Carbon Fiber Reinforced Polymer Repairs of Impact-Damaged Prestressed I-Girders

Carbon Fiber Reinforced Polymer Repairs of
Impact-Damaged Prestressed I-Girders
Ryan J. Brinkman
B.S. University of Cincinnati
Thesis submitted to:
School of Advanced Structures
College of Engineering and Applied Science
Division of Graduate Studies
University of Cincinnati
Dr. Richard A. Miller, Ph.D., P.E., FPCI
For partial fulfillment of the requirements
for the degree of Master of Science
November 2012
Committee:
Dr. Bahram Shahrooz, Ph.D, P.E., FACI
Dr. Kent A. Harries, Ph.D., P.Eng., FACI, University of Pittsburgh

i


UMI Number: 1534443

All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

UMI 1534443
Published by ProQuest LLC (2013). Copyright in the Dissertation held by the Author.
Microform Edition © ProQuest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States Code

ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346


ABSTRACT
Three different Carbon Fiber Reinforced Polymer (CFRP) repair techniques are examined to
determine how effectively each method can restore the ultimate flexural capacity of impact-damaged
prestressed concrete girders. The number of severed prestressing strands and the amount of CFRP
repair material applied are varied to create a large repair matrix. The ultimate moment capacity was
calculated using the program XTRACT. Capacities were also evaluated using AASHTO and ACI
specifications.
The prototype girder’s geometry is based on an impact-damaged I-girder from a bridge in
Eastland County, Texas, which was repaired using CFRP in 2006 using conventional externally bonded
CFRP. The CFRP repair methods examined in this study were near surface mounted (NSM) CFRP,
externally bonded (EB) CFRP, and bonded post tensioned (bPT) CFRP. The area of the girder available
for repair was limited to the bottom soffit for near surface mounted CFRP and bonded post tensioned
CFRP; however, the externally bonded CFRP was applied on the bulb as well as the soffit. The range of
the repairs examined were approximately 25%, 50%, and 75%, and 100% of the maximum practical
amount of CFRP which could be applied based on the girder geometry, ACI guidelines, and
manufacturers’ recommendations.
The geometry of the girder limited the amount of CFRP which could be applied, so the repairs
could only completely restore the ultimate girder capacity when very few strands were damaged and

the maximum amount of CFRP was used. However, the CFRP techniques were shown effective at
restoring some of the lost capacity and thus are an option if only a portion of the lost capacity must be
restored. The study also showed that, in some cases, repairs are completely ineffective and the girder
capacity is not increased beyond the damaged state. Using the results of this study, engineers can
determine when repairing a girder will be effective and the extent to which the repair restores lost
capacity.

ii


This page intentionally left blank

iii


ACKNOWLEDGEMENTS
First I would like to thank my advisor, Dr. Richard Miller for his help and guidance throughout, as well as
giving me the opportunity to assist in this research. I would also like to thank Dr. Bahram Shahrooz and
Dr. Kent Harries for serving on my thesis committee.

I would to thank the National Cooperative Highway Research Program (NCHRP), Program Director Dr.
Waseem Dekelbab, and the Project Panel for funding and providing comments on this research under
NCHRP 20-07/Task 307. I also wish to thank the University of Cincinnati for awarding me a Graduate
Scholarship.

I would like to thank Dr. Harries and Dr. Jarrett Kasan for their help and guidance throughout the whole
project. This would not have been possible without their constant assistance.

Finally, I would to thank my friends, fellow students, and family for all the support over the years. I
would not have made it this far if it were not for you.

iv


TABLE OF CONTENTS

ABSTRACT...................................................................................................................................................... ii
ACKNOWLEDGEMENTS ................................................................................................................................ iv
TABLE OF CONTENTS..................................................................................................................................... v
LIST OF FIGURES .......................................................................................................................................... vii
LIST OF TABLES ........................................................................................................................................... viii
CHAPTER 1: INTRODUCTION ......................................................................................................................... 1
1.1– INTRODUCTION ................................................................................................................................. 1
1.2 – OBJECTIVES ...................................................................................................................................... 2
CHAPTER 2: LITERATURE REVIEW ................................................................................................................. 3
2.1 – GUIDELINES FOR REPAIR OF PRESTRESSED CONCRETE BRIDGE ELEMENTS .................................... 3
2.1.1 - Shanafelt and Horn (1980) ......................................................................................................... 3
2.1.2 - Shanafelt and Horn (1985) ......................................................................................................... 4
2.1.3 - Post NCHRP 280 - Harries, Kasan and Aktas (2009) ................................................................... 7
2.2 - TESTING OF CFRP REPAIR TECHNIQUES ON EXPERIMENTAL GIRDERS............................................. 7
2.2.1 - Quattlebaum, Harries, and Petrou (2005) ................................................................................. 7
2.2.2 - Nordin and Täljsten (2006) ........................................................................................................ 9
2.2.3 - Casadei, Galati, Boschetto, Tan, Nanni, and Galeki (2006)...................................................... 10
2.2.4 - Aram, Czaderski, and Motavalli (2008) .................................................................................... 11
2.3 - TESTING OF CFRP REPAIR TECHNIQUES ON EXTRACTED GIRDERS ................................................. 12
2.3.1 - Aidoo, Harries, and Petrou (2006) ........................................................................................... 12
2.3.2 - Miller, Rosenboom, and Rizkalla (2006) .................................................................................. 13
2.3.3 - Reed, Peterman, Rasheed, and Meggers (2007) ..................................................................... 14
2.4 – FIELD REPAIRS OF IMPACT-DAMAGED GIRDERS............................................................................ 15
2.4.1 - Schiebel, Parretti, and Nanni (2001) ........................................................................................ 15

2.4.2 - Tumialan, Huang, Nanni, and Jones (2001) ............................................................................. 16
2.4.3 - Klaiber and Wipf (2003) ........................................................................................................... 17
2.4.3 - Kim, Green, and Fallis (2008) ................................................................................................... 18
2.4.4 - Yang, Merrill, and Bradberry (2011) ........................................................................................ 19
2.5 - FRP REPAIR GUIDELINES ................................................................................................................. 20
2.5.1 - ACI Committee 440 (2008)....................................................................................................... 20
2.5.2 - Zureick, Nowak, Mertz, and Triantafillou (2010) ..................................................................... 21
2.7 SUMMARY ......................................................................................................................................... 22
CHAPTER 3: PROTOTYPE BRIDGE ................................................................................................................ 24
3.1 – BRIDGE AND GIRDER GEOMETRY................................................................................................... 24
CHAPTER 4: CFRP REPAIR METHODS .......................................................................................................... 26
4.1 – GENERAL REPAIR METHOD INFORMATION ................................................................................... 26
4.2 – NEAR SURFACE MOUNTED (NSM) REPAIR METHOD ..................................................................... 26
4.3 – EXTERNALLY BONDED (EB) REPAIR METHOD ................................................................................ 30
4.4 – BONDED POST TENSIONED CFRP (bPT) REPAIR METHOD ............................................................. 31
v


CHAPTER 5: MODELING .............................................................................................................................. 34
5.1 - XTRACT PROGRAM.......................................................................................................................... 34
5.2 – MODELING THE GIRDER ................................................................................................................. 34
5.3 – MODELING STRAND DAMAGE ....................................................................................................... 38
CHAPTER 6: CALCULATING THE EFFECTIVENESS OF THE REPAIR ............................................................... 39
6.1 - NORMALIZED RATING FACTOR ....................................................................................................... 39
6.2 ANALYSIS OF PROTOYPE GIRDERS .................................................................................................... 40
CHAPTER 7: DISCUSSION ............................................................................................................................. 44
7.1 – GENERAL ANALYSIS ........................................................................................................................ 44
7.2 – NEAR SURFACE MOUNTED METHOD ............................................................................................. 46
7.3 – EXTERNALLY BONDED METHOD .................................................................................................... 47
7.4 – BONDED POST TENSIONED METHOD ............................................................................................ 48

7.5 – METHOD COMPARISON ................................................................................................................. 49
7.6 – UPDATED REPAIR SELECTION CRITERIA TABLES ............................................................................ 54
CHAPTER 8: SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS ........................................................ 59
8.1 - SUMMARY ....................................................................................................................................... 59
8.2 - CONCLUSIONS ................................................................................................................................. 60
8.3- ADDITIONAL RESEARCH RECOMMENDATIONS ............................................................................... 63
LITERATURE REFERENCES ........................................................................................................................... 64
APPENDIX A SAMPLE CALCULATIONS ......................................................................................................... 69
APPENDIX B BRIDGE RATING EXAMPLE ...................................................................................................... 70

vi


LIST OF FIGURES
Figure 3.1.1 Prototype bridge and girder (Yang et al. 2011) ...................................................................... 25
Figure 4.2.1 NSM slot spacing requirements based on ACI 440.2R-08 ...................................................... 27
Figure 4.2.2 NSM spacing for prototype girder .......................................................................................... 29
Figure 4.3.1 EB spacing for prototype girder .............................................................................................. 31
Figure 4.4.1 bPT spacing for prototype girder ............................................................................................ 33
Figure 4.4.2 Schematic representations of CFRP applications (Harries et al. 2012)................................... 33
Figure 5.2.1 XTRACT model of EB method with 17 CFRP strips and 8 prestressed steel strands severed
from the bottom row .................................................................................................................................. 37
Figure 5.2.2 XTRACT model of the girder.................................................................................................... 37
Figure 7.2.1 Normalized rating factors for NSM method ........................................................................... 47
Figure 7.3.1 Normalized rating factor for EB method................................................................................. 48
Figure 7.4.1 Normalized Rating factor for bPT method .............................................................................. 48

vii

LIST OF TABLES
Table 1.1.1 Repair selection criteria (Shanafelt and Horn 1985) .................................................................. 5
Table 4.2.1 Optimization of NSM strip dimensions .................................................................................... 29
Table 5.2.1 XTRACT concrete properties .................................................................................................... 35
Table 5.2.2 XTRACT reinforcing element properties .................................................................................. 37
Table 5.3.1 Repair scenarios considered in this study ................................................................................ 38
Table 6.2.1 Girder capacity and normalized rating factor .......................................................................... 42
Table 6.2.2 Change in girder capacity and normalized rating factor .......................................................... 43
Table 7.5.1 Approximate values of α and β ................................................................................................ 52
Table 7.5.2 Maximum number of severed prestressing strands, nmax, that can be replaced by CFRP.
(Harries et al. 2012) .................................................................................................................................... 53
Table 7.6.1 Repair Selection Criteria – EB-CFRP Techniques. (Harries et al. 2012) .................................... 55
Table 7.6.2 Repair Selection Criteria – PT-CFRP Techniques (Harries et al.2012) ...................................... 56
Table 7.6.3 Repair Selection Criteria – Steel-based Techniques (Harries et al. 2012) ................................ 57

viii


CHAPTER 1: INTRODUCTION
1.1– INTRODUCTION
There are many prestressed bridges currently in service in the United States which have been
subjected to various levels of impact damage. In 1980 it was reported there was an average of 201
damage incidents per year to the nation’s 23,344 prestressed concrete bridges, resulting in an incident
rate of 0.86% (Shanafelt 1980). The actual incident rate is likely to be higher, as not every incident of
damage is reported. Combined with the continuously increasing load demands bridges are required to
support, the damage can cause the bridges to be rated below the American Association of State
Highway and Transportation Officials (AASHTO) inventory rating of 1.0. In these cases, transportation
officials must make the decision to replace the bridge, replace the damaged elements only, repair
damaged elements, post the structure, or otherwise accept the substandard rating.
The benefits of repairing a larger bridge element compared to replacing one are plentiful; it is

quicker, less expensive, less inconvenient for users of the bridge, and usually less complex. However, if
information on bridge repair methods is inadequate, bridges elements may be needlessly replaced
without examining all possible repair techniques. In 1985, a National Cooperative Highway Research
Program Report estimated the cost of repairing a girder with “severe” damage was 15-50% of the cost of
replacing the girder (Shanafelt 1985). The range of this figure demonstrates the importance having
sufficient and updated data on all repair techniques and examining every possible repair technique.
Various repair techniques developed by commercial and academic institutions have been
proven effective. Although these repair methods have been proven as suitable alternatives to replacing
a structure, there has not been extensive research comparing the different methods and their ranges of
applicability. There exists no guideline to inform decision makers whether to replace, repair, or do
nothing to damaged bridge girders, or what techniques are applicable and the most efficient for
restoring damaged capacity to a bridge. There is no “one size fits all” repair technique because
variables, such as the extent of damage and girder geometry, will affect the effectiveness of each repair

1


method. Almost every repair method can be utilized to restore some lost capacity, thus, a table
comparing the selection criterion and comparing the different repair options would be beneficial in
selecting a repair technique.

1.2 – OBJECTIVES
The need for an updated repair selection guidance that compares different repair methods is
great, as it will provide bridge engineers and evaluators with better resources from which to make welleducated decision. A crucial component of such a repair matrix would be the inclusion of data on
recently developed repair techniques utilizing carbon fiber reinforced polymers (CFRP). There has been
successful implementation of these techniques academically and in the field, but the entire range of
applicability for which they can be utilized is still uncertain. The objective of this study is to populate a
matrix involving different types of CFRP repair techniques over various damage ranges to demonstrate
what methods are applicable for restoring the ultimate flexural capacity of damaged I-girders to provide
a reference for engineers considering CFRP repairs. A full load rating example and CFRP repair design

calculations will also be provided as additional reference.

2


CHAPTER 2: LITERATURE REVIEW
2.1 – GUIDELINES FOR REPAIR OF PRESTRESSED CONCRETE BRIDGE ELEMENTS
2.1.1 - Shanafelt and Horn (1980)
The National Cooperative Highway Research Program (NCHRP) Report 226 - Damage Evaluation
and Repair Methods for Prestressed Concrete Bridge Members presents a guide on how to assess
prestressed concrete bridge damage and how to select an appropriate repair. The aim of the report was
to assist the people responsible for the bridge repair in their decision making process by using value
engineering and taking into consideration service life, safety performance, maintenance, cost,
aesthetics, user convenience, speed of repair, and inconvenience to drivers.
To help guide its users, the report defines three different damage classifications: minor damage,
moderate damage, and severe damage. These damage classifications were used in a survey given to the
state departments of transportation to try to establish common definitions and determine each
department’s course of action regarding each classification. Minor damage is described as damage only
to the concrete portion of the girders, with no exposed reinforcing bars or prestressing strands.
Moderate damage is described as damage only to the concrete portion of the girders or extensive
spalling, which results in exposed reinforcing bars, and, or prestressing strands. Severe damage is
described as damage to concrete and reinforcing elements of the girder.
The report goes into detail describing four different repair options: external post tensioning,
metal sleeve splicing, strand splicing, and a combination of external post tensioning and metal sleeve
splicing. External post tensioning is the application of high strength steel rods or strands to the girder by
means of a bolsters (also called corbels) which are cast against the girder. The steel rods are tensioned
by jacking against the bolster drawing the girder into compression to help restore lost prestressing
force. Metal sleeve splicing, also referred to as steel jacketing, is the application of steel plates to
enclose the girders. Shear studs may be required and grouting is usually required to account for
dimensional inconsistencies of the girder. The metal sleeves can help strengthen the girder by providing

3


sufficient additional reinforcement and some degree of confinement. Internal splicing is the process of
reconnecting severed strands with a mechanical splice. Once, spliced prestress can be restored by
heating of the strand, preloading the bridge, or tightening the splice with a torque wrench. These
methods can be used in combination to improve the effectiveness of the repair. The report also
provides a hybrid example of a metal sleeve splice and then installing external post tensioning.
Eleven cases of these repair methods are described in detail and eight of the methods include
some sample calculations. However, though sample calculations are presented, the specific repair
methods were not tested, and there is no evidence of the success or failure of these detailed repairs.

2.1.2 - Shanafelt and Horn (1985)
NCHRP Report 280 - Damage Evaluation and Repair Methods for Prestressed Concrete Bridge
Members, presents Phase II of the project that was published five years earlier. The stated purpose was
to “further evaluate promising methods of repair and to prepare guideline for damage evaluations and
repair techniques.” The report suggests that even though some of the practices were years old upon
the writing of NCHRP 226, not all of the methods were widely accepted. It was believed the lack of
guidance regarding the assessment of damages and the process of repairs, as well as the lack of repair
method testing, were all causes for the repair methods not being implemented. It was also believed
that since replacing a girder is the most conservative approach, some engineers simply preferred this
option.
The report tried to address the lack of guidance by presenting a repair selection criteria table,
Table 2.1.1, to help assist bridge engineers and evaluators. This table provides qualitative assessment
values of four different repair methods: post-tensioning, internal strand splicing, metal splice splicing or
steel jacketing, and replacement. Although somewhat limited in scope, this table does proved a quick
and easy to use reference that is useful when comparing different repair methods.

4

Posttensioning

Repair Method to Consider
Internal
Metal Sleeve
Splicing
Splice

Service and Ultimate Load

Excellent

Excellent

Excellent

Excellent

Overload
Fatigue

Excellent
Excellent
Excellent

Excellent
Limited
N/A

Excellent
Excellent
Excellent

Excellent
Excellent
N/A

Excellent
Limited

Excellent
N/A

Excellent
Excellent

N/A
Excellent

Limited

Limited

Large

Unlimited

Perhaps

Excellent
Good
Excellent
Low
Fair*

Yes
Excellent
Excellent
Excellent
Very Low
Excellent

Probably
Excellent
Good
Excellent
Low
Excellent

No
Excellent
Poor
Excellent
High
Excellent

Damage Assessment Factor

Adding Strength to

Non-Damaged Girders
Combing Splice Methods
Splicing Tendons or
Bundled Strands
Number of Strands Spliced
Preload Required
Restore Loss of Concrete
Speed of Repair
Durability

Cost
Aesthetics
N/A means not applicable.
* Can be improved to excellent by extending corbels on fascia girder

Replacement

Table 1.1.1 Repair selection criteria (Shanafelt and Horn 1985)
This report also tried to address the lack of repair processes and lack of experimental testing to
back up the design recommendations. The methods examined in this were the same repair methods
examined in NCHRP 226: external post tensioning, internal splicing of severed strands, a metal sleeve
splice, and a combination, hybrid example of internal strand splicing and external post tensioning.
For the experimental program, ten different tests were performed on one girder. During the
first nine tests (described below), at least two full loading cycles were applied during the trials, but the
maximum load was limited to 75% of the calculated ultimate moment capacity of the girder. For test 10,
100% of the ultimate load was applied in one cycle. The test load was a single point load located at
midspan of a simply supported girder. Prior to some tests, spalled concrete was patched, however none
of the cracks were repaired by epoxy injection. A record of loads, strains, and deflections were made.
The tests performed were:
Test 1 – Instruments were set up and installed on the undamaged girder and it was loaded until it

5


reached 75% of its calculated ultimate moment capacity.
Test 2 – Concrete corbels and post-tensioned high strength bars were installed and the girder was
loaded until 75% of its calculated ultimate moment capacity.
Test 3 – The high strength bars were removed and the girder was loaded until 75% of its calculated
ultimate moment capacity.
Test 4 – Concrete was broken away in order to sever four of the sixteen prestressed strands and the
girder was loaded until 75% of its calculated ultimate moment capacity.
Test 5 – The four severed strands were each spliced with a single-strand internal splice and a preload
was applied. The girder was patched and the preload was released after the patch regained adequate
strength. The girder was loaded until 75% of its calculated ultimate moment capacity.
Test 6 – The post-tension high-strength bars were reconnected to the girder and the girder was loaded
until 75% of its calculated ultimate moment capacity.
Test 7 – The post-tension bars were disconnected, concrete was broken away, the four strand splices
were severed, and the girder was loaded until 75% of its calculated ultimate moment capacity.
Test 8 – The girder was patched and the external high strength bars were post tensioned. It was loaded
until 75% of its calculated ultimate moment capacity.
Test 9 – The post-tension bars were disconnected, concrete was broken away, two more strands were
severed, for a total of six out of sixteen strands severed. The girder was patched and a metal sleeve
splice was installed. The girder was loaded until 75% of its calculated ultimate moment capacity
Test 10 – The girder was loaded until 100% of its calculated moment capacity.
Although this procedure addressed the need of experimental testing, a problem exists in that
each test was performed one right after the other. The strand splice testing was completed after the
external post tensioning had already been installed, and the metal sleeve was installed after the
completion of eight other tests. The various tests affect the concrete in different ways; even though the

6

tensions bars were removed or the girder was patched, it does not mean the girder did not still exhibit
affects from the previous tests. Such a test protocol generally yields inconclusive results.

2.1.3 - Post NCHRP 280 - Harries, Kasan and Aktas (2009)
This report serves as post NCHRP 280 guidelines for the repair of prestressed concrete girders.
Its contents include updated information on the available repair techniques, survey results on current
practices, and repair examples that now cover FRP. The authors suggest updating the damage
classification and breaking down the “Severe Damage” classification from NCHRP 280 into three
different categories. The Severe I category is described as damage which requires a structural repair but
does not necessitate a prestressed or posttensioned method. The Severe II category is described as
damage which requires prestress or posttensioning with the repair method. The Severe III category is
described as damage too great to be practically repaired and the member should be replaced.

2.2 - TESTING OF CFRP REPAIR TECHNIQUES ON EXPERIMENTAL GIRDERS
2.2.1 - Quattlebaum, Harries, and Petrou (2005)
In this experimental program twelve girders were cast; the girders were 15.6 ft. (4750 mm) long,
10.0 in. (254 mm) deep, and 6.0 in. (152 mm) wide with a clear span of 15.0 ft. (4572 mm). The beams
had one layer with three #4 (13 mm) reinforcing bars. The girders were retrofitted with three different
CFRP systems: conventional adhesive application (CAA), near-surface mounted (NSM), and powder
actuated fastener (PAF). CAA is now more commonly referred to as externally bonded (EB). A single 2.0
in. (51 mm) wide CFRP strip was used for the CAA method. Two - 0.98 in. (25 mm) strips adhered
together and inserted into two - 0.25 in. (6.4 mm) wide by 1.3 in. (32 mm) deep slots were used for the
NSM method. A 3.3 in. (84 mm) wide strip was used for the PAF method and the nails were embedded
1.6 in. (41 mm) into the soffit, staggered with a longitudinal spacing of 2.5 in. (64 mm), and had a
transverse spacing of 1.0 in. (25 mm). A total of six girders were tested under cyclic loading; one of each
method was tested under high stress fatigue loading, and one of each method was tested under low

7

stress fatigue loading. A total of four girders were tested monotonically until failure, one control beam
and one for each retrofit method. Two of the girders were inadvertently damaged and thus there were
no control beams for cyclic loading.
The specimens were loaded using a single point load at midspan. To determine the strain levels
of the steel, three linear displacement transducers were installed on either side of the girder at the level
of the steel and, three strain gauges centered on the midspan were attached to steel. CFRP strains were
also determined using strain gauges, and the deflection was measured using draw wire transducer. The
conclusions drawn were as follows:


All three methods resulted in strength increases over the control specimen. The yield load
increased 21-26% and the ultimate load increased 28-33%.



Under monotonic testing, concrete crushing was the dominate failure mode except for the CAA
specimen, which was controlled by midspan debonding.



The NSM specimen had the best bond after fatigue testing and was the sole specimen which
appeared to have a complete bond after monotonic loading testing.



The PAF specimens had lower strength increases compared to NSM and CAA because of less
effective stress transfer between CFRP and concrete.



The PAF and NSM methods noticeably weakened the concrete.



Care must be taken when using powder actuated faster because the reinforcing steel can be
damaged by nails, as was the case with the low stress fatigue specimen.



PAF outperformed others in fatigue loading, but its long-term performance is still very
questionable because shearing of the fasteners occurred shortly after the initial failure and
there were concerns of corrosion with the metal fasteners.



With cyclic loading, all methods exhibited increases of FRP and steel strain indicating degrading
bond characteristics.

8




PAF is the quickest method since it requires less labor and equipment; hence it may be
applicable for emergency situations. CAA requires more time since the soffit must be prepared,
the epoxy precisely mixed, and the epoxy applied before it cures. NSM requires the most time
since the soffit preparation takes longer than the CAA method because slots for the CFRP must
be cut and epoxy is used as well.

2.2.2 - Nordin and Täljsten (2006)
Fifteen girders were cast for this experimental program. The beams were 13.12 ft. (4 m) long,
11.81 in. (300 mm) deep, and 7.87 in. (200 mm) wide. All the retrofitted girders utilized the near surface
mounted method (NSM). One beam was not strengthened in order to be used as a reference, four
beams were retrofitted with non-prestressed NSM bars, and ten beams were retrofitted with
prestressed NSM bars. Two different types of CFRP material with different moduli of elasticity were
used; six beams were retrofitted with a 36,000 ksi (250 GPa) CFRP and eight with 23,000 ksi (160 GPa)
CFRP. In addition, the length of the CFRP was varied; eight beams had a CFRP length of 13.1 ft. (4.0 m)
that continued under the supports and the six other beams had a CFRP length of 10.5 ft. (3.2 m). The
beams were subject to two point loads placed equidistant from midspan and the beam was loaded to
failure. The beams were fitted with two strain gauges for the concrete at the top of the beam, three
strain gauges for the reinforcing steel, and four strain gauges for the CFRP. The midpoint deflection was
also measured. The conclusions drawn were as follows:


All the methods resulted in flexural strength increases over control specimen.



All the methods resulted in smaller cracks, which will increase durability and service life.



The failure mode of every retrofitted specimen was fiber failure of the NSM rod, thus indicating
efficient stress transfer and good bond.



The non-prestressed beams were more ductile and had a larger deflection at every load than

prestressed beams.

9




The stiffer rods resulted in stiffer behavior from the beam and a higher steel yielding load.

2.2.3 - Casadei, Galati, Boschetto, Tan, Nanni, and Galeki (2006)
For this study three I-girders were cast. The girders were 36 ft. (11 m) long, had a free span of
34 ft. (10.3 m), were 32 in. (810 mm) deep, had a 13 in. (330 mm) wide top flange, and a 17 in. (430 mm)
wide bottom flange. These dimensions fall between standard AASHTO Type I and Type II I-girders. The
concrete deck was 32 in. (810 mm) wide and 6 in. (152 mm) deep. There were two layers of six
prestressing strands. For the retrofitted girders, concrete was chiseled out to expose the strands and
two strands were cut to simulate impact damage. Prior to CFRP application, the loose concrete was
removed and patched. One specimen served as the control and was undamaged, one specimen was
retrofitted using the externally bonding (EB) method, and the third was retrofitted using the prestressed
NSM method. A 0.0065 in. (0.165 mm) thick CFRP sheet was used in the EB method, which was then
covered with U-wraps to prevent delamination. Three - 1.125 in. (2.8 cm) deep and 0.75 in (1.9 cm)
wide slots were cut and a prestressed 0.375 in (9.5 mm) diameter, 12 ft. (3.66 m) long CFRP bar was
inserted into each slot for the NSM method. The NSM bars were prestressed using a steel wedge
anchorage system to achieve the same level of prestress the beam had before the strands were cut. The
beams were subject to two point loads placed equidistant from midspan until failure. Three string
transducers, placed at mid-span and under loading points, measured the vertical displacements. The
conclusions drawn were as follows:


Both systems were capable of restoring the ultimate capacity of the two cut prestressing wires.



The EB CFRP repair failed by debonding of the CFRP and was then immediately followed by the
U-wraps rupturing.



NSM performed in a more ductile behavior and failed when the concrete cover on the sides of
the severed tendons, close to the NSM bars, split open.

10


2.2.4 - Aram, Czaderski, and Motavalli (2008)
In this study four girders were cast; the girders were 7.9 ft. (2.4 m) long, 9.8 in. (250 mm) deep,
and 5.9 in. (150 mm) wide. The CFRP repair method was the externally bonded method. One specimen
served as the control and was undamaged, one specimen was retrofitted using unstressed externally
bonded CFRP, and two specimens were retrofitted using prestressed externally bonded CFRP with the
CFRP prestressed at various levels: one at 36% of maximum value of tensile strength and one at 18% of
maximum value of tensile strength. The strips were 2.0 in. (50 mm) wide and 0.047 in. (1.2 mm) thick.
The gradient method was used to prestress the strips to avoid the use of mechanical anchorage. For the
gradient method, the prestressing force is gradually reduced from the middle toward both ends of the
CFRP strip via heating which cures the adhesive used to bond the CFRP to the beam. The beam was
tested using two point loads placed equidistant from midspan until failure or CFRP debonding. Strain
gauges were applied on the concrete and CFRP. The conclusions drawn were as follows:


Under greater loads, the retrofitted specimens had a smaller deflection and strain at midspan.



CFRP prestressing caused unloading of the steel reinforcement, which in turn increased the load
at which the steel began to yield.



The CFRP failure modes were debonding failure, delamination, and sudden separation of the
CFRP strip.



The load capacity increase of prestressed FRP was less than the unprestressed CFRP because of
the premature debonding of the strips.



Prestressing the strips resulted in lower shear stresses because cracks occurred at higher loads.



The higher prestressed CFRP had higher stresses at the CFRP strip ends.



The gradient method was not effective because the gradient anchorage was in the region of
shear stresses due to loading; however, this can be avoided with a longer span.

11

2.3 - TESTING OF CFRP REPAIR TECHNIQUES ON EXTRACTED GIRDERS
2.3.1 - Aidoo, Harries, and Petrou (2006)
This experimental program examines the CFRP repair and testing of eight reinforced concrete
girders removed from a forty-two year old decommissioned Interstate bridge. The bridge was cast-inplace reinforced concrete, designed in 1957, erected in 1961, and replaced in 2001 because it was not
wide enough to meet traffic demands. The bridge had five simple spans of 30.0 ft. (9.140 m) each, with
girders spaced at 6.50 ft. (1980 mm). The test specimens had a total depth of 32.5 in. (825 mm), a width
of 13.5 in. (343 mm); the test span was reduced to 26.33 ft. (8025 mm). The deck was 6.50 in. (165 mm)
thick and 36.5 in. (927 mm) wide. There were three layers of reinforcing bars; the bottom layer
contained three -#11 (36 mm) bars, the middle layer contained three -#10 (32 mm) bars, and the top
layer contained two -#8 (25 mm) bars. The girders were retrofitted with three different CFRP systems:
conventional adhesive application (CAA), near surface mounted (NSM), and powder actuated fastener
(PAF). Two - 4.02 in. (102 mm) strips placed adjacent to each other on the girder soffit were used in the
CAA method. Four 1.3 in. (32 mm) deep slots were cut and two - 0.98 in. (25 mm) strips adhered
together and inserted into each slot were used for the NSM method. Two - 4.09 in. (104 mm) hybrid
strips were placed on top of each other and fastened to the bottom girder soffit were for the PAF
method. All strip lengths were 25.0 ft. (7620 mm) to avoid going over the beam supports and thus
replicate real world application.
The girders were loaded using a single point load at midspan; four were subjected to monotonic
loading until failure and four were subjected to fatigue loading. To determine the strain levels of the
steel, the girders were fitted with seven horizontally mounted linear variable resistance displacement
transducers centered at midspan on the soffit and five supplementary LVR transducers were fitted on
the both sides of the beam at the level of the steel. The vertical deflections were measured with a draw
wire transducer. Strain gauges were also applied to the CFRP coinciding with the seven horizontal LVRs
to determine the CFRP strain and investigate debonding. The conclusions drawn were as follows:

12




All retrofitted specimens had a higher flexural strength than the control.



All retrofitted specimens showed debonding characteristics in fatigue loading despite the
relatively low fatigue stress levels.



CAA and NSM were much better in monotonic loading than PAF.



Powder actuated fasteners were not a good choice to use on older concrete because significant
cracks and spalling can occur.



NSM showed the best bond characteristics and the best ductility, however took the longest to
install.

2.3.2 - Miller, Rosenboom, and Rizkalla (2006)
In this study, the CFRP repair and testing of an impact-damaged girder extracted from the North
Carolina Bridge 169 in Robeson County is analyzed. The span was 54.8 ft. (16.7 m), but in the test setup,
the clear span between supports was reduced to 53.8 ft. (16.4 m). The girder was an AASHTO Type II Igirder with a total depth of 36.9 in. (913 mm), top flange width of 12.0 in. (305 mm), and bottom flange
width of 17.9 in. (457 mm). The deck was 5.98 in. (152 mm) thick and 14.8 in. (377 mm) wide. The
girder had been impacted close to midspan and one of the sixteen prestressed strands was severed
while others were exposed due to spalling.
The CFRP repair method was EB. Prior to the CFRP application, the loose concrete was removed
and patched. Three layers of CFRP, applied one on top of each other, were adhered to the bottom soffit

and one layer of CFRP was adhered to the rest of bulb. A 5.9 ft. (1.8 m) long CFRP U-wrap was applied to
the damaged area and then five U-wraps spaced at 1.3 ft. (0.4 m) were applied on either side of the
damaged girder. The U-wraps were needed to prevent crack growth at the damaged area and to
mitigate the possibility of CFRP delamination. The beam was instrumented with potentiometers and
strain gauges. The girder was tested in three point bending and was first subjected to fatigue loading
and then loaded until failure. The conclusions are as follows.

13




Girders subjected to impact damage with one of sixteen strand missing can be repaired using
CFRP.



The beam failed as a result of cracks propagating outside of the CFRP repaired areas; thus the
CFRP repaired portion behaved better than undamaged portion even with severed strands.



The CFRP did not delaminate before the concrete began crushing; therefore, it is assumed the
CFRP did not debond prematurely.

2.3.3 - Reed, Peterman, Rasheed, and Meggers (2007)
The CFRP repair and testing of damaged girders extracted from Bridge #56 located in Graham
County, Kansas near Penokee is examined in this report. The girder was constructed in 1969 designed
for H-15 loading, and thus had been subjected to many cases of overloading, which led to significant
cracking and spalling. Damaged girders were replaced and three of the extracted girders were obtained

for testing. The bridge was four spans and the girders were double-T shaped. The girders were then
saw cut to create six single-T shapes which were 40 ft. (12.2 m) long. The total depth of the beam was
23 in. (585 mm) and the top flange was 5 in. (125 mm) deep and 3 ft. (915 mm) wide. Each stem
contained four prestressing strands placed in a single vertical layer. The CFRP repair methods were EB
and a combination of EB and NSM. Five specimens were tested; specimens 1-3 were tested until failure,
and Specimens 4-5 were tested under fatigue loading.
Prior to CFRP application, the loose concrete was removed and patched. Specimen 1 was left
unrepaired and served as a reference, and Specimens 2-5 had two plies of CFRP externally bonded to
the soffit and continuing 3.75 in. (95 mm) up the stem. Specimen 2 had only one 12 in. (305 mm) wide
ply of CFRP U-wrap covering the CFRP on both ends. Specimen 3 and 4 had two plies of CFRP U-wraps
spaced at 18 in. (455 mm) as well as CFRP U-wraps on the ends. In addition to the CFRP externally
bonded to the surface, Specimen 5 had two 60 in. (1500 mm) long #6 (19 mm) NSM CFRP bars inserted
at midspan on both sides of the web. In addition, Specimen 5 had three plies of 20 in. (510 mm) long

14