes—Future Challenges. London, England: Eurocorr.

5. Cederquist, Sally Cole. 2000. “Motor Speedway Bridge Collapse Caused by Corrosion.” Materials Performance 39 (7): 18–19.

6. Rinaldi, Z., S. Imperatore, and C. Valente. 2010. “Experimental Evaluation of the Flexural Behavior of Corroded P/C Beams.” Construction and Building Materials 24 (11): 2267–2278.

7. Ngoc, A. V., A. Castel, and R. Francois. 2009. “Ef-fect of Stress Corrosion Cracking on Stress-Strain Response of Steel Wires Used in Prestressed Concrete Beams.” Corrosion Science 51 (6): 1453–1459.

8. Shahawy, M., and T. E. Beitelman. 1999. “Static and Fatigue Performance of RC Beams Strengthened with CFRP Laminates.” Journal of Structural Engineering 125 (6): 613–621.

9. Soudki, K., E. El-Salakawy, and B. Craig. 2007. “Behavior of CFRP Strengthened Reinforced Concrete Beams in Corrosive Environment.” Journal of Com-posites for Construction 11 (3): 291–298.

10. Kim, Y. J., M. F. Green, G. J. Fallis. 2008. “Repair Of Bridge Girder Damaged by Impact Loads with Pre-stressed CFRP Sheets.” Journal of Bridge Engineering 13 (1): 15–23.

11. Burningham, C. A., C. P. Pantelides, L. D. Reaveley. 2011. “Prestressed Concrete Girder Repair with CFRP Post-tensioned Rods.” ACI Special Publication 2 (275). Farmington Hills, MI: ACI (American Concrete Institute).

12. El Maaddawy, T. A., and K. A. Soudki. 2003. “Effec-tiveness of Impressed Current Technique to Simulate Corrosion of Steel Reinforcement in Concrete.” Jour-nal of Materials in Civil Engineering 15 (1): 41–47.

13. CPCI (Canadian Precast Prestressed Concrete Insti-tute). 2008. CPCI Design Manual. 4th ed. Ottawa, ON Canada: CPCI.

14. ACI Committee 440. 2008. Guide for the Design and Construction of Externally Bonded FRP Systems for Strengthening Concrete Structures. ACI 440.2R-08. Farmington Hills, MI: ACI.

15. Subcommittee G01.05 on Laboratory Corrosion Tests. 2011. Standard Practice for Preparing, Cleaning, and

Evaluating Corrosion Test Specimens. ASTM G1-03. West Conshohocken, PA: ASTM International.

16. ACI Committee 318. 2008. Building Code Require-ments for Structural Concrete (ACI 318M-08) and Commentary. Farmington Hills, MI: ACI.


About the authors

Adham El Menoufy, MASc, is a PhD candidate in the Civil and Environmental Engineering Department at the University of Waterloo in Ontario, Canada. El Menoufy’s research focuses on sustainability and rehabilitation of

prestressed concrete structures and long-term corro-sion effects under fatigue loading.

Khaled Soudki, PhD, PEng, was a professor and Canada Research Chair in innovative structural rehabilitation in the Department of Civil and Environmental Engineering at the University of Waterloo. Soudki was recognized

as an international leader and a world-renowned expert in the rehabilitation of structures. His research and teaching contributions over the past 20 years were in the field of reinforced and prestressed concrete structures with emphasis on the use of fiber-reinforced polymers.

Abstract

This paper presents a study that assessed the effect of corrosion of seven-wire steel strands on the residual capacity of pretensioned concrete T beams subjected to static flexural loading and the viability of carbon-

fiber-reinforced polymer (CFRP) repair in restoring the original capacity. The experimental program com-prised testing six 3.6 m [12 ft] pretensioned concrete T beams, each having 400 mm (16 in.) flange width, 300 mm (12 in.) total height, and 100 mm (4 in.) web thickness. The experimental variables were the mass loss due to corrosion (0%, 2.5%, 5%, and 10%) and the repair condition (unrepaired and repaired beams using adhesively bonded CFRP sheets). Test results showed a reduction in flexure capacity and midspan deflection of up to 76% and 26%, respectively, at 10% mass loss. CFRP repair was able to restore the load capacity to that of the uncorroded beams; however, the reduction in ductility due to corrosion was not revers-ible.

Keywords

Carbon-fiber-reinforced polymer, CFRP, corrosion, girder, repair, strengthening.

Review policy

This paper was reviewed in accordance with the Precast/Prestressed Concrete Institute’s peer-review process.

Reader comments

Please address and reader comments to journal@pci .org or Precast/Prestressed Concrete Institute, c/o PCI Journal, 200 W. Adams St., Suite 2100, Chicago, IL 60606. J

flexural behavior of corroded pretensioned girders

Documents