Isomat Documentatie
Isomat Documentatie
Isomat Documentatie
TRIANTAFILLOU
UNIVERSITY OF PATRAS
DEPARTMENT OF CIVIL ENGINEERING
STRUCTURAL MATERIALS LABORATORY
PATRAS, GREECE
2005
iii
CONTENTS
page
PREFACE
CONTENTS
iii
CHAPTER 1 INTRODUCTION
1.1
General
1.2
2.1
Materials
2.1.1
General
2.1.2
Fibers
2.1.3
Matrix
2.1.4
Composite materials
Example 2.1
2.1.5
2.2
2.3
Adhesives
9
10
Strengthening systems
11
2.2.1
11
2.2.2
Prefabricated elements
12
13
15
3.1
General
15
3.2
15
3.2.1
15
3.2.2
17
3.2.3
17
3.3
17
3.3.1
General, behavior
17
3.3.2
Analytical model
19
Example 3.1
20
iv
page
CHAPTER 4 FLEXURAL STRENGTHENING
21
4.1
General
21
4.2
Initial situation
22
4.3
23
4.4
25
4.4.1
25
4.4.2
27
4.5
Ductility considerations
30
4.6
31
4.7
Example
32
4.8
34
4.9
Columns
35
39
5.1
General
39
5.2
41
5.3
44
Example 5.1
45
Example 5.2
47
Example 5.3
47
Beam-column joints
48
5.4
CHAPTER 6 CONFINEMENT
51
6.1
General
51
6.2
52
6.2.1
Behavior
52
6.2.2
Design model
54
6.3
6.4
6.5
Example 6.1
58
58
Example 6.2
63
Lap-splices
64
6.4.1
64
Example 6.3
66
6.4.2
67
Rebar buckling
68
Example 6.4
69
page
6.6
69
71
7.1
General
71
7.2
Detailing
71
7.2.1
Flexural strengthening
71
7.2.2
Shear strengthening
73
7.2.3
Confinement
74
7.3
Practical execution
76
CHAPTER 8 DURABILITY
79
8.1
General
79
8.2
Temperature effects
79
8.3
Moisture
79
8.4
UV light exposure
80
8.5
80
8.6
Galvanic corrosion
81
8.7
81
8.8
Fatigue
81
8.9
Impact
82
REFERENCES
83
87
vi
PREFACE
ii
INTRODUCTION
CHAPTER 1
INTRODUCTION
1.1 General
The issue of upgrading the existing civil engineering infrastructure has been one of
great importance for over 15 years or so. Deterioration of bridge decks, beams, girders
and columns, buildings, parking structures and others may be attributed to ageing,
environmentally induced degradation, poor initial design and/or construction, lack of
maintenance, and to accidental events such as earthquakes.
The infrastructures
increasing decay is frequently combined with the need for upgrading so that structures
can meet more stringent design requirements (e.g. increased traffic volumes in bridges
exceeding the initial design loads), and hence the aspect of civil engineering
infrastructure renewal has received considerable attention over the past few years
throughout the world. At the same time, seismic retrofit has become at least equally
important, especially in areas of high seismic risk.
Recent developments related to materials, methods and techniques for structural
strengthening and seismic retrofitting have been enormous. One of todays state-of-theart techniques is the use of fiber reinforced polymer (FRP) materials or simply
composites, which are currently viewed by structural engineers as new and highly
promising materials in the construction industry. Composite materials for strengthening
of civil engineering structures are available today mainly in the form of: (a) thin
unidirectional strips (with thickness in the order of 1 mm) made by pultrusion, (b) flexible
sheets or fabrics, made of fibers in one or at least two different directions, respectively
(and sometimes pre-impregnated with resin). Central to the understanding of composites
bonded to concrete is the fact that stresses in these materials are carried only by the
fibers, in the respective directions.
The reasons why composites are increasingly used as strengthening materials of
reinforced concrete members may be summarized as follows: immunity to corrosion; low
weight (about of steel), resulting in easier application in confined space, elimination of
the need for scaffolding and reduction in labor costs; very high tensile strength (both
static and long-term, for certain types of FRP materials); stiffness which may be tailored
to the design requirements; large deformation capacity, which results in substantial
member ductility; and practically unlimited availability in FRP sizes and FRP geometry
and dimensions. Composites suffer from certain disadvantages too, which are not to be
T. C. Triantafillou
INTRODUCTION
potential drawbacks, and final decisions regarding their use should be based on
consideration of several factors, including not only mechanical performance aspects, but
also constructability and long-term durability.
Composites have found their way as strengthening materials of reinforced concrete
(RC) members (such as beams, slabs, columns etc.) in many thousands of applications
worldwide, where conventional strengthening techniques may be problematic (e.g. steel
plating or steel jacketing). For instance, one of the popular techniques for upgrading RC
elements has traditionally involved the use of steel plates epoxy-bonded to the external
surfaces (e.g. tension zones) of beams and slabs. This technique is simple and effective
as far as both cost and mechanical performance is concerned, but suffers from several
disadvantages (Meier 1987): corrosion of the steel plates resulting in bond deterioration;
difficulty in manipulating heavy steel plates in tight construction sites; need for
scaffolding; and limitation in available plate lengths (which are required in case of flexural
strengthening of long girders), resulting in the need for joints. Replacing the steel plates
with FRP strips provides satisfactory solutions to the problems described above. Another
common technique for the strengthening of RC structures involves the construction of
reinforced concrete (either cast in-place or shotcrete) jackets (shells) around existing
elements. Jacketing is clearly quite effective as far as strength, stiffness and ductility is
concerned, but it is labour intensive, it often causes disruption of occupancy and it
provides RC members, in many cases, with undesirable weight and stiffness increase.
Jackets may also be made of steel; but in this case protection from corrosion is a major
issue, as is the rather poor confining characteristics of steel-jacketed concrete. The
conventional jackets may be replaced with FRP in the form of sheets or fabrics wrapped
around RC members, thus providing substantial increase in strength (axial, flexural,
shear, torsional) and ductility without much affecting the stiffness.
T. C. Triantafillou
INTRODUCTION
The
Appendix describes the use of the program Composite Dimensioning for the
dimensioning of RC members strengthened in flexure, shear or through confinement.
T. C. Triantafillou
INTRODUCTION
T. C. Triantafillou
CHAPTER 2
T. C. Triantafillou
this table correspond to monotonic loading and do not account for environmental
degradation and/or sustained loading effects (see Chapter 8).
(MPa)
6000
High
modulus
Carbon
4000
High
strength
Carbon
Aramid
Glass
2000
Mild steel
0
0
0.04
0.02
Fig. 2.1 Typical uniaxial tension stress-strain diagrams for different fibers and comparison with
steel.
Elastic modulus
(GPa)
Tensile strength
(MPa)
Ultimate tensile
strain (%)
215-235
215-235
350-500
500-700
3500-4800
3500-6000
2500-3100
2100-2400
1.4-2.0
1.5-2.3
0.5-0.9
0.2-0.4
70-75
70-75
85-90
1900-3000
1900-3000
3500-4800
3.0-4.5
3.0-4.5
4.5-5.5
70-80
115-130
3500-4100
3500-4000
4.3-5.0
2.5-3.5
Carbon fibers are normally either based on pitch or PAN, as raw material. Pitch
fibers are fabricated by using refined petroleum or coal pitch that is passed through a thin
nozzle and stabilised by heating.
carbonised through burning. The pitch base carbon fibers offer general purpose and high
strength/elasticity materials. The PAN-type carbon fibers yield high strength materials
and high elasticity materials. The density of carbon fibers is 1800-1900 kg/m3. Glass
fibers for continuous fiber reinforcement are classified into three types: E-glass fibers, Sglass and alkali resistant AR-glass fibers. E-glass fibers, which contain high amounts of
T. C. Triantafillou
boric acid and aluminate, are disadvantageous in having low alkali resistance. S-glass
fibers are stronger and stiffer than E-glass, but still not resistant to alkali. To prevent
glass fiber from being eroded by cement-alkali, a considerable amount of zircon is added
to produce alkali resistance glass fibers; such fibers have mechanical properties similar
to E-glass. An important aspect of glass fibers is their low cost. The density of glass
fibers is 2300-2500 kg/m3. Aramid fibers were first introduced in 1971, and today are
produced by several manufacturers under various brand names (Kevlar, Twaron,
Technora). The structure of aramid fiber is anisotropic and gives higher strength and
modulus in the fiber longitudinal direction. Aramid fibers respond elastically in tension but
they exhibit non-linear and ductile behaviour under compression; they also exhibit good
toughness, damage tolerance and fatigue characteristics. The density of aramid fibers is
1450 kg/m3.
2.1.3 Matrix
The matrix for a structural composite material is typically a polymer, of
thermosetting type or of thermoplastic type, with the first being the most common one.
Recent developments have resulted in matrices based on inorganic materials) (e.g.
cement-based). The function of the matrix is to protect the fibers against abrasion or
environmental corrosion, to bind the fibers together and to distribute the load. The
matrix has a strong influence on several mechanical properties of the composite, such
as the transverse modulus and strength, the shear properties and the properties in
compression. Physical and chemical characteristics of the matrix such as melting or
curing temperature, viscosity and reactivity with fibers influence the choice of the
fabrication process. Hence, proper selection of the matrix material for a composite
system requires that all these factors be taken into account.
Epoxy resins, polyester, vinylester and phenolics are the most common polymeric
matrix materials used with high-performance reinforcing fibers. They are thermosetting
polymers with good processibility and good chemical resistance.
Epoxies have, in
general, better mechanical properties than polyesters and vinylesters, and outstanding
durability, whereas polyesters and vinylesters are cheaper. Phenolics have a better
behavior at high temperatures.
2.1.4 Composite materials
Advanced composites as strengthening materials consist of a large number of small,
continuous, directionalized, non-metallic fibers with advanced characteristics, bundled in
the matrix (Fig. 2.2).
(carbon fiber based), GFRP (glass fiber based) or AFRP (aramid fiber based); when
T. C. Triantafillou
different types of fibers are used, the material is called hybrid. Typically, the volume
fraction of fibers in advanced composites equals about 50-70% for strips and about 2535% for sheets. Given also that the elastic modulus of fibers is much higher than that of
the matrix, it becomes clear that the fibers are the principal stress bearing components,
while the matrix transfers stresses among fibers and protects them. Different techniques
are used for manufacturing (e.g. pultrusion, hand lay-up), detailed descriptions of which
are outside the scope of this document. As externally bonded reinforcement for the
strengthening of structures, advanced composite materials are made available in various
forms, which are described in Section 2.2.
Matrix
Fiber
Fig. 2.2 Magnified cross section of a composite material with unidirectional fibers.
E f E fib Vfib + E m Vm
(2.1)
f f f fib Vfib + fm Vm
(2.2)
where:
Ef
E fib
Em
Vfib
Vm
ff
T. C. Triantafillou
f fib
fm
At this point we should note that since E fib / E m >>1 and f fib / fm >>1, the above equations
are approximately valid even if the second terms in the right parts are omitted.
In case of prefabricated strips the material properties based on the total crosssectional area can be used in calculations and are usually supplied by the manufacturer.
In case of in-situ resin impregnated systems, however, the final composite material
thickness and with that the fiber volume fraction is uncertain and may vary. For this
reason the properties of the total system (fibers and matrix) and the actual thickness
should be provided based on experimental testing. Note that manufacturers sometimes
supply the material properties for the bare fibers. In this case a property reduction factor
r1 should apply, to be provided by the supplier of the strengthening system. The above is
better explained in the following example.
Example 2.1
Material supplier provides unidirectional carbon sheets, with a weight of 260 g/m2.
Fiber properties are as follows: E fib = 230 GPa, f fib = 3500 MPa. The nominal thickness
of the sheet, t fib , is calculated based on the fiber material density, say fib = 2000 kg/m3,
as follows: fib : fib x t fib = 260, hence t fib = 0.13 mm. We assume that after resin
impregnation, the composite material reaches a thickness of 0.3 mm, implying a
volumetric fraction of fibers equal to Vfib = 0.13/0.3 = 43%. If the tensile strength and the
elastic modulus of the composite material were measured experimentally, the results
would be lower than 0.43x230 GPa and 0.43x3500 MPa, respectively, say by 10%
(hence r1 = 0.9): 89 GPa 1355 Pa. Therefore, the composite material properties to
be used in calculations should be one of the following:
(a) E f = 89 GPa, f f = 1355 MPa, t f = 0.3 mm, or
(b) E f = 0.9x230 GPa, f f = 0.9x3500 MPa, t f = 0.13 mm.
In a real application the amount of impregnating resin could, in general, be different from
that suggested by the supplier, hence the real thickness of the composite will not be
equal to 0.3 mm. But what is of interest in the calculations is typically the product E f t f
or, sometimes, the product f f t f , hence the above two solutions (a) and (b) are
equivalent. The advantage of solution (a) is that the properties provided by the supplier
are quite close to those expected in-situ and the disadvantage is that those properties are
hypothetical. On the other hand, the advantage of solution (b) is that the material
provided by the supplier is accompanied by a set of properties, which could be combine
T. C. Triantafillou
10
with the proper reduction factor (0.9 in our example) to yield the properties of the in-situ
applied composite.
2.1.5 Adhesives
The purpose of the adhesive is to provide a shear load path between the concrete
surface and the composite material, so that full composite action may develop. The
most common type of structural adhesives is epoxy, which is the result of mixing an
epoxy resin (polymer) with a hardener. Other types of adhesives may be based on
inorganic materials (mainly cement-based). Depending on the application demands,
the adhesive may contain fillers, softening inclusions, toughening additives and others.
When using epoxy adhesives there are two different time concepts that need to be
taken into consideration. The first is the pot life and the second is the open time. Pot
life represents the time one can work with the adhesive after mixing the resin and the
hardener before it starts to harden in the mixture vessel; for an epoxy adhesive, it may
vary between a few seconds up to several years. Open time is the time that one can
have at his/her disposal after the adhesive has been applied to the adherents and
before they are joined together.
Glass-like
Rubbery
Viscous flow
T. C. Triantafillou
11
Typical properties for cold cured epoxy adhesives used in civil engineering
applications are given in Table 2.2 (fib 2001). For the sake of comparison, the same
table provides information for concrete and mild steel too.
Table 2.2
Density (kg/m3)
Elastic modulus (GPa)
Shear modulus (GPa)
Poissons ratio
Tensile strength (MPa)
Shear strength (MPa)
Compressive strength (MPa)
Tensile strain at break (%)
Approximate fracture energy (Jm-2)
Coefficient of thermal expansion (10-6/C)
Water absorption: 7 days - 25 C (% w/w)
Glass transition temperature (C)
Epoxy
adhesive
1100 1700
0.5 - 20
0.2 8
0.3 0.4
9 - 30
10 - 30
55 - 110
0.5-5
200-1000
25 - 100
0.1-3
50 - 80
Concrete
Mild steel
2350
20 - 50
8 - 21
0.2
1-4
2-5
25 - 150
0.015
100
11 - 13
5
---
7800
205
80
0.3
200 - 600
200 - 600
200 - 600
25
105-106
10 - 15
0
---
These
materials are based on cement in combination with other binders (e.g. fly ash, silica
fume, metakaolin), additives (e.g. polymers) and fine aggregates.
adhesive also plays the role of the matrix in the composite material, hence it must be
designed such that compatibility with the fibers (textiles) will be maximized. General
requirements for inorganic binders are high shear (that is tensile) strength, suitable
consistency, low shrinkage and creep and good workability.
T. C. Triantafillou
12
. 2.4
. 2.5
Dry fiber tows (untwisted bundles of continuous fibers) that are wound or otherwise
mechanically placed onto the concrete surface. Resin is applied to the fiber during
winding.
Pre-impregnated fiber tows that are wound or otherwise mechanically placed onto
the concrete surface. Product installation may be executed with or without additional
resin.
Pre-manufactured cured straight strips, which are installed through the use of
adhesives. They are typically in the form of thin ribbon strips or grids that may be
delivered in a rolled coil. Normally strips are pultruded. In case they are laminated,
also the term laminate instead of strip may be used.
T. C. Triantafillou
13
Pre-manufactured cured shaped shells, jackets or angles, which are installed through
the use of adhesives. They are typically factory-made curved or shaped elements or
split shells that can be fitted around columns or other elements.
The suitability of each system depends on the type of structure that shall be
strengthened. For example, prefabricated strips are generally best suited for plane and
straight surfaces (e.g. bottom of beams and slabs), whereas sheets or fabrics are more
flexible and can be used to plane as well as to convex surfaces (e.g. sides of beams,
column wrapping).
Hand lay-up of
CFRP strip
Hand lay-up of
carbon fiber sheets
Column wrapping
using CFRP fabric
Strengthening of cooling tower with carbon fiber sheets
Fig. 2.6 Examples for the application of the basic FRP strengthening technique.
T. C. Triantafillou
14
The basic FRP strengthening technique, which is most widely applied, involves the
manual application of either wet lay-up (so-called hand lay-up) or prefabricated systems
by means of cold cured adhesive bonding. Common in this technique is that the external
reinforcement is bonded onto the concrete surface with the fibres as parallel as
practically possible to the direction of principal tensile stresses. Typical applications of
the hand lay-up and prefabricated systems are illustrated in Fig. 2.6.
Apart from the basic technique there is a number of special techniques with rather
limited applicability: automated wrapping, prestressed FRP, in-situ fast curing using
heating device, near-surface-mounted bars, mechanical fastening etc. The description of
these not so common techniques falls outside the scope of this document.
T. C. Triantafillou
BASIS OF DESIGN
15
CHAPTER 3
BASIS OF DESIGN
3.1 General
The design of RC members strengthened with FRP follows the philosophy of the
relevant design codes (e.g. Eurocodes 2 and 8) and involves the verification for the
ultimate and serviceability limit states, with proper modifications to account for the
contribution of FRP.
For steel
reinforcement fyd = fyk / s , where fyk = representative value of yield stress and s =
safety factor for steel.
The strength of composite materials (added materials) is represented by the
characteristic value if the safety verification is performed in terms of strength, or by the
mean value if the safety verification is performed in terms of deformations.
Their
behavior in uniaxial tension is assumed linear elastic to failure, according to eq. (3.1);
failure is defined at a (design) stress ffd = ffk / f :
f = Ef f
ffd
(3.1)
T. C. Triantafillou
BASIS OF DESIGN
16
c
f
ck
c
concrete
FRP
ffd
Ef
Steel
fyd
0.2%
0.35%
. 3.1
yd
fu
At this point we should point out that that the in-situ tensile strength of FRP is lower
than that measured in a uniaxial tension test, due to stress concentrations, complex
multiaxial states of stress, several layers, environmental degradation effects etc.
All
these reduction factors may be taken into account by assuming that FRP reaches failure
at an effective strain fue , which is less than the mean ultimate strain fum determined
through testing. On the basis of the above, the design value of the effective strength for
FRP, ffde , is given as follows:
f fde =
fue f fk
= e f fd
fum f
(3.2)
More details on the effective strain fue will be given in the sections where this strain
plays an important role (e.g. shear strengthening, confinement).
(1)
(2)
FRP type
CFRP
1.20
1.35
AFRP
1.25
1.45
GFRP
1.30
1.50
Application of prefab FRP systems under normal quality control conditions. Application
of wet lay-up systems if all necessary provisions are taken to obtain a high degree of
quality control on both the application conditions and the application process.
Application of wet lay-up systems under normal quality control conditions. Application of
any system under difficult on-site working conditions.
T. C. Triantafillou
BASIS OF DESIGN
17
Values for the FRP material safety factor are suggested in Table 3.1 (fib 2001). Note
that these values are still a topic of current research and are subject to further
refinements.
relationship between the mean shear stress b at the FRP-concrete interface (equal to
Nf / l bb f in Fig. 3.2, where b f the width of FRP) and the slip s f . This relationship
depends on many factors, including the concrete strength, the type of adhesive, the FRP
characteristics (e.g. thickness, elastic modulus) and the bond length. A typical shear
T. C. Triantafillou
BASIS OF DESIGN
18
stress slip curve is plotted in Fig. 3.4, along with others for deformed and smooth steel
rebars, which are provided for the sake of comparison.
Bond length
slip sf
lb
FRP
debonding
adhesive
Nf
concrete
Nc
Nf
Fig. 3.2 Simplified FRP-concrete bond test (e.g. Zilch et al 1998, Bizindavyi and Neale 1999).
crack
propagation
crack propagation
Fig. 3.3 Cracking in RC beam and possible debonding (the arrows indicate the crack
propagation).
8
CFRP strip
tf = 1.2 mm
6
4
embedded steel bar
12 mm (smooth)
0
0,0
0,2
0,4
0,6
0,8
1,0
slip sf (mm)
Fig. 3.4 (Zilch et al. 1998).
T. C. Triantafillou
BASIS OF DESIGN
19
FRP to cause debonding, that is the maximum anchorable force, N fa , increases with the
bond length l b , until this length reaches a limiting value, beyond which the maximum
anchorable force remains practically unchanged, equal to N fa,max (Fig. 3.5).
Nfa,max
Nfa
Bond length l b
l b,max
if l b < l b,max :
Nfa = Nfa,max
lb
l b,max
l b,max =
2 lb
l b,max
Ef t f
c 2 fctm
(mm)
(N)
(3.3a)
(N)
(3.3b)
(3.4)
with
1.125 2 f
b
kb =
b
1+ f
400
(3.5)
T. C. Triantafillou
BASIS OF DESIGN
20
member cross section (mm), fctm = mean tensile strength of concrete (MPa), E f = elastic
modulus of FRP (MPa) and t f = thickness of FRP (mm). Moreover, c1 = 0.64 (or 0.50, if
the characteristic value of N fa,max is to be calculated) and c 2 = 2.0.
In terms of stresses, the above model results in the following equations for the FRP
design stress ( fd = Nfad / b f t f ) corresponding to debonding:
if l b l b,max :
fd =
0.5k c k b
f ,b
fctmE f
tf
if l b < l b,max :
fd =
0.5k c k b
f ,b
fctmE f l b
t f l b,max
2 lb
l b,max
(MPa)
(3.6a)
(MPa)
(3.6b)
Example 3.1
Consider an FRP strip with width b f = 50 mm, thickness t f = 1.2 mm, elastic modulus
E f = 180 GPa and tensile strength f f = 3000 MPa, epoxy-bonded on a concrete member
with a width b = 100 mm (Fig. 3.6). The mean tensile strength of concrete is assumed
fctm = 1.9 MPa.
N fa
. 3.6
lb
50 mm
100 mm
50
1.125 2
100
kb =
= 1.22 > 1,
50
1+
400
hence from eq. (3.3a) we calculate Nfa,max = 0.64 1.0 1.22 50 1.9 180000 1.2 =
25010 25 kN, corresponding to a stress in the FRP equal to 25010/(50x1.2) = 417
MPa [it is worth noting here that if the strip reached its tensile capacity the respective
force would be Nf = 3000x(50x1.2)/1000 = 180 kN, that is about 7 times higher than that
causing].
In terms of stresses, the design stress in the FRP at debonding (assuming a bond
length at least equal to 238 mm) is given by eq. (3.6) (with material safety factor f,b = 1.5)
fd = 217 MPa.
T. C. Triantafillou
FLEXURAL STRENGTHENING
21
CHAPTER 4
FLEXURAL STRENGTHENING
4.1 General
Reinforced concrete members, such as beams and columns, may be strengthened in
flexure through the use of strips or sheets epoxy-bonded to their tension zones, with the
direction of fibers parallel to that of high tensile stresses (member axis). The concept is
illustrated in Fig. 4.1. Flexural strengthening of columns is, in general, more difficult to
achieve, due to the requirements for anchorage of the FRP through the joints. The latter
is easy to construct if the width of beams is smaller than that of columns (hence sufficient
space is available to bond strips, Fig. 4.2b), but requires small FRP cross sections placed
near the column corners if the column and the beams have similar dimensions (Fig.
4.2c).
The analysis for the ultimate limit state in flexure may follow well-established
procedures for reinforced concrete structures, provided that: (a) the contribution of
external FRP reinforcement is taken into account properly (linear elastic material); and (b)
special consideration is given to the issue of bond between the concrete and the FRP.
T. C. Triantafillou
FLEXURAL STRENGTHENING
22
(b)
(a)
(c)
Fig. 4.2 Flexural strengthening of columns with maximum moment at the ends requires proper
anchorage of the external reinforcement. (a) Incorrect application, (b) continuity of the
FRP through the slab, (c) continuity of the FRP through the joint.
the service moment (no load safety factors are applied) acting on the critical RC section
during strengthening, the strain distribution of the member can be evaluated. As M o is
typically larger than the cracking moment M cr , the calculation is based on a cracked
section (Fig. 4.3). If M o is smaller than M cr , its influence on the calculation of the
strengthened member may easily be neglected.
Based on the transformed cracked section, the neutral axis depth x o can be solved
from:
1 2
bx o + ( s 1)A s2 ( x o d 2 ) = s A s1(d x o )
2
(4.1)
T. C. Triantafillou
FLEXURAL STRENGTHENING
23
co =
Mo x o
E c Io 2
(4.2)
bx 3o
+ ( s 1)A s2 ( x o d 2 ) 2 + s A s1(d x o ) 2
3
(4.3)
Based on strain compatibility, the strain o at the extreme tension fiber can be derived as
follows:
o = co
h xo
xo
(4.4)
The strain o determined by eq. (4.4) is the initial strain at the level of FRP when
strengthening takes place.
Fig. 4.3 Strain distribution in rectangular cross section subjected to moment M o at the time of
strengthening.
T. C. Triantafillou
FLEXURAL STRENGTHENING
24
concrete crushing
steel yields
steel yields
2
FRP fracture
concrete crushing
no steel yielding
T. C. Triantafillou
FLEXURAL STRENGTHENING
25
Mechanisms (1), (2) and (3) in Fig. 4.4 are based on composite action between
concrete and FRP and can be analyzed using standard procedures, whereas the other
mechanisms involve some kind of debonding or peeling-off and will be analyzed
separately.
4.4.1 Full composite action
(1) Steel yielding, concrete crushing
fcd
c=cu=0.0035
d2
As2fsd2
s2
Gx
As2
d
h
As1
As1fyd
s1
tf
Af
Affd
bf
(a)
(b)
(c)
Fig. 4.5 Cross section analysis at the ultimate limit state. (a) Geometry, (b) strain distribution, (c)
internal force distribution.
(4.5)
where =0.8, fcd = design strength of concrete, f yd = design value of tension steel yield
stress, A f = cross section area of FRP, fsd2 = design stress in the top steel
T. C. Triantafillou
FLEXURAL STRENGTHENING
26
reinforcement and fd = design stress in the FRP. Based on strain compatibility, fsd2
and fd are calculated as follows:
x d2
fsd2 = E s c
(4.6)
hx
fd = E f c
o
x
(4.7)
In the above expression c = cu is the ultimate strain in the concrete (=0.0035) and o
is the initial strain given by eq. (4.4). Note that fsd2 should not be taken higher than f yd .
Design bending moment capacity:
MRd =
1
Rd
[A s1fyd (d G x ) + A f fd (h G x ) + A s2 fsd2 (G x d2 )]
(4.8)
where G =0.4, Rd = safety factor for the calculation of the resistance in an existing
member (in general Rd 1, but in the case of flexure Rd =1).
For the equations given above to be valid, the following assumptions should be
checked: (a) yielding of tensile steel reinforcement and (b) straining of the FRP is limited
to the limiting strain, f ,lim (corresponding to fracture or debonding):
s1 = c
f = c
d x f yd
x
Es
hx
o f ,lim
x
(4.9)
(4.10)
where c = cu .
(2) Steel yielding, FRP fracture
T. C. Triantafillou
FLEXURAL STRENGTHENING
1000
c if c 0.002
1000 c 0.5
12
2
1
if 0.002 c 0.0035
3000 c
27
(4.11)
8 1000 c
4(6 1000 ) if c 0.002
c
G =
1000 (3000 4 ) + 2
c
c
if 0.002 c 0.0035
2000 c (3000 c 2)
(4.12)
The resisting moment can be obtained by solving eqs. (4.5) (4.8) (with the above
modifications) for the three unknowns, x , c and MRd .
(3) Concrete crushing
Non-
activation of this mechanism is achieved by limiting the area of FRP below certain limits.
More details are provided in Section 4.5, which describes ductility requirements.
4.4.2 Loss of composite action
(4) Debonding at outermost crack
Using the analytical model described in Section 3.3 one can calculate the bond
length required to prevent debonding. Consider, for example, the beam in Fig. 4.6a, with
a moment diagram as shown in Fig. 4.6b (note the application of the shift rule, resulting
shift of the diagram by a l ). The force distribution in the tension steel ( Nsd ) and in the
FRP ( N fd ) is provided in Fig. 4.6c. As an approximation, the total tensile force (in steel
and FRP), Nsd + N fd , equals MSd / z , where z = 0.95 d = lever arm.
Based on Fig. 4.6c, the FRP anchorage length is calculated beyond the location
(section A) where the total tension force envelope MSd / z intersects the line
corresponding to the maximum force carried by the steel only, NRsd = A s1f yd . At this
location the FRP tension force is N fad and the corresponding anchorage length is l b .
The anchorable force (design value) N fad can be estimated based on internal force
equilibrium as follows:
MSd
A E
A E
= N fad 1 + s1 s s1 N fad 1 + s1 s
z
A fEf f
A fEf
(4.13)
T. C. Triantafillou
FLEXURAL STRENGTHENING
28
(d)
(a)
al
al
(b)
Myd
lb
MSd
NRsd = As1fyd
NSd , NRd
(f)
(c)
Nfad
lb
MSd
Nsd
Nfd
(e)
Myd
NRsd = As1fyd
Nfad,max
NRfd
NRfd
NSd , NRd
MSd/z
MSd/z
It is clear that the force N fad is limited by Nfad,max [eq. (3.3a), with safety factor f ,b ]
and that sufficient space should be provided for the anchorage length l b . If this is not
the case, section A must be re-positioned (in the direction where the bending moment
decreases, that is towards the support), Fig. 4.6d-f, so that N fad will be reduced to
Nfad,max or so that a lower l b will be required (as seen in Fig. 3.5, a small reduction in
N fad results in substantial reduction in l b ). If the anchorage length is still not adequate,
then the FRP width should be increased and the thickness decreased, or mechanical
anchorages should be provided.
(5) Debonding at intermediate flexural crack
The analytical model described in Section 3.3 applies here too, provided that a
proper correction is made to account for the fact that the true state of stress and strain at
the concrete-FRP interface near vertical cracks in a real beam is not identical to that in
the experimental setup of Fig. 3.2.
experimental evidence suggest that the maximum shear stress at the interface is much
lower than the one found in the test setup. Based on the literature, it is proposed here to
modify the model of Section 3.3.2 by increasing the debonding force by 150%. Hence,
T. C. Triantafillou
FLEXURAL STRENGTHENING
29
the FRP strain corresponding to debonding in the vicinity of flexural cracks (where the
shear force is practically zero) is calculated as follows:
if l b l b,max :
f ,b, fl = fl
0.5k c k b
f ,b
fctm
Ef t f
if l b < l b,max :
f ,b, fl = fl
0.5k c k b
f ,b
fctm l b
E f t f l b,max
(4.14a)
2 lb
l b,max
(4.14b)
where fl = 2.5 .
The calculations for the resisting moment are performed as in the above case (2),
with fd = E f f ,b,fl .
(6) Debonding at intermediate shear crack
The statements made above apply here too, except that the increase o fthe
debonding force is about 100% compared with the experimental setup. Hence, the FRP
strain corresponding to debonding in the vicinity of shear cracks is:
if l b l b,max :
f ,b, fl sh = fl sh
0.5k c k b
f ,b
fctm
Ef t f
if l b < l b,max :
f ,b, fl sh = fl sh
0.5k c k b
f ,b
fctm l b
E f t f l b,max
(4.1a)
2 lb
l b,max
(4.15b)
where fl sh = 2 .
The calculations for the resisting moment are performed as in the above case (2),
with fd = E f f ,b,fl sh .
(7) FRP end shear failure peeling-off
Investigations by several researchers (e.g. Oehlers 1992, Ziraba et al. 1994, Jansze
1997, Raoof and Hassanen 2000), have indicated that when externally bonded plates
stop at a certain distance from the supports (as is typically the case in strengthening
applications) a nearly vertical crack might initiate at the plate end (plate end crack) and
then grow as an inclined shear crack (Fig. 4.7). However, by virtue of internal stirrups,
the shear crack may be arrested and the bonded-on plate separated from the concrete at
the level of the longitudinal reinforcement in the form of spalling (Fig. 4-10 right). The
latter failure mode is also called concrete peeling-off, and is attributed to a critical
T. C. Triantafillou
FLEXURAL STRENGTHENING
30
combination of shear and vertical tensile stresses at the plate end. A simple, yet reliable
and conservative approach for the verification of FRP end shear failure involves the
following checks:
VSd, end 1.4 VRd,c
MSd,end
(4.16)
2
MRd
3
(4.17)
where VSd,end and MSd,end is the acting shear force and bending moment (design values)
at the FRP end, VRd,c is the member shear resistance neglecting the contribution of
stirrups and MRd is the moment resistance [minimum value calculated based on
mechanisms (1), (2), (5) and (6)]. It is noted that the verification of (4.17) is rather easy
to achieve, e.g. by adjusting the FRP end. However, is (4.16) is not satisfied, then the
member should be strengthened near the FRP ends in shear (see next chapter).
h
cu o
d xy d
h
h
(4.18)
T. C. Triantafillou
FLEXURAL STRENGTHENING
31
cu
xlim
su
yd
fu
fu,c fu
f,min
f
Fig. 4.8 Strain distribution at critical cross section.
A last point to be made here is that large ductility values are not always achievable,
especially when the FRP quantity is controlled by serviceability, in which case the
member is under-designed in terms of strength.
A summary of the verifications for the ultimate limit state is provided next:
1.
Determine the resisting moment for the member before strengthening ( Mo,Rd ).
2.
From the service moment Mo prior to strengthening determine the initial strain o at
the extreme tension fiber.
Calculate the required FRP area A f (corresponding to MRd ) for cases (1), (2) and
(6) [or (5), in the absence of shear force] at the critical section, based on eqs. (4.5)(4.12). Note that these equations with c cu and f f ,lim = min( fu , f ,b, fl sh )
describe three failure modes simultaneously (steel yielding concrete crushing, steel
yielding FRP fracture, steel yielding debonding at intermediate crack). As an
approximation, f ,lim may taken equal to 0.004-0.005.
verification.
4.
Calculate the anchorage length and finalize the FRP configuration based on the
anchorage verification [mechanism (4)].
5.
T. C. Triantafillou
FLEXURAL STRENGTHENING
6.
32
Verify the shear resistance of the member (given that the flexural resistance has
been increased). If not satisfied, shear strengthening is required.
4.7 Example
216
150
500
fu
= 0.01.
40
50
Fig. 4.9
320
250
strengthening is Mo = 47 kNm. Rd = 1.
Geometric data: A s1 = 940 mm2, A s2 = 400 mm2, h = 500 mm, d = 450 mm, d1 =
50 mm, d2 = 40 mm and b = 1200 mm. The ratio s = E s / E c equals 200/29 = 6.9.
Solving eq. (4.1)-(4.4) we find o = 0.00066.
Assuming k c = 1 and k b 1, eq. (4.14a) for debonding near the mid-span (where
the moment is maximum and the shear equals zero) gives:
f ,b, fl = 2.5
0 .5 1 . 0
2 .2
= 0.003 , hence f ,lim = min(0.01, 0.003) = 0.003,
1.5
150000 1.1
which is the FRP strain at the critical section (mid-span) for the ultimate limit state
(debonding).
Next, with MRd = 203 kNm (and Rd = 1) from eq. (4.5) (4.12) we calculate x =
104 mm, c = 0.00071 and A f = 245 mm2. Each strip has a cross section area equal to
88 mm2, hence the use of 3 strips is required, with a total cross section area of 264 mm2,
which corresponds to MRd = 206 kNm, x = 105 mm and c = 0.00071. These strips will
be placed one next to the other, in order to avoid multiple layers.
The next step is the verification of the end anchorage (Fig. 4.10), which results in a
total length of strips equal to 4.10 m.
Finally, the FRP end shear calculations give:
5
5 0.45
T. C. Triantafillou
FLEXURAL STRENGTHENING
33
It is concluded that (4.17) is satisfied but (4.16) is violated, hence the ends should be
strengthened in shear for a shear force equal to 133.25 56.62 = 76.63 kN (according to
the procedure described in Chapter 5).
(a)
(d)
a l 0.40 m
a l 0.40 m
(b)
203 kNm
(e)
203 kNm
MSd
l b = 0.20 m
MSd
1.35+0.50 m
1.35 m
(c)
NRsd = As1fyd
NRsd = As1fyd
(f)
= 408.7 k
= 408.7 k
MSd/z
Nfad,max=48.20 kN
Nfad=71.11 kN
203/0.95d=475 kN
NSd , NRd
475 kN
MSd/z
NSd , NRd
940 200
408.7 N fad 1 +
N fad = 71.11 kN
264 150
1
.
5
1
.
1
l b,max =
150000 1.1
= 193.6 mm 0.2 m
2 2 .2
Fig. 4.10 Verification of anchorage.
T. C. Triantafillou
FLEXURAL STRENGTHENING
34
From the equilibrium of forces and strain compatibility, the depth of the neutral axis
x e is obtained from the following:
1 2
bx e + ( s 1)A s2 ( x e d 2 ) = s A s1(d x e ) + f A f h 1 + o x e
2
c
Ec c =
x
1
bx e h e + ( s 1)A s2
2
3
(4.19)
Mk
(x e d2 ) (h d ) A d x e (h d)
2
s s1
xe
xe
(4.20)
where f = E f / E c and Mk is the characteristic value of the acting moment. The last two
equations can be solved for the unknown x e and c .
The moment of inertia of the cracked section is given by:
I2 =
bx 3e
+ ( s 1)A s2 ( x e d2 ) 2 + s A s1(d x e ) 2 + f A f (h x e ) 2
3
(4.21)
whereas that of the uncracked section may be approximated as follows (for rectangular
cross sections):
I1
bh 3
12
(4.22)
T. C. Triantafillou
FLEXURAL STRENGTHENING
35
Regarding stress verification, apart from limiting stresses in the concrete and steel, it
is required to limit the stress in the FRP, f , under the rare load combination, as follows:
h xe
f = E f c
o f fk
xe
(4.22)
where the reduction coefficient < 1 accounts for the poor behavior of some composites
(e.g. GFRP) under sustained loading. Based on creep rupture tests (e.g. Yamaguchi et
al. 1998), indicative values of are 0.8, 0.5 and 0.3 for CFRP, AFRP and GFRP,
respectively. Note that as the design is often governed by the serviceability limit state,
relative low FRP strains at service load may be expected, so that FRP creep rupture is
typically not of concern.
The verification of deflections and crack widths is performed in analogy to the case of
reinforced concrete members (e.g. fib 2001).
4.9 Columns
The analysis of cross sections where bending develops in combination with an axial
force is performed according to the principles presented above, the basic difference
being the addition of one more term in the force and moment equilibrium equations: NSd
in the right part of eq. (4.5) and NSd [(h / 2) G x] in the right part of eq. (4.8), where NSd
is the acting axial force (design value). Furthermore, the contribution of FRP in carrying
compression should be neglected. Assuming that debonding is prevented (e.g. through
proper anchorage inside slabs or joints, Fig. 4.2b-c, the failure mechanism will be one of
the following:
of
tension
( s1 f yd / E s ),
steel
debonding
or
FRP
fracture
concrete crushing ( c = cu )
The bending moment axial force interaction at failure is best demonstrated through
the so-called interaction diagrams, such as those given in Fig. 4.12a-b. Those diagrams
have been constructed for various equivalent geometric ratios of steel and FRP
reinforcement, eq , defined as:
eq = s + f
where
A s,tot = 2A s1 = 2A s2
A s,tot A f ,tot E f
Ef
=
+
Es
bd
bd E s
(symmetrically
placed
(4.23)
steel
reinforcement)
and
T. C. Triantafillou
FLEXURAL STRENGTHENING
36
simplicity it has been assumed that f ,min =0.008. The interaction diagrams in Fig. 4.12
show that the effectiveness of FRP in increasing the flexural capacity decreases
substantially as the axial load increases.
0,8
d=NSd / bhfcd
eq=0,006
eq=0,007
eq=0,008
eq=0,009
C16/20
S400
b/h=1
d1/h=0,10
Ef=180 GPa
Es=200 GPa
1,0
eq=0,010
eq=0,011
eq=0,012
0,6
0,4
0,2
0,0
0,00
(a)
0,05
0,10
0,15
0,20
0,25
0,30
0,35
0,40
0,45
0,50
0,55
0,60
d=MSd / bh fcd
0,8
d=NSd / bhfcd
eq=0,012
C16/20
S400
b/h=1
d1/h=0,10
Ef=180 GPa
Es=200 GPa
1,0
eq=0,014
eq=0,016
eq=0,018
eq=0,020
eq=0,022
eq=0,024
0,6
0,4
0,2
0,0
0,00
(b)
0,05
0,10
0,15
0,20
0,25
0,30
0,35
0,40
0,45
0,50
0,55
0,60
d=MSd / bh fcd
Fig. 4.12 Axial force bending moment interaction diagrams for square cross sections (b=h)
under uniaxial bending. Concrete C16/20, steel S400, d1/h=0.10, Ef =180 GPa. (a)
As,tot=0.006, (b) As,tot=0.012.
As a general conclusion one may state that flexural strengthening of columns is not
always feasible (and easy as in the case of beams); and certainly the FRP contribution is
of rather low effectiveness, unless the axial load is kept at low levels (e.g. d < 0.2).
T. C. Triantafillou
FLEXURAL STRENGTHENING
37
T. C. Triantafillou
FLEXURAL STRENGTHENING
38
T. C. Triantafillou
SHEAR STRENGTHENING
39
CHAPTER 5
SHEAR STRENGTHENING
5.1 General
Shear strengthening of RC members using FRP may be provided by bonding the
external reinforcement (typically in the form of sheets) with the principal fiber direction as
parallel as practically possible to that of maximum principal tensile stresses, so that the
effectiveness of FRP is maximized (see Fig. 5.1 for the dependence of the FRP elastic
modulus on the fiber orientation). For the most common case of structural members
subjected to lateral loads, that is loads perpendicular to the member axis (e.g. beams
under gravity loads or columns under seismic forces), the maximum principal stress
trajectories in the shear-critical zones form an angle with the member axis which may be
taken roughly equal to 45o. However, it is normally more practical to attach the external
FRP reinforcement with the principal fiber direction perpendicular to the member axis
(Fig. 5.2). Photographs of typical applications are shown in Fig. 5.3.
30
60
90
T. C. Triantafillou
SHEAR STRENGTHENING
40
(b)
(a)
tf
C
(c)
(d)
(e)
(f)
sf
bf
E
(g)
(h)
bf
(k)
(i)
Fig. 5.2
(j)
sf
Shear strengthening of: (a)-(h) beams, (i)-(k) columns and shear walls.
T. C. Triantafillou
SHEAR STRENGTHENING
41
(a)
(b)
Fig. 5.3 Shear strengthening (a) of beam end with CFRP, (b) of column with GFRP.
df
0.1d
Inclined crack
VRd,f =
2t f b f
df fe,d (cot + cot ) sin
sf
(5.1)
where d f = height of FRP crossed by the shear crack, measured from the longitudinal
steel reinforcement (equals 0.9 d in the case of fully wrapped members, e.g. Fig. 5.2g, k),
= angle of diagonal crack with respect to the member axis (assumed equal to 45,
based on the classical Mrsch-Ritter truss analogy), = angle between principal fiber
T. C. Triantafillou
SHEAR STRENGTHENING
42
orientation and longitudinal axis of member, fe,d = design value of mean stress in the
FRP crossing the shear crack, in the principal fiber direction (effective stress).
Note that the differences between eq. (5.1) and that for the contribution of internal
stirrups to shear resistance ( VRd,s ) are: use of 2t f b f instead of A sw (cross section area
of stirrups), s f instead of sh (spacing of stirrups) and fe,d instead of f ywd (yield stress
of stirrups).
For the most common case of continuous sheets or fabrics (instead of equally
spaced strips) b f = s f sin and eq. (5.1) gives:
VRd,f = 2t f df fe,d (cot + cot ) sin2
(5.2)
Furthermore, for the typical case where the FRP is applied with the fibers perpendicular
to the member axis ( = 90), we obtain:
VRd, f = 2t f d f fe,d cot
(5.3)
The exact calculation of the effective stress fe,d is not a straightforward task. In
approximation, this stress varies linearly with the crack opening, which may be taken as
minimum at point A in Fig. 5.4 and maximum at point B. Hence the stress increases
linearly up to a maximum value, fd,max , which, in approximation, controls failure of the
FRP material. On the basis of the above assumptions, one may write:
0 .5 d f
fe,d = D f fd,max = 1
fd,max
0 .9 d
(5.4)
The value of fd,max at the ultimate limit state in shear depends on the definition of
failure, which can be one of the following:
FRP fracture
Fracture of the FRP is most likely the case in fully wrapped and properly anchored
jackets (e.g. Fig. 5.2g-k). In this case
fd,max = f fde
(5.5)
where f fde is the design value of FRP strength given by eq. (3.2) (note that this is lower
than the tensile strength of jacket in uniaxial tension).
consider that the strength reduction coefficient, e , for determining f fde is 0.80.
T. C. Triantafillou
SHEAR STRENGTHENING
43
FRP debonding
For open-type jacketing (e.g. U-shaped or side-bonded, Fig. 5.2c-e or Fig. 5.2b,
respectively), fracture of the jacket is not likely to occur (except for the case of Fig. 5.2e,
where anchorage conditions are slightly improved). In this case debonding of the FRP is
expected to be the dominant failure mode (Fig. 5.5), which can be described with the
analytical model presented in Section 3.3.2. This model can be adopted with k c = 1 ,
k b = 1 and an empirical coefficient sh (in analogy to fl and fl sh described in Section
4.4.2):
if l b l b,max :
fd,max = sh
0. 5
f ,b
fctm f
tf
if l b < l b,max :
fd,max = sh
0.5
f ,b
fctmE f l b
t f l b,max
(5.6a)
2 lb
l b,max
(5.6b)
where sh = 1.25 ,
lb =
df
sin
(5.7a)
lb =
df
2 sin
(5.7b)
l b,max =
Ef t f
c 2 fctm
(5.8)
T. C. Triantafillou
SHEAR STRENGTHENING
44
Note that in the case of U-shaped (three-sided) jackets the best-anchored part of the
FRP is that at the maximum crack opening, with a bonded length d f / sin (Fig. 5.6a),
where in the case of two-sided jackets the best-anchored part of the FRP is at the middle
of the shear crack (Fig. 5.6b); hence the factor 2 in eq. (5.7b).
(a)
(b)
Fig. 5.6 Bond length of (a) U-shaped FRP jacket, (b) two-sided FRP.
The improved anchorage shown in Fig. 5.2e, where the FRP end is rolled around a
rod and inserted into grooves is an interesting solution, which may be considered of
effectiveness in between that for open (Fig. 5.2c) and closed (Fig. 5.2g-k) jackets. This
case could be treated using the expressions for U-shaped jackets [eq. (5.6)-(5.7a], with
fd,max increased by approximately 30%.
Limiting strain
Some researchers have proposed that the effective strain in the FRP be limited to a
maximum value, in the order of 0.006, to maintain the integrity of concrete and secure
activation of the aggregate interlock mechanism. With this limitation, fd,max should not
be taken higher than 0.006 E f .
The contribution of FRP to shear resistance is provided through the term VRd,f in the
well-known equation for the design shear resistance:
VRd =
1
Rd
(5.9)
struts and Rd = safety factor (>1) for the determination of the shear resistance in
existing members ( Rd =1.20).. The FRP contribution VRd,f in eq. (5.9) is given by eq.
T. C. Triantafillou
SHEAR STRENGTHENING
45
(5.1) or (5.2) for FRP in the form of strips at equal spacing or continuous jackets,
respectively, with fe,d calculated from eq. (5.4), in which fd,max is determined as
follows:
cyclic (e.g. seismic) loading depends on the target ductility factor: high values of ductility
result in reduced shear resistance (e.g. Moehle et al. 2001), which affects (reduces) the
terms of eq. (5.9), but not the one regarding the FRP contribution ( VRd,f ). Hence the
reduced shear capacity due to cycling does not affect the equations presented above.
Finally it should be noted that if shear strengthening is provided by means of equally
spaced strips, the spacing should be such that the shear crack intersects at least two
strips, that is s f s f ,max = 0.5 min(df , 0.9d) (for = 45 and = 90).
Example 5.1
1200
216
150
500
40
320
250
Fig. 5.7
We assume the following properties for the CFRP: thickness of one layer = 0.12 mm,
elastic modulus E f = 230 GPa, effective design tensile strength f fde =2560 MPa. The
jacket will be applied according to the configuration shown in Fig. 5.2c (U-shaped).
Eq. (5.5):
Eq. (5.7):
The problem will be solved trying different numbers of layers, in order to illustrate their
relative effectiveness in carrying shear.
Jacket with one layer:
T. C. Triantafillou
SHEAR STRENGTHENING
46
230000 0.12
= 83 mm < l b
22
Eq. (5.8):
l b,max =
Eq. (5.6a):
fd,max = 1.25
0 .5
1.5
2 230000
= 815 MPa
0.12
0.5 310
fe,d = D f fd,max = 1
815 = 0.62 815 = 505 MPa
0
.9 460
Eq. (5.3):
The above value is the shear force carried by a one-layered jacket. This value is quite
low, hence we try three layers:
230000 (3 0.12 )
= 144 mm < l b
22
Eq. (5.8):
l b,max =
Eq. (5.6a):
fd,max = 1.25
0.5 2 230000
= 471 MPa
1 .5
3 0.12
0.5 310
fe,d = D f fd,max = 1
471 = 0.62 471 = 292 MPa
0.9 460
Eq. (5.3):
Eq. (5.8):
l b,max =
Eq. (5.6a):
fd,max = 1.25
0.5 2 230000
= 408 MPa
1 .5
4 0.12
0.5 310
fe,d = D f fd,max = 1
408 = 0.62 408 = 253 MPa
0
.9 460
Eq. (5.3):
T. C. Triantafillou
SHEAR STRENGTHENING
47
Example 5.2
1200
216
an
appropriate
shear
strengthening
150
500
40
320
250
Fig. 5.8
We assume that the strips have a width b f = 40 mm, thickness t f = 1.4 mm, elastic
modulus E f =120 GPa and effective design strength f fde =1360 MPa.
Eq. (5.5):
df = 0.9d
Eq. (5.4):
Calculation of spacing:
Eq. (5.1):
VRd, f =
2 1.4 40
0.9 460 360 10 3 > 75 kN s f < 223 mm
sf
s f ,max = 0.5 0.9 460 = 207 mm. Finally we propose the use of strips at a spacing of
200 mm.
Example 5.3
365 mm
We assume the following properties for the CFRP: thickness of one layer = 0.12 mm,
elastic modulus E f = 230 GPa, effective design tensile strength f fde =2560 MPa. The
jacket will be applied according to the configuration shown in Fig. 5.2i (full wrapping).
T. C. Triantafillou
SHEAR STRENGTHENING
Eq. (5.5):
48
Typical shear failures of (exterior) beam-column joints are shown in Fig. 5.10.
Studies on joints strengthened with FRP in shear demonstrated that even very thin FRP
jackets (e.g. 2-3 layers of carbon fiber sheets with layer thickness in the order of 0.12
mm) properly anchored outside the joints can provide an increase in shear capacity by
well above 80-100% (Antonopoulos 2001, Antonopoulos and Triantafillou 2002,
Antonopoulos and Triantafillou 2003). This is feasible provided that the sheets will be
made of fibers primarily in the beam direction, but if possible, also in the column (Fig.
5.11).
(a)
(b)
Fig. 5.10 Shear failure of exterior joints: (a) Hyogo-ken Nanbu earthquake, Japan, 1995. (b)
Kalamata earthquake, Greece, 1986 (fib 2003).
T. C. Triantafillou
SHEAR STRENGTHENING
49
(a)
(b)
Fig. 5.11 Typical configurations for shear strengthening of beam-column joints and anchorage
outside the joint. (a) Exterior joint, (b) Interior joint.
60
50
40
30
20
10
0
-10
-20
-30
-40
-50
-60
P,
C1
P (kN)
P (kN)
10 20 30 40 50
(mm)
60
50
40
30
20
10
0
-10
-20
-30
-40
-50
-60
P,
F22
(a)
10 20 30 40 50
(mm)
(b)
Fig. 5.12 Load-displacement loops for poorly detailed (lack of stirrups) beam-column joints. (a)
Non-strengthened specimen, (b) Strengthened specimen, which shows a 70%
increase in shear strength.
An approximate and simple method to account for the contribution of FRP to the
shear resistance of joints is to assume that the fibers in the beam direction are activated
up to a strain equal to 0.004.
T. C. Triantafillou
SHEAR STRENGTHENING
50
T. C. Triantafillou
CONFINEMENT
51
CHAPTER 6
CONFINEMENT
6.1 General
Confinement is generally applied to members in compression (Fig. 6.1), with the aim
of enhancing their load carrying capacity or, in cases of seismic upgrading, to increase
their ductility. FRP, as opposed to steel that applies a constant confining pressure after
yielding, has an elastic behavior up to failure and therefore exerts a continuously
increasing confining action. The confining stresses applied by the FRP result in one or
more of the following:
1.
2.
Increase of chord rotation after flexural yielding of columns (that is, increase of
ductility).
3.
4.
(a)
(b)
Fig. 6.1 Confinement of columns with FRP jackets: (a) CFRP, fibers in the horizontal direction,
(b) helically applied GFRP.
T. C. Triantafillou
CONFINEMENT
6.2
52
6.2.1 Behavior
Consider a concrete cylinder (Fig. 6.2a) with diameter D , fully wrapped with an FRP
jacket with thickness t f and elastic modulus E f (in the direction of the fibers, that is
circumferentially).
(a)
tf
(b)
Fig. 6.2 (a) Axially loaded column. (b) Lateral stresses due to confinement.
The lateral stresses l (in the radial direction, due to dilation of the concrete) exerted in
the jacket (equal but of opposite sign act on the concrete) are calculated as follows:
l =
2t f
2t
1
f = f Ef f = f Ef f
D
D
2
(6.1)
where f and f = FRP tensile stress and strain, respectively, and f = volumetric ratio
of FRP. The result of confining stresses l is control of lateral expansion and hence
increase of deformability, until the tensile stress f (corresponding strain f ) in the FRP
reaches its tensile strength f fde (corresponding strain fue ); at this point the jacket
fractures (Fig. 6.3) and the member fails. Of course the mechanism described above is
possible only provided that premature debonding of the FRP (at its ends) will not occur.
Note here that the circumferential tensile strength of the jacket is, in general, lower
than the tensile strength of FRP measured in a uniaxial tension test. This is attributed to
the multiaxial state of stress in the FRP, stress concentrations, the use of many layers,
the quality of application etc., and may be taken into account through the reduction factor
e , with values in the of 0.7-0.9:
f fde = e f fd
(6.2)
T. C. Triantafillou
CONFINEMENT
53
Fig. 6.3 Tensile fracture of FRP jacket in the circumferential direction when the tensile stress f
reaches the design FRP strength f fde .
c
fcc=ccu
Compressive
stress
tf increases
b
a
fc
(ccu, ccu)
(cc, fcc)
Unconfined concrete
co cu
Compressive strain
c
ccu
Fig. 6.4 Compressive stress-strain curves for concrete confined with FRP.
The stress-strain relationship for concrete confined with FRP is given schematically
in Fig. 6.4.
conclusions:
Jackets of very low thickness increase only the ultimate strain ccu (curve a in Fig.
6.4).
Jackets of low thickness result in confined concrete strength fcc which corresponds
to strain cc lower than that at ultimate ( ccu ) (curve b in Fig. 6.4).
For a given type of FRP, the strength fcc and ultimate strain ccu of confined
concrete increase with the thickness of the jacket.
T. C. Triantafillou
CONFINEMENT
54
For jackets of equal thickness but with different types of fibers (e.g. carbon versus
glass) the confined strength fcc increases with the jacket strength f fde (carbon is
better than glass in this case), whereas the ultimate strain ccu increases with the
jacket strength f fde but also, mainly, with its ultimate strain fue (glass is better than
carbon in this case).
For jackets of equal stiffness (expressed by the product E f t f ), the confined strength
fcc increases with the ultimate strain of FRP fue .
As far as the design of FRP jackets for confinement is concerned, typically we aim at
calculating the required thickness t f (for a given type of FRP) for a target confined
strength fccd (design value) and/or for a target ultimate strain ccu . The international
literature on FRP-concrete confinement models is vast.
described next (fib 2001). The model applies to columns with rectangular cross section
(dimensions b and d , b d ), rounded at the corners with a radius rc .
fccd = E sec,ud ccu fcd
E sec,ud =
E sec,Md =
(6.3)
E
(E E sec,ud )
1)] sec,Md c
E sec,ud (E c E sec,Md )
Esec,Md
Ec
(6.4)
Ec
(6.5)
E
1 ffde
1 + 2 c
fcd 0.002 E f
1d 2d fcd
0.002[1 + 5(1d 2d 1)]
1d = 2.254 1 + 7.94
lud,b
fcd
lud,b
fcd
(6.6)
1.254
(6.7)
d 2
d
2d = 1 0.6 1.4 + 0.8 lud,b
b
b
fcd
lud,b = f
In
the
above
expressions
Ec
(6.8)
2t f
f fde
d
initial
modulus
(6.9)
of
elasticity
for
concrete
[ E c = 1.05 9500 ( fck + 8)1/ 3 ] and f = confinement effectiveness coefficient for the
T. C. Triantafillou
CONFINEMENT
55
specific jacket used, depending on: (a) the cross section geometry (aspect ratio, radius at
corners, Fig. 6.5), (b) the degree of concrete coverage (Fig. 6.6b) and (c) the fiber
orientation with respect to the member axis (Fig. 6.6c). Specificallly:
f = n s a 1
Shape coefficient:
n =
Ae
= 1
Ag
(6.10)
(b 2rc )2 + (d 2rc )2
b 2 + d 2
1
3bd
A
3 A g 1 s
A g
(6.11)
1 f
2
sf
2d
s =
1
A
2d
1 s
Ag
Coverage coefficient:
a =
(6.12)
(6.13)
1 + (tan f )
where A g = area of cross section, A s = cross section area of longitudinal steel, sf = clear
space between strips, for the case of partial coverage (Fig. 6.6a), d = smallest dimension
of the cross section (or diameter, in the case of circular columns) and f = fiber
orientation with respect to member axis (Fig. 6.6b). For circular cross sections n =1, for
fully covered members s =1 and for fibers in the direction perpendicular to the member
axis a =1.
rc
d
b'=b-2rc
b
d'=d-2rc
Confined
concrete
Fig. 6.5 Confinement of rectangular cross sections is achieved by rounding the corners.
T. C. Triantafillou
CONFINEMENT
56
(a)
(b)
bf
d sf / 2
sf
Confined
concrete
Fig. 6.6 Confinement (a) with equally spaced strips, (b) with helically applied fibers.
Other confinement models found in the international literature are much simpler,
typically in the form:
fccd
= 1 + k 1 lud
fcd
fcd
ccu = cu
+ k 2 lud
fcd
(6.14)
n
(6.15)
In eqs. (6.14)-(6.15) lud is the mean confining stress (at failure of the jacket),
approximately equal to (Fig. 6.7):
lud,d
lud,b
Fig. 6.7 Mean confining stress in each direction of rectangular cross section.
lud =
=
lud,b + lud,d
2t
1 2t f
f fde + f f f fde
f
2
d
b
(b + d) t f
1
f ( f ,b + f , d )f fde = f
f fde
2
bd
(6.16)
T. C. Triantafillou
CONFINEMENT
57
where lud,b and lud,b are the mean confining stresses in the direction of sides b and
(E c E 2 )2 2
4fcd
cd = fcd + E 2 c
if 0 c t
(6.17a)
if t c ccu
(6.17b)
where
Compressive
stress
t =
2fcd
(E c E 2 )
(6.18)
E2 =
fccd fcd
ccu
(6.19)
cd
fccd
E2
Unconfined
concrete
fcd
Ec
co=0.002
Compressive strain
t
cu = 0035
FRP-confined
concrete
c
ccu
Finally, one may rely on the simpler, but not so accurate for the case of FRPconfined concrete, models described in Eurocodes 2 or 8.
T. C. Triantafillou
CONFINEMENT
58
Example 6.1
fcd = 20 Pa and elastic modulus E c = 33.5 GPa. The column is to be jacketed with
either CFRP or GFRP, aiming at increasing the compressive strength to fccd = 35 MPa
and the ultimate strain to ccu = 0.025: (a) For CFRP we assume E f = 230 GPa, ffd =
2590 MPa, thickness of one layer = 0.12 mm. (b) For GFRP we take E f = 70 GPa, ffd =
1400 MPa and thickness of one layer 0.17 mm. Finally we assume that the tensile
strength of the jacket is reduced by 5% with respect to tension testing specimens (that is
e =0.95).
For CFRP f fde = e f fd = 0.95 2590 = 2460 MPa and for GFRP f fde = 0.95 1400 = 1330
MPa. The results for the required fiber sheet thickness and the corresponding number of
layers are calculated in Table 6.1, based on the analytical model of eq. (6.3) (6.9), for
three different cross sections.
confinement is to increase strength then the required CFRP is much less than GFRP,
whereas the opposite is the case if the aim is to increase deformability.
Table 6.1 Required fiber sheet thickness for various types of cross sections.
Cross
rc
Ag
section
(cm)
(cm2)
(effectiveness)
( b , d m)
d=0.3
for ccu =
0.025
for fccd =
35 MPa
for ccu =
0.025
896.5
0.50
0.39 (4)
0.31 (3)
0.82 (7)
0.12 (1)
1246.5
0.32
0.74 (7)
0.56 (5)
1.56 (13)
0.22 (2)
886.2
0.64
0.31 (3)
0.24 (2)
0.64 (6)
0.10 (1)
b=0.3
0.25
0.5
0.3
0.3
T. C. Triantafillou
CONFINEMENT
59
at failure u (Fig. 6.9a), which is more or less equivalent to increasing the ductility. The
ductility may be quantified through the member chord rotation ductility factor,
= u / y , or through the curvature ductility factor, = u / y , where: y = chord
rotation at yielding, u = curvature at failure and y = curvature at yielding. Note that,
essentially, the chord rotation ductility factor member (relative end) displacement
ductility factor, = u / y , where u and y the relative displacement of member
ends at ultimate and yielding, respectively (Fig. 6.9). In the above definitions failure is
considered when either there is an abrupt fall in the members response (e.g. load
displacement curve) or the response parameter (e.g. force) has been reduced by 20%
with respect to its peak (Fig. 6.9b).
y
Pu
0.2Pu
Ls
u
Lpl
(a)
(b)
(c)
Fig. 6.9 (a) Lateral loading of RC member. (b) Load-displacement diagram. (c) Curvature.
u = y + u y L pl 1 0.5
L
s
(6.20)
where L s = shear span (distance from base of column to the point where the bending
moment is zero, equal to the ratio of moment to shear at the column end) and L pl =
plastic hinge length. The chord rotation at yielding, y , is not affected by FRP jacketing
and equals:
For beams or columns:
y = y
fy
Ls + aV z
h
db
+ 0.00131 + 1.5 + 0.13 y
3
Ls
fc
(6.21)
T. C. Triantafillou
CONFINEMENT
60
y = y
fy
Ls + aV z
L
(6.22)
where db = mean diameter of tension steel rebars, h = height of cross section, f y = yield
stress of longitudinal steel (MPa) and fc = concrete strength (Pa). The above material
data are taken as mean values of in-situ assessed properties, divided by a data reliability
factor (1.0, 1.2, 1.35), as per Eurocode 8. The term a V z is the tension shift of the
bending moment diagram a l for shear cracking at 45 and expresses the effect of
tension forces shifted by a l to the members flexural deformations. The coefficient a V ,
which multiplies the internal force lever arm z at the end cross section, equals 0 if the
shear force at flexural yielding, VMy = M y / L s , is less than the shear cracking force Vcr ,
or 1 otherwise. Note that the shear cracking force may be taken as the shear resistance
of the member without shear reinforcement, VR,c , as calculated by Eurocode 2 with a
safety factor c = 1 .
The plastic hinge length L pl may be estimated from the following expression:
L pl = 0.1L s + 0.17h +
0.24f y
fc
db
(6.23)
where f y and fc are in Pa. The curvatures y u are calculated based on section
analysis at yielding and failure. u is calculated as u = ccu / x u , where x u = depth of
compression zone at failure and ccu = ultimate strain of concrete, as provided by the
confinement model, e.g. eq. (6.15) (it is this term that is mainly affected by the properties
of the FRP jacket!).
The chord rotation u (or the curvature at failure u ) can increase by jacketing the
RC member at its critical regions (member ends), Fig. 6.10, where strains in concrete and
steel are expected to be high. In these regions the confinement exerted by the FRP
increases the ultimate strain of concrete (in addition to delaying rebar buckling and bond
failure at lap-splices) and hence the ductility (Fig. 6.11).
T. C. Triantafillou
CONFINEMENT
61
C4_XB Load v Deflection
150
150
100
100
50
50
0
-120
-80
-40
200
40
80
120
Load kN
Load kN
250
200
0
-120
-80
-40
-50
-50
-100
-100
-150
-150
-200
-200
-250
40
80
120
-250
Deflection mm
Deflection mm
(a)
(b)
Fig. 6.11 Load-displacement loops for RC column of 0.25x0.50 m cross section under cyclic
loading. (a) Unretrofitted member. (b) Member retrofitted with two layers of carbon
sheet (thickness of each layer = 0.12 mm) at 0.60 m of the column base.
In summary, the design of FRP jackets for a given chord rotation at failure u (which
is introduced in the compliance criteria for the performance levels specified in Eurocode
8) requires the expression of u in terms of the jacket properties. This is achieved
through the following steps:
Determine the plastic hinge length L pl from eq. (6.23).
Calculate the yield curvature y , based on cross section analysis.
Calculate the chord rotation at yielding from eq. (6.21) or (6.22).
Solve eq. (6.20) for the required jacket characteristics.
An alternative approach for relating the FRP jacket characteristics to the ultimate
chord rotation (mean value) at flexural failure of beams or columns designed according to
old provisions for seismic design is based on the use of the following empirical
relationship (urocode 8):
um
max (0.01, )
fc
= 0.016 0.3
max (0.01, )
( )
0.225
Ls
h
0.35
f
sx yw + f fx fde
f
fc
c
25
(1.25
100 d
(6.24)
where:
= mechanical reinforcement ratio of tension longitudinal reinforcement (including any
T. C. Triantafillou
CONFINEMENT
62
s
= 1 h
2b o
s
1 h
2ho
b2
1 i
6b oho
(6.25)
In eq. (6.25) b o and ho are the dimensions of confined concrete core to the centerline of
the stirrups and bi is the centerline spacing of longitudinal rebars supported by stirrups.
It is strongly recommended that if the stirrup ends are not bent towards the concrete core
( 135 at corners, 90 on the sides), the confinement provided by stirrups should be
neglected ( = 0).
The corresponding to eq. (6.24) formula for the mean value of the plastic part of the
ultimate chord rotation ( pl
u = u y ) is:
pl
um
max (0.01, )
= 0.0145 0.25
max (0.01, )
0 .3
0.2 L s
(fc )
0.35
f
sx yw + f fx fde
fc
fc
25
(1.275
100 d
) (6.26)
For shear walls designed according to old seismic design code provisions the right part of
eq. (6.24) (6.26) should be multiplied by 0.625 and 0.6, respectively (0.016 and
0.0145 are replaced by 0.01 and 0.0087).
A careful examination of eq. (6.24) and (6.26) reveals that the contribution of FRP
lies only in the exponent of 25.
Another alternative approach to deal with the design of FRP jackets for a target
ductility is to use the following simple but highly conservative equation proposed by
Tastani and Pantazopoulou (2002):
(6.27)
lud in eq. (6.27) is the confining stress at the ultimate limit state, given e.g. by eq. (6.9),
which neglects the contribution of stirrups. Note that the use of eq. (6.9) in rectangular
columns applies with d taken as the cross section dimension perpendicular to the plane
of bending. The application of this approach is illustrated in the next example.
T. C. Triantafillou
CONFINEMENT
63
Example 6.2
Longitudinal reinforcement:
18, f yd = 350 Pa
3m
(a)
(b)
Fig. 6.12 (a) Loading of column and (b) retrofitting for ductility.
Consider a column with cross section 0.30x0.40 m, subjected to strong axis bending
(Fig. 6.12).
strength is 11 Pa; and the carbon fiber sheets to be used have an elastic modulus 230
GPa, tensile strength 3000 MPa and thickness 0.12 mm (one layer). We assume that the
FRP strength reduction coefficient is e = 0.90. The objective is to design the jacket
(that is to calculate the required number of layers) for a target displacement (or chord
rotation) ductility factor ( = ) = 4.
Tensile strength of the jacket: 0.90 3000 = 2700 MPa.
Confinement effectiveness coefficient, eq. (6.11): A g = 1195 cm2, A s = 15.25 cm2.
n = 1
35 2 + 25 2
= 0.48
15.25
3 1195 1
1195
0.48 f 2700
300
4 = 1.3 + 12.4
0.1
11
hence
t f = 0.40 mm
T. C. Triantafillou
CONFINEMENT
64
that is 0.40/0.12 = 3.3 4 layers (if repeated with t f = 4 0.12 mm, the calculations give
= 4.75 ).
6.4 Lap-splices
6.4.1 Behavior and design
FRP jackets in regions with straight lap-spliced rebars provide confinement which
increases the friction between lap-splices and prevents slippage (typically this is not of
concern in lap-splices with 180 hooks, in which case slippage is not activated). The
improved behavior in FRP-confined lap-spliced regions has been demonstrated in many
studies, including those of Ma and Xiao (1997), Saadatmanesh et al. (1997), Seible et al.
(1997), Restrepo et al. (1998), Osada et al. (1999), Haroun et al. (2001) etc. Typical
results are shown in Fig. 6.13.
(a)
(b)
Fig. 6.13 Cyclic loading response of column with rectangular cross section: (a) unretrofitted
member, (b) member retrofitted at lap-splices (Saadatmanesh et al. 1997).
F = A b fs = p c b l s
Lateral stress l
Diagonal struts
ls
Bond stress b = l
T. C. Triantafillou
CONFINEMENT
65
According to the friction model of Fig. 6.14 and the possible failure patterns of Fig.
6.15 (details are omitted), it can be shown that lap-splice failures may be prevented using
fiber sheets with a thickness t f as follows:
t f = Rd
l s
bd1
A b f yd
l s,min
(b + d)p c ffde l s
(6.28)
where A b = cross section area and diameter of one spliced rebar, l s = available lapsplice length, l s,min = lap-splice length required to prevent slippage, p c = perimeter of
crack at lap-splice failure (Fig. 6.15b,c), f yd = yield stress of longitudinal rebars, b and d
= dimensions of rectangular cross section, = friction coefficient, f fde = effective FRP
jacket strength in circumferential direction and Rd = safety factor.
An additional
condition to met in order to prevent lap-splice failure according to Seible et al. (1997) is
that the radial concrete strain should be kept below a critical value, in the order of 0.0010.002. Hence, f fde in eq. (6.28) should exceed the value
f fde 0.0015 E f
(6.29)
D
n lap-splices
s
db
ls
(a)
db
pc=(D/2n)+2(db+c)
(b)
c
pc=(s/2)+2(db+c)
2 2 (db+c)
(c)
Fig. 6.15 (a) Column confinement at lap-splice region. (b) Cracking of circular section in the
tension zone due to bond failure and definition of critical crack path. () Similarly for
rectangular columns.
T. C. Triantafillou
CONFINEMENT
66
Closing this section we should point out that the effect of FRP confinement at lapspliced rebars is favorable only for the corner rebars (in rectangular cross sections),
where confining stresses are substantial due to rounding of the corners.
Example 6.3
Consider the column of Fig. 6.12a (0.30x0.40 m cross section) with 16 rebars and
f yd = 230 Pa, under lateral loading which causes bending with respect to either the
strong or the weak axis. We assume that the radius at column edges is rc = 25 mm and
that the concrete cover is c = 30 mm. The concrete strength is 11 Pa, the friction
coefficient is taken = 1.4, the lap-splice length is l s = 0.25 m and l s,min = 0.40 m.
Assuming that confinement at the lap-splice region is provided with carbon fiber sheets
with elastic modulus E f = 230 GPa, tensile strength 2600 Pa and thickness of one
layer 0.12 mm, determine the required number of layers to prevent lap-splice failure.
Take Rd = 1.5.
(a) Strong axis bending
d
220
Fig. 6.16a
150
Fig. 6.16b
T. C. Triantafillou
CONFINEMENT
67
The calculations are as above, but note that FRP jacketing will prevent lap-splice failure
only at the corner rebars.
6.4.2 Effect of lap-splices on chord rotation
The effect of lap-splices on chord rotation is taken into account by computing the
yield chord rotation y and the plastic part of the ultimate chord rotation pl
u with twice
as high compared to that outside the lap-splice region. The same applies for y and M y .
Moreover, if l s < l s,min , then pl
u , u , M y and y should be computed by multiplying the
yield stress of longitudinal rebars by l s / l s,min . Moreover, the 2nd term in eq. (6.21)
(6.22) should be multiplied by the ratio of the reduced yield moment to that outside the
lap-splice region. Finally, the right part of eq. (6.26) should be multiplied by l s / l su,min .
For lap-splices without FRP jacketing:
l s,min =
l su,min =
0.3 f y
fc
db
(6.30)
fy
f
1.05 + 14.5 l sx yw
fc
fc
db
(6.31)
where
s
l = 1 h
2b o
s
1 h
2ho
nrestr
(6.32)
n = total number of longitudinal rebars in the column perimeter and nrestr = number of
rebars supported at corners of stirrups or by cross ties.
For lap-splices with FRP jacketing at a height at least equal to 2 l s /3:
l s,min =
l su,min =
0.2 f y
fc
db
(6.33)
fy
f
1.05 + 14.5 l, f fx fde
fc
fc
db
(6.34)
where l, f = 4/ n (because confinement is effective only in the vicinity of the four corner
rebars). Note here that in order to avoid accounting for the FRP contribution twice in the
T. C. Triantafillou
CONFINEMENT
68
correction for pl
u , f in the power of 25 in eq. (6.26) should be taken as zero. Finally, all
strength parameters in the above equations are given in MPa.
0.45nfs2 d
4E dsE f f
(6.35)
where n = total number of longitudinal rebars in the cross section, f s = stress in the
rebars at a strain equal to 0.04 and E ds = double modulus of rebars, defined as follows
(Fig. 6.17):
E ds =
(E
4E s E i
s
+ Ei
(6.36)
Stress
s
fu
fs
0.04
Strain
T. C. Triantafillou
CONFINEMENT
69
approximately (and conservatively) equal to 40 MPa. Hence, with the introduction of the
safety factor we have:
t f = Rd
10nd
Ef f
( E f in MPa)
(6.37)
Example 6.4
rebars
18, S500s
18.
Fig. 6.18
For carbon fiber sheets with E f = 230 GPa and thickness of one layer equal to 0.12 mm,
the required sheet thickness to delay rebar buckling is:
tf =
1.5 10 10 400
= 0.54 mm
230000 0.48
It must be made clear that FRP jacketing in RC columns: (a) increases the axial load
capacity (strength), if the predominant loading is axial and (b) increases substantially the
deformability (ductility, chord rotation) and/or the shear resistance, if the predominant
loading is lateral (seismic forces). Contrary to the case of steel jacketing, the stiffness is
not affected by FRP jacketing, implying that very flexible structures (e.g. buildings with
pilotis) may remain vulnerable and may require stiffening in addition to strengthening, as
per the structural analysis results.
Under the condition that the intervention does not aim to increase the stiffness (or
the flexural resistance!), any given seismic excitation will provide (through the structural
analysis) (a) the target chord rotation (or ductility) and (b) the design shear (accounting
for capacity design, that is flexural yielding before shear cracking).
The required
T. C. Triantafillou
CONFINEMENT
70
T. C. Triantafillou
71
CHAPTER 7
7.1 General
This chapter summarizes basic detailing and practical execution rules for the
application of composites as externally bonded reinforcement.
7.2 Detailing
Detailing rules are summarized here for the three basic cases: (a) flexural
strengthening, (b) shear strengthening and (c) confinement.
7.2.1 Flexural strengthening
According to the fib bulletin 14 (2001), the following rules should be respected (for
beam strengthening):
Fig. 7.1
the
concrete
cover
of
the
longitudinal reinforcement.
internal
cover
Lap joints of strips should be avoided; they are absolutely not necessary, because
FRP can be delivered in the required length. Nevertheless, if needed, lap joints
should be made in the direction of the fibers with an overlap that will ensure tensile
fracture of the FRP prior to debonding at the lap joint.
Crossing of strips is allowed (e.g. strengthening of two way slabs) with bonding in the
crossing area.
If strips or sheets are to be applied in several layers, the maximum number or layers
should not exceed 3 or 5 for prefabricated strips or in-situ cured sheets, respectively.
In the case of applying FRP strips over supports of continuous beams or slabs, the
strips should be anchored at a distance in the order of 1 m in the compression zone
(Fig. 7.2).
T. C. Triantafillou
72
FRP
1m
Shift rule
Anchoring of FRP (especially if the strips are staggered) can be ensured by applying
bonded FRP stirrups that enclose the longitudinal strips at their ends (Fig. 7.3, 7.4).
The use of such stirrups is strongly recommended. Note that these stirrups are not
considered to be part of the shear reinforcement but are responsible to keep the
longitudinal strips in their position and to prevent peeling-off.
Section A - A
T. C. Triantafillou
73
Fig. 7.4 Improved anchorage at FRP strip ends using transverse FRP.
Fig. 7.5 Typical configurations for the anchorage of FRP ties in the compression zone.
non-metallic
(e.g.
FRP);
if
shear
150 mm
to the potential
of galvanic
Fig. 7.6
corrosion).
Minimum permissible radii at corners of rectangular cross sections are in the order of
20 mm for carbon or glass fibers and 10 mm for aramid fibers.
T. C. Triantafillou
74
Shear strengthening of columns between partial height infill walls should be done
along the full column height, not just in the free part (Fig. 7.7).
WRONG
CORRECT
Fig. 7.7 Shear strengthening of column between partial height infill walls.
Fig. 7.8
7.2.3 Confinement
Rounding of the corners in columns should be done at the maximum possible radius
(typically determined by the concrete cover).
Typical
Fig. 7.9
gap
Fig. 7.10
Concerning the application of FRP on rectangular columns or pier walls with large
aspect ratio, the FRP does not actually confine the internal concrete structure if just
applied to the surface.
constrained on both sides along the length through the use of dowels or bolts or
spike anchors (Fig. 7.11) that anchor the jacket to the existing structure, thereby
T. C. Triantafillou
75
creating shorter distances. Spike anchors provide a low cost solution, which has
been tested with very good results for the attachment of FRP jackets at the reentrant
corners of L-shaped cross section columns, Fig. 7.12 (Karantzikis et al. 2005).
First layer
Spike anchor
Concrete
Final layer
(a)
(b)
(c)
Fig. 7.12 Fixing the jacket at reentrant corner: (a) typical configuration, (b) spike anchors, (c)
photograph of anchors at reentrant corner.
As in the case of columns strengthened in shear, full wrapping with several pieces of
FRP along the height should be done with the lap joints in different sides (Fig. 7.8).
When jackets are provided to prevent lap-splice failures (e.g. at the bottom of
columns), the FRP should extend at a height equal to at least 2/3 of the lap splice.
T. C. Triantafillou
76
aramid fibers. Bonding on concrete surfaces is achieved with two-part epoxy adhesives.
Details about specific systems as far as material properties and practical execution are
concerned are given by the supplier of the strengthening system. In this section we
provide general rules, applicable to most of the commercially available systems.
The concrete should be sound and free from serious imperfections (e.g. cavities,
wide cracks, protrusions), roughened (e.g. by means of sand blasting or water jet
blasting) and made laitance and contamination free. Surface moisture in excess of
4% requires the use of special resins. Typical surface preparation steps are given in
Fig. 7.13.
(a)
(b)
(c)
Fig. 7.13 Surface preparation: (a) Grinding, (b) cleaning (c) leveling.
Use
of
prefabricated
strips
requires
Selection of the appropriate resin should be made on the basis of in-situ temperature
and humidity requirements.
FRP strips should be cut to proper size using an electric or manual saw. Depending
on the type of strips, cleaning (e.g. with acetone) or removal of a surface veil may be
required prior to bonding. Handling of strips by workers should be performed with
care (the use of gloves is strongly recommended).
T. C. Triantafillou
77
Prepared substrate
Resin
Air
Strip with resin in trapezoidal
configuration
Plastic roller
(b)
(a)
Fig. 7.15 (a) Application of resin on concrete and FRP, (b) application of pressure during
rolling.
Use of roller.
Placement of strip.
Sheets should be applied with special care to ensure that wrinkles are avoided and
that the fibers are as straight as practically possible. Impregnation of sheets with
resin is achieved using a plastic roller (Fig. 7.17)
T. C. Triantafillou
(a)
78
(b)
(d)
(c)
(e)
Fig. 7.17 In-situ impregnation of sheet: (a) Prime, (b) placement of first layer of sheet, (c)
impregnation of sheet on concrete. (d) Pre-impregnation of sheet and (e) application
of pre-impregnated sheet.
The average thickness of resin layer between strips and the concrete substrate
should be in the order of 1.5 mm. The resin used to impregnate sheets must have
an appropriate viscosity and used at the proper quantity, to ensure full impregnation
without entrapped air.
Last, but certainly not least, the FRP strengthening system should be applied by
properly trained and qualified personnel.
T. C. Triantafillou
DURABILITY
79
CHAPTER 8
DURABILITY
8.1 General
This chapter provides a brief overview of the durability of FRP-based strengthening
systems with regard to a number of factors, namely:
Temperature effects
Moisture
Galvanic corrosion
Fatigue
Impact
8.3 Moisture
FRP materials are, in general, highly resistant to moisture. Occasionally, extremely
prolonged exposure to water (either fresh or salt) may cause problems with some
fiber/resin combinations. The resin matrix absorbs water, which causes a slight reduction
T. C. Triantafillou
DURABILITY
80
in strength and the glass transition temperature. However, most structural adhesives
(high quality epoxy resins) are extremely resistant to moisture (Blaschko et al. 1998). As
far as the fibers are concerned, the high susceptibility of aramid to moisture deserves
special attention; carbon fibers are practically unaffected, whereas glass fibers have an
intermediate behavior.
At this point it is worth pointing out that full jacketing of RC with FRP provides a
moisture/vapor/air barrier which increases the longevity of members by protecting them
from harsh conditions (e.g. chlorides, chemicals). On the hand, in case of poor concrete
conditions, the encapsulation is at risk if the member is exposed to extreme climate
cycling and/or excessive moisture. Applications of FRP to a structural member that is at
risk of water pooling should not involve fully encapsulating the concrete. Good internal
and surface concrete conditions, proper surface preparation, adequate concrete
substrate exposure and proper application of an adequate FRP system may substantially
reduce this risk.
T. C. Triantafillou
DURABILITY
81
The contact of carbon fibers with steel may lead to galvanic corrosion, a problem
which is not of concern in the case of glass or aramid fibers.
8.8 Fatigue
In general, the fatigue behavior of unidirectional fiber composites is excellent,
especially when carbon fibers are used, in which case the fatigue strength of FRP is even
higher than that of the steel rebars (e.g. Kaiser 1989, Deuring 1993, Barnes and Mays
1999).
T. C. Triantafillou
DURABILITY
82
8.9 Impact
The strength of composites under impact loading is highest when aramid fibers are
used (hence the use of these materials in bridge columns that may suffer impact loading
due to vehicle collision) and lowest in the case of carbon fibers. Glass gives intermediate
results.
T. C. Triantafillou
REFERENCES
83
REFERENCES
International
du
Beton
fib
(2001),
Externally
onded
FRP
T. C. Triantafillou
REFERENCES
84
T. C. Triantafillou
REFERENCES
85
Osada, K., Yamaguchi, T. and Ikeda, S. (1999), Seismic performance and the
retrofit of hollow circular reinforced concrete piers having reinforcement cut-off planes
and variable wall thickness, Transactions of the Japan Concrete Institute, 21, 263-274.
Plevris, N. and Triantafillou, T. C. (1994), Time-dependent behaviour of RC
members strengthened with FRP laminates, Journal of Structural Engineering, ASCE,
120(3), 1016-1042.
Priestley, M. J. N., Seible, F. and Calvi, G. M. (1996), Seismic Design and Retrofit of
Bridges, John Wiley & Sons, New York, USA.
Raoof, M. and Hassanen, M. A. H. (2000), Peeling failure of reinforced concrete
beams with fibre-reinforced plastic or steel plates glued to their soffits, Proceedings of
the Institution of Civil Engineers: Structures and Buildings, 140, 291-305.
Restrepo, J. I., Wang, Y. C., Irwin, R. W. and DeVino, B. (1988) Fibreglass/epoxy
composites for the seismic upgrading of reinforced concrete beams with shear and bar
curtailment deficiencies, Proceedings 8th European Conference on Composite Materials,
Naples, Italy, 59-66.
Saadatmanesh, H., Ehsani, M. R. and Jin, L. (1997) Repair of earthquake-damaged
RC columns with FRP wraps, ACI Structural Journal, 94(2), 206-215.
Seible, F., Priestley, M. J. N., Hegemier, G. A. and Innamorato, D. (1997) Seismic
retrofit of RC columns with continuous carbon fiber jackets, Journal of Composites for
Construction, ASCE, 1(2), 52-62.
Tastani, S. and Pantazopoulou, S. (2002), Design of seismic strengthening for brittle
RC members using FRP jackets, Proceedings of 12th European Conference on
Earthquake Engineering, London, Paper 360.
Teng, J. G.; Chen, J. F.; Smith, S. T. and Lam, L. (2001), FRP Strengthened RC
Structures, John Wiley & Sons Inc.
Triantafillou, T. C. (1998), Shear strengthening of reinforced concrete beams using
epoxy-bonded FRP composites, ACI Structural Journal, 95(2), 107-115.
riantafillou, T. C. (2004), Structural Materials, Papasotiriou Bookstores (in Greek).
Triantafillou, T. C. (2004), Strengthening and Seismic Retrofitting of RC Structures
with Fiber Reinforced Polymers (FRP), Papasotiriou Bookstores (in Greek).
Triantafillou, T. C. and Plevris, N. (1992), Strengthening of RC beams with epoxybonded fibre-composite materials, Materials and Structures, 25, 201-211.
Yamaguchi, T., Nishimura, T., and Uomoto, T. (1998), Creep model of FRP rods
based on fibre damaging rate, Proceedings of 1st International Conference on Durability
of Fibre Reinforced Polymer (FRP) Composites for Construction, Eds. B. Benmokrane
and H. Rahman, Sherbrooke, Canada, 427-437.
Zilch, K., Niedermeier, R. and Blaschko, M. (1998), Bericht ber versuche zum
verstrken von betonbauteilen mit CFK (Test report on retrofitting concrete members with
T. C. Triantafillou
REFERENCES
CFRP).
86
T. C. Triantafillou
87
The Program
Composite Dimensioning
This Appendix gives information on the use of the program Composite Dimensioning,
which may be used for the dimensioning of concrete members strengthened with FRP in
flexure, shear or through confinement. The program makes use of the composite materials
provided by ISOMAT S.A, a major producer of chemicals and mortars for construction in
Greece.
Click .
Click setup.
By clicking OK the following window allows the user to choose one of the following three
options: FLEXURAL STRENGTHENING, SHEAR STRENGTHENING, CONFINEMENT.
88
FLEXURAL STRENGTHENING
First the type of cross section is selected (Beam or Column) and then the cross section
geometry is defined, followed by selection of concrete class or design strength, material
data for composite materials (pre-selected products or user-defined elastic modulus, limiting
strain) and data for the longitudinal reinforcement. Next, the user defines the moment Mo
acting in the critical section during strengthening (initial situation) and the axial force No , in
the case of columns. Finally a selection of the target moment capacity MRd (and axial force
NRd in the case of columns) in the critical section at the ultimate limit state is made.
By clicking Solve a new window appears, which gives the results in terms of the FRP
cross section area ( A f ), the resisting moment ( MRd,o ) of the unstrengthened cross section
and the degree of strengthening.
89
After selecting Return, a click on Input FRP dimensions opens a new window where the
user inputs the width b f and thickness t f of FRP strips (or sheets). Next, the Solve
button yields the required number of strips and the corresponding cross section area.
90
Finally the user has two options (apart from exiting the program): (a) Return without
Solving, which displays again the window of results, or (b) Return with Solving, which
displays the window of results updated with new values for A f and MRd (as well as the
updated degree of strengthening), those corresponding to the specific FRP geometry
chosen.
SHEAR STRENGTHENING
First the user selects the type of jacket, that is the FRP anchorage conditions. A
Closed jacket is typically the case in columns/shear walls (with full access) or beams with
fully anchored FRP in the compression zone, whereas an Open jacket is typically the case
in T-beams strengthened with U-shaped sheets.
defined, followed by selection of concrete class or design strength and material data for
composite materials (elastic modulus, design strength and effective strength).
In the
following the use of either continuous jacketing or strips (of width b f ) at equal spacing ( s f )
is specified and, finally, the shear to be carried by the FRP, VRd, f , is introduced.
91
click on Input FRP dimensions opens a new window where the user inputs the thickness
t fib of each layer of sheet to be used in shear strengthening. Next, the Solve button yields
the required number of layers and the corresponding total thickness of the fiber sheet, t f .
Finally the user has two options (apart from exiting the program): (a) Return without
Solving, which displays again the window of results, or (b) Return with Solving, which
displays the window of results updated with new values for t f and VRd, f , those
corresponding to the number of layers calculated.
92
CONFINEMENT
The type of cross section (rectangular or circular) is selected first and the cross section
geometry is defined, followed by selection of concrete class or design strength and material
data for composite materials (elastic modulus, design strength and effective strength). Next
the user inputs the data regarding the existing stirrups (strength of steel, spacing, cross
section of ties in each direction, concrete cover) and defines the solution requirements,
which can be one of the following: (a) increase the concrete strength and/or the ultimate
strain, e.g. for columns where axial loading is predominant; (b) increase the displacement
ductility factor (equal to the chord rotation ductility factor ), for columns subjected to
lateral (seismic) loading causing bending either in the strong or in the weak axis (that is
parallel to either the larger or the smaller side of the cross section.
93
At this point we must emphasize that use of eq. (6.27) is made, which gives quite
conservative results (thicker jackets).
thickness t f is calculated.
Apart from the thickness t f , the strength of concrete fcc1d confined with the existing stirrups
is also calculated in this case. Moreover, for a given target confined concrete strength the
corresponding ultimate strain is calculated and vice-versa. Note that if the user specifies
both a target strength and a target ultimate strain, the thickness t f returned by the program
is the one corresponding to the maximum of these two cases (hence the ultimate strain and
strength values after strengthening correspond to this thickness).
For increase in ductility:
94
Next, as in the case of shear strengthening, a click on Input FRP dimensions opens a
new window where the user inputs the thickness t fib of each layer of sheet to be used for
confinement.
The Solve button yields the required number of layers and the
Finally the user has two options (apart from exiting the program): (a) Return without
Solving, which displays again the window of results, or (b) Return with Solving, which
displays the window of results updated with new values for t f and (or ccu and fccd ,
depending on the requirements), those corresponding to the number of layers calculated.
OTHER INFORMATION ABOUT THE PROGRAM
By clicking Options on the data entry form, the user may specify its data, printing
details, and, for the case of Flexural strengthening, whether a failure mode that would not
involve Steel yielding would be acceptable or not. Finally, Print gives a printout of all the
input and output parameters.
95
Finally, by clicking Products Data Sheets on the data entry form, a new form is
presented in which all relative to repairing and strengthening issues products of ISOMAT
could be presented. The products are divided by the use, as those related to substrate
preparation (mortars or epoxy resins) and those related to FRP application (fabrics/plates or
epoxy resins). The technical data sheets are in .pdf format so the user should have already
install Acrobat Reader.