A .
$33.:
,5”; 9.

.1 :A-Imn

.. .

33w». mu. 3.,

‘ v
.12‘343 .A.

Warm

 

a.
. 3 2.3.

.vﬂﬁhr. suit...

.1-

«.

.
3......A...la.u.ﬂ..r.1h
:5»..::.;. I.

2. 59.qu... e

.
a i: 3...: A.
.0. $5.2: Mai? .

“£315.11. , M.

a
3..

 

.1"
.Pl....
. .. 11!.
.9 liiéihﬁva:

.9 1.!
3:55;;
thﬁnthrLal-Jsllx.
K11...A§.I:7t{:'1f..
.7. 11:1. .

. .. A...
gs.
a.

 

. 11.1....1v i: at . t 3h”
L) liars}. .1..o?£.(33i..4§lu in?!
an 9. .r‘! . A
. . Vii—3‘22: .
pick}?
H ill
.0! ii!

 

33:323. , ..
i..s!£x.ra 11.3...) 91"“. .3.-
.c 331.35.. 61.17....» .1! 91.. 2
5. h. 13.7 2.. .igl’fr 0‘
. ignitinﬁ‘sxlal:
. i; ‘35!!!
3:21;)..617 1 A

.s . :53:
{3.33:}!
x3 .5... ..

ad! :59 :6

14...: {32,552}...
1.5%. 33,335}.-
A... 7 2......

 

        

 

 

 

 

 
    

.
V A f :21..- .. 3..., 4
. 5.... A: , A A A . A . .
y L: i:
A. .. A h .: , ,. :3: =23 :.: s :31 « . Z
. o.

     
     

x

.
s

55.3...

 
   
 

s. F vi a.

.
f
3. .

‘|

E; .i...

.. . r x...

   

 

 

THESlS

2 NC

This is to certify that the I
dissertation entitled

Repair of Corrosion-Damaged
Columns Using FRP Wraps

presented by

Mohamad Imad Baiyasi

has been accepted towards fulﬁllment
of the requirements for

Ph . D . degree in Civil Engineering ‘ ,

Wake/W

Major professor

Date June 2000

MS U is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

 

 

 

LIBRARY

Mlchlgan State
University

 

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDue.p65-p.15

 

Repair of Corrosion-Damaged Columns
Using FRP Wraps

Mohamad Imad Baiyasi

A Dissertation

Submitted to
Michigan State University
In partial fulfillment of the requirement
for the degree of

DOCTOR OF PHILOSOPHY
Department of Civil and Environmental Engineering

2000

Abstract

Repair of Corrosion-Damaged Columns
Using FRP Wraps

By
Mohamad Imad Baiyasi

Many bridge columns in Michigan are damaged by chloride contamination
resulting in the corrosion of the steel reinforcement, and swelling and spalling of the
concrete and use of the bridges is typically continued. This in itself may not be a serious
problem since most columns in Michigan are over-designed and the loss of strength is not
a signiﬁcant issue. However, the lack of any method to minimize or prevent corrosion of
the steel results in continued deterioration and unsightly columns. Polymer composite
(also known as ﬁber-reinforced polymer or FRP) jackets offer a possible remedy to this
problem. They offer a rapid repair technique with the potential to enhance the long-term
durability and compression strength of damaged columns due to the conﬁnement that is
provided when ﬁbers are oriented in the hoop direction. Fibers oriented in the vertical
direction can enhance the bending strength.

Experiments were conducted to assess the effects of using FRP wraps with ﬁbers
oriented in the hoop direction for rehabilitating corrosion-damaged columns. Issues that
were explored are: (1) effect of freeze-thaw and wet-dry cycles on the properties of FRP
panels; (2) freeze-thaw durability of concrete square and cylindrical specimens wrapped
with glass and carbon FRP and subjected to an internal expansive force; and (3) effect of

wrapping on the rate of corrosion in an accelerated corrosion test.

The results of the freeze-thaw experiment indicate that freeze-thaw cycles have no
statistically signiﬁcant effect on the compressive strength of glass and carbon wrapped
specimens. For round specimens, glass and carbon wraps increased the strength by a
factor of about 2.3 and 2.6, respectively. For square specimens, glass and carbon wraps
increased the strength by a factor of 1.4-1.5. Freeze-thaw conditioning generally reduced
the longitudinal failure strain of wrapped specimens.

The square wrapped specimens had lower compressive strength compared to the
round specimens, even though the cross sectional area of the square prisms is higher than
that of the round cylinders. This is due to the reduced conﬁnement provided by the wraps
for square cross sections and stress concentrations that develop at the comers. Wrapped
square prisms always failed by rupture of the wrap at a corner. A reduction of
approximately 30% to 40% in failure stress was noted between round and square
wrapped specimens.

The results of the accelerated corrosion experiment indicate that wrapping
reduced the corrosion depth in the reinforcing bars by 46% to 59% after 190 days of
testing. Both glass and carbon wraps are equally effective in slowing down corrosion.
Although unbonded wraps do reduce stress concentrations in the FRP, they are less
effective in reducing the corrosion rate than the bonded wraps. It is postulated that this is
due to the ingress of water along the unbonded FRP-concrete interface.

Wrap strains for bonded specimens with both types of wraps tend to level off with
time indicating that corrosion slows down signiﬁcantly after some time. .One explanation
could be that the stress concentration near the anodes in the bonded wraps is more

effective in containing the corrosion-induced crack and reducing the corrosion rate. The

slip of unbonded wraps and the resulting redistribution of strain along the entire wrap

may be less effective at containing the large corrosion-induced crack near the anodes.

Dedicated to

my father, M. Fayez Baiyasi

Acknowledgements

The author would like to thank the following people:

Dr. R. Harichandran (Ron) for his support, guidance, and understanding. I am unable to

ﬁnd the words to express my gratitude. Thank you for everything.

Dr. P. Soroushian (CEE), Dr. L. Segerlind (AGE) and Dr. V. Zeidan (Math) for serving

as Ph.D. guidance committee members.

Dr. G. Cloud (MSM Department) for providing the data acquisition system used in this

work.

Dr. D. Liu (MSM Department) for allowing use of the impact machine used in this work.

Mr. J. Grove (MSU Radiology Department) for repeatedly taking X-rays of the concrete

specimens and not complaining once. Thank you John.

Mr. Roger Till (MDOT) for his invaluable input and guidance.

Mr. John Janiszewski, student helper, (aka LB), I will remember your help and hard work

always. Thanks a whole lot.

vi

TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION AND LITERATURE REVIEW

INTRODUCTION ................................................................................................... 1
PROJECT OBJECTIVES ....................................................................................... 4
LITERATURE REVIEW ........................................................................................ 4
1.3.1 Fiber Reinforced Polymers (FRP) for Infrastructure ..................................... 4
1.3.1.1 Fibers .................................................................................................. 5
1.3.1.2 Matrices .............................................................................................. 6
1.3.2 Durability of Concrete ..................................................................................... 7
1.3.3 Environmental Effects on FRP Composites ................................................... 9
1.3.4 Corrosion of Reinforcing Steel ..................................................................... 13
1.3.4.1 Factors Affecting Corrosion ............................................................ 14
1.3.4.2 Volume Expansion Due to Corrosion of Steel ............................... 16
1.3.4.3 Localized Corrosion ........................................................................ 17
1.3.5 Mechanical Properties of FRPs ..................................................................... 18
1.3.5.1 Stress Corrosion and Stress Rupture ............................................... 18
1.3.6 Effect of Conﬁnement ................................................................................... 19
1.3.7 Repair of Corrosion Damaged Columns Using FRP ................................... 25

CHAPTER 2: DESCRIPTION OF EXPERIMENTS

2.1

LABORATORY TESTING ...................................................................................... 30
2.1.1 Stiffness and Strength of Glass and Carbon FRPs ....................................... 30
2.1.2 Strain Expected in Wraps Due to Corrosion ................................................ 31
2.1.3 Freeze-Thaw Test .......................................................................................... 32
2.1.3.1 Mold Fabrication ............................................................................. 34
2.1.3.2 Bristar Calibration ........................................................................... 35
2.1.3.3 Chloride Content .............................................................................. 38
2.1.3.4 Strain Gage Placement .................................................................... 39
2.1.3.5 Compression Testing ....................................................................... 42
2.1.4 Accelerated Corrosion ................................................................................... 44
2.1.4.1 Corrosion Prior to Wrapping ........................................................... 46
2.1.4.2 Construction of the Corrosion Tank and Appurtenances ................ 49
2.1.4.3 Monitoring Progress of Corrosion During Test ............................ 50
2.1.4.4 Corrosion Test Matrix ................................................................... 54

CHAPTER 3: DATA COLLECTION AND ANALYSIS OF RESULTS

3.1

EFFECT OF FREEZE-THAW AND WET-DRY CYCLING ON THE
PROPERTIES OF FRP PANELS ......................................................................... 56

vii

3.2 FREEZE-THAW TEST ......................................................................................... 57

3.2.1 Strain Gage Readings ..................................................................................... 57
3.2.2 Results of Compression Testing .................................................................... 69
3.2.3 Statistical Analysis ......................................................................................... 82
3.2.4 Effect of Sustained Loads on Freeze-Thaw Durability of Wraps .............. 88
3.2.5 Comparison of Measured and Predicted Compression Strength ............... 88
3.3 ACCELERATED CORROSION ........................................................................... 89
3.3.1 Mass Loss Results ...................................................................................... 89
3.3.2 Statistical Analysis ..................................................................................... 94
3.3.3 Strain Measurements .................................................................................. 96

CHAPTER 4: SUMMARY, CONCLUSIONS, AND

RECOMMENDATIONS
4.1 FREEZE-THAW TEST ....................................................................................... 102
4.2 ACCELERATED CORROSION TEST .............................................................. 104
4.3 RECOMMENDATION FOR FIELD INSTALLATION .................................... 105
4.4 COST SAVING OF USING FRP WRAPS ......................................................... 106

CHAPTER 5: FIELD INSTALLATION AND FUTURE STUDIES

5.1 FIELD INSTRUMENTATION ........................................................................... 108

5.1.1 Corrosion Monitoring of Field Columns ................................................. 108

5.1.2 Results of Field Monitoring ..................................................................... 113

5.1.2.1 Calculation of Corrosion Rate .................................................... 113

REFERENCES ............................................................................ 114
APPENDICES

APPENDIX A: Aerospace Corporation’s FRP Panel Environmental Durability Data...118

APPENDIX B: Calculation Details ................................................................................. 120

B.1 Strain in Column Wrap After 10 Years ................................................... 120

B2 Conﬁning Pressure and Strain in Steel Jacket ......................................... 121

APPENDIX C: Specimen, Bar and Gage Numbering ..................................................... 123

APPENDIX D: Provisions for Field Installation ............................................................. 125

viii

Table 1.1

Table 2.1
Table 2.2
Table 2.3
Table 2.4
Table 2.5
Table 2.6

Table 2.7
Table 2.8
Table 3.1

Table 3.2
Table 3.3
Table 3.4

Table 3.5

Table 3.6

Table 3.7

Table 3.8

Table 4.1

Table 4.2
Table 5.1

LIST OF TABLES

Capillary porosity of Portland cement paste as a function of

water/cement Ratio and the degree of hydration ......................................... 8
Vender recommended and measured wrap properties .............................. 31
Volume expansion for some rust products ................................................ 31
Strain in column wrap due to steel corrosion after 10 years ..................... 32
Freeze-thaw laboratory test matrix ............................................................ 34
Internal pressure generated by corrosion for wrap strain of 0.531% ........ 36

Corrosion level required in two bars to induce a strain of

0.531% in the wrap .................................................................................... 46
Number of wrapped and unwrapped specimens in corrosion groups ....... 48
Accelerated corrosion laboratory test matrix ............................................ 55
Mean properties for FRP panels ................................................................ 57
Ductility enhancement under compression for wrapped specimens ......... 72
Freeze-thaw summary data ........................................................................ 82

Results of hypothesis tests (95%) on specimens exposed to

freeze-thaw cycles ..................................................................................... 84
Comparison of measured and predicted conﬁned compression

strength ...................................................................................................... 89
Mass loss and average corrosion depth for specimens exposed to

130 days of accelerated corrosion ............................................................. 92
Mass loss and average corrosion depth for specimens exposed to

190 days of accelerated corrosion ............................................................. 93
Results of hypothesis tests (95%) on specimens exposed to

accelerated corrosion ................................................................................. 95

Estimated material and installation cost for Tyfo-S glass and

MBrace carbon wrap systems .................................................................. 106
Estimated total cost of different repair techniques over ten years .......... 107
Initial weight of corrosion bars installed in ﬁeld columns ...................... 109

ix

Fig. 1.1

Fig. 1.2

Fig. 1.3
Fig. 1.4
Fig. 1.5
Fig. 1.6
Fig. 2.1
Fig. 2.2

Fig. 2.3
Fig. 2.4
Fig. 2.5

Fig. 2.6
Fig. 2.7
Fig. 2.8
Fig. 2.9

Fig. 2.10
Fig. 2.11

Fig. 2.12
Fig. 2.13
Fig. 2.14
Fig. 2.15
Fig. 2.16

Fig. 3.1

Fig. 3.2

Fig. 3.3

LIST OF FIGURES

Load vs. deﬂection curve for glass FRP composite specimens
subjected to 300 cycles of freezing and thawing ......................................... 9

Expansion and cracking of concrete due to corrosion of the embedded
steel ....................................................................................................... 16

Reduction of tensile strength of E-glass ﬁbers under sustained loads ...... 18

Stress/strain curves for conﬁned and unconﬁned concrete ....................... 20
Stress-strain curves for A36 steel, E-glass and carbon ﬁbers ................... 21
Illustration of effectively conﬁned area of a rectangular cross section ..... 25
Prism mold used for freeze-thaw test specimens ...................................... 34
Cylindrical mold used for freeze-thaw test specimens .............................. 35
Strain in steel tube for water/Bristar ratio of 400g/1000g ......................... 37
Strain in steel tube for water/Bristar ratio of 500g/1000g ......................... 37
Strain measurement instrument and switch box ........................................ 42
Capping ﬁxture .......................................................................................... 43
New compressometer ................................................................................ 43
Wiring diagram for accelerated corrosion specimens ............................... 45
Severely corroded specimens prior to patching ........................................ 48
Patching of the severely corroded specimens ............................................ 49
Corrosion tank ........................................................................................... 50
Corrosion specimens in the tank ............................................................... 50
Sample X-ray taken in the beginning of the accelerated corrosion test 52
Sample X-ray taken after 90 days of accelerated corrosion test ............... 53
Sample X-ray taken after 105 days of accelerated corrosion test ............. 53
Reinforcing bars after removal from a corrosion specimen ...................... 54

Hoop strains in glass wrap of round control specimen #3 before
correcting for thermal contraction of dummy FRP panel .......................... 59

Hoop strains in glass wrap of square control specimen #7 before
correcting for thermal contraction of dummy FRP panel .......................... 60

Hoop strains in carbon wrap of round control specimen #11 before
correcting for thermal contraction of dummy FRP panel .......................... 60

Fig. 3.4
Fig. 3.5
Fig. 3.6

Fig. 3.7
Fig. 3.8

Fig. 3.9

Fig. 3.10

Fig. 3.11
Fig. 3.12

Fig. 3.13
Fig. 3.14

Fig. 3.15
Fig. 3.16

Fig. 3.17
Fig. 3.18

Fig. 3.19
Fig. 3.20

Fig. 3.21
Fig. 3.22

Fig. 3.23

Hoop strains in carbon wrap of square control specimen #15 before

correcting for thermal contraction of dummy FRP panel .......................... 61
Hoop strains in glass wrap of round specimen #1 during freeze-thaw
cycles ....................................................................................................... 61
Hoop strains in glass wrap of round specimen #2 during freeze-thaw
cycles ....................................................................................................... 62
Hoop strains in glass wrap of control round specimen #3 ........................ 62
Hoop strains in glass wrap, round specimen #4 during freeze-thaw

cycles ....................................................................................................... 63
Hoop strains in glass wrap, square specimen #5 during freeze-thaw

cycles ....................................................................................................... 63
Hoop strains in glass wrap, square specimen #6 during freeze-thaw

cycles ....................................................................................................... 64
Hoop strains in glass wrap of control square specimen #7 ....................... 64
Hoop strains in glass wrap of square specimen #8 during freeze-thaw
cycles ....................................................................................................... 65
Hoop strains in carbon wrap of round specimen #9 during freeze-thaw
cycles ....................................................................................................... 65
Hoop strains in carbon wrap of round specimen #10 during freeze-thaw
cycles ....................................................................................................... 66
Hoop strains in carbon wrap of control round specimen #11 .................... 66

Hoop strains in carbon wrap of round specimen #12 during

freeze-thaw cycles ..................................................................................... 67
Hoop strains in carbon wrap of square specimen #13 during freeze—thaw
cycles ....................................................................................................... 67
Hoop strains in carbon wrap of square specimen #14 during freeze-thaw
cycles ....................................................................................................... 68
Hoop strains in carbon wrap of control square specimen #15 .................. 68

Hoop strains in carbon wrap of square specimen #16 during
freeze-thaw cycles ..................................................................................... 69
Compressive stress-strain curves for plain, round, control specimens ...... 73
Compressive stress-strain curves for plain, round specimens
subjected to 150 freeze-thaw cycles .......................................................... 73
Compressive stress-strain curves for plain, round specimens

subjected to 300 freeze-thaw cycles .......................................................... 74

xi

Fig. 3.24
Fig. 3.25
Fig. 3.26
Fig. 3.27

Fig. 3.28
Fig. 3.29

Fig. 3.30
Fig. 3.31
Fig. 3.32
Fig. 3.33

Fig. 3.34

Fig. 3.35

Fig. 3.36
Fig. 3.37
Fig. 3.38
Fig. 3.39
Fig. 3.40
Fig. 3.41
Fig. 3.42

Fig. 3.43
Fig. 3.44

Fig. 3.45
Fig. 3.46
Fig. 3.47

Compressive stress—strain curves and tensile hoop strain for glass-
wrapped, round, control specimens ........................................................... 74

Compressive stress-strain curves for glass-wrapped, round specimens
subjected to 150 freeze-thaw cycles .......................................................... 75

Compressive stress-strain curves and tensile hoop strain for glass-
wrapped, round specimens subjected to 300 freeze-thaw cycles .............. 75

Compressive stress-strain curves and tensile hoop strain for glass-
wrapped, square, control specimens .......................................................... 76

Compressive stress-strain curves for glass-wrapped, square specimens
subjected to 150 freeze-thaw cycles .......................................................... 76

Compressive stress-strain curves and tensile hoop strain for glass-
wrapped, square specimens subjected to 300 freeze-thaw cycles ............. 77

Compressive stress-strain curves and tensile hoop strain for carbon-
wrapped, round, control specimens ........................................................... 77

Compressive stress-strain curves for carbon-wrapped, round specimens
subjected to 150 freeze-thaw cycles .......................................................... 78

Compressive stress-strain curves and tensile hoop strain for carbon-
wrapped, round specimens subjected to 300 freeze-thaw cycles .............. 78

Compressive stress-strain curves and tensile hoop strain for carbon-
wrapped, square, control specimens .......................................................... 79

Compressive stress-strain curves for carbon-wrapped, square

specimens subjected to 150 freeze-thaw cycles ........................................ 79
Compressive stress-strain curves and tensile hoop strain for carbon-

wrapped, square specimens subjected to 300 freeze-thaw cycles ............. 80
Failure modes for square specimens under compression testing .............. 80
Failure modes for round specimens under compression testing ................ 81
Average compressive strength of round glass-wrapped specimens .......... 85
Average compressive strength of round carbon-wrapped specimens ....... 86
Average compressive strength of square glass-wrapped specimens ......... 86

Average compressive strength of square carbon-wrapped specimens ...... 87

Average compressive strength of round plain specimens ......................... 87
Average corrosion depths due to accelerated corrosion ............................ 91
Hoop strains in bonded, glass-wrapped specimens ................................... 99
Hoop strains in unbonded, glass-wrapped specimens ............................... 99
Hoop strains in bonded, carbon-wrapped specimens .............................. 100
Hoop strains in unbonded, carbon-wrapped specimens .......................... 100

xii

Fig. 3.48 Hoop strains in glass and carbon—wrapped specimens,

gages at the anodes .................................................................................. 101
Fig. 5.1 Column condition before surface repairs ................................................ 110
Fig. 5.2 Column condition after surface repairs ................................................... 110
Fig. 5.3 Corrosion probe and reinforcing bars for monitoring mass loss ............. 111
Fig. 5.4 Field installation of glass wrap to selected columns ............................... 111
Fig. 5.5 Field installation of carbon wrap to selected columns ............................ 112
Fig. 5.6 Completed installation of glass and carbon wraps .................................. 112

xiii

Chapter 1
Introduction and Literature Review

1.1 Introduction

Retroﬁtting bridge columns with jackets encasing the concrete has been
undertaken extensively in the West Coast, primarily for seismic rehabilitation. Steel
jackets have been used for the most part, but are expensive to install and require several
days for each column. The use of ﬁber reinforced polymers (FRP) wraps also has been
developed and numerous columns have been retroﬁtted with different kinds of wraps.
FRP wraps can be installed quickly (four to six columns per day), with minimal
interruption to trafﬁc flow. FRP wraps can be made of low-cost glass ﬁber, medium-cost
aramid (also known as Kevlar) ﬁber, or high-cost high- performance carbon ﬁber, and
polyester, vinyl ester or epoxy resins can be used as a matrix. Hexcel Fyfe Co. (Del Mar,
Calif.) is the ﬁrst composite fabricator to install a ﬁeld demonstration with the California
Department of Transportation, but other suppliers such as XXsys Technologies Inc.,

C. C. Myers Inc., Master Builders Inc., Mitsubishi Corporation, and Hardcore DuPont
Composites L.L.C. also are performing ﬁeld installations now.

The Pennsylvania Department of Transportation selected the Tyfo-S ﬁber wrap
system (by Fyfe Company) for a concrete-column repair project on Interstate 84 in
Scranton (Tanicone 1995). The New York State Department of Transportation used FRP
wraps on six severely deteriorated concrete columns of the Court Street Bridge (Tioga
County) in the summer of 1998 to protect against failures and to improve their durability

in a cost effective manner. Six suppliers of FRP column wraps participated in this project

(Alampalli et al. 1999). In California numerous columns have been wrapped with the

Tyfo-S ﬁber wrap system, and XXsys Technologies’ﬁlarnent winding for the bridges

under the jurisdiction of the California Department of Transportation. This was done

mainly for seismic strengthening. XXsys Technologies in conjunction with the Utah

Department of Transportation used its composite wrapping system in 1996 to strengthen

and extend the useful life of an existing bridge. The bridge has deteriorated from the

effects of more than thirty years of corrosion of the steel reinforcing bars. Structural

Preservation Systems of Baltimore, Maryland has used carbon wraps in 1997 to

strengthen parking garage columns in Charleston, South Carolina.

While advanced composite materials are relatively expensive, labor costs are
approximately 30% less than for conventional strengthening methods primarily because
of the light weight of the materials. The quick turnaround and the reduced labor costs
associated with FRP have reduced the overall cost making its use attractive for
rehabilitation and strengthening of concrete infrastructure.

Four types of FRP jackets are currently available:

0 Composite fgb_rics: The fabrics must be thoroughly saturated usually with two-part
resin during installation and are cured at room temperature. The mechanical
properties of the ﬁnal product are somewhat variable. The TyfoTM S Fiberwrap
System by Fyfe Company is of this type.

0 Prepreggated commsite fabrics (Prepregs): Prepregs are pre-impregnated, usually
with an epoxy resin, which results in good control over the mechanical properties of
the ﬁnal product. However, most prepregs must be stored in cold storage before

installation, and need to be thermally cured after installation. For column

applications, thermal curing can be achieved using heating blankets. There are few
commercial carbon ﬁber prepregs, such as MBrace and REPLARK marketed by
Master Builders and Mitsubishi respectively, that may be air cured.

o Filament winding: Jackets are produced by winding a continuous composite ﬁber
onto the resin coated column. Typically automated winding equipment is used and the
jacket is heat cured. XXsys Technolgies has retroﬁtted several columns in the West
Coast using this technology.

0 Precured s_hells: Precured jackets are formed around a mandrel of the diameter
matching the column to be jacketed, typically using a multi-axial, stitched, nonwoven
E-glass ﬁber fabric. The jackets have an axial slit that allows them to be opened and
placed around a column for installation. Field crews install the jacket by ﬁrst spraying
a urethane adhesive onto the column. Two workers can then snap on a 132 kg, 1.2 m
tall by 1.2 m diameter jacket segment. The adhesive and jacket installations are
repeated to obtain three to ﬁve plies of composite, with the axial slit in each
additional ply being staggered from that of the previous ply to avoid overlap. The
whole multi-layer jacket system is bound to the column while the adhesive cures,
creating an efﬁcient, labor-saving system. AS an alternative to the 3600 bands, smaller
arc segments can be installed by gluing each segment around the column with
sufﬁcient overlap. The smaller arc segments are easier to package and ship.

All of the jacketing systems described above are acceptable, and the advantage of
one over another would depend on their performance in Michigan’s harsh climate, their
cost and their availability. Precured shells are presently available only for Circular

columns.

1.2 Project Objectives

The main objective of this project was to investigate the suitability of using glass and
carbon FRP wraps to repair concrete columns damaged by corrosion in Michigan. The
following issues were investigated:

0 Strength loss of wrapped columns subjected to expansive forces, due to freezing and
thawing.

0 Strains in FRP wraps during freeze-thaw conditioning.

0 Increased strength due to conﬁnement provided by wraps.

o Magnitude of conﬁning pressure generated by wraps due to corrosion.

o Localization of strains near reinforcing bars in wraps bonded to the concrete and the
merit of using unbonded wraps.

0 Reduction in corrosion rate due to the use of FRP wraps.

1.3 Literature Review

1.3.1 Fiber Reinforced Polymers (FRP) for Infrastructure

Newly developed composite materials have been used recently in civil
engineering structures because of their superior mechanical properties as well as their
resistance to aggressive environmental conditions. In general composites can be deﬁned
as a combination of two or more materials, that are insoluble into one another, without
chemical interaction such that the properties of the combination is better than the
individual constituents (De Wilde 1988). Fiber reinforced polymers are made of two

constituent materials: polymer ﬁbers and polymer matrices.

1.3.1.1 Fibers

Fibers have the largest volume and are the load-carrying element of FRP
composites. Proper selection of the amount, type and orientation results in a composite
with the desired mechanical properties.

0 Glass ﬁbers: These are widely used. Molten glass can be drawn into ﬁne continuous
ﬁlaments. These can be fabricated into continuous ﬁbers, chopped strands, woven
fabrics and rrrilled ﬁbers. The strength of glass ﬁber is highly dependent on the form
in which the ﬁbers are used. Continuous ﬁbers have the highest strength and chopped
ﬁbers have the lowest strength. The average tensile strength for freshly drawn glass
ﬁbers may exceed 3500 MPa. Surface ﬂaws tend to reduce this value to between
1750-2100 MPa. The internal structure of a glass ﬁber is a three dimensional network
of different atoms. They behave in a linearly elastic manner until failure (Mallik
1993). Glass ﬁbers are available in a variety of forms suited for different applications.
The most common type is E-glass (which was initially developed for use in electrical
applications). Other types of glass ﬁbers that are used include S-glass (which has
approximately 25% greater tensile strength than E—glass but is more expensive), C-
glass (which was developed for application in corrosive environments), D-glass
(which has lower density and dielectric constant than the other types of glass ﬁbers),
AR-glass (alkali resistant), and ECR-glass (modiﬁed E-glass which provides
improved acid resistance).

0 Carbon and ggaphite ﬁbers: The difference between carbon and graphite is in the
molecular structure. In carbon, the bonding between layers is weak, so it has two-

dimensional ordering. Graphite is formed from carbon atoms, which are arranged in

crystallographical parallel planes of regular hexagons. Carbon ﬁbers are
commercially available in three basic forms: long and continuous tow, chopped (6-50
mm long), and milled (30-300mm long). Carbon ﬁber can also be woven into two-
dimensional fabrics of various styles. Graphite ﬁbers are mainly considered in high
strength composite applications. They show very high speciﬁc strength and stiffness.
Graphite has a higher tensile modulus than carbon. Generally, as the modulus of
elasticity increases, ultimate load and elongation decreases. Therefore, high modulus
graphite ﬁbers exhibit a lower strain at failure than high strength carbon. The tensile
strength and the modulus of elasticity of graphite and carbon are not temperature
dependent. These ﬁbers behave elastically to failure and are highly resistant to
aggressive environment. Their diameter is in the range of 5 to 10 microns.

0 Aramid ﬁbers: Made from aromatic polyamides, these have the lowest speciﬁc
gravity and highest speciﬁc tensile strength among all type of ﬁbers (Mallik 1993).
Due to its high tensile strength and modulus of elasticity, aramid was the ﬁrst organic

ﬁber to be used as a reinforcing ﬁber.

1.3.1.2 Matrices

The matrix is considered the secondary material in FRPs. Its major roles are
transferring stresses between the ﬁbers and protecting ﬁbers against the environmental
and mechanical conditions. The importance of the matrix in a composite is its effect on
interlaminar and in-plane shear strengths. It also provides support against buckling of the
ﬁbers under compressive loads.

Polymer matrices are divided into two categories:

0 Thermoplastic Polymers: Individual molecules are in a linear structural form. Weak
secondary bond holds these molecules together. Heat or pressure temporarily breaks
the bonds, which causes movement between the molecules. After cooling, the
molecules set into their new position. Thermoplastics have higher impact strength,
fracture and microcracking resistance compared to thermosetting polymers. Examples
of thermoplastic polymers include nylon and polyethylene.

o Thermosetting Polymers: Also known as resin. The molecules are joined together by
cross-links which leads to a more stable three-dimensional form that can not be
reshaped by heat or pressure. Epoxy, polyester, and vinyl ester are the most common

types of thermosetting polymers (Malek and Saadatmanesh 1996).

1.3.2 Durability of Concrete

Concrete is a porous material consisting of cement, water, ﬁne and coarse
aggregates and, possibly, admixtures. Cement and water react to form a hardened paste
binding together the coarse and the ﬁne aggregates. Voids are left in the originally water-
ﬁlled space between the cement grains, which are not ﬁlled with the hydration products
of the chemical reactions. These voids are known as capillary pores. They range in size
from approximately 5 nm to 1 mm. Capillary forces in such small volumes play an
important role in the durability characteristics of concrete.

The capillary pore volume is a function of two parameters: the water/cement ratio
of the paste, and degree of hydration of the cement (Pigeon and Pleau 1988). Table 1.1
gives the approximate capillary porosity of Portland cement paste as a function of the

water/cement ratio and the degree of hydration. The water/cement ratio and the degree of

hydration also have an inﬂuence on the average size of capillary pores. The average size
decreases signiﬁcantly with the degree of hydration, and the number of very large pores

decreases signiﬁcantly with lower water/cement ratios.

Table 1.1. Capillary porosity of Portland cement paste as a function of
the water/cement ratio and the degree of hydration.

 

 

 

 

 

 

Water/cement ratio Capillary POTOSitY (% volume)
(by mass) At 50% hydration At 75% hydration
0.4 31 13
0.5 39 28
0.6 46 36

 

 

 

 

Cement paste also contains a signiﬁcant volume of smaller pores that are called
gel pores. The hydrants have a very large speciﬁc surface area, which is covered with a
few layers of absorbed water. The gel pores correspond to the surface occupied by this
absorbed water. Unless the temperature is high or the relative humidity is very low, the
gel pores are always ﬁlled with water because the forces that bind water to the surfaces of
the hydrates are strong.

When ice forms in the pores, a 9% increase in the volume of water takes place
when water changes from liquid to solid. This volume expansion and the ﬂow of water,
as it is forced out of the pores, cause tensile stress to be generated in the paste. This is the
basic cause of damage to concrete due to freezing.

When concrete is air entrained, which is achieved by adding admixtures to the
mix, a very large number of closely spaced air voids develop. If these air voids are

sufﬁciently close, the pressure generated by the ﬂow of water out of the pores does not

cause any damage and water can freeze in these voids without generating internal

pressures in the concrete (Pigeon and Pleau 1988).

1.3.3 Environmental Effects on FRP Composites

Environmental factors such as extreme temperature ﬂuctuation and water
absorption can adversely affect the behavior of some polymer composite material. Water
absorption reduces the strength and stiffness of some polymeric composites by as much
as 30%. compared to dry material. Water absorption breaks down the interface between
the reinforcing ﬁber and resin matrix leading to loss of strength and rigidity. Cycles of
freezing and thawing tend to magnify the effect of water absorption (Gomez and Casto
1996). The Strength loss in glass FRP due to 300 cycles of freezing and thawing is

depicted in Figure 1.1

 

 

 

 

 

 

 

900
750 wand.ESUn:0 Cychﬁs
HT”

600 ,4 Isophthalic Polyester
ﬁg /// 0 Qwﬂes
:1 450 /,-_..- W——-»vmy1 Ester 3a) Cycles
*3 /%—-
.4 300 / “metrsopminue Polyester

//F KIleﬂes
150
0

 

 

 

 

 

 

 

 

 

0. 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Deflection (Inches)

Figure 1.1 Load vs. deﬂection curve for glass FRP compo’site specimens subjected to 300
cycles of freezing and thawing (reproduced from Gomez and Casto 1996)

9

While several studies have been conducted on the strength of columns wrapped
with FRPs, studies on durability under harsh environmental conditions such as freeze-
thaw, exposure to chloride, and degradation from exposure to ultraviolet light are much
fewer. Questions related to durability under harsh environmental conditions are
extremely important in Michigan, and it is important to provide answers to these
questions based on laboratory research prior to ﬁeld installation.

Chajes et al. (1994) investigated the durability of several composite systems
externally attached to concrete beams. One set of beams was exposed to cycles of
freezing and thawing, the second set was exposed to cycles of wetting and drying, while
the third set was unconditioned. Their results indicated that ﬂexural strength was lost due
to a degradation of the bond between the concrete and the external reinforcement.
Degradation of the composite material was not reported.

Sen et al. (1993) investigated the durability concrete beams pretensioned with
glass/epoxy FRP subjected to cycles of wetting and drying. Several specimens were
initially cracked to simulate pile—driving damage. Their results indicated extensive
damage in the glass FRP, leading to an unacceptable level of strength loss.

Gomez and Casto (1996) studied the effect of chloride and freeze-thaw on two
pultruded ﬁberglass all-composite systems. Both systems used glass ﬁbers, but one used
a vinyl ester resin while the other used a polyester resin. Samples were exposed to freeze-
thaw cycles while immersed in a 2% sodium chloride and water solution. Their results
indicated a loss of 22-32% in the ﬂexural strength and stiffness of the composite

materials.

10

Fyfe et al. (1996) studied the effect of prolonged ( 1000 hours) exposure of the
TyfoTM S System to ozone, salt water immersion, fresh water immersion, alkaline soil
burial, high temperature, low temperature, and Weatherometer aging. All tests were
performed according to ASTM standards, but only on the composite material, not on
wrapped concrete specimens. In general, no serious degradation was observed due to
prolonged exposure. However, a fundamental limitation of this study is that the
environmental conditions were not cycled, and hence durability against cyclic freeze-
thaw and chloride immersion cannot be assessed.

Toutanji and Balaguru (1998) studied the effect of wet-dry and freeze-thaw
conditions on the performance of concrete columns wrapped with two layers of carbon
and two layers of glass FRP composites. Three types of FRP wraps were used; two types
of carbon and one glass. Twenty four concrete specimens (76 mm diameter and 305 mm
long cylinders) were divided into three groups. Each goup consisted of eight specimens:
six conﬁned (two with each type of the FRP sheets) and two unconﬁned. The ﬁrst group
was used as virgin samples and the second was exposed to wet-dry cycling and the third
group was exposed to freeze-thaw cycling. Three hundred wet-dry and freeze-thaw cycles
were performed in a salt water solution. CFRP experienced no reduction in strength or
ductility due to wet—dry exposure, whereas samples with GFRP experienced reduction of
10% and 20% in strength and ductility, respectively. In the case of freeze-thaw exposure,
both CFRP and GFRP wrapped specimens experienced reductions in strength and
ductility. Strength losses were 28% and 19% for glass and carbon, respectively.

Ductility losses were 65% and 30% for glass and carbon, respectively. The specimens

11

subjected to freeze-thaw cycling also exihibited more catastrophic brittle failure as
compared with the unconditioned and the wet-dry conditioned specimens.

Rivera and Karbhari (1999) conducted tests on concrete cylinders wrapped with
glass and carbon FRP after subjecting them to 201 freeze-thaw cycles (between 225° C
and ~20 ° C). Three layers of carbon fabric and seven layers of glass fabric were used.
Wrapped specimens subjected to freeze-thaw cycling showed increased stiffness and
strength and more catastrophic failure compared to control specimens under ambient
temperature. In order to isolate the effects of temperature from those of excessive
moisture absorption, thawing was in air.

Murphy et al. (1999) investigated the effects of alkali exposure on the
performance of glass ﬁber reinforced composites. Glass ﬁber reinforced vinylester
coupons (two, four and six layers) were placed in solutions with pH and salt content
predicted by leaching out the concrete itself, new concrete (28 days) and old concrete (ten
years). In addition, a cementitious extract was prepared by collecting the solution that
formed after settling of the aggregates (while preparing the new concrete). The starting
pH level was approximately 12 and dropped to 8.5 in about 60-80 days for the new and
old concrete solutions. For the cementitious extract solution, the pH dropped to 8.5 in
about 250 days. The strength was degraded by 17 to 32% over a period of one year. The
coupons placed in the cementitious extract was degraded the most (32%) indicating that
reduction in strength cannot be attributed to pH levels alone, but rather to a combination
of alkaline salts from concrete and the pH levels present.

Almusallam et al. (2000) conducted tests on concrete specimens wrapped with

three layers of bi-directional glass FRP. Each group of specimens contained three GFRP

12

wrapped specimens and three unwrapped specimens. It was concluded that the
compressive strength of wrapped cylinders subjected to alkaline solution and alkaline
solution at elevated temperature (60 °C) exhibited lower increase in strength than the
other groups. The increase in strength was about 23% compared to 54% for wrapped
specimens at room temperature.

The Aerospace Corporation (Los Angeles, California) conducted extensive FRP
panel durability testing on various wrap systems (Steckel 2000). For the glass and carbon
systems used in this research study, they concluded that carbon panels are not affected by
conditioning while glass panels showed strength and strain reduction of up to 35% and
15% - 20% under 10,000 hours of humidity and salt water (or alkaline solution)
exposures, respectively. It should be noted that these conditions were not cycled. Only 20
freeze-thaw cycles were conducted and those had no effect on the FRP panels was noted.
Appendix A shows the results obtained by of the Aerospace Corporation for glass and

carbon FRP panels, respectively.

1.3.4 Corrosion of Reinforcing Steel

Corrosion is a natural process and is a result of the inherent tendency of metals to
revert to their more stable compounds, usually oxides. Most metals are found in nature in
the form of various chemical compounds called ores. In the reﬁning process, energy is
added to the ore, to produce the metal. It is this same energy that provides the driving

force causing the metal to revert back to the more stable compound.

13

Corrosion of the reinforcement reduces strength, durability, and service life of the
reinforced concrete structure. As the reinforcement corrodes, it expands causing cracking
of concrete and spalling.

An ASTM-sponsored study (Guttman and Sereda 1968) found the corrosion rate
in steel exposed to air at various locations varied from 0.033 to 0.058 mm per year in
Cleveland to 0.030 to 0.043 mm per year in Ottawa. For a #25 bar, these translate to
5.3% to 9.1% in Cleveland and 5.0% to 6.7% in Ottawa in 10 years. Similar rates have

been observed in Michigan steel bridges (McCrum 1994).

1.3.4.1 Factors Affecting Corrosion

The presence of chlorides, temperature, relative humidity, cover depth, and
concrete quality are the major factors affecting the rate of corrosion.

Chlorides can come from several sources. They can be cast into the concrete to
promote rapid hardening or they can diffuse from the outside. Chlorides can diffuse into
concrete as a result of sea salt spray and deicing salt. Chloride contamination of bridge
columns resulting from winter maintenance chloride (deicing salt) applications results in
the continuing deterioration of the steel reinforcement, which in turn causes bursting
forces emanating from the steel location outward to the periphery of the columns.

Minimum concrete cover and low quality of concrete (high water/cement ratio)
decrease the time needed for chlorides to reach the reinforcement. The time for corrosion
to start will therefore be decreased and the rate of corrosion will increase (Allen 1995).

Environmental conditions such as ambient temperature and relative humidity also

affect corrosion. The concentration of free chloride ions in the pore water increases with

14

temperature. In addition, corrosion reactions occur at a much faster rate with an increase
in temperature. The corrosion rate of steel was found to vary linearly with temperature.
The corrosion rates at 40° C was found to be almost double that at 0 ° C (Lopez et al
1993).

Concrete is alkaline. It contains microscopic pores with high concentrations of
soluble calcium, sodium and potassium oxides. These oxides form hydroxides, which are
highly alkaline when water is added. This alkaline condition leads to a passive layer
forming on the steel surface in the form of a dense, impenetrable ﬁlm, which if fully
established and maintained, prevents further corrosion of the steel. Chlorides act as
catalysts to promote corrosion. When there is sufﬁcent chloride concentration at the
reinforcing bar surface to break down the passive layer of oxide on the steel, the
corrosion process proceeds quickly.

Brockenbrough et al. (1985) conducted tests on stacks of Cor-Ten A steel and
carbon steel compressed together with spring washers and exposed (to the atmosphere) at
Monroeville, Pennsylvania and Kure Beach, North Carolina. The stacks clamped at low
pressures (21.5 and 43.5 kPa) showed large increase in thickness because of corrosion
product pressure after ﬁve and a half years of exposure. In addition, it was concluded that
there is an initial pressure threshold of about 1035 to 1380 kPa above which the rate of
corrosion is extremely small and no signiﬁcant expansion due to corrosion product

pressure takes place.

15

1.3.4.2 Volume Expansion Due to Corrosion of Steel

Corrosion of steel is an electrochemical process. The electrochemical potentials
to form the corrosion cells may be generated when cells are formed due to differences in
concentration of dissolved ion in the vicinity of steel, such as alkalies, chlorides, and
oxygen. As a result, some parts of the metal become anodic and the others cathodic.
The fundamental chemical changes occurring at the anodic and cathodic areas are as
follows (see Fig. 1.2-a).

Anode: Fe -> 2e ' + Fe 2*

Cathode: V202 +H20 +2e’-)2(OH)°

Cathode Process Anode Process
02+ ZHZO +4e‘»40t-l‘ Fe-bFe”+ 2e"

    

Moist Concrete 0:

 

 

 

 

 

 

 

 

 

on Electrolyte Fe (0le

F32 03 Surface Film F. (OH): 1

Steel
“(0141331120 1
1 1 1 1 1 1 1 1
O l 2 3 4 5 6 7

Current Flow Volume. cm’
(a) (b)

Figure 1.2. Expansion and cracking of concrete due to corrosion of the embedded steel
(reproduced from Mehta 1996)

16

The transformation of metallic iron to rust can result in increases in volume of up
to 600% (Mehta 1996), depending on the ﬁnal rust form (see Fig. 1.2-b). Since the
presence of both air and water is required for the corrosion activity to continue, column
wrapping may provide adequate conﬁnement that would minimize the entry of air and

water, thereby slowing down the continuation of corrosion.

1.3.4.3 Localized Corrosion

Corrosion of steel reinforcement inside bridge columns is not uniform. It is
dependent on the location of cracks and surface exposure (i.e. salt contact due to deicing
of roads). Further, the volume expansion due to corrosion is localized near the
reinforcement bars. This may exert strains on the wrap at localized areas. All or most of
the metal loss occurs at discrete areas (www.Corrosionsource.com 2000).

Pitting corrosion is highly localized corrosion occurring on a metal surface.
Pitting is commonly observed on surfaces with little or no general corrosion.

Pitting typically occurs as a process of local anodic dissolution where metal loss is
exacerbated by the presence of a small anode and a large cathode.

Crevice corrosion is another form of localized corrosion which may occur in
small areas of stagnant solution in crevices, joints and under corrosion deposits. Crevice
corrosion is the localized corrosive attack that occurs as a result of the occluded cell that
forms under a crevice on the metal surface. To prevent this type of corrosion, it is
recommended that crevices be closed with non-absorbent materials or a barrier to prevent

moisture penetration into the crevice be incorporated.

l7

1.3.5 Mechanical Properties of FRPs

Many FRPs have tensile strengths that exceed the strength of steel, but their
stiffness is generally lower than that of steel. When loaded along the ﬁber direction they

behave essentially linearly until fracture, and are therefore brittle by nature.

1.3.5.1 Stress Corrosion and Stress Rupture

The average ultimate tensile strength of freshly drawn glass ﬁbers may exceed
3500 MPa. However, surface ﬂaws tend to reduce the tensile strength to values in the
range of 1750 to 2100 MPa. Strength degradation is increased as the surface flaws grow
under cyclic loads. This is one of the major disadvantages of using glass ﬁbers in
applications where fatigue may be an issue. Sustained loads also cause surface flaws to
grow, resulting in reduced tensile strength. Figure 1.3 shows reduction of strength with

time for 15- glass ﬁber under different temperatures.

 

3000 “r’ 1 l

 

 

 

O

2 N
g 2500 .- -
.1: O
5
u
u
”3 1500' 500°C 4
Q)
3
m 1000- a
5 600°C
H ,

500 l 1 J

100 101 102 1o3 104

Load Duration (MIN)

Figure 1.3 Reduction of tensile strength of E-glass ﬁbers under sustained loads
(taken from Mallick 1993)

18

Stress corrosion refers to the characteristic property of FRPs in which the failure
strength under long term sustained loads in a chemical environment is lower than its short
term tensile strength. In air, this phenomenon is referred to as “stress rupture” (Sen et al.
1993). Stress rupture tests are usually performed by applying a constant tensile stress to a
specimen until it fractures completely. The time at which fracture occurs is termed the
“lifetime”. Creep, on the other hand, is deﬁned as the increase in strain with time at a
constant load level (Mallick 1993).

Glass, Aramid, and Boron ﬁbers and their composites exhibit failure by stress
rupture. Carbon ﬁbers, on the other hand, are relatively less prone to stress rupture
failure. In order to prevent stress rupture in glass FRP for a period of 10, 30, and 50
years, the sustained strains in the GFRP should be less than about 0.35 Eu, 0.32 an, 0.30
8“, respectively (Sen et al.1993, AC12000).

The relationship between the sustained stress (or strain) and the logarithm of time
is approximately linear as shown in ﬁgure xx. ACI Committee 440R recommends the use
of a safety factor of 1.67 and hence a safe level of sustained strain to prevent stress

rupture is about 0.2 8,.

1.3.6 Effect of Confinement

Lateral conﬁning pressure increases the strength and ductility of concrete in the
axial direction. The stress-strain curves of conﬁned concrete show a remarkable energy
dissipation characteristic. Such behavior is of great importance as it can prevent
catastrophic failure of highway bridges or high-rise buildings under overload conditions.

Figure 1.4 shows stress-strain curves for conﬁned and unconﬁned concrete.

19

Compressive Stress

 

 

 

Compressive Strain

Figure 1.4 Stress/strain curves for conﬁned and unconﬁned concrete (Mander, et 31.1988)

The relationship between conﬁned compressive strength (f 65'), unconﬁned
compressive strength (f c'), and the lateral stress in core concrete produced by conﬁning
.pressure (f,) is
fcc'=fc'+kfr (1.1)
The average value of the conﬁnement effectiveness coefﬁcient k was found to be equal to
4.1 (Richart, et al. 1928).

For circular column with spiral steel, the conﬁning pressure f, is given by (Nilson and

Winter 1991)

2.4wa (1.2)
d s

 

20

where

A,,, = the cross-sectional area of the spiral steel

dC = the outside diameter of the spiral

S = the pitch of the spiral.

fy = spiral steel yield stress
Eq. (1.2) is calculated assuming that the spiral steel reaches its yield stress fy before the
column eventually fails.

FRP materials are essentially linear elastic up to the point of fracture, while steel
shows an elastic-plastic behavior. The stress-strain curve of the conﬁning materials
affects the failure mode of the conﬁned core. Figure 1.5 shows stress-strain curves for

A36 steel, E-glass and carbon ﬁbers.

 

 

 

 

Figure 1.5 Stress-strain curves for A36 steel, E-glass and carbon ﬁbers

21

Conﬁnement of concrete columns with FRP wraps offers many advantages in
comparison to other conﬁnement methods. Composite material with their high strength
and high stiffness to density ratios allow for ﬁeld installation with minimal workforce
and disruptions to trafﬁc. Preliminary testing of concrete columns wrapped with FRPs
has shown that the conﬁnement provided results in improved compressive strength and
ductility. Picher at el. (1996) conducted a series of tests on conﬁned circular, square and
rectangular concrete specimens. Axial loads were applied to concrete specimens
wrapped with different orientation of carbon FRP wraps. It was concluded that
conﬁnement of concrete cylinders with CFRP wraps improves their compressive strength
and ductility (up to 41% axial strength increase and about 500% axial strain increase for
cylinders conﬁned with three layers of carbon sheets with ﬁbers oriented in the hoop
direction) compared to unconﬁned specimens. Although axial stiffness decreases with
the increase of ﬁber angle orientation, ductility and modes of failure remain the same.
Wrapping of square and rectangular specimens improves ductility but to a lesser degree
than that observed for cylindrical specimens. In the case of square and rectangular
specimens, it was found that rounding the comers greatly improved the compressive
strength.

Tests on round and rectangular specimens and full scale columns wrapped with
glass and carbon FRPs was conducted by Kestner at e1. (1997). They found that
enhancement in axial strength and deformation are proportional to jacket strength and
stiffness. It was found that due to the ineffectively conﬁned concrete region in the square
cross sections, the jackets provided to square cross sections were not as effective as those

provided to circular cross sections. A shape factor, K}, was used to account for the

22

ineffectively conﬁned regions of concrete within the rectangular and square cross
sections.
From mechanics of thin walled cylinders, the conﬁning pressure in a conﬁned
column can be determined to be
f,=2(f,tn)/D (1.3)
where
I: the thickness of the wrap per layer
f, = the circumferential stress of wrap
n = number of wrap layers
D = the diameter of the concrete cylindrical column
The maximum conﬁning pressure, fm is determined by the ultimate tensile strength of the
wrap fm given by
fm=2(ﬁurn)/D=2(E£mtn)/D (1.4)
where
E = modulus of elasticity of the wrap
8n, = wrap ultimate strain
Substituting fm from Eq. (1.4) into Eq. (1.1) yields the maximum compressive strength
due to conﬁnement with FRP wraps.
More generally, for round and rectangular cross sections, the maximum conﬁning

pressure may be expressed as (Restrepol and DeVino 1996)

fm=0.5 Kgijaut (1.5)

23

where
p,- = 4n/D for circular columns
p,- = 2n(d+b)/db for rectangular columns
n = number of layers of wrap
D = overall diameter of circular column
b = overall width of rectangular column
d = overall depth of rectangular column
K} = shape factor determined as the ratio of effectively conﬁned concrete to the
gross area of the section
The shape factors for circular and rectangular are

Circular: K} = 1

1-[(b—2r)2+(d—2r)2]- p
1"P

 

Rectangular: K} =

where
r = radius of rounded corners
p = longitudinal reinforcement ratio of cross section
Fig. 1.6 shows the effectively conﬁned area of a rectangular cross section.
For circular columns, substituting K, = 1 in Eq. (1.5) will yield Equation (1.4), while for
rectangular columns

fm = x; n Eem t (d+b)/db (1.6)

24

      
   
 

  

A: 2 area of effectively
confined concrete
core

ineffectively conﬁned
concrete (typ)

 

FRPC jacket

An = total cross-sectional area

Fig. 1.6 Illustration of effectively conﬁned area of a rectangular cross section

1.3.7 Repair of Corrosion-Damaged Columns Using FRP

A review of research conducted to study the effectiveness of FRP wraps in
repairing corrosion-damaged columns indicates that FRP wraps have the following
advantages:

0 They provide a barrier to oxygen, moisture and chlorides.
0 They have high strength and stiffness and provide conﬁnement and ductility to the
concrete structure.

Debaiky and Green ( 1999) are investigating the suitability of using FRP wrap to

rehabilitate corroded concrete structures. The focus of the experiment is to assess what

25

happens to the corrosion process after the structure has been rehabilitated with FRP
wraps. The experiment consists of 12 reinforced concrete columns (300x 1200 mm) with
chlorides cast in the concrete cover. The columns are placed in a water bath to initiate
corrosion. After the initial corrosion phase, the columns will be wrapped with FRP
sheets. The “natural” accelerated corrosion environment will then be continued.
Monitoring corrosion will be through the use of a half-cell potential. This project is still
underway.

Pantazopoulou et al. (1996), Michniewicz (1996), and Lee (1998) from the
University of Toronto conducted several projects using FRP for repair of corrosion
damaged columns since 1993.

Two concrete mixes were used to construct the 150x300 mm cylinders. High
density concrete was used for the end caps, and high porosity concrete contaminated with
2.6% N aCl by weight of cement was used for the middle 210 mm of the cylinder height.
The corrosion was electrically accelerated using a 6 volt potential while the cylinders
were placed in 50 mm of 2% chloride solution. Damage was signiﬁcant after 150 days.
Several repair techniques were considered such as conventional patching, epoxy coating
and wrapping with plastic foil. In addition, glass FRP wrap also was used. The most
effective repair method was the use of a conventional patch with two layers of glass FRP
(GFRP) wrap. The strength was fully restored and ductility was doubled compared with
an undamaged specimen.

Ten large-scale circular columns also were built and corrosion was accelerated in
a similar manner (voltage ranged from 3 to 15V). The columns were repaired using

different methods ranging from:

26

0 Surface cleaning, non shrinkage grout, epoxy coat, and 2 layers of GFRP wrap to

o Epoxy coat and 2 layers of GFRP wrap only.

Compression tests results showed only columns that were repaired using the ﬁrst method,

surface cleaning, non shrinkage grout, epoxy coat, and 2 layers of GFRP wrap, had the

capacity of an undamaged column.

Another project was also conducted at the University of Toronto. Seven large
scale columns were corroded using a 12 V ﬁxed potential and wet-dry cycles consisting
of one day wet and 2.5 days dry. After signiﬁcant corrosion, they were wrapped with one
layer of carbon FRP (CFRP) sheet. The results showed that using the CFRP wrap
increased the strength of the corroded specimens by 28% compared to the unwrapped
specimens. The axial deformation at failure was six times that of corroded unrepaired
specimens.

It should mentioned that the above research projects conducted at the University
of Toronto did not investigate the effect of continuing corrosion on the FRP
rehabilitation.

At the University of Sherbrooke, the following ﬁeld applications have been
conducted (Rochette et al. 1996, Demers et al. 1996, and Kenneth et al. 1998):

0 Corrosion-damaged circular building columns were repaired with GFRP after the
column section was restored with cement grout. The repair took place in November
1995, the resin cured properly despite the cold temperature.

0 Eighteen (18) circular bridge columns were repaired in August 1996. Five columns

were wrapped with GFRP, four with CFRP, and three were repaired using

27

conventional material. Axial deformation and circumferential expansion were
monitored.

0 The concrete pier on the Champlain bridge in Montreal was repaired in 1997. The
pier received nine layers of GRFP wrap in the four meters above its base.

The University of Minnesota is currently monitoring the rehabilitation of
corrosion-damaged columns for a bridge near Minneapolis (Debaiky and Green 1999).
The set up is as follows:

0 Electrochemical chloride extraction (ECE) was used on three columns and then one
column was wrapped with CFRP sheets (supplied by Hexcel-Fyfe Company), the
second was wrapped with CFRP sheets (supplied by Tonen Corporation), and the
third was wrapped with chopped glass sheets.

0 ECE was used on three columns and were then sealed without wrapping.

0 Three columns were wrapped similar to the ﬁrst group without the ECE.

0 Three columns were left as control specimens.

Corrosion will be monitored for ﬁve years. Chloride sampling will be conducted once a

year. This project started in 1997 and is still underway.

In order to study the deterioration and evaluate different repair techniques, an
FRP column wrap project was initiated by the New York State Department of
Transportation in 1998 (Halstead et al. 2000). Six severely deteriorated concrete columns
of the Court Street Bridge (Tioga County) were wrapped using six different wrap systems
(by different suppliers). A comprehensive testing program to evaluate the effectiveness of
FRP column wrapping was implemented. Prior to installation of the FRP wraps, three

corrosion probes were embedded in each column. Additional monitoring equipment was

28

installed on each column and data collection started in September 1998. Strain sensors
were installed on the surface to measure the effect of continued corrosion on the wraps.
Concrete humidity and temperature are also being monitored. The corrosion probes use
linear polarization to monitor instantaneous corrosion rates. Data is being collected at

three month intervals. This non-destructive testing is scheduled to continue for ﬁve years.

29

Chapter 2
Description of Experiments

2.1 Laboratory Testing

2.1.1 Stiffness and Strength of Glass and Carbon FRPs

A 4-ply composite Tyfo-S ﬁber glass/epoxy sheet and a 2-ply Tonen
carbon/epoxy sheet were fabricated at MSU on 9/30/97 and 10/21/97 under the
supervision of the respective composite vendors. After the vendor-recommended curing
periods of ﬁve to seven days, these specimens were tested under direct tension at the
MSU Composite Material and Structures Center to check the moduli against the vendor-
recommended values. Table 2.1 shows a comparison between actual tested and
recommended moduli, thickness, effective axial stiffness per unit width (equal to
modulus x thickness) and ultimate strain. Although the thickness and the moduli varied
from the vender recommended values, especially for carbon, the effective axial
stiffnesses which control conﬁnement and behavior were within 96% and 88% of the
recommended values for glass and carbon FRPs, respectively. Master Builders specify
the thickness and modulus of their carbon FRP based on the ﬁber properties only which

is why their values differ signiﬁcantly from the measured values.

30

 

 

 

 

 

 

 

 

 

 

 

Table 2.1 Vendor recommended and measured wra 3 properties
Thickness Modulus Effective Ultimate
Wrap (mm) (MPa) Stiffness Strain
Type (N/mm)
Vendor Recommendations
Glass 1.016 27600 28042 .02
Carbon 0. 165 227700 37590 .015
As Measured
Glass 1.227 2201 1 27000 .020
Carbon 0.625 53061 33150 .014

 

 

 

 

 

 

2.1.2 Strain Expected in Wraps Due to Corrosion

The transformation of metallic iron to rust can result in increases in volume of up

to 600% (Mehta 1996), depending on the ﬁnal rust form. (Table 2.2 shows sample

calculations for the volume expansion for some rust products.)

Rust Forms: FeO, R304, R203, Fe(OH)2, Fe(OH)3, Fe(OH)3 3H20

Table 2.2. Volume expansion for some rust products

 

 

 

 

 

 

 

 

 

Rust Form Density (D) Mol.Wt. Volume(MW/D) Vol.Rust/Vol.Fe
Fe 7.86 56 7. 12
FeO 5.7 71.85 12.6 12.6/7.12:1.7
Fe3O4 5.18 231.54 44.70 44.7/(3x7. 12) =2
Fe(OH)2 3.4 89.86 26.43 26.43/7.12:3.71

 

 

Table 2.3 shows the strain developed in the wrap due to corrosion in a 915 mm

diameter column with longitudinal steel ratios of 3%, 2%, and 1%, and ties spaced at 152

mm and 305 mm. Corrosion rates by cross section of 5% in the longitudinal

reinforcement and 20% in the lateral reinforcement over a 10 year period are assumed
(Martin and Schieles 1969). For cases one and two it is assumed that the volume of rust is

three times and six times, respectively, the volume of the corroded steel. The table shows

31

 

that the strain in the wrap is within 0.30 a“ = 0.6% for GFRP (which is the sustained
strain limit in order to prevent stress rupture in glass FRP for a period of 50 years) for all
situations considered, indicating that stress rupture should not be a problem. If carbon

FRP is used, then stress rupture is not an issue and the strains in Table 2.3 are below the

rupture strain of 8,, = 1.5%. Hence the use of CFRP is also feasible.

Table 2.3 Strain in column wrap due to steel corrosion after 10 years

 

 

 

 

 

 

 

 

 

 

 

 

 

Column Spacing of Steel ratio of Strain in wrap
dia. (mm) tie (mm) cross section Case 1 I Case 2
#13 tie

3% 0.394% 0.450%

305 2% 0.260% 0.327%."

915 1% 0.145% 0.203%

3% 0.414% 0.530%

152 2% 0.290% 0.410%

1% 0.160% 0.284%

 

 

 

 

 

 

TSample calculation for this row is shown in Appendix 3.1.

2.1.3 Freeze-Thaw Test

Strength and durability tests were carried out on wrapped circular (diameter of
152 mm by 305 mm high) and square cylinders (152 mm by152 mm by 305 mm high).
The primary purpose of the tests was to determine the endurance of the jackets under
simulated cyclic environmental conditions, with strength considerations being secondary.
An internal bursting force similar to that produced by corroding steel was induced. This
was done by fabricating cylinders with a hole in the longitudinal direction and ﬁlling it
with an expanding cement known as Bristar (used for silent demolition). Chloride was

impregnated into the cylinders during casting in order to simulate deteriorated concrete

32

(11 kg of NaCl/m2 was used). Strength tests were carried out on plain control cylinders as
well as wrapped test specimens before and after freeze-thaw conditioning.

Climate data for Lansing, Michigan indicates that there were 58 days in the
1993/ 1994 year when the temperature cycled above and below 32° F and 78 days for the
1992/1993 year. Based on this and ASTM C666 speciﬁcations, 150 and 300 freeze/thaw
cycles were used in the freeze-thaw conditioning. Subsequent to freeze/thaw cycles, the
compressive strengths were compared against those of wrapped control specimens that
were not subjected to freeze-thaw cycles. In addition, unwrapped plain concrete
specimens also were exposed to freeze-thaw cycles to establish loss of strength in
concrete alone due to freeze/thaw conditioning.

Table 2.4 shows the test matrix used for the freeze/thaw laboratory testing. A
total of 60 specimens were involved in the testing, of which 30 were subjected to
freeze/thaw. Specimen and gage numbering is shown in Table CI in Appendix C.

A water/cement ratio of 0.4 was used; the 28-day strength for the freeze-thaw specimens
was 37700 kPa.

A half bridge conﬁguration was used for reading the strain gages, and temperature
correction was done by using dummy gages mounted on glass and carbon FRP panels
which also were located in the freeze-thaw machine. However, thermal contraction and
expansion of the FRP panels on which the dummy gages were mounted had to be
determined and compensated for. The dummy gages also were used when reading strains

on control specimens not subjected to freeze-thaw.

33

Table 2.4 Freeze-thaw laboratory testing matrix

 

 

 

 

 

 

 

 

 

 

 

No. of Specimens
Specimen Type Conditioning Glass Carbon
Unwrapped Wrap Wrap
Round None 3 3 3
Square None 3 3
Round None 3 3 3
Swan None 3 3
Round 150 cycles of freeze-thaw 3 3 3
Sgyge 150 cycles of freeze-thaw 3 3
Round 300 cycles of freeze-thaw 3 3 3
Square 300 cycles of freeze-thaw 3 3

 

2.1.3.1 Mold Fabrication
Prism molds

Nine 152 mm x 152 mm x 305 mm PVC molds were fabricated. Each is
composed of ﬁve panels (four sides and one bottom). The short side panels were
fabricated with a center hole that was 38 mm in diameter. A 38 mm steel rod was placed
in the hole during casting. Dow Corning release agent was applied to the steel rod and a
plastic sheet was then wrapped around it to aid removal after the concrete sets. The
center hole was later ﬁlled with Bristar (a form of an expanding grout). Fig. 2.1 provides

a picture of the prism mold.

 

Fig. 2.1 Prism mold used for freeze-thaw test specimens

34

 

Cylindrical Molds

Nine 152 mm x 305 mm steel cylindrical molds were fabricated. Steel base plates
and wooden top plates with a 38 mm diameter center hole were used. As with the prism
molds, a steel rod was inserted at the middle of each cylinder to create a hole in which

Bristar was later inserted. Fig. 2.2 provides a picture of the cylindrical mold.

 

Fig. 2.2 Cylindrical mold used for freeze-thaw test specimens

2.1.3.2 Bristar Calibration

The Bristar mix used to create an internal bursting force in cylinders had to be
calibrated to yield the appropriate pressure when set. Nine 152 mm x 305 mm steel tubes
were ﬁlled with concrete and a 38 mm diameter hole was fabricated in the center of each.
Each steel tube was mounted with two strain gages located diametrically opposite each
other at mid-height on the exterior surface.

After the concrete was allowed to set, Bristar mixes with different water/Bristar
ratios were poured into the center hole. The intent was to calibrate the water/Bristar ratio

so that a conﬁning pressure in the steel tube similar to that developed by corrosion-

35

induced expansion could be generated. The desired conﬁning pressure was based on the
strains in composite wraps due to the expected volume expansion in a bridge column
caused by corrosion (see Table 2.3).

Table 2.5 shows the strains that would be induced in the steel tube used for
calibration by conﬁning pressures expected to be generated by corrosion in composite
wrapped columns (for a wrap strain of 0.531%). These strain, selected from Table 2.3,
corresponds to a wrap strain that would be generated in a 915 mm diameter column due
to steel corrosion after 10 years when the steel ratio by cross section and tie spacing are
3% and 152 mm, respectively. A steel ratio of 3% and tie spacing of 152 mm are
conservative. Sample calculations on how the values in Table 2.5 were obtained are

shown in Appendix B.2.

Table 2.5 Internal pressure generated by corrosion for wrg) strain of 0.531%

 

 

 

 

 

 

 

 

Wrap Number Pressure Strain in
of layers kPa steel jacket

2 3763.6 0.029%

Glass 3 5644.9' 0.045%‘
4 7527.2 0.060%

1 23 10.4 0.018%

Carbon 2 4620.8 0.037%
3 693 1 .2 0.055%

 

 

 

 

 

 

Sample calculation for this row is shown in Appendix B.2.

Figs. 2.3 and 2.4 show the strain generated in the steel jacket when the
water/Bristar weight ratio was 400g/1000g and 500g/1000g, respectively. The maximum
strain of over 0.062% generated in the steel calibration jacket by the 400g/1000g
water/Bristar ratio corresponds to an internal pressure of over 7600 kPa, which is larger

than all the pressures shown in Table 2.5. However, the maximum strain of about 0.038%

36

generated in the steel calibration jacket by the 500g/1000g water/Bristar ratio
corresponds to an internal pressure of about 4830 kPa. This is closer to the pressures
expected due to corrosion in columns (with wrap strain of 0.531%) wrapped with 3 layers
of ﬁberglass and 2 layers of carbon. The 500g/1000g water/Bristar ratio is quite dilute

and a higher water content is not feasible. Therefore, a water/Bristar ratio of 500g/1000g

was used for all specimens requiring Bristar.

 

0

§
0
0

Strain us
§

 

 

 

01234567891011121314
runway)

Fig.2.3 Strain in steel tube for water/Bristar ratio of 400g/1000g

 

 

 

 

 

 

o 2 4 6 e 10
Time (day)

Fig.2.4 Strain in steel tube for water/Bristar ratio of 500g/1000g

It was decided that three layers of glass and two layers of carbon would be used in
the freeze/thaw tests, and that the Bristar would be prepared to generate a pressure of

about 4830 kPa. The strain in the glass wrap would then be about 0.45% while that in the

37

carbon wrap would be about 0.55%. Variations in these values occurred because it was
not possible to control the Bristar pressure precisely. It was felt that using more than three
layers of glass and two layers of carbon would unnecessarily increase the cost of the
wraps.

Since Bristar is highly porous, and water absorption with subsequent freezing and
thawing within the hole containing Bristar was undesirable, the ends of the specimens

were coated with epoxy prior to the freeze/thaw tests.

2.1.3.3 Chloride Content
Sodium chloride (NaCl) was used exclusively in freeze/thaw and accelerated
corrosion tests to contaminate concrete with chloride ions. Some examples of NaCl
concentrations found in the literature are:
1. Arya and Said-Shawaqi (1996) conducted tests on concrete prisms. Concrete was
dosed throughout by either 2%, 3% or 4% Cl' ions by weight of cement.
For concrete mix data used by the Michigan Department of Transportation (MDOT)
and 4% Cl' ions by weight, this translates to 22.12 kg NaCl/m3 of concrete:
0 336 kg cement/m3 of concrete X 0.04 = 13.44 kg Cl'/m3 of concrete
. 13.44 kg cu m3 of concrete x (58.5 NaCl/35.5 cr) = 22.12 kg NaCl/m3 of
concrete.
For 3%c1' and ma this translates to 16.59 and 11.06 kg NaCl/m3 of concrete
respectively.
2. Yamato, et al. (1987) found that the chloride content investigated in an off-shore

concrete bridge was about 3.0% N aCl by weight of cement at a point 2 cm in from the

38

vertical sides of a girder. Following calculations as in item 1, this translates to 22.69
kg N aCl/m3 of concrete.
A 2% Cl' ion by weight of cement was used in the freeze/thaw and accelerated

corrosion tests. This translates to 11 kg NaCl/m3 of concrete.

2.1.3.4 Strain Gage Placement

Strain gages were used to monitor wrap hoop strains during freeze/thaw tests on
six round specimens and six square specimens (three specimens for each type of wrap),
and four control specimens (one for each type of wrap and specimen shape). Each
specimen was ﬁtted with two strain gages oriented in the circumferential direction and
placed opposite each other at mid-height.

Half of the specimens were wrapped with three layers of Tyfo-SEH Glass
composite and the other half were wrapped with two layers of MBrace (Tonen) Carbon
composite. A total of 16 specimens (eight with each type of wrap system) that were to
undergo 300 cycles of freeze-thaw were ﬁtted with strain gages (two strain gages per
specimen). The gages were coated with wax and silicon for moisture and mechanical
protection.

The expansive nature of Bristar caused the specimens to expand in the hoop
direction as desired. An undesirable side effect was simultaneous expansion in the
longitudinal direction. This caused the specimens with the carbon wrap to split across the
cross sectional area since the carbon wrap contained no longitudinal ﬁbers. The glass
wrap had Kevlar strands embedded in the longitudinal direction, which prevents these

specimens from splitting.

39

Considerable effort was devoted to devising a system for releasing the vertical
expansion of Bristar using a greased aluminum tube. However, after several unsuccessful
trials it was determined that the vertical stresses caused by Bristar could not be totally
eliminated. In order not to risk having the carbon-wrapped specimens fail while in the
freeze-thaw machine, additional longitudinal reinforcement was provided to the carbon-
wrapped specimens. This was done by strengthening with strips of carbon in the
longitudinal direction. The strain gage reading should not be affected since the strips are
placed adjacent to the gages but not above them, and the longitudinal strips should not
provide any additional conﬁnement. Carbon-wrapped specimens subjected to freeze-thaw
as well as carbon-wrapped control specimens were ﬁtted with vertical strips.

Initial strain gage readings were taken prior to pouring Bristar in the center hole
of each specimen. After the initial expansion period of the Bristar, which is about one
week, an epoxy compound was used to cap the top and the bottom of the center hole. It
was not possible to control the Bristar pressure precisely. The average strain in the wraps
after the addition of Bristar to the specimens measured before starting the freeze/thaw
testing varied from:

o 0.31% to 0.60% for Glass with an average of 0.47%
o 0.24% to 0.68% for Carbon with an average of 0.48%

The freeze/thaw machine was set to subject the specimens to freeze/thaw cycles
according to ASTM C666 Procedure B, with freezing in air and thawing in water. Strains
were monitored during the freeze/thaw tests. A half bridge conﬁguration was used for
strain measurements, with dummy gages mounted on FRP panels located inside the

freeze/thaw machine so that strains due to temperature variations were eliminated.

4O

Considerable effort was required to properly adjust the freeze/thaw machine.
Since some of the specimens were wrapped with composite wrap systems and others
were not, careful and precise calibration was needed to control the freeze and thaw
temperatures. The wrapped specimens took longer to reach -17.8°C (end set point for the
freeze cycle) and 4.4°C (end set point for the thaw cycle) than the unwrapped specimens.
A plus or minus 1.7° tolerance is allowed at the upper and lower set points by ASTM
C666. After a few trials, it was established that an ideal sump water temperature of 7.2°C
would ensure that all specimens attain temperatures of
-17.78i1.7°C at the end of the freeze cycle and 4.4i1.7°C at the end of the thaw cycle
according to ASTM C666. Temperatures at the center of control specimens were
monitored for both unwrapped and wrapped specimens.

Specimens prepared as mentioned above were then placed in the freeze/thaw
chamber for 150 and 300 freeze-thaw cycles. The strains were measured throughout this
period. Two switch boxes were fabricated to facilitate reading of the strain gages during

the freeze/thaw and accelerated corrosion tests. A switch box is shown in Fig. 2.5.

41

 

Fig. 2.5 Strain measurement instrument (left) and switch box (right) used for
strain gage reading
2.1.3.5 Compression Testing

Considerable preparation was required prior to compression testing of the freeze-
thaw specimens. Two special ﬁxtures were manufactured to facilitate the testing.

Both end surfaces of each specimen had to be capped in order to provide two
perfectly parallel contact surfaces for load application. Sulphur is commonly used for
capping. The standard ﬁxture used to align specimens vertically and cap the ends is not
effective for wrapped specimens. The standard ﬁxture requires specimens with smooth
sides, but FRP wraps make the sides of wrapped specimens uneven. The standard ﬁxture,
therefore, does not assure parallel end surfaces after capping. A special ﬁxture was
fabricated to enable capping of wrapped specimens. The new ﬁxture could be used with
cylinders and square prisms, and minimized the physical labor required to lift up a
specimen, pour melted sulphur on a plate, and lower the specimen onto the plate to install

the end cap. This ﬁxture is shown in Fig. 2.6.

42

The standard ASTM compressometer (ﬁxture used to measure the axial strain
during compression testing) cannot be used with square prisms. A new compressometer
was fabricated for use with the square prisms. This new compressometer is shown in

Fig. 2.7.

 

Fig. 2.6 Capping Fixture Fig. 2.7 New Compressometer

The following data was gathered every six seconds using a data acquisition
system:
0 The compressive load and axial strain on all specimens using a load cell and a
compressometer ﬁtted with an LVDT, respectively.

0 The hoop strain in wraps for specimens ﬁtted with strain gages

43

2.1.4 Accelerated Corrosion

Accelerated corrosion tests using an electrochemical cell were conducted to study
the effect of conﬁnment on the progression of corrosion in the reinforcing steel within a
reasonable time frame. In addition, the hypothesis that FRP wraps slow down corrosion
by reducing permeability of water and oxygen or that they inhibit corrosion by
developing sufﬁcient conﬁning pressure (Brockenbrough at e1. 1985) was to be
evaluated. The conﬁning pressure was monitored using the strain readings of the wraps
during the accelerated corrosion tests. In addition, the rate of corrosion was indirectly
measured to determine if conﬁnement had any effect on corrosion activity.

The test used by Detwiler (1991) on lollipop specimens was adopted for use with
four #13 steel reinforcing bars cast in 152 mm diameter by 305 mm high concrete
cylinders. The specimens were initially immersed in salt water (with 3% NaCl) at room
temperature (approximately 20 °C) and connected to a power source so that two of the
steel rods became anodic and the other two became cathodic. Figure 2.8 shows the wiring
diagram used for connecting specimens to the power supply in order to accelerate
corrosion of the reinforcement. The exposed ends of the steel bars were protected against
crevice corrosion by using a Teﬂon tube tightened with a nut and by covering the
exposed end with silicon rubber. This forms a barrier to prevent moisture penetration into
the crevice (the interface between the steel bars and the concrete surface).

Two bars were used as anodes and two bars were used as cathodes to keep the
corrosion products within the specimens as in natural corrosion. When an external
cathode is used, the corrosion products tend to migrate out of the cylinder. In addition,

since a conductive medium must be provided during accelerated corrosion testing, the

44

cathodes must be placed inside the wraps, because the specimens were not continually
immersed in water. Corrosion was induced on only two of the four bars in each specimen
(i.e., at the anodes). Table 2.6 shows the corrosion level required in two bars for various
volume ratios (i.e., vol. of rust/vol. of corroded steel) to induce a hoop strain of 0.531%,
which is the anticipated strain in the wrap due to steel corrosion after 10 years (see Table
2.3). Two layers of carbon and three layers of glass were used for the accelerated

corrosion test. This was the same number used for the freeze-thaw test.

 

 

 

 

 

7Cathodes

\L‘\‘1 ‘U

’Anodes
(site of corrosion)

305 mm ._~~»o~-""

 

 

 

 

 

 

 

 

 

 

Figure 2.8 Wiring diagram for accelerated corrosion specimens

In order to simulate road column exposure to rain and water spray due to passing
trafﬁc, the concrete specimens were subjected to cycles of wetting and drying. The
specimens were soaked in salt water for one hour each day and the water was then

drained.

45

Table 2.6 Corrosion level required in two bars to induce a strain of 0.531% in the wrap*

 

 

 

 

 

 

Wrap Strain % Vol. Ratio Percent Corrosion °/o
3 38.1
Any number of layers 0.531 4 25.4
Carbon or glass 5 19.1
6 15.3

 

 

 

 

 

*Two #13 bars in 6" (152 mm) diameter cylinder

The rate of corrosion was measured to determine if conﬁnement has any effect on
corrosion activity. The corrosion rate was measured using the ASTM Gl (ASTM 1990)
test. The hoop strain generated in the wraps due to corrosion induced expansion was
monitored using strain gages oriented in the circumferential direction.

A water/cement ratio of 0.6 was used for the accelerated corrosion specimens to
promote capillary porosity, which in turn would aid the corrosion process (see Table 1.1).
The 28-day strength for the corrosion specimens was 20353 kPa.

Four unwrapped dummy specimens were subjected to accelerated corrosion using
different resistors to vary the current intensities. These specimens were closely monitored
to determine when cracking initiated. Based on these trials, resistors of 8 Q were used

with a 12 V power supply.

2.1.4.1 Corrosion Prior to Wrapping

Samples were exposed to accelerated corrosion while submerged in salt water
prior to applying the composite wrap. This simulated initial corrosion in ﬁeld columns
before wrapping is applied. The initial corrosion period was closely monitored. When
cracking started to develop, the specimens were taken out of the water and dried

thoroughly before wrapping them with the two types of wrap systems (glass and carbon).

46

One approach to prevent stress concentrations on FRP wraps due to localized
volume expansion is not to bond the wrap directly on the column but only provide bond
between the different layers of the wrap. Thus localized volume expansion is contained
by the entire wrap system. As with bonded wraps, volume expansion due to corrosion
will strain the wrap inducing conﬁning pressure. About half of the wrapped specimens
contained a plastic sheet between the concrete and the wrap in order to prevent the wrap
from bonding to the concrete.

A total of 24 specimens were subjected to the initial phase of accelerated
corrosion for 13 days. Although subjected to the same conditions, the specimens had
signiﬁcant variation in corrosion level as observed from concrete cracking. The
specimens were divided into three groups based on the severity of cracking - severe,
moderate and light.

A total of ﬁve severely corroded specimens, in which some concrete had spalled
off, were patched. Patching was done using a sand/cement mortar (that was contaminated
with 2% Cl' ion by weight of cement. Specimens for which spalls occurred at an edge
were placed in a mold and mortar was placed within the mold. These severely corroded
specimens before and after patching are shown in Figs. 2.9 and 2.10, respectively.
Specimens were selected systematically from the three groups for wrapping as shown in
Table 2.7. Three layers of glass wrap and two layers of carbon wrap were used to be
consistent with the number of layers used in the freeze-thaw test. A total of 16 specimens
were ﬁtted with strain gages oriented in the circumferential direction (two gages per
specimen located diametrically opposite each other at mid-height). All specimens were

then ready for the next phase of accelerated corrosion.

47

Table 2.7 Number of wgpped and unwrapped specimens in corrosion groups

 

 

 

 

 

 

 

 

 

. . Number Wrapped with
Corrosron Group Wrap Adhes1on Glass Carbon Nothing
Severe Bonded 1 l 1
Unbonded 1 1
Bonded 1 1
M°derate Unbonded 1 1 1
. Bonded 2+2 2
”gm Unbonded 2 2 2+2

 

 

 

 

 

g I "A l ’5‘“ ‘

 

48

Fig. 2.9 Severely corroded specimens prior to patching

 

 

Fig. 2.10 Patching of the severely corroded speciemns

2.1.4.2 Construction of the Corrosion Tank and Appurtenances

A special wood tank coated with ﬁberglass to accommodate the corrosion
specimens was constructed. The tank was ﬁtted with a marine pump, a ﬂoat shut-off
mechanism, ball valves, and a timer control. In automatic mode, the pump was activated
once a day to ﬁll the tank with salt water (3% NaCl) from a holding tank located below
the ﬁberglass tank. The ﬂoat shut-off mechanism would turn the pump off when the
water level covered the top of the specimens. After one hour of soaking, the timer opened
the ball valve and the water was drained into the holding tank. Photographs of the
corrosion tank and the corrosion specimens placed in the tank are shown in Figs. 2.11 and

2.12, respectively.

49

 

Fig. 2.12 Corrosion specimens in the tank

2.1.4.3 Monitoring Progress of Corrosion During Test

Monitoring corrosion levels during the accelerated corrosion test was important in
order to know when to remove specimens. Unwrapped specimens were expected to
corrode faster than wrapped specimens. Also, some specimens were to be removed

50

approximately mid-way through testing when the anodic reinforcement in unwrapped
specimens lost about half their cross sectional area. The total length of time for the
corrosion test could not be predicted in advance. A method was needed to monitor
corrosion levels during the test.

Two dummy specimens were originally fabricated to monitor corrosion levels
through destructive means. The plan was to cut off sections of the dummy specimens at
regular intervals and visually examine the cross section of corroded bars. This approach
was error prone because corrosion occurs unevenly and the dummy specimen size would
be altered each time a section was sliced off.

A non-destructive method of monitoring corrosion levels was sought, and an
approach utilizing X-rays was identiﬁed. Some unwrapped and wrapped pre-corroded
specimens were subjected to X-rays in a standard radiology laboratory. The X-ray
negatives clearly show the uncorroded parts of the anodic steel reinforcement.

Prior to beginning the accelerated corrosion process after specimens were
wrapped, representative specimens were subjected to X-rays to ascertain the level of
corrosion during the pre-corrosion phase. The X-ray images showed the reinforcing bars
inside the specimen and the approximate level of corrosion in them. This was necessary
to establish a reference point and assess the progress of corrosion.

Periodically, the specimens selected for observation were transported to the
radiology laboratory and subjected to X-rays. By comparing the state of corrosion in the
wrapped and unwrapped specimens, it was evident that the corrosion rate in wrapped
specimens was signiﬁcantly lower than that in unwrapped specimens. Fig. 2.11, 2.12, and

2.13 show views of the X—rays of a typical unwrapped specimen taken after 0, 90 and 105

51

days of accelerated corrosion after the precorrosion stage, respectively. To facilitate
reproduction, the X-ray images were outlined to clearly show the edges of the reinforcing
bars and the specimen, and the photographic image was reduced to create the line
drawings shown in the ﬁgures. Fig. 2.16 shows typical reinforcing bars removed from a
specimen. The middle two bars are the cathodes while the outer two bars are the anodes
(the site of corrosion).

Specimens were removed for X-ray exposure when there was no water in the
tank. The current used to accelerate corrosion was shut-off while the specimens were
removed and transported to the radiology laboratory, and turned back on when the

specimens were returned to the corrosion tank approximately two hours later.

 

 

 

 

 

 

 

 

 

 

 

C— _ i ___Anode __‘"

‘.

[ cm _
L _i:_ ‘ A Anode *‘I—

 

Figs. 2.13 Sample X-ray taken in the beginning of the accelerated corrosion test

52

 

 

 

 

 

 

 

 

 

 

 

 

 

__ f L *1

if __A __ M L...—
l “m“

_ e r -mi T:

RI, J

 

 

 

 

 

 

 

 

 

 

E—‘* W? —i 1 ~ 17;. ~11:
f _ _ Cathode 7 '
R; T _ AM “‘7 IT

—2 .—-——_]--

 

 

Figs. 2.15 Sample X-ray taken after 105 days of accelerated corrosion test

53

 

Fig. 2.16 Reinforcing bars after removal from a corrosion specimen. The middle two bars
are the cathodes while the outer two bars are the anodes (the site of corrosion)
2.1.4.4 Corrosion Test Matrix
The total number of specimens was 24, including the two extra specimens
originally planned for corrosion monitoring through destructive means. The numbers of
samples of the various types of specimens used are given in Table 2.8. Specimen, bar and
strain gage numbering is shown in Table B2 in Appendix B. Specimens were removed
from the corrosion process and the amount of corrosion was measured as follows:
1. Four unwrapped, four carbon wrapped (two bonded, two unbonded), and ﬁve
glass wrapped (three bonded, two unbonded) specimens were removed when the
X-ray technique indicated that the diameter of corroded bars in unwrapped
specimens was reduced to about 70% of the initial diameter (which corresponds to
about 50% reduction in the cross sectional area). This corresponded to 130 days

of accelerated corrosion.

54

2. Two unwrapped, four carbon wrapped (two bonded, two unbonded), and ﬁve
glass wrapped (three bonded, two unbonded) specimens were removed when the
accelerated corrosion test could not be effectively continued for the unwrapped
specimens. This was due to the upper tips of the anodes breaking off and occurred
after 190 days of accelerated corrosion.

Comparisons of corrosion levels at each of the two stages mentioned above was used

evaluate the effectiveness of the different wrapping systems in reducing corrosion.
Wrap strains were monitored until specimens were removed from the corrosion

process. These strains were used to estimate the amount of conﬁning pressure built up

due to corrosion.

Table 2.8 Accelerated corrosion laboratory test matrix

 

 

 

 

 

Wra No. of Specimens No. of Specimens
p Tested for 130 days Tested for 190 days
None (control) 4 2
Carbon bonded 2 2
Glass bonded 3 3
Carbon unbonded 2 2
Glass unbonded 2 2

 

 

 

 

55

Chapter 3
Data Collection and Analysis of Results

3.1 Effect of Freeze-Thaw and Wet-Dry Cycling on
the Properties of FRP Panels

Table 3.1 provides the mean mechanical properties of FRP panels after 300
freeze-thaw cycles and 190 wet-dry cycles with 3% NaCl solution. By comparing these
results to the unconditioned panels tested (see Table 2.1), the following observations are
made:

0 The freeze-thaw conditioning had little effect on the effective stiffness of glass panels
while carbon panels’ effective stiffness was increased by 21%. (Note that effective
stiffness = modulus x thickness.) The ultimate strains were reduced by approximately
20% and 28% for glass and carbon panels, respectively.

0 The wet-dry cycles with 3% NaCl solution had no effect on the effective stiffness of
glass panels while the effective stiffness for carbon panels was increased by
approximately 39%. The ultimate strains were reduced by approximately 20% and
36% for glass and carbon panels, respectively.

The results in Table 3.1 are based on samples cut from one glass panel and one carbon

panel. Investigations involving several panels may yield more reliable results.

56

Table 3.1 Mean properties for FRP panels

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Thickness Modulus Effective Ultimate
Wrap (mm.) (MPa) Stiffness Strain
300 Freeze-Thaw Cycles
Glass 1.092 23805 26000 0.016
Carbon 0.508 79012 40138 0.010
190 Wet-Dry Cycles
Glass 0.914 29539 27000 0.016
Carbon 0.571 81 109 46313 0.009

 

 

 

 

 

3.2 Freeze-Thaw Test

3.2.1 Strain Gage Readings

Specimens subjected to 300 freeze-thaw cycles were equipped with strain gages,
while those subjected to 150 freeze-thaw cycles were not equipped with strain gages. The
FRP hoop strains were monitored about once a day during the entire testing period for
specimens ﬁtted with strain gages. Two readings were made each day, one during the
freeze phase and the other during the thaw phase. All strain gages survived the freeze-
thaw test.

By observing the difference in the strain reading of the control specimens
(wrapped, no F/T conditioning) between freeze and thaw cycles, the thermal contraction
of the FRP panels from thaw to freeze cycles could be determined. Figures 3.1-3.4 show
the uncorrected strain readings on round and square control specimens that were not
subjected to freeze-thaw. On any given day, the reading of a gage during a freeze or thaw
cycle should be approximately the same, since the control specimens were not subjected
to freeze-thaw. The difference in gages readings from the thaw to freeze curves observed

in Figures 3.1-3.4 is therefore due to contraction of the dummy FRP panel. The

57

compensation strains for glass and carbon were determined by the average difference
between the thaw and freeze readings for control specimens from day 15 onward, when
the strains were stable. Note that for carbon (Figures 3.3-3.4), there is only a slight
difference between thaw and freeze readings since its coefﬁcient of thermal expansion is
close to zero. The compensation strains computed in this manner are 372 micro-strain for
glass FRP and —37 micro-strain for carbon FRP. Figures 3.7, 3.11, 3.15 and 3.19 show
the compensated strains on the same four control specimens, respectively. The strains
during the thaw and freeze cycles are now approximately the same from day 10 onward
as expected.

The compensated FRP strains on round and square specimens wrapped with glass
and subjected to freeze-thaw are shown in Figures 3.5. 3.6, 3.8, 3.9, 3.10 and 3.12. In
general, the strain during the freeze cycle is 100—200 micro-strain higher than that during
the thaw cycle. This is most likely due to the thermal contraction of the glass wrap during
freezing. Since the concrete specimens prevent the contraction, the tensile strain in the
glass wraps increase. An exception is Figure 3.6, in which the strains during the thaw
cycle is slightly but consistently higher than that during the freeze cycle.

The compensated FRP strains on round and square specimens wrapped with
carbon and subjected to freeze-thaw are shown in Figures 3.13, 3.14, 3.16-3.18 and 3.20.
The results are much less consistent for carbon than for glass. In Figures 3.13 and 3.15-
3.17 the thaw strains are slightly higher than the freeze strains, while the freeze strains
are higher than the thaw strains in Figures 3.14 and 3.20.

A reason for some of the variability in the wrap strains could be the ingestion of

water into the Bristar. Although epoxy caps were used on both ends of each specimen to

58

prevent water penetration, during freeze-thaw cycling the caps of several specimens
ruptured due to expansive pressure from Bristar. The loss of strain with time in Figure

3.17 also is likely to be due to the loss of pressure in the Bristar.

 

 

        

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

4500
4000
3500 - . , ‘9~9-"3‘--5--.a-n‘5‘9'°'9 ---- Bio-6-3-...a
"4",: ...... \‘..‘,-&-*--b--...*-..--§-....*,*. _. ..~b..‘,.......... __ ” _.___”_
:Ix '”'*‘*'*""X~.-*..—-*'*'3.:A‘u-mm-*..:.:,_:.*.-._§
3000
I:
“i 2500 —O—Geoe$lreeze
8 --G-Geoeafreeze
2 2°00 "'¢"‘GQ95MQW
'3 --I--Gegeathew
1500
1000
500
0 1 . , Y T Y
O 5 10 15 20 25 30 35

Test Dilation (Days)

Fig. 3.1 Hoop strains in glass wrap of round control specimen #3 before correcting for
thermal contraction of dummy FRP panel

59

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7500
7000
.---.er
5.. “g...“ ,A.‘
55110 A s»...- L _ .'_‘"O'-ﬁr-A' et-e~-----k-n--L.Ae-t---~A
‘.*'o'
11 X'\.‘
6000 I .‘\
3’9 ----- .‘x ’,a-a~B-e—"9--.3--.G.o""‘“‘9‘"""‘°‘G“~a-—-GI
I, ‘s“9= r
c 5500 ' “'x-*-x".~"%. .x-ﬁ-n-R'A‘Wm-w-*-ee-K--te-*:--
._¢ ,. 'K +Gsoe 13 freeze
g n --G- Gage 14 freeze
g ---o---Geoe131hew
, 5000 --x-—Gage14thaw
4500
4000
3500
mo V Y r v I r
0 5 10 15 20 25 30 35
Teeth-ileum

Fig. 3.2 Hoop strains in glass wrap of square control specimen #7 before correcting for
thermal contraction of dummy FRP panel

 

    

 

 

 

 

 

 

 

 

 

 

 

8000
”A" . A
.I..‘.*_‘ A A A A ...' ﬁ * * t
7500 - .. .x ~ ;
”3:37:11": 'ﬂ-G-G'L‘M=RZL£PL‘G‘B
7000
g ——e—Ge9e 21 freeze
--o-- Ge9e22 freeze
6500 .......Gm21m.w
--x--Gege 22 thew
6000
5500
SW I v r r T T
0 5 10 15 20 25 30 35
TeetDumlonmeye)

Fig. 3.3 Hoop strains in carbon wrap of round control specimen #11 before correcting for
thermal contraction of dummy FRP panel

60

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3000
JR ,--'3‘—‘ x‘\ a"*.*1‘ - . .*--.
2000 ﬁzz. ---------- fade-ﬁan-Jg—ai—ua-a s‘iﬂ-‘-’-"-’*““'~a--
g —e—Gage 29 freeze
--a-- One 30 freeze
8 1500 we“ Gage29thaw
i —~x-- gageao thaw
1000
500
o Y T l V V Y
0 5 10 15 20 25 30 35
Tenors-donate”)

Fig. 3.4 Hoop strains in carbon wrap of square control specimen #15 before correcting for
thermal contraction of dummy FRP panel

 

5000

 

 

4800

 

4600-

 

4400 ‘ ."-

 

 

4200

 

-O—Gegetfreeze
4ooo --a--Gegeereeze
""°"Gm1 M
3800 —-X--GageZthew

 

More“

 

 

 

 

3600

 

 

3400

 

3200

 

 

3000

 

o 5 10 15 20 25 30 35
Test Dinettes! (Days)

Fig. 3.5 Hoop strains in glass wrap of round specimen #1 during freeze-thaw cycles

61

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6500
.... “.~‘~ 'x‘s * ﬂ -* *
."' ‘ . . ' .---*-__ ‘ _ _ - ‘ ._._,._ _ u. .- _
'*" n...- A-t-gm_—z--.n-rI-:vg'g'a‘a"---g-g-G"G'B--‘6
m0 \ '_"v — f
\.__—v
5500 ”HQ...“
...,.--" ."'-‘,.*.‘--““"--A---uQ..-.--*-A--Ang“‘*°"”-~-A-'A"A-'A"*"---A
tK:::;,fa¢,av0——-—e—e—e—e———0———e—-—e—o’0—°’°“—‘———9—0—°—0—0—-4
C 5000 -—e—Gage$freeze
g --o-- Gage4lreeze
~~e-vGage31hm~
4500 - . K-- Gage 4 thaw
4000
3500
3000 T ’r Y Y Y .
0 5 10 15 20 25 30 35
Tuulhladon(0emn

Fig. 3.6 Hoop strains in glass wrap of round specimen #2 during freeze-thaw cycles

 

4500

 

4000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2500 t—o—Gagesmze
uo- Gageelreeze
"-5-- Gagesthaw
2000 mun-Gnosthaw
1500
1000
500
0 T V T Y 1 T
0 5 1o 15 20 25 30 35

Ted Dilation (Days)

Fig. 3.7 Hoop strains in glass wrap of control round specimen #3

62

 

5000

 

4500

’r‘~.o"B ‘\

 
   
    

 

4000 .— "‘x

a.’ .” ...... ”.“‘“.*.-._x-‘-*."

. .--- ‘Il . “.‘I X.
[4‘ o.*.‘ .c*...no‘-oa~"‘... x, *

 

3500 .45)?
12:?

 

2500

 

-—e—Gage7freeze
--a>- Gageafreeze

 

IcroStrstn

2000

.......G$7m“
--x--Gage81haw

 

 

 

 

 

1500

 

1000

 

 

 

 

remote-source”)

35

Fig. 3.8 Hoop strains in glass wrap, round specimen #4 during freeze-thaw cycles

 

6000

 

 

 

 

 

—+—Gageslreeze
--a-- Gage wheeze
0.....WQM
--s(-- Gage 101nm

 

 

 

 

3500

 

 

 

3000

Team-Ion (Days)

Fig. 3.9 Hoop strains in glass wrap, square specimen #5 during freeze-thaw cycles

63

 

5000

 

 

g —+—Gage11lroeze
350° 1 --O-- Gage121reeze
g +“ c..‘...6m1‘maw

 

 

 

 

—-x--Gaget2thaw

 

 

 

 

 

 

3000
f
2500
II
2% r V T I ﬁr T
0 5 1 O 15 20 25 30 35

TsdDuratlonmays)

Fig. 3.10 Hoop strains in glass wrap, square specimen #6 during freeze-thaw cycles

 

7500

 

 

 

 

 

-—-0-—Gaoe131reeze
~o-Gaos14treeze
menu-onshore
--It--Gagef4lhaw

 

 

 

 

 

 

4500

 

 

 

 

 

TedDInIaruDays)

Fig. 3.11 Hoop strains in glass wrap of cotrol square specimen #7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

n“""'a M m .__—a—-H
m ,..I - - -.....‘..,..-_v — -’ '1‘.._.....-r —W
.-" ‘m- -I - .Q... .-. - - .
as.-- ”.4" “i" s--a'"*sx~——--Iczﬁ~.:;-..a
ssoo “"'
0000
g 5500 —O-—Gage151reeze
-—O--Gage161rssze
35000 ----e GeostSthaw
--x~-Gage161haw
4500
4000
3500
3000 f r
0 5 10 15 20 25 30 35

Test Duals!- (Days)

Fig. 3.12 Hoop strains in glass wrap of square specimen #8 during freeze-thaw cycles

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7000
5500
i —o—Geoe17teeze
--o-- Gage 10 has
5000 ---e-- Gage17thaw
g mun-Gagetathaw
4500
4000
3500
N Y T Y V Y Y
0 5 10 15 2O 25 30 35

Ted Mon (Days)

Fig. 3.13 Hoop strains in carbon wrap of round specimen #9 during freeze-thaw cycles

65

6000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5500
5000
g —+-—-Ga9e19 freeze
‘500 f —-o-- Gage 20 freeze
§ 1 men-Gagetsmaw
‘ --x--Gage201haw
1
4000
3500
3000 . . , Y
0 5 10 15 20 25 30 35

Fig. 3.14 Hoop strains in carbon wrap of round specimen #10 during freeze-thaw cycles

 

 

 

 

 

 

 

 

 

 

 

 

8000
..-*‘*- _-‘"'." “ c . ' . . 't“"‘
7500 - M -' ' . -. ' - . :f .DI...
€:":T:'::-3T"“°“‘r:'::°:'.:;’i I )e-Ot‘
7000
g +Gaoe 21 Freeze

"0' 3.90 22 feels

0500
---.-..G.°.2‘ m“
""°' Gage 22 thaw

6000

5500

5000 . - .

Test Duration (Days)

Fig. 3.15 Hoop strains in carbon wrap of control round specimen #11

66

7000

 

 

 

 

 

—-o-- Gage24 haze
---e---Gego23m
~~st--Ga9e24 thaw

: t—e—Gege231reeze
5000

 

 

 

 

4500

 

4000

 

 

 

 

 

Fig. 3.16 Hoop strains in carbon wrap of round specimen #12 during freeze-thaw cycles

5000

 

4500 -

 

4000

 

 

 

—e—Gege251reeze
--o- Gaoe261reeze
whamzsm
--st--Geoe2611‘1aw

 

3500

 

 

 

 

 

 

 

 

 

3000
2500
20m V f 7 V V Y
0 5 1 0 1 5 20 25 30 35

Test Dilation (Days)

Fig. 3.17 Hoop strains in carbon wrap of square specimen #13 during freeze-thaw cycles

67

 

 

 

 

 

 

 

 

 

 

 

 

5500
I
5m 5 ‘ ue’ e
"s .- -. ..... a ----- ts . .. at
ﬁnk .- A t
A ‘* "e~.,qr 4r A
24‘ .01 so u-x ,ar 51- at so - .x
I e ‘ \. x-'.
.x' ’8 r 4s-aril---o
(w‘ gngf'ﬂ--’C~ﬂ —_-‘ ‘ — 4- — _ JJ-—---.‘.
45m '8',J' "ar— —‘w---"' ' —1v-w
li' —-e—Gaoe 27 freeze
--o-- Gageza freeze
m". Gage27thew
--st--gage281haw
4000
3500
3000 v 4
0 5 10 15 20 25 30 35

Teetnhaalen(Days)

Fig. 3.18 Hoop strains in carbon wrap of square specimen #14 during freeze-thaw cycles

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g w—e—Geoe291reeze
‘500 "re-Gageaoteeze
men Gne291haw
"“Z:§PO°3010IV
1000
500
0 v T . , , .r
0 5 10 15 20 25 30 35

Fig. 3.19 Hoop strains in carbon wrap of control square specimen #15

68

 

 

 

 

k
ﬁ
\

V’D'ai'M-G-mg—ua

e .6 ‘----
.X X _-.*-__ 31??" i‘zfﬂ-Bsa-—-ﬂ
.- ’ m-r‘” " "war ----- m' “’*""~--x

 

c
3000 —-e—Gage 31 freeze

--o- Gage 32 taste
:

 

no.1...6”. 3‘ m
--x~-Gege 32 crew

 

 

 

 

 

 

 

 

0 5 10 15 20 25 30 35
Test Milan (Days)

Fig. 3.20 Hoop strains in carbon wrap of square specimen #16 during freeze-thaw cycles

3.2.2 Results of Compression Testing

Figures 3.21 to 3.35 show results of the compression tests for plain and wrapped
control specimens and those subjected to 150 and 300 cycles of freeze-thaw. For wrapped
specimens ﬁtted with strain gages in the hoop direction, longitudinal compression strains
are given along the positive x-axis while the average tensile hoop strain is shown along
the negative x-axis. The compression stress for wrapped specimens was computed by
excluding the concrete cross sectional area lost due to the presence of the hole in which
Bristar was inserted. The following observations are made:

0 Plain round specimens (Figures 3.21-3.23): Only one of three specimens survived
freeze-thaw conditioning for 300 cycles. This specimen had approximately the same
compression strength as the control specimens (~35000-45000 kPa). One of the

specimens subjected to 150 freeze-thaw cycles displayed low stiffness and strength,

69

and a progressively hardening behavior— it is not apparent what contributed to this
behavior. There is no signiﬁcant reduction in strength due to freeze-thaw conditioning
except for the anomalous specimen.

Round glass-wrapped specimens (Figures 3.24-3.26): The ultimate strength values
for two of the six control specimens were unreliable because these specimens could
not be crushed in the MSU compression testing machine and were retested at MDOT.
In general, conditioning had little effect and compression strength and failure strains
were approximately the same for control and conditioned specimens. Strength of
wrapped specimens (~105000-114000 kPa) was approximately 2.6 times larger than
the strength of unwrapped specimens.

Square glass-wrapped specimens (Figures 3.27-3.29): Again conditioning had little
effect on the compressive strength (~62000-66000 kPa), but it reduced the
longitudinal strain at failure from about 0.01-0.017 to ~0.007. Strength of wrapped
specimens was approximately 1.5 times larger than the strength of unwrapped
specimens.

Round carbon-wrapped specimens (Figures 3.30-3.32): Conditioning reduced
compression strengths from about 92000 kPa (unconditioned) to about 80000 kPa
(300 cycles) representing about a 15 % strength loss. One specimen each in the 150
and 300 cycle batches had unusually high strengths, indicating that one batch of
specimens prepared might have had a different strength level. Longitudinal failure
strains reduced from about 0.015 to 0.01 (~33 %). Strength of wrapped specimens
(~95000 kPa) is approximately 2.3 times larger than the strength of unwrapped

specimens.

70

0 Square carbon-wrapped specimens (Figures 3.33-3.35): Conditioning reduced
compression strengths slightly from about 58000-65000 kPa to about 55000-63000
kPa. Longitudinal failure strains reduced from about 0007-00] to about 0.005.
Strength of wrapped specimens (~60000 kPa) is approximately 1.4 times larger than
the strength of unwrapped specimens. Note that for square specimens, glass and
carbon wraps increased the strength by about the same amount.

The square wrapped specimens had lower compressive strength compared to the
round specimens, even though the cross sectional area of the square prisms is higher than
that of the round cylinders. This is due to the reduced conﬁnement provided by the wraps
for square cross sections and stress concentrations that develop at the comers. Wrapped
square prisms always failed by rupture of the wrap at a comer (see Figure 3.36). Note that
a reduction of approximately 30% to 40% in failure stress exists between the round and
the square specimens. Figures 3.36—3.37 show the failure modes under compression
testing.

It was also noted that the square wrapped specimens demonstrated a sudden loss
of strength after the peak stress was reached. However, the wraps were undamaged
during this loss of strength. The loss of strength is most likely due to the failure of the
ineffectively conﬁned regions of concrete. These regions do not experience capacity
enhancement resulting from conﬁnement.

A measure of ductility enhancement under compression is the ratio of the mean

longitudinal failure strain of wrapped specimens (cu, wrappm) to the mean longitudinal

failure strain of unwrapped specimens (cu, unwrapped).

71

The mean failure strains and strain ratios are given in Table 3.2. The ultimate

strain for unwrapped specimens is difﬁcult to determine accurately because of the rapid

unloading near failure. Therefore, for unwrapped specimens, the strain at peak stress is

used. As expected, ductility under compression is enhanced more for round specimens

than for square specimens. In general F/T conditioning reduces the ductility. Glass

wrapped specimens are more ductile than the carbon wrapped specimens because the

ultimate failure strain of glass is higher than that of carbon.

Table 3.2 Ductility enhancement under compression for wrapped specimens

 

Mean Longitudinal

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Shape Wrap No. of F/I‘ Cycles Failure Strain (%) 8,, W/E“ MW
0 0.2 1.0
None 150 0.2 1.0
300 0.2 1.0
0 1.8 9.0
Rom“ Glass 150 1.2 6.0
300 1.7 8.5
0 1.3 6.5
Carbon 150 1.0 5.0
300 0.9 4.5
0 0.2"I 1.0
None 150 0.2* 1.0
300 0.2* 1.0
Square 0 1.3 6.5
Glass 150 1.2 6.0
300 0.7 3.5
0 0.8 4.0
Carbon 150 0.5 2.5
300 0.5 2.5

 

 

 

 

 

* Unwrapped square specimens were not used. It is assumed that the failure strain for unwrapped square
specimens is approximately the same as that for unwrapped round specimens

72

 

50000

 

 

 

 

 

 

 

 

 

 

 

 

 

40000 .
—— Specimen 55
: 30000 *Specimen 55
; *Specimen 57
c3 _.—Specimen 58
I- .
.. —.—S salmon 59
‘0 20000 p
_._.Specirnen 80
l
1
10000
O Y I I T
0.000 0.005 0.010 0.015 0.020

Longitudinal Stre in

Fig. 3.21 Compressive stress-strain curves for plain, round, control specimens

 

50000

40000

 

MP

 

 

--— Specimen 32
—— Specimen 33
_._ Specimen 34

“

 

 

 

 

 

 

 

 

 

T

0.000 0.005 0.010 0.015 0.020
W! Strah

Fig. 3.22 Compressive stress-strain curves for plain, round specimens subjected to 150
freeze-thaw cycles

73

 

 

 

 

—~— Specimen 17

 

Stress (kPa)

 

 

 

 

 

T

0.005 0.010 0.015 0.020
Longitudinal aein

Fig. 3.23 Compressive stress-strain curves for plain, round specimens subjected to 300
freeze-thaw cycles

 

-o.015 -o.010 -o.005 0.000 0.005 0.010 0.015 0.020 0.025
-t-toop Strain +Longitudlnal Strain

Fig. 3.24 Compressive stress-strain curves and tensile hoop strain for glass-wrapped,
round, control specimens

74

120000

110000

100000

 

-—— Specimen 20
-— Specimen 21
—-—- Specimen 22

 

 

Stress (kPa)

 

40000

20000

10000

 

0.000 0.005 0.010 0.015 0.020
Longitudnal Strain

Fig. 3.25 Compressive stress-strain curves for glass-wrapped, round specimens subjected
to 150 freeze-thaw cycles

Stre- (kPa)

 

-0.015 -0.010 0.005 01!!) 0.0(5 0.010 0.015 0.020

Fig. 3.26 Compressive stress-strain curves and tensile hoop strain for glass-wrapped,
round specimens subjected to 300 freeze-thaw cycles

75

 

 

11
a

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

-— Specimen 40
A —-- Specimen 41
5' 79°99 -— Specimen 42
V _... Specimen 7
8 -— Specimen 7
g -- Specimen 43
-— Specimen 44
£0909—
20000—
%09004

 

 

 

 

 

IL
1 V7

0.015 0.010 0.005 0.000 0.005 0.010 0.015 0.020
-HoopStraln+l.ongitudlnal Strain

 

Fig. 3.27 Compressive stress-strain curves and tensile hoop strain for glass-wrapped,
square, control specimens

 

 

 

 

 

120000
110000
100000
90000
00000
E 7°°°°
6
8 °°°°°
. *swamm 23
8 50000 -— Specimen 24
-— specimen 25
40000
30000
20000
10000
0
0.000 0.005 0.010 0.015 0.020

Longitudinal err-1n

Fig. 3.28 Compressive stress-strain curves for glass-wrapped, square specimens subjected
to 150 freeze-thaw cycles

76

 

~0.015 -o.010 -o.005 0.000 0.005 0.010 0.015 0.020
-Hoop Strain +Longltudlnal Strain

Fig. 3.29 Compressive stress-strain curves and tensile hoop strain for glass-wrapped,
square specimens subjected to 300 freeze-thaw cycles

   

  

—— Specimen 45
-—— Specimen 48
--— Specimen 47
-— Specimen 1 1
_.... Specimen 1 1
—— Specimen 48
_._ Specimen 49

 
    
 
   
   
    

Stress (kPa)

-0.015 -0.010 -0.005 0.000 0.005 0.010 0.015 0.020
-t-loop Strain +Longltudlnal Strain

Fig. 3.30 Compressive stress-strain curves and tensile hoop strain for carbon-wrapped,
round, control specimens

77

120000

 

1 10000

 

1 00000

90.0. 1”“

 

 

 

 

 

 

 

 

--— Specimen 28

 

—~— Specimen 26
W -—— Specimen 27

Stress (kPa)

 

 

 

 

 

 

 

 

7 T 1

0.005 0.010 0.015 0.020
Longitudinal Strain

 

Fig. 3.31 Compressive stress-strain curves for carbon-wrapped, round specimens
subjected to 150 freeze-thaw cycles

 

A --—epecmn 9
g *spectren 9
: -—spech'en 10
2 —--spech'en 10
iii

*spech'en 12
——-spech'en 12

 

 

 

 

-0.015 -0.010 -0.005 0.000 0.005 0.010 0.015 0.020
-l-ioop Strain +LongltudlnslStrsln

Fig. 3.32 Compressive stress-strain curves and tensile hoop strain for carbon-wrapped,
round specimens subjected to 300 freeze-thaw cycles

78

 

—— Specimen 50
--- Specimen 51
—- Specimen 15
._.._ Specimen 15
_._ Specimen 52
—— Specimen 53
—— Specimen 54

Stress (kPa)

 

 

 

 

-0.015 0.010 -o.005 0.000 0.005 0.010 0.015 0.020
-HoopStrsln +LongltuclnalStrsln

Fig. 3.33 Compressive stress-strain curves and tensile hoop strain for carbon-wrapped,
square, control specimens

120000
110000

100000

 

_._ Specimen 29
-— Specimen 30
—-— Specimen 31

 

arses (kPa)

 

 

 

0.000 0.005 0.010 0.015 0.020
Longitudinal Strait

Fig. 3.34 Compressive stress-strain curves for carbon-wrapped, square specimens
subjected to 150 freeze-thaw cycles

79

Stre- (kPa)

   

-0.015 -0.010 -0.005 0.000 0.005 0.010 0.015 0.020
Hoop Strain +Longltudlnal Strain

Fig. 3.35 Compressive stress-strain curves and tensile hoop strain for carbon-wrapped,
square specimens subjected to 300 freeze-thaw cycles

 

Fig. 3.36 Failure modes for square specimens under compression testing

80

 

(a) Glass wrap

 

(b) Carbon wrap

Fig. 3.37 Failure modes for round specimens under compression testing

81

3.2.3 Statistical Analysis

Table 3.3 provides the ultimate compression strength, mean, standard deviation
and 95% conﬁdence margin for each category of specimens. The cross sectional area lost

by the cavity containing Bristar was deducted when calculating the stresses.

Table 3.3 Freeze-thaw summary data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

No. of Specimen Type Ultimate Compressive Strength (kPa)
F/l‘ . . . Standard 95% Conf.
Cycles Shape Wrap Indrvrdual Specrmens Mean Deviation Mary
109801,106451
Glass 108858 108370 1727 14291
94459,79721
Round Carbon 79962 84714 8440 120967
300 Plain 42875, *, * ' 42875 0 NA
62370, 64362
Glass 64551 63761 1208 13002
Squaw 57870 57279
Carbon 63004 59384 3149 17822
106575,112621
Glass 109316 109504 3027 17520
Round Carbon 97675: 84705 88519 7966 119788
83178
. 41080, 29825
150 Plain 36926 35944 5691 114138
64881. 61500
Glass 66988 64456 2769 16877
Square 55332 58231
Carbon 61421 58328 3046 17566
109504, 104977
Glass 1 14289, 110873 109911 3856 16136
92200”, 89000b
Round 93254, 90494, 95130
Carbon 86555, 93174, 96738 92558 3612 13791
0 . 41911.41696. 40932
Plarn 39973, 37114, 44818 41074 2531 12656
65820,65776,64148
Glass 61856, 62216, 61792 63601 1907 12002
Square 61623 58034 60294
cam“ 65542. 63967. 62082 61924 2652 12783

 

 

 

 

 

 

 

 

 

 

 

 

' Specimens that did not survive 300 Ffl‘ cycles are denoted with ‘
" Data unreliable because specimens did not fail when tested at MSU and were retested at MDOT

82

The 95% conﬁdence margin is calculated as

s . . . . .
t‘ — where s 18 sample standard devratlon and n is sample srze

J;

The average compressive strength for the control, 150 freeze-thaw cycle, and 300
freeze-thaw cycle specimens is displayed in Figures 3.38-3.42. A 95% conﬁdence
interval is also provided. Due to the small sample sizes and unknown population
variances, the t-distribution was used for all hypotheses tests in this report.

The null and alternate hypotheses are:

0 Null hypothesis (Ho): There is no signiﬁcant difference between the means of control
and freeze-thaw specimens, i.e., Ll comm. = 11 m
0 Alternate hypothesis (Ha): There is a signiﬁcant difference in means of control and

freeze-thaw specimens, i.e., p. comm. at 11 m

Sample calculations

Comparing results for carbon wrapped, control, round specimens and carbon wrapped,

300 F/T cycle, round specimens, for the former

1?, = 92558 kPa, s. = 3612 kPa, n. = 6, and for the latter

E; = 84714kPa, 52:8440 kPa , n; = 3

where :1 , S, and n are sample average, standard deviation, and size, respectively.

A conservative degree-of-freedom is (Neter at e1. 1992)

d.o.f.= smaller ofn1 - 1 or n; — 1 = 3-1=2 = t‘ = 4.303 ( 2 tail test)

The 95 % conﬁdence interval for 11 mm. - p. m (difference between the mean strength of

control specimens not subjected to F/T and the mean strength of specimens subjected to

F/T) is
83

_ _ 2 2
,u — ﬂ = x — x i t. s—1 + s_, where is ulation mean (Neter at el. 1992)
control F IT 1 2 n! n '1 pop
2

=7844 :t: 4.303 (5091) = (-14063, 29751)

Since this conﬁdence interval spans zero, the means are not signiﬁcantly different and

H0 is not rejected.

The results of the hypothesis tests for the various comparisons are given in

Table 3.4. The standard deviations and number of samples for each case are given in

Table 3.1. The value of 1* was 4.303 for all comparisons.

Table 3.4 Results of hypothesis tests (95%Lon specimens exposed to freeze-thaw cycles

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Shape Wrap No. of F/T C.I. (kPa) for Outcome of
Cycles 11 gm] - p m Test
Plain :13?) 0969:3119952) Don’t Reject H0
150 (-10789, 11603) Don’t Reject H0
Rmmd Glass 300 (-7799, 10881) Don’t Reject 11.,
Carbon 150 (- 16745, 24821) Don’t Reject H0
300 (-l4064, 29751) Don’t Reject H0
Glass 150 (-8506, 6796) Don’t Reject H0
Square 300 (-4339, 4659) Don’t Reject H0
Carbon 150 (-5290, 12482) Don’t Reject H0
300 (-6565, 11645) Don’t Reject H0

 

At the 95% conﬁdence, means of the compressive strength of freeze thaw

 

specimens are not signiﬁcantly different from those of control specimens. Similarly, the
freeze thaw cycles have no statistically signiﬁcant effect on the compressive strength of
round specimens.

At the 95% conﬁdence level, means of the compressive strength of freeze thaw

specimens are not signiﬁcantly different from those of control specimens. Similarly, the

84

freeze-thaw cycles have no statistically signiﬁcant effect on the compressive strength of

square specimens.

It should be noted that a reduction in mean compressive strength was observed for
carbon-wrapped specimens after freeze-thaw conditioning. This difference is not

statistically signiﬁcant for the sample size used in this study.

Stress (kPa)

 

Control 150 FIT cycles 300 F/T cycles

Fig 3.38 Average compressive strength of round glass-wrapped specimens

85

120000
100000

Stress (kPa)

 

Control 150 F/T cycles 300 F/T cycles

Fig 3.39 Average compressive strength of round carbon-wrapped specimens

120000
100000
80000

40000
20000

Stress (kPa)

 

Control 150 PH cycles 300 HT cycles

Fig 3.40 Average compressive strength of square glass-wrapped specimens

86

120000
100000
80000

Stress (kPa)

40000
20000

 

Control 150 F/T cycles 300 F/T cycles

Fig 3.41 Average compressive strength of square carbon-wrapped specimens

 

120000

 

 

iii

 

 

 

 

‘5’

 

 

Stress (kPa)

 

 

 

 

'2’

 

 

 

 

 

 

 

 

 

 

O
.1

Control 150 F/T cycles 300 F/T cycles

Fig 3.42 Average compressive strength of round plain specimens

87

3.2.4 Effect of Sustained Loads on Freeze-Thaw Durability of

Wraps

Bristar was used in the wrapped specimens to investigate the durability of glass
and carbon wraps under sustained load subjected to freeze-thaw cycling. The sustained
load simulated the load generated in wrapped columns by corrosion products.

The compression strength of wrapped specimens subjected to freeze-thaw cycling
was not signiﬁcantly different than that of wrapped control specimens. This indicates that
the wraps did not sustain any signiﬁcant damage due to freeze-thaw cycling under

sustained load.

3.2.5 Comparison of Measured and Predicted Confined
Compression Strength

Equations for predicting the conﬁned compression strength of wrapped circular
and rectangular columns are outlined in Section 1.3.6. Here the conﬁned compression
strengths predicted by Eq. 1.1 and 1.5 are compared to the measured strengths. Measured
FRP panel properties in Tables 2.1 and 3.1 were used in the predictions for unconditioned
and conditioned specimens, respectively. The observed and predicted compression

strengths are shown in Table 3.5.

88

Table 3.5 Comparison of measured and predicted conﬁned compression strength

 

 

 

 

 

 

 

 

 

 

No. of F/T Predicted Measured Predicted/
Shape Wrap Type
Cycles Strength (kPa) Strength (kPa) Measured
Glass 127,707 109,910 1.16
Round
0 Carbon 90,542 92,577 0.98
Glass 87,614 63,601 1.38
Square
Carbon 67,649 61,924 1.09
Glass 108,224 108,370 1.00
Round
300 Carbon 85,131 84,714 1.00
Glass 77,147 63,761 1.21
Square
Carbon 64,742 59,384 1.09

 

 

 

 

 

 

 

Table 3.5 indicates that the Restrepol-DeVino model for rectangular sections over
predicts the measured strengths of unconditioned specimens by about 38% and 9% for
glass and carbon wrapped specimens, respectively. For unconditioned round glass and
round carbon wrapped specimens, the Richart model over predicts the observed strength
by about 16% and 2%, respectively. For conditioned specimens, the Restrepol-DeVino
model for rectangular sections over predicts the measured strengths by about 21% and
9% for glass and carbon wrapped specimens, respectively. For conditioned round glass
and round carbon wrapped specimens, the predicted and the measured strength are almost

the same.

3.3 Accelerated Corrosion

3.3.1 Mass Loss Results

Corrosion specimens were removed from the corrosion tank in two phases and
measurements of corrosion-induced mass loss were determined. Mass loss data and

corrosion depths for the ﬁrst batch (exposed to 130 days of accelerated corrosion) and the

89

second batch (exposed to 190 days of accelerated corrosion) is given in Tables 3.6 and
3.7, respectively. The standard deviation and 95% conﬁdence margin for the corrosion
depth also are provided. The corrosion depth for each individual bar varied signiﬁcantly
over the length of the bar. The depth reported in Tables 3.6 and 3.7 is the average depth
calculated from the total mass loss for each bar. The average corrosion depth over the

entire bar can be calculated from the fractional mass loss (FML) of the bar.

 

 

FML = " ’
W1
where
w, = initial weight of bar
wf = ﬁnal weight of bar
Note that
2 2 2 2
pm, _ pm f ’1' — rf
FML = 2 = 2
P7”.- "1

where p is the density.

Average corrosron depth 1. "r _ rf

= r,(1—J1-FML)

90

The average corrosion depths for all categories and 95% conﬁdence intervals are
displayed in Fig. 3.43.

 

 

 

 

3500
E
c: 3000
E.
D. 25m
8 lGlass bonded
g 2000 lGlass unbonded
g lCalbon bonded
3 150° ECarbon unbonded
‘3 100° lPlain
U)
S
g 500
<
0

 

130 days 190 days
Duration of Accelerated Corrosion

Fig. 3.43 Average corrosion depths due to accelerated corrosion

The following observations are made:

1. Corrosion depths for reinforcement in specimens with unbonded wraps were
approximately 20% more than those in specimens with bonded wraps after 190 days
of testing. This may be due to water seepage between the concrete and the plastic
sheet used to create the unbonded condition.

2. Wrapping reduced the corrosion depth by 46% — 59% after 190 days of testing.

It should be noted that the specimens removed after 130 days had corrosion levels
of “medium” to “severe” prior to wrapping. The specimens removed after 190 days, on
the other hand, had a “low” corrosion level prior to wrapping. This accounts for the

higher variability in the corrosion mass loss in the specimens removed after 130 days.

91

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.« «.8. 2.8 m. 16..
8«« «.8. 2.8 m.
1.8 m.«t 1.8« ..
. . 51.8.2
8..« m S. m 8« «
«3n 8. 58 B8 8.8. 3.8 9. 26.. an:
2.8 o8... 3.8 8
28 3.. 3.8 8
«.8 ago. 0.88 8 298
8.... o.\.«. 1.8 «5
8a.. .8. «8. 8«. m.«o. ..8« 8 51.8.2 828....
mom. ..o«. o.8« 8 298 8&8
4...... m..«. 3.8 8
8o. «.3 3.8 «n
can m8 .8. «.m. .. .«. 3.8 on 898 .828
8.. com. 38 .m 51.8.... 898
8o. ...8 $8 8
8... an. . ..8« «8
«Sn 8 «8. 88. Q8. ....8« 8 51.8.2 888....
on... 3. . 2.8 8 898 88.0
$9 «.8. «.8« «c
8.. 1.8 .88 a.
3... an. . m.8« we Beam
8«. m. .o. o.8« 8 8.1.8.
8m. 5. 8.. 8.. 0.8 9.8 8 16.. v.86
m8 o. .o e.8« ..
«8 118 I138 m 15.82
Emma—2 5:860 :35. 20......on a. a. «was. .02 am means; 88mm on»...
.80 $8 2888 .1828. 98.. «no.2 .8510 .96.. 8.818 82.88
.2... £80 8.818 case)...

coﬁoboo 3.8288 .0 mane om. 8 .3838 808.8% .8 Snow 8.8.80 emcee; e5. 82 83>. 9m 03.8.

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

88 «.8 38 m. 16..
38 8.88 .88 n.
.84 8. 88 $8 3.8 8.1.8 . 16.. so...
88 .8. ..8« m
5.. «.8. 6.68 8 16..
1.8. 8.8. .38 8 8.28....
«««n 8. 8.... «3.. no. . ..8« 8 so. 8&8
8... «.1... 6.68 ««
68. 8.8. 5.1.8 5. 16..
« . .n o. .8. «8. 8.8. 8.88 .. 86.6..
8«. ..«o. .88 8 16.. .688
c.«. 6.8. «.8« 8
m .1. 3... 8.88 3. so.
.8 .. 3.... 83 8.... 3.8 «c 688......
8... no. . «.8« .« 8.. 68.6
8... or... .88 8
8«. «.8 6.68 .8 8..
.8. 1.8. .88 «m
8... QB 38 8 oooeon
..n m. 5.. 68. «.6. 38 .m 26.. 820
8.. m8 ..8« m.
8.. 6.8 38 .. 26..
1:66.... 8.658 ecos. .8668... .m. 3.82.. .oz .8 means; 8.8 on...
.80 4.8 2888 683...... one. one... .8610 .26.. 8.880 85.88
.11.. £80 5.880 owes...

aoﬁoboo 3.8282. we 93.. ea. c. @8098 368.60% .8 ﬁne“. commohoo 08.02. .65. 82 82>. 5m 03....

93

3.3.2 Statistical analysis
The following hypotheses are tested for the specimens exposed to 190 days of

accelerated corrosion.

1. Null hypothesis (Ho): There is no signiﬁcant difference between the mean
corrosion depth for bonded and unbonded specimens with the same type of wrap.
i°e°9 Ho: l-l bonded = ll unbonded
Alternate hypothesis (H.): There is a signiﬁcant difference between the mean

corrosion depth for bonded and unbonded specimens with the same type of wrap.

i.e., . H3: [1 bonded at u unbonded

2. Null hypothesis (Ho): There is no signiﬁcant difference between the mean
corrosion depth for carbon bonded and glass bonded specimens or between
carbon unbonded and glass unbonded specimens, i.e, H0: p. m = [.1 glass
Alternate hypothesis (Ha): There is a signiﬁcant difference between the mean
corrosion depth for carbon bonded and glass bonded specimens or carbon

unbonded and glass unbonded specimens, i.e., Ha: u carbon at u glass

Sample calculations

Comparing results for carbon wrapped bonded specimens and carbon wrapped unbonded

specimens, for the former

f, = 1297 um, S1 = 70 um, n1 = 4, and for the latter

I, = 1550 um, 52:140 um, n2 = 4

94

where E , S, and n are sample average, standard deviation, and size, respectively.
A conservative degree-of-freedom is

d.o.f.= smaller of m - l or In - 1 = 4-1=3 z) t“ = 3.182 ( 2 tail test)

The 95% conﬁdence interval for u bonded - u unbonded, using a 2-sample t is

2 2

— — o S S . .
u bonded - u unbonded = x, - xzit —' + —2 where u 18 population mean
n n
l 2

= -252 :I: 3.182 (78) = (-500, -3.4)
Since this conﬁdence interval does not span zero, the means are signiﬁcantly different
and H0, is rejected.
The results of the hypothesis tests for the various comparisons are given in
Table 3.8. The standard deviations and number of samples for each case are given in

Table 3.7. The value of t* was 3.182 for all comparisons.

Table 3.8 Results of hypothesis tests (95%) on specimens exposed to accelerated
corrosion

 

 

 

 

 

 

 

 

 

 

Wrap Type CI. for u beaded - u mded (um) Outcome of Test
Glass (-343, -144) Reject Irl0
Carbon (300, -3.4) Reject Ho
FRP/Concrete Adhesion CI. for p ca_rbon - ll glas_s (11m) Outcome of Test
Bonded (~41, 253) Don’t reject Ho
Unbonded (-108, 340) Don’t reject Ho

 

The following observations are made:
0 The mean corrosion depths for bonded and unbonded specimens are different
at the 95% signiﬁcance level. The bonded wrap is more effective in reducing

the rate of corrosion than the unbonded wrap.

95

 

o The mean corrosion depths for specimens with glass and carbon wraps for
either the bonded or unbonded conditions are not signiﬁcantly different. Both
wrap systems, glass and carbon, are equally effective in reducing the

corrosion rate.

3.3.3 Strain Measurements for Accelerated Corrosion Test

The wrapped corrosion specimens were ﬁtted with strain gages. Two gages
oriented in the circumferential direction were mounted at mid-height were mounted on
each specimen. On some specimens, the strain gages were installed near the anodes, the
site of corrosion and subsequent volume expansion. For others, strain gages were
installed between the anode and cathode. The purpose of this arrangement was to
investigate the variation in strain gage readings with respect to the site of corrosion. The
observations below are for the specimens exposed to 190 days of accelerated corrosion:
o For glass bonded specimens (Figure 3.44), it is evident that the strain reading is
considerably higher for specimen 7 (approximately 4200 micro strain) for which the
gages are located at the anodes. For other specimens the gages (installed between the
anodes and cathodes) are consistent at about 900 micro strain. Based on the average
strains developed at gages not located at the anodes, the conﬁning pressure is
estimated to be about 950 kPa away from the anodes, but much higher near the
anodes.

o For glass unbonded specimens (Figure 3.45), all strain gage readings are about 1500
micro strain regardless of the gage location. This yields a conﬁning pressure of about

1600 kPa (scaled directly from Table 2.5 as 5644.9 x 1500/5310). The unbonded

96

condition, created by a plastic sheet located between the wrap and the specimens,
allows the wrap to expand more freely instead of concentrating the strains near the
anode as in the bonded wraps.

For carbon bonded specimens (Figure 3.46), the results are similar to those for the
glass bonded specimens. The strains developed for specimen 13 (for which the gages
are located at the anodes) are higher than the gage readings for the other specimens
(for which gages are located between the anodes and the cathodes) and is
approximately 1800 micro strain. The average strain value where gages are located
between the anodes and the cathodes is about 1300 micro strain, and yields a
conﬁning pressure of about 1150 kPa.

For carbon unbonded specimens (Figure 3.47), all strain gage readings are about the
same regardless of the gage location and approximately 1300 micro strain. This yields
a conﬁning pressure of about 1150 kPa.

Figures 3.44 to 3.47 indicates that wrap strains for bonded specimens with both

types of wraps tend to level off with time. One explanation could be that the stress

concentration near the anodes in the bonded wraps is more effective in containing the

corrosion-induced crack and reducing the corrosion rate. The slip of unbonded wraps and

the resulting redistribution of strain along the entire wrap may be less effective at

containing the large corrosion-induced crack near the anodes.

Figure 3.48 shows strains at gages placed on the anodes, and indicates the

following:

Even though the corrosion rate for the bonded specimens is lower than that for the

unbonded specimens, hoop strains developed near the anodes for specimens with

97

bonded wraps are higher than those developed with unbonded wraps. In the case of
unbonded specimens, the entire wrap (to some degree) absorbs the volume expansion
associated with corrosion of the reinforcement, while for bonded wraps the strain is
localized near the corrosion-induced crack at the anode.

o Corrosion of reinforcement seems to have a more direct effect on strain values
generated in glass bonded wraps compared to those with carbon bonded wraps. The
maximum strain for glass bonded wraps is approximately 4000 micro strain compared
to 1800 micro strain for carbon bonded wraps. It should be noted that variations in
strain readings also could be inﬂuenced by initial wrap tightness around the
specimens and by crack width and direction. Although the wrap strain near the anode

is large, there is no danger of stress rupture in the glass since the stress rupture limit
(with a safety factor of 1.67) is about 0.2 8,. = 4000 micro strain. Further, in real

columns the strain near corroding bars will be signiﬁcantly smaller because of the
larger column diameter and concrete cover.
Note that the wrap strains measured in the accelerated corrosion test away from
the anodes are lower than the wrap strains generated by Bristar in the freeze-thaw test.

Thus the internal expansive force used in the freeze-thaw test was very conservative.

98

Stub (us)

4500
4200
3900
3600
3300
3000
2700
2400
21 00
1 800
1 500
1 200

600
300

 

O 2 4 6 810121416162022242628
Tlmc(wcclts)

Fig. 3.44 Hoop strains in bonded, glass-wrapped specimens

4500
4200
3900
3600
3300
3000
2700

5 2100

 

0 2 4 6 810121416182022242626
Tllnﬂwcolu)

Fig. 3.45 Hoop strains in unbonded, glass-wrapped specimens

99

4500
4200
3900
3600
3300

2700
32400
5 2100
1800
1500
1200

600
300

 

-300
0 2 4 6 810121416162022242628

Tina (woolen)

Fig. 3.46 Hoop strains in bonded, carbon-wrapped specimens

4500
4200
3900
3600
3300

2700
‘3 2400
5 2100
g 1800

 

0 2 4 6 810121416182022242628
Tlmcmulu)

Fig. 3.47 Hoop strains in unbonded, carbon-wrapped specimens

100

4500 .. _. _, .. '
4200 ' " °‘ ‘
3900

.. .. *0...

2700
2400
2100
1800
1500
1200

Stilt! (pg)

600
300

 

O 2 4 6 8 1O 12 14 16 18 2O 22 24 26 28
Time (weeks)

Fig. 3.48 Hoop strains in glass and carbon-wrapped specimens, gages at the anodes

101

Chapter 4
Summary, Conclusions and
Recommendations

Experiments were conducted to assess the effects of using glass and carbon FRP
wraps in rehabilitating corrosion-damaged columns. Issues that were explored are: (1)
effect of freeze-thaw and wet-dry cycles on the properties of FRP panels; (2) freeze-thaw
durability of concrete square and cylindrical specimens wrapped with glass and carbon
FRP and subjected to an internal expansive force; and (3) effect of wrapping on the rate

of corrosion in an accelerated corrosion test.

4.1 F reeze-Thaw Test

Strength and durability tests were carried out on wrapped circular (diameter of
152 mm by 305 mm high) and square cylinders (152 mm by152 mm by 305 mm high).
The primary purpose of the tests was to determine the endurance of the jackets under
simulated cyclic environmental conditions. Creation of an internal bursting force similar
to that produced by corroding steel was attempted. This was done by fabricating
specimens with a hole in the longitudinal direction and ﬁlling it with an expanding
cement known as Bristar (used for silent demolition). Chloride was impregnated into the
cylinders during casting in order to simulate deteriorated concrete. Compression strength
tests were carried out on plain control cylinders as well as wrapped test specimens after

150 and 300 cycles of freeze-thaw conditioning. A total of sixty specimens were utilized

102

in the freeze-thaw test. The strength of FRP panels subjected to 300 freeze-thaw cycles
also was investigated.

Bristar was used in the wrapped specimens to investigate the durability of glass
and carbon wraps under sustained load and subjected to freeze-thaw cycling. The
sustained load simulated the load generated in wrapped columns by corrosion products.
The means of the compressive strength of freeze thaw specimens are not signiﬁcantly
different from those of control specimens at the 95% conﬁdence level. This holds both
for carbon and glass wraps, and for specimens with round and square cross sections. It
should be noted that a reduction in mean compressive strength was observed for carbon-
wrapped specimens after freeze-thaw conditioning. Based on our sample size and
statistical analysis, this difference is not signiﬁcant. The results indicate that the wraps
did not sustain signiﬁcant damage due to freeze-thaw cycling under sustained load.

The square wrapped specimens had lower compressive strength compared to the
round specimens, even though the cross sectional area of the square prisms is higher than
that of the round cylinders. This is due to the reduced conﬁnement provided by the wraps
for square cross sections and stress concentrations that develop at the corners. Wrapped
square prisms always failed by rupture of the wrap at a corner. A reduction of
approximately 30% to 40% in the failure strength was observed for the square specimens
compared to the round specimens. This loss of strength due to reduced conﬁnement in
square specimens is reasonably accounted for by Restrepo and Devino’s (1996) model of
conﬁnement. Richart’s model of conﬁnement (1928) predicts the strength of round

wrapped specimens reasonably well.

103

Compression strength of wrapped specimens is 1.4 to 2.6 times larger than the
strength of unwrapped specimens for square and round sections, respectively. Ductility of
wrapped specimens under compression is 4 to 9 times larger than that of unwrapped

specimens for square and round sections, respectively.

4.2 Accelerated Corrosion Test

Tests were conducted on twenty-four 152 mm x 305 mm concrete cylindrical
specimens. A water/cement ratio of 0.6 and 2% Cl' ion by weight of cement (intended to
simulate chloride contaminated columns) were used in the mix. Initially all specimens
were partially submerged in 3% NaCl solution and subjected to electrically induced
corrosion until cracks were visible on the exterior surfaces. After the initial corrosion
stage, specimens were wrapped with glass and carbon FRP. Specimens were then placed
in a tub and soaked in 3% NaCl solution for one hour each day while the electrically
induced accelerated corrosion continued for several weeks. Some specimens were
removed after 130 days and others after 190 days. The merit of using unbonded wraps to
minimize localized wrap strains near reinforcing bars was investigated. The ASTM 61
mass loss test was performed to determine the total corrosion of reinforcement in
unwrapped and wrapped specimens. The strength of FRP panels exposed to 190 wet-dry
cycles with salt water also was determined.

The mean corrosion depths for glass and carbon wraps are not signiﬁcantly
different for either the bonded or unbonded conditions. Both wrap systems, glass and
carbon, are equally effective in reducing the corrosion rate. Wrapping reduced the

corrosion depth by 46% — 59% after 190 days of testing.

104

The mean corrosion depths for bonded and unbonded specimens are signiﬁcantly
different after 190 days of accelerated corrosion. The bonded wrap is more effective in
reducing the rate of corrosion than the unbonded wrap. Corrosion depths for
reinforcement in specimens with unbonded wraps were approximately 20% more than
those in specimens with bonded wraps after 190 days of testing. This may be due to water
seepage between the concrete and the plastic sheet used to create the unbonded condition.

Wrap strains for bonded specimens with both types of wraps tend to level off with
time. One explanation could be that the stress concentration near the anodes in the
bonded wraps is more effective in containing the corrosion-induced crack and reducing
the corrosion rate. The slip of unbonded wraps and the resulting redistribution of strain
along the entire wrap may be less effective at containing the large corrosion-induced

crack near the anodes.

4.3 Recommendation for Field Installation

It is evident from the experimental study conducted that both carbon and glass
wrap systems are equally resistant to freeze-thaw cycles and reduce the corrosion rate by
about the same rate. Therefore, three layers of glass wrap or two layers of carbon wrap
may be used to repair Michigan bridge columns. Reducing the number of layers may also
be feasible, but it is not possible to provide any recommendation about this without
additional studies.

The preferred wrap system will most likely depend on the material and
installation cost rather than performance issues. However, it should be noted that many

studies indicate strength degradation of glass FRP in an alkaline and/or humid

105

environment under elevated temperature. Thus in regions with long periods of hot and
humid conditions, carbon FRP may be preferable to glass FRP.
It is also recommended that a non-destructive technique or coring be used every

ten years to monitor the condition of the concrete inside the wrap.

4.4 Cost Saving of using FRP Wraps

The estimated cost for the conventional chip and patch repair technique that is
currently used by the MDOT is approximately $500-$725/m2 of repaired column surface.
The estimated cost of the glass and carbon wrap systems used in this research study as

provided be the respective suppliers is provided in Table 4.1.

Table 4.1 Estimated material and installation cost for Tyfo-S glass and MBrace carbon
wrap systems

 

 

 

 

Wrap Type Matej'ial Cost Installation Cost No. of *Surfacez Total Cost/
lm llayer /m llayer Layers Prep./ m m

Glass $54 $54 3 $101 $425

Carbon $75 $54 2 $101 $360

 

 

 

 

 

 

* Estimated cost of surface preparation prior to wrap installation was provided by MDOT

Another cost factor that should be considered in this analysis is the expected life
span of the repair. It is estimated that the conventional chip and patch repair currently
used has a lifetime of approximately ﬁve years. The FRP wrap system on the other hand,
should last at least twice as long.

Factoring in life expectancy and assuming an inﬂation rate of 3% per year for
materials and labor, Table 4.2 provides total costs estimated for conventional repair (chip

and patch) and FRP wraps systems (glass and carbon) over a ten-year period beginning

106

 

now. Assuming 10 m2 of repaired surface area for a typical bridge column, the cost
saving over ten years by using an FRP wrap is estimated to be between $6,550 and
$12,050 per column. With the large number of bridge columns in need of repair in

Michigan, this cost saving clearly justiﬁes the use of FRP wraps for the repair.

Table 4.2 Estimated total cost of different repair techniques over ten years

 

 

 

 

 

 

Conventional Repair $1080-$1565/ m2
Glass FRP Wrap (3 layers) $425/ m2
Carbon FRP Wrap (2 layers) 8360/ m2

 

107

Chapter 5
Field Installation and Future Studies

5.1 Field Instrumentation

5.1.1 Corrosion Monitoring of Field Columns

Six corrosion probes were installed on six corrosion damaged ﬁeld columns in the
summer of 1999. These columns had considerable surface spalling and reinforcement was
exposed at several locations. Each column was also ﬁtted with two pre-weighed #13
reinforcing steel bars approximately 305 mm long to measure mass loss. The corrosion
probes and the steel bars were located at the same level of the existing column
reinforcement steel and about two meters above the roadway surface. The initial weights
of the bars are given in Table 5 .1. These columns are located on Lansing Road in
Lansing, Michigan under the I-96 overpass (Bridge ID SO9&S10 of 23 (52)). Columns
one through three are located under the westbound overpass while columns four through
six are located under the eastbound overpass. After superﬁcially repairing the column
surface spalls by patching (chloride was added to the patch to match existing chloride
content of the column), two columns were wrapped with two layers of carbon ﬁber
sheets, two were wrapped with three layers of glass ﬁber sheets and two were left

unwrapped.

108

Table 5.1 Initial weight of corrosion bars installed in ﬁeld columns

 

 

Column No Column 1 Column 2 Column 3
’ Glass wrapped Carbon wrapped Control
Bar No. 1 2 3 4 5 6

 

BarWt. (g) 252.09 276.14 270.70 263.80 263.58 271.13

 

 

 

 

 

Col mn No Column 4 Column 5 Column 6
u ' Control Glass wrapped Carbon wrapped
Bar No. 7 8 9 10 11 12

 

 

Bar Wt. (g) 270.76 270.70 275.87 270.26 270.23 275.85

 

 

 

 

 

 

 

 

The corrosion probes manufactured by Rohrback Cosasco Systems, Inc. (Santa Fe
Springs, California) are based on an electrical resistance measurement system. This
system is simple to install, directly measures the total corrosion, does not need regular
measurements, is smaller, easy to transport, and costs about $5,500 for six probes and the
readout measurement device.

Corrosion data is being collected about twice a month. Due to the short duration
since the corrosion probes were installed, the data is not signiﬁcantly different from the
initial readings. Corrosion monitoring is scheduled to continue for about 10 years. At that
time, the reinforcing bars installed to monitor mass loss will be cleaned and mass loss
analysis will be conducted.

Figs. 5.1 and 5 .2 show a column condition before and after minor surface repairs,
respectively. Fig. 5.3 shows a corrosion probe and pre-weighed reinforcing bars used for
monitoring mass loss. Figs. 5 .4 and 5.5 show the ﬁeld installation of glass and carbon
wraps, respectively. Fig. 5.6 shows columns after repairs were completed. The carbon
wrapped column (nearest to camera) is yet to be painted while the glass wrapped column

(middle one) has been painted.

109

 

Figs. 5.1 Column condition before surface repairs

 

Figs. 5.2 Column condition after surface repairs

110

 

Fig. 5.3 Corrosion probe and reinforcing bars for monitoring mass loss

 

Fig. 5.4 Field installation of glass wrap to selected columns

111

 

Fig. 5.5 Field installation of carbon wrap to selected columns

 

Fig. 5.6 Completed installation of glass and carbon wraps (glass wrapped column in the
background with top coat and ﬁnal paint layers applied).

112

Appendix D provides provisions for ﬁeld installation as recommended by the
wrap manufacturers — glass wrap system by Tyfo-S and carbon wrap system by Master

Builders.

5.1.2 Results of Field Monitoring

Wrapped and unwrapped columns in the ﬁeld have been monitored for 10 months

using corrosion probes. So far no signiﬁcant corrosion activity has been detected.

5.1.2.1 Calculation of Corrosion Rate
The corrosion rate for the concrete monitoring corrosion probes manufactured by
Rohrback Cosasco is calculated as follows:

ADzal Re admg x 0.365 x Span
ATime(Days)

 

Corrosion Rate (mils/year) =

The corrosion probe used in the ﬁled columns (Model 650-0-T50) has a span of 25 mils.

113

References

ACI (2000). “Guide for the Design and Construction of Externally Bonded FRP Systems
for Strengthening Concrete Structures.” ACI Committee 440 R.

Alampalli, S., O’Connor, J., and Yannotti, A. (1999). “Advancing Composites.” Journal
of Civil Engineering, December Al-A7.

Allen, M. L. (1995). “Probability of Corrosion Induced Cracking in Concrete.” Cement
and Concrete Research, Vol.25, No. 6, 1179-1190.

Almusallam, T. H., Al-Salloume, Y. A., and Alsayed, S. H. (2000). “Durability of
Concrete Cylinders Wrapped with GFRP Sheets at Different Environmental

Conditioning.” Seventh Annual International Conference on Composites Engineering,
27-28.

Arya C. and Sa'id-Shawqi, Q. (1996). " Factors Inﬂuencing Electrochemical Removal of
Chloride from Concrete.” Cement and Concrete Research, Vol. 26, No. 6, 851-860.

ASTM, 1990. “Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test
Specimens.” AST M Designation: G 1-90 (Reapproved 1994),
9-15.

Brockenbrough, R. L. and Gallagher, W. P. (1985). “Effect of Clamping Pressure and
Joint Geometry on Corrosion Induced Bowing and Distortion of Bolted Joints in
Weathering Steel.” Journal of Construction Steel Research, 213-238.

Chajes, Mertz, Thomson, and Farschman (1994). “Durability of Composite Material
Reinforcement.” Proceedings, Third Material Engineering Conference, ASCE, 598-605.

De Wilde, WP. (1988). Proceedings of the International Conference on Computer Aided
Design in Composite Material Technology. Computational Mechanics Publication,
Springer-Verlag Berlin Heidelberg.

Debaiky, A. and Green, G., and Hope, B., (1999). “FRP Rehabilitation of Corrosion-
Damaged Concrete Structures.” Proceedings of the 44"I International SAMPE
Symposium, May 23-27.

Demers, M. et al. (1996). “The Strengthening of Structural Concrete with an Aramid

Woven ﬁber/Epoxy Resin Composites”, Proceedings of the 2nd International Conference,
ACMBS, Montreal, PQ, Canada, pp.435-442.

114

Detwiler, R., Kjellsen K., and Gjorv, O. (1991). “ Resistance to Chloride Intrusion of
Concrete Cured at Different Temperatures.” ACI Materials Journal, Vol. 88, 19-24.

Fyfe, E. R., Watson, R. J. and Watson, S. C. (1996). “Long-Term Durability of
Composites Based on Field Performance and Laboratory Testing.” Proceedings of the
ICCI’96 Conference, Tucson, Arizona, 982-995.

Gomez, J., and Casto, B. (1996). “Freeze Thaw Durability of Composite Materials.”
Proceedings of the IC CI ’96 Conference, Tucson, Arizona, 947-955.

Guttman, H. and Sereda, P.J. (1968). “ Measurement of Atmospheric Factors Affecting
the Corrosion of Metals.” Metal Corrosion in the Atmosphere, ASTM STP 435, ASTM.

Halstead, O’Connor, Alampalli, and Minser. (2000). “Evaluating FRP Wrap with N DT
Methods.” Proceedings of the NDT Conference.

Hyun, A. (1995). “Fire Endurance and Hose Stream Tests of Three 48” high by 56” Wide
Non-Symmetrical Walls.” Technical Report, Inchape Testing Services of Warnock
Hersey, Inc., Pittsburgh, CA.

Kenneth, W. N. and Labossiere, P. “Fiber Composite Sheets in Cold Climate Rehab.”
Concrete International, V. 20, No. 6, June 1998, pp. 22-24.

Kestner, Harries, Pessiki, Sause, and Ricles (1997). “Rehabilitation of Reinforced
Concrete Columns using Fiber Reinforced Polymer Composite Jackets.” Report No. 97-
07. Advanced Technology for Large Structural Systems, Lehigh University, Bethlehem,
PA.

Lee, C. (1998). “Accelerated Corrosion and Repair of Reinforced Concrete Columns
Using CFRP Sheets”, M. Eng. Thesis, Department of Civil Eng., University of Toronto,
Toronto, ON, Canada, 106 p.

Lopez-Anido, R. (1993). “Inﬂuence of Temperature on the Service Life of Rebars.”
Cement and Concrete Research, Vol.23, No. 5, 1130-1190.

Malek, A. and Saadatmanesh H. (1996). “Physical and Mechanical Properties of Typical
Fibers and Resins.” Proceedings of the ICCI ’96 Conference, Tucson, Arizona, 68-79.

Mallick, P. K. (1993). Fiber Reinforced Composites—Materials, Manufacturing and
Design. Second Edition, Marcel Dekker Inc., New York, 326-327.

Mander, J. B., Priestley, M. J. N., and Park R. (1988). “Theoretical Stress-Strain Model

for Conﬁned Concrete.” Journal of Structural Engineering, ACSE, Vol. 114, No. 8,
1 804- 1 826.

115

Martin, H. and Schieles, P. (1969). “ The Inﬂuence of Time and Environmental
Condition on Corrosion of Deformed Bars in Cracked Concrete.” Preliminary Report of
RILEM International Symposium on Durability of Concrete, Vol. H, Prague.

Martin, H. and Schieles, P. (1969). “The Inﬂuence of Cracks on Corrosion of Steel in
Concrete.” Preliminary Report of RILEM International Symposium on Durability of
Concrete, Vol. H, Prague.

McCrum, R. (1994). MDOT ofﬁce memorandum to Sonny Jadun, June 22.

Mehta, P., and Monteiro, J. (1993). Concrete, structure, properties, and materials.
Second Edition, Prentice-Hall, Englewood Cliffs, 160-164.

Meier, U. (1996). “Composites for structural repair and retroﬁtting.” Proceedings,
ICCI’96 Conference, Tucson, Arizona, 1202-1216.

Michniewicz, J. (1996). “Repair and Rehabilitation of Reinforced Concrete Columns
with Fiber-Reinforced Plastics”, M. Eng. Thesis, Department of Civil Eng, University of
Toronto, Toronto, ON, Canada, 97 p.

Murphy, K., Zhang, S., and Karbhari, V. M. (1999). “Effect of Concrete Based Alkaline
Solutions on Short Term Response of Composites.” Proceeding of the 44”I International
SAMPE Symposium, May 23-27.

Neter, J., Wasserman, w., and Whitmore, GA. (1992). Applied Statistics, 4‘“ Edition,
Allyn and Bacon, Boston, MA.

Nilson A. and Winter G. (1991). Design of Concrete Structures. Eleventh Edition,
McGraw-Hill, New York, New York.

Pantazopolou et al. (1996). “Repair of Corrosion-Damaged Concrete Using ACM”,
Proceedings of the 2'“! International Conference, ACMBE, Montreal, PQ, Canada,
pp.287-298.

Picher, F., Rochette, P., and Labossiere, P. (1996). “Conﬁnement of Concrete Cylinders
with CFRP.” Proceedings of the IC CI ’96 Conference, Tucson, Arizona, 829-841.

Pigeon, M. and Pleau, R.(1998). Durability of Concrete in Cold Climates. Modern
Concrete Technology Series. E & FN Spon, London, UK, 1-5.

Restrepol, J. and DeVino, B. (1996). “Enhancement of the Axial Load Carrying Capacity

of Reinforced Concrete Columns by means Fiberglass Epoxy-Jackets.” Proceedings of
Advanced Composite Materials in Bridges and Structures II, Montreal, August, 547-553.

116

Richart F. E., Brandtzaeg A, and Brown R. L. (1928). “ A Study of the Failure of
Concrete under Combined Compressive Stresses.” University of ILL. Eng Exp. Stn. Bull
185.

Rivera, J. and Karbhari, V. (1999). Effects of Extended Freeze-Thaw exposure on
Composite Wrapped Concrete Cylinders.” Proceedings of the 44’” SAMPE Symposium,
May 23-27.

Rochette, P. and Labossiere, P. (1996). “ A Plasticity Approach for Concrete Columns

Conﬁned with Composite Materials”. Proceedings of the 2'"! International Conference,
ACMBS, Montreal, PQ, Canada, pp.359-366.

Sen, R., Mariscal, D., and Shahaway M. (1993). “Durability of Fiberglass Pretensioned
Beams.” ACI Structural Journal, Vol. 90, 525-533.

Sen, R., Mariscal, D., and Shahaway M. (1993). “Investigations of S2 Glass Epoxy
Strands in Concrete.” Proceedings of the F RP Components Structures, International

Symposium, 15-33.

Steckel, G. (2000). Personal Communication, The Aerospace Corporation, El Segundo,
CA.

Tarricone P. (1995). “Composite Sketch.” Civil Engineering Magazine, May, 52-55.
Toutanji, H. and Balaguru, P. (1998). “Durability Characteristics of Concrete Columns
Wrapped with FRP Tow Sheets.” Journal of Materials in Civil Engineering, February,
52-57.

Wrobel, P. (1994). “Laboratory Measurements of Corrosion Activity of Steel
Reinforcement in Concrete using Simple Equipment.” Cement, Concrete, and

Aggregates, CCAGPD, Vol. 16, No. 2, Dec. 1994, 100-103.

www.corrosionsource.com (2000).

Yamato, T., Emoto, Y., and Soeda, M. (1987). "Freezing and Thawing Resistance of
Concrete Containing Chloride." Concrete Durability, ACI SP100-50, Vol. 1, 901-917.

117

Appendix A

Aerospace Corporation’s FRP Panel Environmental Durability Data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fyfe Company E-Glass/Epoxy
SEH SllTny S Epoxy
SHORT
YOUNG’S TENSILE FAILURE BEAM GLASS WEIGHT
Emu MODULUS, STRENGTH. STRAIN, SHEAR TRANSITION “333' CHANGE,
msi IISI % TRENGTH, TEMP., "C %
ksl

CONTROL 3.96 1 0.13 80.5 1 5.1 2.10 1 0.18 5.9 1 0.5 65, 64, 68, 68 83 1 3

100%

HUMIDITY/38° C

1000 Hour 4.04 1 0.13 71.6 1 2.8 1.82 1 0.08 6.0 1 0.4 72 83 1 2 0.56
3000 Hour 3.94 1 0.10 67.9 1 1.9 1.77 1 0.05 5.8 1 0.3 73 84 1 2 0.82
10,000 Hour 3.93 1 0.18 51.4 1 2.1 1.31 1 0.08 4.5 1 0.3 73 82 1 2 1.09
ISALT WATER

1000 Hour 4.03 1 0.09 80.8 1 2.2 2.07 1 0.06 6.0 1 0.9 65 85 1 2 0.46
3000 Hour 4.02 1 0.04 81.7 1 1.2 2.09 1 0.03 5.6 1 0.2 63 84 1 3 0.57
10,000 Hour 4.09 1 0.07 66.0 1 1.9 1.64 1 0.04 4.6 1 0.2 63 82 1 2 0.91
pH 9.5 C8C03

SOLUTION

1000 Hour 3.85 1 0.03 83.2 1 2.8 2.25 1 0.11 5.9 1 0.3 65 83 1 2 0.36
3000 Hour 4.00 1 0.13 80.8 1 4.1 2.11 1 0.11 6.0 1 03 61 85 1 2 0.53
10,000 Hour 3.88 1 0.06 62.4 1 2.5 1.63 1 0.08 5.1 1 0.3 64 84 1 2 0.88
DRY HEAT AT

60° C

1000 Hour 3.89 1 0.06 82.0 1 1.7 2.17 1 0.08 6.4 1 0.4 95 85 1 2 -0.33
3000 Hour 4.05 1 0.06 84.8 1 2.4 2.16 1 0.09 6.7 1 0.8 87 85 1 2 -0.44
20 4.02 1 0.06 78.0 1 2.1 2.1” 1 0.06 5.2 1 0.3 68 82 1 3 0.59
FREEZWTHAW
{CYCLES

UV/CONDENSAT 4.03 1 0.08 84.0 1 3.1 2.18 1 0.11 6.5 1 0.2 36 83 1 3 -0.42
ION, 100 CYCLES
IDIESEL FUEL, 4 4.01 1 0.06 334 1 2.6 2.16 1 0.06 5.9 1 0.2 67 81 1 2

Hour

 

 

 

 

 

 

 

 

118

 

Master Builders "MBRACE" Carbon/Epoxy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CF-130/MBI Epoxy
SHORT
YOUNG’S TENSILE FAILURE BEAM GLASS WEIGHT
15%st MODULUS STRENGTH, STRAIN, SHEAR TRANSITION Hggggf’ CHANGE, %
, uni ks! 95 NGTH. TEMP. 'C (2 FLY/6 PLY)
ks!
ICONTROL 32.8 1 1.8 636 1 27 1.75 1 0.09 7.8 1 0.3 67, 67, 67, 70 92 1 2
100%
HUMIDITY/38° C
1000 Hour 34.0 1 1.4 591 1 25 1.59 1 0.08 7.6 1 0.1 75 91 1 1 1.13/0.95
3000 Hour 33.2 1 0.4 540 1 17 1.51 1 0.06 7.2 1 0.1 74 92 1 1 1.41/1.03
10,000 Hour 33.1 1 0.8 596 1 22 1.67 1 0.07 6.9 1 0.2 70 93 1 2 1.51/1.46
ISALT WATER
1000 Hour 33.6 1 0.5 619 1 25 1.70 1 0.05 7.5 1 0.2 65 90 1 3 1.14/0.65
3000 Hour 33.9 1 1.1 623 1 23 1.74 1 0.07 7.6 1 0.4 65 91 1 2 1.24/0.88
10,01” Hour 32.1 1 1.6 610 1 23 1.75 1 0.08 6.8 1 0.1 63 91 1 3 1.48/137
pH 9.5 CaCO,
OLUTION
1000 Hour 32.9 1 1.3 597 1 27 1.70 1 0.11 7.6 1 0.1 65 92 1 1 1.24/0.“
3000 Hour 31.8 1 0.8 585 1 35 1.70 1 0.09 7.2 1 0.6 67 91 1 2 1.27/ 1.02
10,000 Hour 33.1 1 1.5 615 1 39 1.70 1 0.12 6.7 1 0.2 62 92 1 1 131/0.78
RY HEAT AT
0° C
I1000 Hour 33.4 1 1.2 637 1 23 1.73 1 0.08 9.5 1 0.2 84 94 1 1 -0.47/-0.20
P000 Hour 32.6 1 0.9 582 1 12 1.67 1 0.05 8.6 1 0.4 85 93 1 1 - /-0.33
20 33.3 1 1.7 561 1 29 1.57 1 0.06 7.5 1 0.1 72 91 1 1 1.3”037
FREEZMHAW
CYCLES
ICONDENSAT 33.6 1 1.2 644 1 37 1.76 1 0.09 8.4 1 0.3 79 91 1 2 -0.63/-0.33
ION, 100 CYCLES
DIESEL FUEL, 4 34.1 1 1.5 589 1 9 1.61 1 0.08 8.2 1 0.1 66 93 1 3 0.0211100
Hour

 

 

 

 

 

 

 

 

119

 

Appendix B: Calculation Details

8.1 Strain in Column Wrap After 10 years

1.

3.

Geometric and Material Promrties

Column diameter, D = 91.44 cm

Column area, A, = 6563.60 em2

Initial volume of col. = A, (100 cm) = 656360 em3
Core diameter = dc = 80.01 cm

Longitudinal steel Ratio, p = 2%

Steel area, A,= p Ac: 131.27 cm2

Tie steel diameter, d = 1.27 cm 2

Tie Steel cross sectional area: ASp = ‘
Tie Spacing = 30.48 cm

Volume Of rust / volume of corroded steel = 6

= 1.266 cm2

 

Rust Volume

Corrosion rate for steel rebar (deformed) = 5%

Corroded Steel area, Am” = 0.05 A, = 6.56 cm2

Rust area, Am, = 6 Am, = 39.38 cm2

Rust Volume, Vm, = Am, (100 cm) = 3938 cm3

Corroded rate forties (smooth steel) = 20%

Corroded tie steel area, A6,“, = 0.2 A3,, = 0.2532 cm2

Rust area Spiral steel, Amy,“ = 6AM”, = 6 (0.2532) = 1.519 cm2

Rust volume tie steel, me-e = Arm“, (length of tie) = 1.519 ”dc 100/30.48 = 1252.0

C1113

Strain in Wrap
Change Of volume in vertical steel: Chg. V,, = %Vm, = %(3938) = 3281 .66cm3
Change of volume in tie steel: Chg. V,,, =%vm,_,,, = %(1252) = 1043.33 cm3

Final column volume: Final V = Initial column volume +Chg. Vs, +Chg.V,,,
= 656360 + 3281.66 + 1043.33 = 660684.99 cm3

 

2
Final column diameter: Vﬁw, = @100
Dﬁm, =9l.74 cm
Strain generated in wrap:
. D final - D initial
Stram = z .327 %
D initial

120

8.2 Confinlng Pressure and Strain in Steel Jacket

Conﬁning Pressure

From mechanics of thin walled cylinders, the conﬁning pressure can be
determined to be equal to:

f,=2(f,tn)/D=2(E£,tn)/D

 

For glass FRP,

f,= 2 (22011 x 0.00531 x 0.1227 x 3) / 15.24 = 5.645 MPa
where:

t = the thickness of the wrap per layer = 0.127 cm

f, = the circumferential stress of wrap

n = number of wrap layers = 3

D = the dia. of the concrete cylindrical column = 15.24 cm

E = modulus of elasticity of the wrap (glass/epoxy) = 22011 MPa
8, = wrap strain = 0.531

Strain in Steel Jacket
frD = Zfs: to
fu = (frD)/(2 to)
£51=frt/Est= (frD)/(2 tlesr)

= (5.645 x 15.24) / (2 x 0.477 x 200100) = 0.045%

121

where

f, = the conﬁning pressure generating the strain in the jacket = 5.645 MPa
(1 = diameter of steel jacket = 15.24 cm

f,, = stress in steel jacket

E 3, = modulus of elasticity of steel jacket = 200100 MPa

t,, = thickness of steel jacket =0.477 cm

8,, = strain in steel jacket

122

Appendix C: Specimen, Bar and Gage Numbering

Table C .l and for ﬁ'eeze—thaw test

            

reeze

Specimen Number Gage Number
1

2

123

Table C.2 bar and

 

 

Specimen Number
1

2

11
12
13
14
15
16
17
18
19
20
21
22
23

24

WCI‘C

 

for accelerated corrosion test

Ber Numbert

Gage Number
1

 

" Gage located on the anode-others are located in between anode and cathode

124

 

Appendix D
FIELD INSTALLATION OF FIBER REINFORCED
PLASTIC (FRP) WRAPS
ON CONCRETE COLUMNS

Materials-The following is a list of material required for each type of composite
system:

1. TYFO SHE-51 composite system (glass):
Vendor: FYFE Co.
6044 Cornerstone Court West, Suite C
San Diego, CA 92121-4730
Tel: 619-642-0694

Local Rep.: Kurt Baron
Tel: 847-706-9230

Fabric: Tyfo SEH-51

Epoxy: Tyfo-S, parts A & B

Top coat: Tyfo-WS, parts A&B

Paint: Tyfo A (acrylic), or Tyfo U (polyurethane). Color:
Concrete gray

2. MBrace composite system (carbon):
Vendor: Master Builders, Inc.
23700 Chagrin Boulevard
Cleveland, OH 44122
Tel: 800-MBT-9990

Local Rep: Brad Costello
Tel: 248-683-3554

Fabric: MBrace CF130

Primer: MBrace Primer, parts A&B
Saturant: Mbrace Saturant, parts A&B
Filler: Mbrace Putty, parts A&B

Top Coat: Mbrace Top Coat. Color: Concrete gray

125

Construction Methods.-

Do not apply when ambient temperatures are lower than 4°C.

Do not apply to wet surface or when rainfall is anticipated.

Do not apply when dew point is within 3°C of the concrete surface
temperature.

Do not apply when humidity is 90% or higher.

Manufacturer’s representative shall be on site for initial placement.
Directions of the manufacturer’s representative shall be followed.

1. TYFO SEH composite system:

A.

Remove loose concrete from the column surface and ﬁll all voids to a
smooth surface using Type S-F or Type S-M patching mixture,
depending on the depth of the patch. The area of the patch shall be
blown out with oil-free compressed air. The cleaned area for patching
shall be ﬂushed out with clean water under pressure immediately prior
to application of the patching mixture. Forming methods used to retain
the patching mixture shall not leave holes in the concrete surface. The
patch shall cure for at least three days prior to application of wrap.
Uneven surfaces, protrusions, and sharp edges shall be ground smooth.
Dust from surface grinding shall be removed by using an oil-free air
blower or other suitable means.

Pre-cut desired number of layers of fabric to a length exceeding the
column perimeter by at least 150 mm. Off site labor shall be used
where possible. An overlap of 150 mm shall be provided in the ﬁber
direction when terminating the wrap.

Round or bevel comers by grinding to a radius of at least 20 mm.

The concrete surface shall be free of any moisture at the time of
application.

Mix parts A and B of Tyfo-S epoxy at a ratio of 100:42 by volume (or
100:34.5 by weight) with a tolerance of 10%. Stir with a mechanical
mixer, typically 5 minutes at 400 to 600 rpm until uniformly blended.
Pot life is 3 hours at 20°C and 45 minutes at 38°C.

Apply one coat of Tyfo-S epoxy by brush or roller to prime the
surface. Volume to be applied may vary depending on the porosity of
the concrete surface. Wait 2 to 4 hours and then apply wrap (as
described below) while primer is still tacky.

Tyfo-S epoxy shall be applied to TYFO fabric as follows:

126

(1) Place dry fabric sheets in a saturation bath and add epoxy. Work
epoxy into fabric using gloved hands, a paint roller, or similar tool.
Alternatively, an automatic saturating machine may be used.

(2) After the fabric has been completely saturated (both sides), remove
excess epoxy by squeegying it out with a plastic trowel or by
blotting the excess resin with the next dry fabric to be saturated.

(3) Use a PVC pipe to spool the saturated fabric prior to wrapping
column.

H. The fabric shall be placed on the column entirely by hand assuring a
smooth, uniform, mat ﬁnish. The (white) glass ﬁbers shall be oriented
horizontally. The (yellow) Kevlar ﬁbers will then be oriented
vertically. Pull the layer so that it is taut and free of bubbles. A lap
length of at least 150 mm is required in the ﬁber longitudinal direction
when terminating the layer.

I. Apply additional layers while the wrapped column surface is still tacky
to the touch, ensuring that overlaps are staggered.

J. Mix parts A and B of Tyfo-WS epoxy at ratio of 100:42 by volume (or
100:34.5 by weight) with a tolerance of 10%.

K. Apply ﬁnal coat of Tyfo-WS top coat with a minimum thickness of 0.4
mm. Epoxy shall be tacky to touch before ﬁnal coat is applied.

L. The system shall be protected from damage, debris, and moisture
during the initial curing period of 24 hours. Final curing is completed
in 72 hours.

M. Apply ﬁnish of two coats of Tyfo A or Tyfo U paint with a minimum
thickness of 0.1 mm per coat.

2. MBrace composite system

A. Remove loose concrete from the column surface and ﬁll all voids to a
smooth surface using Type S-F or Type S-M patching mixture,
depending on the depth of the patch. The area of the patch shall be
blown out with oil-free compressed air. The cleaned area for patching
shall be ﬂushed out with clean water under pressure immediately prior
to application of the patching mixture. Forming methods used to retain
the patching mixture shall not leave holes in the concrete surface.
Uneven surfaces to receive FRP shall be ﬁlled with Type S-F mixture
or other approved material. Surface irregularities must be rounded and
smoothed to less than 1 m using a grinder. Dust from surface
grinding shall be removed by using an oil free air blower or other

127

suitable means.

. Pre-cut desired number of layers of fabric to a length exceeding the
column perimeter by at least 100 mm. Off site labor shall be used
where possible. An overlap of 100 mm shall be provided in the ﬁber
direction when terminating the wrap. The length of the carbon sheet
should preferably be less than 3 m for manageability.

. Round or bevel corners by grinding to a radius of at least 13 mm.

. The concrete surface shall be free of any moisture at the time of
application.

. Mix parts A and B of the primer at a ratio of 3:1 by volume (or 100230
by weight). Stir with a mechanical mixer for at least 3 minutes at 400
to 600 rpm until uniformly blended. Pot life of the epoxy is
approximately one-half to one hour after mixing.

. Prime the concrete surface with the primer using a brush or a roller at a
rate of 0.25 to 0.4 kg/mz. Volume to be applied may vary depending
on the porosity of the concrete surface.

. Mix parts A and B of the ﬁller at a ratio of 3:1 by volume (or 100:30
by weight).

. Apply ﬁller to primed surface preferably within 1 or 2 days of, and no
more than one week after, applying the primer. The ﬁller may be
applied immediately after applying the primer. Applying the ﬁller is
optional.

Wait 2 to 4 hours before mixing and applying the saturant. The
surface must be tacky to touch when applying the saturant, otherwise it
must be roughened using sandpaper. If sandpaper is used, then the
surface shall be cleaned using an air blower before applying the
saturant.

Mix parts A and B of the saturant at a ratio of 3:1 by volume (or
10034 by weight). Pot life of the saturant is approximately one-half
to one hour.

. Apply one coat of the saturant to primed surface using a medium nap
(9.5 mm) roller at a rate of 0.25 to 0.4 kg/mz.

. Place carbon sheet on a ﬂat horizontal surface so that the backing
paper is on top, smooth down by hand and peel away backing paper,
and then wrap the sheet around the column. The surface that originally

128

contained the backing paper shall be placed against the column and the
ﬁbers shall be oriented horizontally (i.e., the corners containing the
obtuse angles of the diamond stitch pattern shall be aligned
horizontally).

. Squeeze the surface of adhered carbon sheet in the ﬁber direction in
order to impregnate the saturant into the sheet. Remove excess epoxy
by squeegying it out with a plastic trowel (without sharp edges) and
roll out bubbles.

. Apply additional saturant over the bonded carbon sheet at the overlap.

. Wait a minimum of 30 minutes and then roll on an overcoat of
saturant. (This is done for each layer.)

. Repeat saturant mixing and rolling for consecutive carbon sheets,
waiting 1 to 2 hours after applying each layer. Apply while previous
coat of saturant is tacky to touch and stagger wrap overlaps for each
layer.

. The system shall be protected from damage, debris, and moisture
during the curing period of not less than 24 hours.

. Apply ﬁnish of two coats of Mbrace Top Coat using a 10 mm nap
roller at a rate of 4.9 m2/liter per coat.

129