Chi Chiu Lam: Finite Element Study of Bond-Slip Behaviour of CFRP and GFRP Laminates On Brick Masonry

C
h
i

C
h
i
u

L
a
m
F
i
n
i
t
e

E
l
e
m
e
n
t

S
t
u
d
y

o
f

B
o
n
d
-
S
l
i
p

B
e
h
a
v
i
o
u
r

o
f

C
F
R
P

a
n
d
G
F
R
P

L
a
m
i
n
a
t
e
s

o
n

B
r
i
c
k
M
a
s
o
n
r
y
Finite Element Study of Bond-Slip Behaviour of CFRP and
GFRP Laminates on Brick Masonry
Chi Chiu Lam 2009 Italy
Chi Chiu Lam
Finite Element Study of Bond-
Slip Behaviour of CFRP and
GFRP Laminates on Brick
Masonry
Italy 2009
Finite element study of bond-slip behaviour of CFRP and GFRP Laminates on Brick Masonry

Erasmus Mundus Programme
ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS i
DECLARATION

Name: Chi Chiu LAM
Email: fstccl@umac.mo

Title of the
Msc Dissertation:
Finite element study of bond-slip behaviour of CFRP and GFRP Laminates on
Brick Masonry
Supervisor(s): Prof. M. R. Valluzzi and Dr. E. Garbin
Year: 2009

I hereby declare that all information in this document has been obtained and presented in accordance
with academic rules and ethical conduct. I also declare that, as required by these rules and conduct, I
have fully cited and referenced all material and results that are not original to this work.

I hereby declare that the MSc Consortium responsible for the Advanced Masters in Structural Analysis
of Monuments and Historical Constructions is allowed to store and make available electronically the
present MSc Dissertation.

University: University of Padova
Date: 21
st
, July, 2009
Signature:
___________________________


ii ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS

This page is left blank on purpose.


ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS iii
ACKNOWLEDGEMENTS

The author would like to thank Prof. M. R. Valluzzi for her persistent technical guidance and
encouragement throughout the period of the thesis work. The author would also like to thank Dr. E.
Garbin and M. Panizza for the valuable information and suggestion to the thesis work.
The author would like to express his gratitude to the European Commission for the financial support
through out the masters program. The support from the University of Macau is also much appreciated.
The author would like to thank Prof. P. Roca and all the professors from Barcelona, Padova, Guimares
and Prague who have given lectures in Barcelona, Spain. The author would like to thank all of his
masters colleagues for their consistent mental support.
The author would like to express his gratitude to his parents and his wife for their continuous support and
encouragement.


iv ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS



ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS v
ABSTRACT

Rehabilitation and preservation of historic monuments and ancient structures is attracting more and
more interest in the world. There are a certain amount of historic monuments which were built by using
brick masonry. As the development of fiber reinforced polymer (FRP) is getting more popular now,
applying bonded FRP to strengthen historic brick masonry monuments becomes one of the alternative
strengthening method. In the past decade, many researches have been carried out to investigate the
behaviour of bonded FRP to reinforced concrete structures. However, still a few contributions are
available concerning debonding problem on masonry. In this report, the experimental study which was
carried out by Panizza et al. (2009) about the bond-slip behaviour of CFRP and GFRP laminates on
brick masonry were briefly summarized. Based on the experimental results and the proposed bond-slip
equations, finite element analysis of the test specimens were carried out. Comparison of the results
obtained from different finite element models were made, they are: (1) coarse mesh versus fine mesh,
(2) exponential bond-slip versus bilinear bond-slip and (3) FRP with concentrated fibre property versus
FRP with distributed fibre property. The finite element results were also compared with the results
obtained from the analytical solution proposed by Yuan et al. (2004). With the stiffness of FRP assigned
as the average value of stiffness obtained from the test, the results obtained from both finite element
method and analytical solution compared well with the test results in term of both maximum load and
deflection. Parametric studies were carried out based on the analytical solution for different bond length
and width ratio of FRP to brick. It is found that the minimum required bond length for the current CFRP
and GFRP specimens studied are about 70 mm and 59 mm, respectively. When the bond length is
longer than the minimum required bond length, increasing the bond length does not increase the
maximum load capacity significant, however, the maximum deflection increases with increasing bond
length. For the current specimens studied, the width ratio of FRP to brick does not have significant affect
to the maximum load capacity and deflection, especially for specimen with lower stiffness ratio of FRP to
brick.


vi ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS



ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS vii
ASTRATTO

La riabilitazione e la conservazione dei monumenti e delle strutture storiche sta divenendo sempre pi di
maggiore interesse nel mondo. Molti monumenti e strutture storiche sono stati costruiti in muratura di
mattoni e necessitano di adeguate tecniche dintervento. Dato che lo sviluppo dei materiali fibro
rinforzati (Fiber Reinforced Polymers FRP) sta diventando sempre pi comune, la loro applicazione
mediante incollaggio si configura come uno dei possibili metodi di rinforzo dei monumenti e delle
strutture storiche in muratura. Nel corso dellultima decade, sono state effettuate molte ricerche per
studiare il comportamento dellincollaggio degli FRP a strutture in cemento armato. Tuttavia, sono
ancora pochi i contributi disponibili sul comportamento di adesione alla muratura. Nel presente lavoro
brevemente riportato lo studio sperimentale di base, condotto da Panizza et al. (2009), per
lidentificazione della legge di adesione locale di laminati in fibra di carbonio (CFRP) e di vetro (GFRP)
applicati su muratura di mattoni. Sulla base dei risultati sperimentali e delle leggi di adesione identificate,
sono state effettuate una serie di analisi di dettaglio agli elementi finiti delle prove sperimentali. I risultati
numerici ottenuti da diversi approcci di modellazione sono stati tra di loro confrontati. I confronti hanno
riguardato: (1) la suddivisione del dominio con elementi di media o piccola grandezza, (2) lutilizzo di una
legge locale di adesione esponenziale o bi-lineare, (3) la modellazione del materiale composito FRP
secondo un approccio con fibre concentrate o distribuite. I risultati degli elementi finiti sono stati anche
confrontati con i risultati ottenuti dalla soluzione analitica proposta dal Yuan et al. (2004). Utilizzando la
rigidezza assiale del composito FRP pari a quella ottenuta dalle prove sperimentali, i risultati ottenuti dal
metodo agli elementi finiti e dalla soluzione di analisi sono in ottimo accordo sia in termini di carico
massimo che di deformazione. Attraverso lutilizzo della soluzione analitica, sono state eseguite analisi
parametriche che hanno considerato differenti lunghezze di ancoraggio e differenti rapporti di larghezza
FRP/mattone. Si sono quindi determinate la minima lunghezza di ancoraggio per i compositi in CFRP e
GFRP utilizzati, le quali risultano rispettivamente pari a circa 70 e 59 mm. Si inoltre osservato che
quando la lunghezza di ancoraggio maggiore di quella minima richiesta, ogni suo eventuale
incremento non aumenta il carico massimo, che rimane pressoch costante, ma aumenta solo lo
spostamento ultimo. Infine, sulla base degli attuali modelli di studio, si constatato che il rapporto di
larghezza FRP/mattone non influenza significativamente il carico e lo spostamento ultimo, in particolare
nel caso di prove di adesione con un basso rapporto di rigidezza tra FRP e mattone.


viii ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS



ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS ix
TABLE OF CONTENTS
Page
1. Rehabilitation and preservation of historic clay brick structures............................. 1
1.1 A brief overview of historic clay bricks .................................................................... 1
1.2 Example of clay bricks structures in Italy ................................................................ 1
1.3 Traditional strengthening and modern strengthening techniques............................ 3
1.4 Scope and objective............................................................................................... 4
1.5 Outline of thesis...................................................................................................... 4
2. Strengthening of structures with FRP: bond-slip behaviour..................................... 5
2.1 Bond-slip of FRP to RC structures.......................................................................... 5
2.2 Experimental study of bond-slip behaviour of FRP-to-brick by Panizza et al. (2009).
............................................................................................................................... 7
2.2.1 Experimental tests description and test results....................................................... 8
2.2.2 Fracture energy and bond-slip curve .................................................................... 10
3. Finite element analysis of FRP laminates on brick masonry................................... 13
3.1 Material properties, boundary condition and analysis procedure........................... 13
3.2 Finite element mesh study of FRP laminates on brick masonry............................ 16
3.2.1 Comparison of FE load versus deflection results with test results......................... 18
3.2.2 Comparison of FE strain results with test results .................................................. 20
3.3 Exponential bond-slip curve versus bilinear bond-slip curve................................. 24
3.3.1 Comparison of FE strain results with test results (Exponential versus Bilinear
bond-slip models) ................................................................................................. 28
3.4 Finite element model of FRP (Concentrated fiber versus Distributed fiber) ........... 31
3.4.1 Model with exponential bond-slip behaviour (Concentrated fiber versus Distributed
fiber) ..................................................................................................................... 32
3.4.2 Model with bilinear bond-slip behaviour (Concentrated fiber versus Distributed fiber)
............................................................................................................................. 38
4. Comparison of finite element resutls with analytical solution ................................ 45
4.1 Full-range analytical solution proposed by Yuan et al. (2004)............................... 45
4.2 Comparison of finite element results with analytical solution................................. 49
4.3 Comparison of analytical solution results and finite element results with test results
............................................................................................................................. 52
4.3.1 Comparison of FE strain results with test results .................................................. 55

x ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS
5. Parametric study of bond-slip behaviour of FRP-to-brick ....................................... 67
5.1 Effect of bond length of FRP to the capacity and deflection of FRP-to-brick joint .. 67
5.2 Effect of width ratio of FRP-to-brick to the capacity and deflection of FRP-to-brick
joint....................................................................................................................... 71
6. Summary and conclusion .......................................................................................... 75
6.1 Conclusion............................................................................................................ 75
6.2 Suggestion for further study.................................................................................. 76
References ......................................................................................................................... 77


ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS xi
LIST OF FIGURE
Page
Figure 1 Torrazzo brickwork Tower adjacent to the Cathedral of Cremona........................... 2
Figure 2 Various type of bond test methods.......................................................................... 5
Figure 3 Bond-slip curves from existing bond-slip models for FRP-to-concrete..................... 7
Figure 4 Geometry of specimen and location of strain gauge (Panizza et al. 2009) .............. 9
Figure 5 Bond-slip curves for CFRP specimens.................................................................. 12
Figure 6 Bond-slip curves for GFRP specimens.................................................................. 12
Figure 7 Expontiental bond-slip curve and data points used in finite element analysis (CFRP)
............................................................................................................................................ 15
Figure 8 Expontiental Bond-slip curve and data points used in finite element analysis (GFRP)
............................................................................................................................................ 15
Figure 9 Boundary conditions of FE model ......................................................................... 16
Figure 10 Typical (a) coarse mesh and (b) fine mesh FE models........................................ 17
Figure 11 Percentage of nodal forces assigned to the loading end of FRP (concentrated fiber
model).................................................................................................................................. 18
Figure 12 Comparison of the load versus deflection of CFRP specimens (coarse versus fine
mesh models) ...................................................................................................................... 19
Figure 13 Load versus deflection of fine mesh models of GFRP specimen......................... 20
Figure 14 Comparison of strain results of CFRP fine mesh models and test results............ 22
Figure 15 Comparison of strain results of GFRP fine mesh models and test results ........... 24
Figure 16 Bilinear bond-slip curve and data points used in finite element analysis (CFRP). 25
Figure 17 Bilinear bond-slip curve and data points used in finite element analysis (GFRP). 25
Figure 18 Comparison of the load versus deflection of CFRP specimens (exponential versus
bilinear bond-slip behaviour) ................................................................................................ 27
Figure 19 Comparison of the load versus deflection of GFRP specimens (exponential versus
bilinear bond-slip behaviour) ................................................................................................ 27

xii ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS
Figure 20 Comparison of strain results of CFRP specimens (exponential versus bilinear
bond-slip)............................................................................................................................. 29
Figure 21 Comparison of strain results of GFRP specimens (exponential versus bilinear
bond-slip)............................................................................................................................. 31
Figure 22 Percentage of nodal forces assigned to the loading end of FRP (Distributed fiber
model).................................................................................................................................. 32
Figure 23 Comparison of the load versus deflection of CFRP specimens with exponential
bond-slip behaviour (concentrated fiber versus distributed fiber).......................................... 33
Figure 24 Comparison of the load versus deflection of GFRP specimens with exponential
bond-slip behaviour (concentrated fiber versus distributed fiber).......................................... 34
Figure 25 Comparison of strain results of CFRP specimens with exponential bond-slip
behaviour (concentrated versus distributed fiber)................................................................. 36
Figure 26 Comparison of strain results of GFRP specimens with exponential bond-slip
behaviour (concentrated versus distributed fiber)................................................................. 37
Figure 27 Comparison of the load versus deflection of CFRP specimens with bilinear bond-slip
behaviour (concentrated fiber versus distributed fiber) ......................................................... 39
Figure 28 Comparison of the load versus deflection of GFRP specimens with bilinear bond-slip
behaviour (concentrated fiber versus distributed fiber) ......................................................... 39
Figure 29 Comparison of strain results of CFRP specimens with bilinear bond-slip behaviour
(concentrated versus distributed fiber) ................................................................................. 41
Figure 30 Comparison of strain results of GFRP specimens with bilinear bond-slip behaviour
(concentrated versus distributed fiber) ................................................................................. 43
Figure 31 Single-lap FRP bonded joint................................................................................ 45
Figure 32 Local bilinear bond-slip curve.............................................................................. 47
Figure 33 Interfacial shear stress distribution and propagation of debonding ...................... 47
Figure 34 Typical full-range analytical load versus displacement curve............................... 48
Figure 35 Comparison of FE results and analytical solution (CFRP specimens) ................. 51
Figure 36 Comparison of FE results and analytical solution (GFRP specimens) ................. 52
Figure 37 Load versus deflection results of CFRP specimens with test average FRP stiffness
(E
cfrp
t
cfrp
) .............................................................................................................................. 54

ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS xiii
Figure 38 Load versus deflection results of GFRP specimens with test average FRP stiffness
(E
gfrp
t
gfrp
) .............................................................................................................................. 54
Figure 39 Comparison of strain results of CFRP specimens (with mean value of E
cfrp
t
cfrp

assiged) ............................................................................................................................... 60
Figure 40 Comparison of strain results of GFRP specimens (with mean value of E
gfrp
t
gfrp

assiged) ............................................................................................................................... 65
Figure 41 Load versus deflection curve of CFRP specimens with different bond length
(Analytical versus finite element results) .............................................................................. 68
Figure 42 Load versus deflection curve of GFRP specimens with different bond length
(Analytical versus finite element results) .............................................................................. 68
Figure 43 Analytical load versus deflection curve of CFRP and GFRP specimens with different
bond length.......................................................................................................................... 70
Figure 44 Maximum deflection versus bond length (b
frp
/ b
brick
= 0.42)................................. 71
Figure 45 Load per width of CFRP specimens with different bond length and width ratio.... 72
Figure 46 Load per width of GFRP specimens with different bond length and width ratio.... 72
Figure 47 Summary of load capacity, bond-length and width ratio of FRP of CFRP specimen
............................................................................................................................................ 73
Figure 48 Summary of load capacity, bond-length and width ratio of FRP of GFRP specimen
............................................................................................................................................ 74


xiv ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS



ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS xv
LIST OF TABLE
Page
Table 1 Summary of bricks mechanical properties and reinforcement components properties
.............................................................................................................................................. 8
Table 2 Summary of the experimental failure load, P
u
(Panizza et al. 2009) ....................... 10
Table 3 Significant values for local bond of CFRP and GFRP bonded to brick specimens.. 11
Table 4. Summary of material properties of brick and FRP used in finite element analysis . 14
Table 5. Summary of FE results of coarse mesh and fine mesh model of CFRP specimens19
Table 6 Summary of FE results of fine mesh model of CFRP and GFRP specimens.......... 20
Table 7. Summary of FE results of exponential and bilinear bond-slip curves of CFRP
specimens ........................................................................................................................... 26
Table 8 Specimens with exponential bond-slip behaviour (concentrated versus distributed
fiber) .................................................................................................................................... 33
Table 9 Specimens with bilinear bond-slip behaviour (concentrated versus distributed fiber)
............................................................................................................................................ 38
Table 10 Comparison of finite element resutls and analytical results................................... 51
Table 11 Summary of results from finite element analysis, analytical solution and test (with
equivalent stiffness of FRP equals to the average value of test results) ............................... 53
Table 12 Maximum deflection versus bond length (b
frp
/ b
brick
= 0.42) .................................. 70


xvi ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS



ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS 1
1. REHABILITATION AND PRESERVATION OF HISTORIC CLAY BRICK
STRUCTURES

Rehabilitation and preservation of historic monuments and ancient structures is attracting more and
more interest in the world. However, the conservation of architectural heritage is a difficult task due to
the complex geometry of buildings and large variability of construction materials. It is a multidisciplinary
work that requires the contribution of different scientists and professionals in order to collect all the data
necessary for the intervention. In the last decades, the strategic importance of historic buildings due to
cultural and economical reasons caused a large increase in studies dealing with historic structures and
materials. In the case of ancient clay brick masonry, studies have been focusing on the main mechanical
properties, retrofitting techniques, seismic vulnerability, etc.
1.1 A brief overview of historic clay bricks

Due to the aggressions of the environment (snow, rain, cold, heat, etc.) human being started to protect
themselves by using the natural materials such as natural caves, tree trunks, animals fur, straw, clay,
etc. With the appearance of the first civilization, around 9000 to 7000 BC, the construction techniques
evolved stone, adobe, wood and clay brick begun to be used. The first vestiges of brick masonry
buildings were found in the region of Israel (Mesopotamia) and dated from 9000 to 8000 BC (Fernandes
et al 2006). Clay brick masonry is, effectively, one of the finest and most durable construction techniques
ever invented by human. Masonry consist of building stable bonded stacks of small pieces by hand
(Vekey, 1998). Used since the time of the first villages and cities built by human, masonry application
has been growing and evolving to new uses all over the entire civilized world. It was a fundamental
building material in Mesopotamian, Egyptian and Roman periods. During Roman period, the use of clay
brick increased and become specialized in order to maximize its benefits. Despite several modifications
of the clay brick uses, shape and manufacture along thousands of years of constant evolution, the
simplicity that made its success remained.
1.2 Example of clay bricks structures in Italy
There are many historical brick masonry structures in Italy. One of the examples is the Torrazzo which is
a medieval brickwork tower adjacent to the Cathedral of Cremona, a town 90 km far from Milan, Italy
(Binda et al. 2000). The height of the Tower is around 112 m and it is the highest medieval masonry
tower in Europe. The real date of construction is not known but declared around the 13
th
century. It
belongs to a group of monuments, the Cathedral, the Bapistery, the Town Hall Palace, the Militia Loggia
and the Torrazzo itself, which form one of the most beautiful Italian squares. The Bell-Tower consists of

2 ADVANCED MASTERS IN STRUCTURAL ANALYSIS OF MONUMENTS AND HISTORICAL CONSTRUCTIONS
three architectural elements: the Romanesque tower, the Ghirlandina and the Bertazzola lodge at the
base (Fig. 1). The Romanesque tower has a square shape 13 m for each side large with load bearing
masonry walls 3.3 m thick at the ground floor. It has four vertical reinforcing built-in strips at the four
corners, two semi-cylindrical ones on the facade on each side and six horizontal cornices, made of
simple little arches or hanging ones. The external load-bearing walls of the tower show since years
several cracks. Since the crack pattern has developed along the years a possible time dependent
behaviour of the material can be supposed. This phenomenon, together with the effects of cyclic loads
as wind and temperature variations can eventually cause to the structure long-term heavy damages.
Binda et al. (2000) applied systematically georadar, sonic tests and flat jack tests on the walls of the
Torrazzo. It is shown that an external thin leaf is partially detached in some part of the tower base.

Figure 1 Torrazzo brickwork Tower adjacent to the Cathedral of Cremona

Another example of brick masonry structures in Italy is the Basilica of S. Lorenzo at Cremona, Italy. The
Basilica of S. Lorenzo shows all the formal contradictions deriving from centuries of continuous
modifications of the load-bearing structures. The Basilica is a three naves construction built in the 12
th

century on the remains of a previous church destroyed by a fire. In 15
th
century, an important
intervention introduced vaulted ceilings to the lateral naves and a monumental three-lobes chapel on the
north-west corner. Several damages were reported by Anzani et al. (2007).

It is shown from the above examples that nowadays ancient buildings exhibit often serious damages,
which are the result of many years of abandon, of the weathering of materials, poor or inadequate
structural behaviour and damage due to earthquake. Therefore, it is necessary to carry out modification
or strengthening of those structures which show potential problems.
1.3 Traditional strengthening and modern strengthening techniques
Since the beginning of structural restoration, architects and engineers have envisaged and actually
applied a wide variety of repair or strengthening interventions to improve the structural response of
ancient masonry structures. Many historic constructions are structurally inadequate for current use.
Additional material deterioration due to environmental factors and lack of maintenance has caused
significant weakening of historically important structures (Drysdale et al. 1993). A few decades ago,
strengthening of structures was accomplished by the materials available at that time. The criteria for
choosing a particular solution must take into account not only its structural effectiveness and cost, but
also the compatibility with the techniques and materials used in the construction of the monument
regarding its original conception and historical value. Different techniques, such as pre-stressing cables,
injected mortar, inserted steel bars, wooden planks, iron cramps and synthetic polypropylene fibers, for
retrofitting historical constructions have been adopted. Some of the oldest techniques are still in use,
such as dismantling and remounting with possible improved material substitution. In the process
masonry element or structures contain parts that have to be removed, substituted or repaired, if a local
intervention is not feasible. The main objective is to recover the functionality of a structure maintaining its
historical and cultural value, modifying an erroneous design. Application of strengthening to arched
structures in the last 20-25 years allowed developing several methods. Such methods include the
installation of stainless steel reinforcing bars in the near surface zones of the masonry (Sumon 1997).
As an example the use of near-surface mounted reinforcement (NSM) in masonry arch bridges can
enhance the load carrying capacity, delaying the formation of cracks and hinges and, and at the same
time, minimize any disruption to the bridge users (Garrity 1995).
Traditional techniques employ the materials and building processes used originally for the construction
of ancient structures. Besides the usage of traditional techniques, modern techniques such as the
application of modern materials for strengthening are getting more attention nowadays. Modern
approaches are based on the idea that the strengthening should be light and removable and, if possible,
it should not change the structural scheme or the construction. This objective can be achieved by using
advance composite materials. Strengthening of existing masonry structures can be carried out by
applying bonded fiber reinforced polymer (FRP) such as glass fiber reinforced polymer (GFRP) or
carbon fiber reinforced polymer (CFRP). These materials present several advantages, such as low
specific weight, corrosion immunity and high tensile strength. Their flexibility and easy application allow
a wide range of intervention scenarios. The FRP materials have the advantage that they are easily adapt
to the surface to be strengthened, on the external face of the element locally (as a strips arrangement) or
to the whole surface of the structure (as a grid reinforcement arrangement). The bond between the

masonry and fiber is normally obtained with the use of epoxy resins or mortar (Sanchez et al. 2007). An
effective use of this technique requires certain regularity in the masonry surface.
1.4 Scope and objective
It is shown previously that strengthening of brick masonry structures can be done by applying the
modern materials such as fiber reinforced polymer (FRP). In the past decade, many researches have
been carried out to investigate the behaviour of bonded FRP to reinforced concrete structures. However,
still a few contributions are available concerning debonding problem on masonry. Recently, Panizza et
al. (2009) have carried out double-lap push-pull shear tests to investigate the bond-slip behaviour of
externally bonded FRP to brick. Based on the test results, bond-slip equations (one continuous and one
bi-linear equation) are proposed for FRP bonded to brick element. In this thesis, finite element analysis
of those test specimens is carried out in order to achieve the following goals:
1. Set up a suitable finite element model based on the bond-slip equations proposed by
Panizza et al. (2009).
2. Compare the finite element results with different assumption of FRP material
properties.
3. Based on the verified finite element model, compare the finite element results with the
available analytical solution in literature.
4. Carry out parametric studies of the bond-slip behaviour of FRP bonded to brick by using
the verified analytical solution.
1.5 Outline of thesis
The outline of thesis is as following. A brief introduction of clay brick, clay brick masonry structure and
strengthening techniques is shown in Chapter 1. Discussion of bond-slip tests of external bonded FRP
to concrete and masonry elements is shown in Chapter 2 together with the test results of Panizza et al.
(2009). In Chapter 3, finite element analysis of the test specimens of Panizza et al. (2009) is presented.
Followed to the finite element analysis of the test specimens, the finite element results were compared
with the results obtained from current available analytical solution in Chapter 4. Parametric study based
on the analytical solution of the FRP-to-brick bond strength with different parameters, such as bond
length of FRP and width ratio of FRP to brick, is shown in Chapter 5. Summary, conclusion and
suggestion for further study is shown in Chapter 6.

2. STRENGTHENING OF STRUCTURES WITH FRP: BOND-SLIP BEHAVIOUR
In the past decade, the application of advance materials, such as fiber reinforced polymer (FRP) for
strengthening structures become more popular. One of the main reasons is due to the great effort of the
searches in this area and more understanding about the behaviour of FRP to the substrate is achieved.
Due to the popularity of reinforced concrete (RC) structures nowadays, most of the researches are
focused on the bonded behaviour of FRP applied to RC structures. In fact, similar technique can be
applied to historical brick masonry structures. However, more studies about the applied of FRP on brick
masonry structures need to be carried out in order to obtain better understanding of the behaviour.
2.1 Bond-slip of FRP to RC structures
A fairly large amount of bond tests for the FRP sheet-concrete interfaces under shear have been carried
out in the past decades. As it is shown in Fig. 2 that test methods include single lap pullout test method
(Chajes et al. 1996), double lap pullout bond tests (Sato et al. 2001) and bending tests (Lorenzis et al.
2001).

(a) single lap pullout test (b) double lap pullout test (c) bending test
Figure 2 Various type of bond test methods

Through those experimental studies the bond mechanisms of FRP sheet-concrete interfaces, the
important aspects such as the (1) bond strength, (2) interfacial fracture energy, (3) effective bond length
and (4) bond stress-slip relationship have been clarified. A summary of those important aspects is
reported by Dai et al. 2005 and they are listed in following.
Bond strength: the FRP sheet-concrete interface fails mostly at a thin layer beneath the concrete
surface. As a result, the concrete surface condition and strength are critical factors affecting the
interfacial bond strength. At present, sandblasting is the most common surface treating method. Chajes
et al. (1996) and Sato et al. (2000) studied the effects of concrete strength f
c
and concluded that the
average interface bond strengths, which are the ultimate pullout forces divided by bond areas between
FRP sheets and concrete, are linearly proportional to f
c
1/2
, f
c
2/3
, and f
c
1/5
, respectively. Besides the
concrete property, the FRP and adhesive properties affect the interface bond strength as well. In
general, using higher FRP stiffness (Nakaba et al. 2001) and softer adhesives (Dai et al. 2002) can
increase the average bond strength.

Interfacial fracture energy: the interfacial fracture energy G
f
, which is the area underneath the interfacial
bond stress-slip curve, is an important parameter for the bond characteristics. Based on different types
of interfacial bond stress-slip relationships, Yuan et al. (2001) proved that the maximum interfacial bond
force can be expressed as a function of the G
f
and FRP stiffness (E
frp
t
frp
). Due to the clear physical
meaning of the G
f
, it is very useful to apply it in numerical analysis for deriving bond strength and
anchorage length models as well as for clarifying the debonding failure mechanisms of FRP
sheet-concrete interfaces in more comprehensive ways (Wu and Yin 2002).
Effective bond length: there exists an active bonding zone named as the effective bond length L
e
, along
which most of the interfacial load is transferred between FRP sheets and concrete. When the bond
length of FRP sheet-concrete interfaces exceeds the L
e
, the bond strength will not increase significantly
any longer. In general, it was reported that the effective bond length increases with the stiffness of FRP
sheets. However, due to the different materials used in various researches, the effective bond length
was reported in a fairly big range (from 45 mm to 275 mm).
Bond stress-slip relationship: the bond-slip relationship relates on the interfacial fracture energy, G
f
, as
discussed previously. The shape of the bond-slip model determines the predicted distribution of axial
strains in the plate. Lu et al. (2005) summarized four existing bond-slip models for normal-adhesive
interfaces for an FRP-to-concrete bonded joint with the following properties: f
c
= 32 MPa, f
t
= 3.0 MPa, b
f

= 50 mm, b
c
= 100 mm, E
f
t
f
= 16.2 GPa mm. The curves of the four bond-slip models are shown in Fig.
3 and it can be seen that the shapes of the predicted bond-slip curves differ substantially. In particular,
the linear-brittle model of Neubauer and Rostasy (1999) is very different from the other three models.
The fact that the bond stress reduces to zero at the ultimate slip dictates that there exists an effective
bond length beyond which an increase in the bond length will not increase the ultimate load. Nakaba et
al. (2001) and Savioa et al. (2003) have shown that the bond-slip curve should have an ascending
branch and a descending branch. A bilinear model of bond-slip curve is proposed by Monti et al. (2003)
and it can be used as an approximation but the linear-brittle model proposed by Neubauer and Rostasy
(1999) is unrealistic. Apart from the general shape, the slip at maximum stress and the ultimate slip at
zero bond stresses, determine the accuracy of the model. It is interesting to know that the models by
Nakaba et al. (2001), Monti et al. (2003) and Savioa et al. (2003) are in reasonably close agreement,
and the linear-brittle model of Neubauer and Rostasy (1999) predicts a similar maximum bond stress. Lu
et al. (2005) used another approach to obtain the bond-slip curve of FRP-to-concrete. Their new
bond-slip models are not based on axial strain measurements on the FRP plate; instead, they are based
on the predictions of a meso-scale finite element model, with appropriate adjustment to match their
predictions with the experimental results for a few key parameters. The bond-slip curves proposed by Lu
et al. (2005) (the precise model and the bilinear model) are shown in Fig. 3 as well. By comparing both
the bond strength and strain distribution in the FRP plate, Lu et al. (2005) concluded that their new
bond-slip models which is based on a combination of finite element results and the test results give
better prediction than the other bond-slip models mentioned.


Figure 3 Bond-slip curves from existing bond-slip models for FRP-to-concrete

2.2 Experimental study of bond-slip behaviour of FRP-to-brick by Panizza et al.
(2009)

As it is shown in the previous section that many researches have been focused on the bond-slip
behaviour of FRP-to-concrete elements, however, only a few investigations concerning about the
debonding on masonry substrate are available. Recently, some researchers such as Aiello et al. (2005)
investigated the bond behaviour of FRP to natural stones and Briccoli Bati et al. (2007) investigated the
bond behaviour of FRP to solid clay bricks. They adopted double-lap push-pull shear tests which consist
in loading in tension two reinforcement strips, symmetrically connected to the support, in order to create
shear stresses at the interface; the brittle support is subjected to compressive stresses. This set-up is
based on the assumption that the applied load is equally distributed on the two strips, but it is also
particularly simple and suitable for the usual common available test machine.
Experimental study of the bond behaviour of CFRP and GFRP laminates on brick masonry was carried
out by Panizza et al. (2009). In their study, five samples of clay brick bonded by high-strength carbon
reinforcement and five samples of clay brick bonded by alkali-resistant glass reinforcement were tested
to examine the bond strength of the FRP to the clay brick. Solid clay bricks (nominal dimension 250 x
120 x 55 mm) were used as substrate, and the MBrace Wet lay-up system as reinforcement. In their
study, double-lap Push-pull Shear Tests were performed to examine the bond-slip behaviour of CFRP
and GFRP laminates on brick masonry. The experimental results, in terms of failure load, were
compared with predictive bond-strength models proposed in literature, mainly available for concrete.
Also, the interface fracture energy which based on the experimental strength was calibrated for FRP
bond to brick. Bond-slip equations, an exponential equation based on simplified analytical model and a

bilinear equation, were proposed. Brief descriptions of the test program and test results are shown in
the following section.
2.2.1 Experimental tests description and test results
In the experimental program descript by Panizza et al. (2009), the solid clay bricks (nominal dimension
250 x 120 x 55 mm) were used as substrate, and the MBrace wet lay-up system was used as
reinforcement. There are two types of fibers used in the experimental program: (1) high strength carbon
fibers (CFRP) and (2) alkali-resistant glass fibers (GFRP). Five specimens were prepared for each fiber,
respectively. Therefore, ten specimens were tested in total. The main bricks properties and the
reinforcement system properties, which are obtained from producers datasheets, are summarized in
Table. 1.
Table 1 Summary of bricks mechanical properties and reinforcement components properties
Clay Brick
Compressive strength (cubic) f
c
= 50.94 MPa
characteristic value f
ck
= 41.17 MPa
Direct tensile strength f
t
= 2.37 MPa
Elastic modulus (secant) E
b
= 16110 MPa
Shear modulus (estimated using = 0.16) G
b
= 6944 MPa
Adhesive
Compressive strength > 80 MPa
Direct tensile strength > 50 MPa
Tensile elastic modulus (tangent) E
a
> 3000 MPa
Shear modulus (estimated using = 0.33) G
a
= 1128 MPa
Carbon Fiber
Characteristic tensile strength 3430 MPa
Characteristic tensile strength per unit width 565 N/mm
Ultimate tensile strain 1.5 %
Tensile elastic modulus E
carb
= 230000 MPa
Thickness of one ply t
carb
= 0.165 mm
Glass Fiber
Characteristic tensile strength 1700 MPa
Characteristic tensile strength per unit width 390 N/mm
Ultimate tensile strain 2.8 %
Tensile elastic modulus E
glass
= 65000 MPa
Thickness of one ply t
glass
= 0.23 mm

The fiber reinforcement was applied to the both sides of the clay brick. The bonded length of fiber
reinforcement on each side is 200 mm and an unbonded length from the end of the brick, equal to 30
mm, was imposed next to the loaded end in order to minimize edge effects. Each strip of reinforcement,

made by a single layer of fibers and two layers of epoxy resin, was bonded, at the loaded end, to a steel
support connected to the testing machine. Brick was connected to the machine through a steel frame,
made by two plates linked by bolts. Strain gauges were mounted on the fiber reinforcement, one on the
unbonded zone next to the loaded end of the reinforcement and six on the bonded length. The strain
gauges were applied in a non-equally distributed pattern with less spaced near the loaded end where
higher strain is expected in order to optimize the number of instruments. The geometry of specimens
and location of strain gauges are shown in Fig. 4.

(a) Geometry of specimen

(b) Location of strain gauge
Figure 4 Geometry of specimen and location of strain gauge (Panizza et al. 2009)

The failure loads, P
u
, are summarized in Table 2. It is shown from the table that the specimens bonded
with CFRP have higher failure loads compared to the specimens bonded with GFRP. The mean failure
load in the first case is around 35% higher than the latter.


Table 2 Summary of the experimental failure load, P
u
(Panizza et al. 2009)
Specimens C1 C2 C3 C4 C5 Average
E
frp
t
frp
(N/mm) 27,129 55,512 47,024 45,789 55,845 46,260
P
u
, (N) 31,884 34,233 35,325 39,210 40,301 36,191

Specimens G1 G2 G3 G4 G5 Average
E
frp
t
frp
(N/mm) 11,715 20,013 18,525 23,598 19,514 18,673
P
u
, (N) 23,380 27,940 27,300 26,400 28,360 26,676

2.2.2 Fracture energy and bond-slip curve
The interface fracture energy mode II, G
f
, is defined by Equation 1 as the definite integral of the
tangential stress, , expressed as function of the mutual slip of composite and substrate, s:

=
0
f
ds ) s ( G
(1)
Therefore, it is possible to relate the failure load (P
u
) with the fracture energy (G
f
) and one common
equation which has been reported by Savoia et al. (2003) and Dai et al. (2005) is as following:

f frp frp frp u
G t E 2 b 2 P =
(2)
By applying the above equation to the experimental results, the fracture energy (G
f
) was calibrated
through the mean values of failure load (P
u
), width of FRP (b
frp
), the elastic modulus of FRP (E
frp
) and the
thickness of FRP (t
frp
). It is reported by Panizza et al. (2009) that the fracture energy (G
f
) of the CFRP
and GFRP specimens is 1.42 N/mm and 1.91 N/mm, respectively.
In order to set up a suitable bond-slip curve according to the strain results and the predicted fracture
energy, it is proposed to use a single function which considered an ascending segment and a softening
behaviour as following:

Bs
e s A ) s (

=
(3)
where A and B are regression constants, t is the interface tangential stress and s is the composite slip.
Applying the calibrated fracture energy value, it was possible to have a fitting function that depends on
just one parameter, as shown in following equation.


Bs
f
2
f
2
0
e s G B ) s ( ; G B A
B
A
ds

= = =

(4)
By applying the above equations, an UniPd curve in a normalized form is proposed as following.

|
|
\
|
|
|
\
|
=
0
s
s
1
0
max
e
s
s
) s (

(5)
where s
0
= 1 / B and
max
= (s
0
) are the coordinates of the point of maximum tangential stress.
After the optimization of the Unipd curves, in case of carbon reinforcement and glass reinforcement,
equations were calibrated according to bilinear law, whose analytical form is shown in following.

|
|
\
|
=
0
s s
s s
) s / s (
) s (
0 f
f
max
0 max

0
f 0
0
s s
s s s
s s 0

(6)
where s
f
is the ultimate strain, related to null stress. This form is commonly proposed by some guidelines
(fib Bulletin 14 2001; CNR-DT 200 2004).
Based on the test results, the parameters which are needed for Equations 5 and 6 for the CFRP and
GFRP specimens are summarized in Table 3. The plot of the bond-slip curves are shown in Figs. 5 and
6 together with the bond-slip data obtained from literatures (Panizza et al. 2009).

Table 3 Significant values for local bond of CFRP and GFRP bonded to brick specimens
Specimen Curve G
f

N/mm

max

MPa
s
o

mm
s
f

mm
CFRP UniPd fitting 1.42 7.22 0.072 ---
Bilinear fitting 1.42 7.22 0.034 0.392
GFRP UniPd fitting 1.91 6.33 0.111 ---
Bilinear fitting 1.91 6.33 0.048 0.603


0.0
2.0
4.0
6.0
8.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Slip s (mm)
T
a
n
g
e
n
t
i
a
l

s
t
r
e
s
s

(
M
P
a
)
Data from literatures
UniPd curve
Bilinear
CARBON

Figure 5 Bond-slip curves for CFRP specimens

0.0
2.0
4.0
6.0
8.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Slip s (mm)
T
a
n
g
e
n
t
i
a
l

s
t
r
e
s
s

(
M
P
a
)
Data from literatures
UniPd curve
Bilinear
GLASS

Figure 6 Bond-slip curves for GFRP specimens


3. FINITE ELEMENT ANALYSIS OF FRP LAMINATES ON BRICK MASONRY
The bond behaviour of CFRP and GFRP laminates on brick masonry of the test specimens was studied
by mean of finite element method. The finite element program, Diana (2008), is used for modeling and
analyzing the test specimens. In order to reduce the computational effort, the three-dimensional
FRP-to-brick specimen was modeled as a plane stress problem by using eight-node quadrilateral
isoparametric plane stress element (CQ16M) for both FRP and brick materials. For the finite element
modeling of wet lay-up FRP sheets, three options may be used (Lu et al. 2004): (a) a plate with its
thickness being equal or similar to the actual thickness of the laminate including all adhesive, but with
the fibers concentrated in a thickness equal to the nominal thickness of the fiber sheet sitting in the
middle of the plate, (b) a plate with its thickness being equal or similar to the actual thickness of the
laminate including all adhesive, with the fibers assumed to be evenly distributed across the plate
thickness; and (c) a plate with a nominal thickness (generally the thickness of the fiber sheet) without
considering the adhesive. In general, option (a) and (b) are more suitable for modeling the laminated
FRP since both material properties of fiber and matrix are considered in the finite element model. In
order to examine the different between options (a) and (b), comparison of the finite element results were
carried out in the finite element analysis and the results are discussed in the later section. Between the
interface of the FRP and brick, six-node (3+3 nodes) interface elements (CL12I) were used for modelling
the bond-slip behaviour according to the bond-slip equations proposed by Panizza et al. (2009). Details
of the finite element model are shown in the following section.
3.1 Material properties, boundary condition and analysis procedure
In the finite element model, both FRP and brick are considered as isotropic elastic material. As it is
discussed in the previous section, the FRP was modelled according to two options. The first option
considered the fiber as reinforcement and it was concentrated in the middle of the matrix. The second
option considered the FRP as a homogenous material with a corresponding elastic modulus calculated
according to the following equation.

frp
fiber frp
matrix
frp
fiber
fiber frp
t
) t t (
E
t
t
E E
+ =
(7)
where E
frp
= elastic modulus of composite material, E
fiber
= elastic modulus of fiber, E
matrix
= elastic
modulus of matrix, t
frp
= thickness of composite material and t
fiber
= thickness of fiber. The thickness of
the composite material (t
frp
) is assumed to be 2 mm for both CFRP and GFRP composite. Summary of
the material properties of brick and FRP used in the finite element analysis is shown in Table 4.


Table 4. Summary of material properties of brick and FRP used in finite element analysis
Brick Elastic modulus E
b
16110 MPa
Poisson ratio 0.16
Matrix Elastic modulus E
matrix
3,000 MPa
Poisson ratio 0.33
Carbon fiber Elastic modulus E
carb
230,000 MPa
Thickness t
carb
0.165 mm
CFRP composite Elastic modulus E
cfrp
21,728 MPa
Thickness t
cfrp
2.0 mm
Glass fiber Elastic modulus E
glass
65,000 MPa
Thickness t
glass
0.23 mm
GFRP composite Elastic modulus E
gfrp
10,130 MPa
Thickness t
gfrp
2.0 mm

All the non-linear behaviour was modelled by the interface element. The exponential bond-slip
behaviour of the interface element was assigned according to the data obtained from Table 3 and
Equation 5. In the finite element analysis, the continuous bond-slip equations were converted to a set of
pair of data points (shear stress versus slip) as an input for the bond-slip behaviour. The corresponding
bond-slip curves for CFRP and GFRP specimens are shown in Figs. 7 and 8 together with the data
points used in the finite element analysis, respectively.

0
1
2
3
4
5
6
7
8
0 0.2 0.4 0.6 0.8 1
Slip, s (mm)
T
a
n
g
e
n
t
i
a
l

s
t
r
e
s
s

(
M
P
a
)
UNIPD
Data for FEA

Figure 7 Expontiental bond-slip curve and data points used in finite element analysis (CFRP)

0
1
2
3
4
5
6
7
0.0 0.2 0.4 0.6 0.8 1.0
Slip, s (mm)
T
a
n
g
e
n
t
i
a
l

s
t
r
e
s
s

(
M
P
a
)
UNIPD
Data for FEA

Figure 8 Expontiental Bond-slip curve and data points used in finite element analysis (GFRP)

Due to symmetry, only half of the specimen is modeled. Therefore, suitable boundary conditions were
assigned to the brick and the FRP as it is shown in Fig. 9. The initial stiffnesses in the normal and
tangential direction of the bond interface are assigned as 16,110,000 N/mm
3
and 240.56 N/mm
3
. The
initial stiffness in the normal direction is about 1,000 times of the elastic modulus of the brick per mm and
it is believed that this high value of stiffness is good enough to avoid the FRP element to penetrate to the
brick element. The initial stiffness in the tangential direction is according to the initial slope of the
bond-slip curve. For the un-bonded surface between the FRP and brick, similar interface elements as
used in the bonded interface are used. However, the stiffness in the tangential direction is assigned
close to zero in order to simulate the free slip behaviour in this region.

Figure 9 Boundary conditions of FE model

Nodal forces are applied to the loading end of the FRP. In order to simulate the uniform displacement at
the loading end of the FRP, suitable nodal forces should be assigned to the nodes according to their
position and modeling option. Detail of the arrangement of nodal forces to achieve uniform displacement
at the loading end of FRP is discussed in the later section. Non-linear analysis was performed by using
an energy based method together with arc-length control method. With a monotonic incremental load
assigned to the loading end of the FRP, equilibrium of the system was calculated and the corresponding
reactions, internal stress/strain and deformation of the structures were obtained.
3.2 Finite element mesh study of FRP laminates on brick masonry
For the finite element modeling of FRP-to-brick, one of the important issues is the size of the mesh.
Therefore, mesh study was carried out in order to examine the effect of mesh size to the results of the
finite element analysis. Two finite element meshes, the coarse mesh and the fine mesh, of the

FRP-to-brick model were prepared based on the dimension and material properties of the CFRP
specimen. For the coarse mesh model, the size of element is about 2 x 2.5 mm. The number of elements
in the coarse mesh model is about 1490 plus 115 interface elements. In order to have better resolution of
the stress and strain near the interface between FRP and brick, a fine mesh model was created. The fine
mesh model is a modification of the coarse mesh model by refining the element size within the interface
region. It was done by introducing a transfer layer of elements near the interface region in order to
reduce the size of element at that location to around 1 x 1 mm. Also, the FRP was modeled by using two
layers element of around 1 x 1 mm. In this case, the number of elements in the fine mesh model is 3085
plus 230 interface elements. Therefore, the number of elements of the fine mesh model is about double
of that of coarse mesh model. The corresponding coarse mesh and fine mesh finite element models are
shown in Fig. 10. In the finite element mesh study, the concentrated FRP option is used for modeling the
composite material. The material properties of the matrix were assigned to the element of the composite
material. Then, the fiber was modeled by using the reinforcement element function given in the finite
element program. The reinforcement was located at the middle height of the composite material for all
cases. By introducing the reinforcement function, the stiffness of the elements representing the
composite materials will be increased accordingly.
2 mm
2.5 mm
FRP as reinforcement

(a) coarse mesh model (Total number of elements = 1490 + 115 interface elements)

(b) fine mesh model (Total number of elements = 3085 + 230 interface elements )
Figure 10 Typical (a) coarse mesh and (b) fine mesh FE models

As it is stated previously, nodal forces are applied to the loading end of the FRP in order to simulate the
uniform displacement at the loading end of the FRP. Since the number of elements at the loading end of
FRP are different for coarse mesh and fine mesh models, different percentage of nodal loading should
be assigned. For the case in which fiber are considered as reinforcement and concentrated in the middle
height of the composite materials, the percentage of nodal forces assigned to the loading end of the FRP
for coarse mesh and fine mesh model are shown in Fig. 11. It is noticed that due to the higher stiffness of
the fiber, more loading was attracted to the middle height of the composite materials.

(a) coarse mesh model (b) fine mesh model
Figure 11 Percentage of nodal forces assigned to the loading end of FRP (concentrated fiber model)

3.2.1 Comparison of FE load versus deflection results with test results

Comparison of the results of the maximum end displacement of FRP () and the corresponding load (P)
obtained from the finite element analysis of the coarse mesh and fine mesh models are shown in Table
5 and the corresponding load versus displacement curves of FRP loading end is shown in Fig. 12. As it
is shown in Table 5, the maximum end displacement of FRP and the maximum load for the coarse and
fine mesh models are almost the same. However, it is shown in the figure that the load versus
displacement curves for the fine mesh and coarse mesh models are almost the same in the ascending
part but not in the region near the descending part. The results obtained from the fine mesh model are
smoother compared to the results obtained from the coarse mesh model. Compare to the average test
ultimate load (P
u, ave
), the finite element results are about 95% of P
u, ave
for both cases.


Table 5. Summary of FE results of coarse mesh and fine mesh model of CFRP specimens
Max. end displacement of FRP
() mm
Max. Load
(P) N
P / P
u, ave

Coarse mesh model 1.67 34,473 0.95
Fine mesh model 1.67 34,537 0.95

0
5000
10000
15000
20000
25000
30000
35000
40000
45000
0 0.3 0.6 0.9 1.2 1.5 1.8
Deflection in loading direction ( ) mm
L
o
a
d
i
n
g

(
P
)

N
Fine mesh
Coarse mesh
Test range of P
u
0.3P
0.9P
1.0P
Test Average of P
u

Figure 12 Comparison of the load versus deflection of CFRP specimens (coarse versus fine mesh
models)

Similar finite element analysis was carried out for the GFRP specimens using the fine mesh model. The
load versus deflection curve of the FRP loading end is shown in Fig. 13 and the results of the maximum
end displacement and maximum load of FRP is shown in Table 6 together with the results obtained from
the CFRP finite element analysis. Compare to the average test result of the failure load (P
u, ave
) of GFRP
specimens, the finite element results is about 4% higher than the average values of test results. It is also
shown from that finite element analysis that the maximum end displacement of the FRP of GFRP
specimen is larger than that of the CFRP specimen due to the lower elastic modulus of the GFRP
composite.



Table 6 Summary of FE results of fine mesh model of CFRP and GFRP specimens
Fine mesh model Max. end displacement of FRP
() mm
Max. Load
(P) N
P
FE
/ P
u, ave

CFRP 1.67 34,560 0.95
GFRP 2.77 27,843 1.04

0
5000
10000
15000
20000
25000
30000
0 0.5 1 1.5 2 2.5 3
L
o
a
d
i
n
g

(
P
)

N
Fine mesh - GFRP
Test range of P
u
0.3P
0.9P
1.0P
Test Average of P
u

Figure 13 Load versus deflection of fine mesh models of GFRP specimen

3.2.2 Comparison of FE strain results with test results
The finite element strain results of the fine mesh model were compared with the test results as shown in
Fig. 14. As it is shown in the figure, the strain results were compared in three different load levels (30%,
90% and 100% of maximum loading level as shown in Fig. 12). For the load levels under 90% of
maximum loading level, the strain results of the fine mesh model compared well with the test results. For
the maximum load levels, the strain results compared well with the test results near the free end region.
However, near the loading end, the strain results over predicted the strain values compared to the test
results of C1 to C4, but the finite element strain results are compared well with test results of C5.
Nevertheless, it is shown that the strain results obtained from the fine mesh model compared well with
the test results.


0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
Distance along bond length (mm)
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE

(a) Load level = 30% of maximum loading
0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE

(b) Load level = 90% of maximum loading

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE

(c) Load level = 100% of maximum loading
Figure 14 Comparison of strain results of CFRP fine mesh models and test results

Similar comparison of the finite element strain results to the test results was carried out for the GFRP
specimen as well and the results are shown in Fig. 15. As it is shown in the figure, the strain results were
compared in three different load levels as well (30%, 90% and 100% of maximum loading level as shown
in Fig. 13). The behaviour of the strain results of GFRP specimens is similar to the CFRP specimens. It
is also found that for the load levels under 90% of maximum loading level, the strain results of the fine
mesh model compared well with the test results. For the maximum load levels, the strain results
compared well with the test results near the free end region. However, near the loading end, the strain
results over predicted the strain values compared to the test results. Due to the lower stiffness of the
GFRP composite, the strain results were all higher than that of the CFRP specimens.


0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(
m
m
/
m
)
G1
G2
G3
G4
G5
FE

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(
m
m
/
m
)
G1
G2
G3
G4
G5
FE


0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(
m
m
/
m
)
G1
G2
G3
G4
G5
FE

Figure 15 Comparison of strain results of GFRP fine mesh models and test results

3.3 Exponential bond-slip curve versus bilinear bond-slip curve
In the previous study, the continuous exponential bond-slip curves which are proposed by Panniza et al.
(2009) for CFRP or GFRP bonded to brick element is used in the modelling of the bond-slip behaviour in
the finite element analysis. As it is mentioned in Chapter 2 that bilinear bond-slip curves are also
proposed by Panniza et al. (2009) for CFRP or GFRP bonded to brick element. In order to compare the
results of the model using the continuous exponential bond-slip curves and the bilinear bond-slip curves,
finite element analysis were carried out by using the bilinear bond-slip curves and the results are shown
in the following section.
In the finite element analysis, the fine mesh model and the same non-linear analysis procedures were
used and all the material properties of brick and FRP are the same as those used in the previous study
except the bond-slip data for the interface element was changed from exponential curve to bilinear
curve. The bond-slip behaviour of the interface element was assigned according to the data obtained
from Table 3 and Equation 6. The bilinear bond-slip curves for CFRP and GFRP specimens are shown
in Figs. 16 and 17 together with the data points used in the finite element analysis, respectively.


0
1
2
3
4
5
6
7
8
0 0.2 0.4 0.6 0.8 1
Slip, s (mm)
T
a
n
g
e
n
t
i
a
l

s
t
r
e
s
s

t

(
M
P
a
)
Bilinear
Data for FEA

Figure 16 Bilinear bond-slip curve and data points used in finite element analysis (CFRP)

0
1
2
3
4
5
6
7
8
0 0.2 0.4 0.6 0.8 1
Slip, s (mm)
T
a
n
g
e
n
t
i
a
l

s
t
r
e
s
s

(
M
P
a
)
Bilinear
Data for FEA

Figure 17 Bilinear bond-slip curve and data points used in finite element analysis (GFRP)


obtained from the finite element analysis of models with exponential bond-slip and bilinear bond-slip
curves respectively are shown in Table 7 and the corresponding load versus displacement curves of
FRP loading end is shown in Figs. 18 and 19. For the CFRP specimens, as it is shown in Table 7, the
maximum end displacement of FRP and maximum load for the model with exponential bond-slip
behaviour and the model with bilinear bond-slip behaviour are almost the same. However, the model
with bilinear bond-slip behaviour showed a slightly increase in maximum displacement and maximum
load. Compare to the average test ultimate load (P
u, ave
), the predicted of the model with exponential
bond-slip behaviour and the model with bilinear bond-slip behaviour are 95% and 96 % of P
u, ave
,
respectively. Similar behaviour of the maximum end displacement is observed for the GFRP specimens.
Compare to the average test ultimate load (P
u, ave
), the predicted of the model with exponential bond-slip
behaviour and the model with bilinear bond-slip behaviour are 104% and 106 % of P
u, ave
, respectively.
Therefore, it is shown that the results obtained in both models only showed very little different in term of
both maximum end displacement and maximum load.

Table 7. Summary of FE results of exponential and bilinear bond-slip curves of CFRP specimens
() mm
Max. Load
(P) N
P / P
u, ave

CFRP specimens
Exp. Bond-slip 1.67 34,537 0.95
Bilinear Bond-slip 1.69 34,650 0.96
GFRP specimens
Exp. Bond-slip 2.77 27,843 1.04
Bilinear Bond-slip 2.79 28,237 1.06


0
5000
10000
15000
20000
25000
30000
35000
40000
45000
0 0.3 0.6 0.9 1.2 1.5 1.8
L
o
a
d
i
n
g

(
P
)

N
Exp BS
Bilinear BS
Test range of P
u
0.3P
0.9P
1.0P
Test Average of P
u

Figure 18 Comparison of the load versus deflection of CFRP specimens (exponential versus bilinear
bond-slip behaviour)
0
5000
10000
15000
20000
25000
30000
0 0.5 1 1.5 2 2.5 3
L
o
a
d
i
n
g

(
P
)

N
Exp BS
Bilinear BS
Test range of P
u
0.3P
0.9P
1.0P
Test Average of P
u

Figure 19 Comparison of the load versus deflection of GFRP specimens (exponential versus bilinear
bond-slip behaviour)


3.3.1 Comparison of FE strain results with test results (Exponential versus
Bilinear bond-slip models)
The finite element strain results of the model with exponential bond-slip behaviour and model with
bilinear bond-slip behaviour were compared with the test results as shown in Figs. 20 and 21. As it is
shown in the figures, the strain results were compared in three different load levels (30%, 90% and
100% of maximum loading level as shown in Figs. 18 and 19). It is shown that for both CFRP and GFRP
specimens, the finite element strain results are almost the same for models with exponential and bilinear
bond-slip behaviour, respectively. For the load levels under 90% of maximum loading level, the finite
element strain results compared well with the test results. Similar to the previous studies, for the
maximum load levels, the strain results compared well with the test results near the free end region for
the CFRP specimens. However, near the loading end, the strain results over predicted the strain values
compared to the test results of C1 to C4, but the finite element strain results are still compared well with
test results of C5. Nevertheless, it seems that the strain results obtained from both models compared
well with the test results especially in the lower load levels. Similar results are also observed for the
GFRP specimens.

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Bilinear Bond-slip
Exp. Bond-slip



0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Bilinear Bond-slip
Exp. Bond-slip

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Bilinear bond-slip
Exp. Bond-slip

Figure 20 Comparison of strain results of CFRP specimens (exponential versus bilinear bond-slip)

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Bilinear Bond-slip
Exp. Bond-slip

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Bilinear Bond-slip
Exp. Bond-slip


0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Bilinear Bond-slip
Exp. Bond-slip

Figure 21 Comparison of strain results of GFRP specimens (exponential versus bilinear bond-slip)

3.4 Finite element model of FRP (Concentrated fiber versus Distributed fiber)
As it is mentioned in the beginning of this chapter, there are three options for modeling the FRP: (a) a
plate with its thickness being equal or similar to the actual thickness of the laminate including all
adhesive, but with the fibers concentrated in a thickness equal to the nominal thickness of the fiber sheet
sitting in the middle of the plate, (b) a plate with its thickness being equal or similar to the actual
thickness of the laminate including all adhesive, with the fibers assumed to be evenly distributed across
the plate thickness; and (c) a plate with a nominal thickness (generally the thickness of the fiber sheet)
without considering the adhesive (Lu et al. 2004). In general, option (a) and (b) are more suitable for
modeling the laminated FRP since both material properties of fiber and matrix are considered in the
finite element model. In the above study, option (a), in which the fiber is considered as reinforcement
and is concentrated in the middle of the composite material, was applied for the modeling of FRP. In
order to examine the different between option (a) and (b), finite element analysis of the test specimens
were carried out by following option (b) for modeling the FRP. The fine mesh model is used and both
exponential and bilinear bond-slip behaviour are considered in the following finite element study.

3.4.1 Model with exponential bond-slip behaviour (Concentrated fiber versus
Distributed fiber)
In general, all the properties (such as mesh size, element type, boundary condition, bond-slip data and
material properties) for the finite element models of FRP with distributed fiber are the same as previous
study, except the equivalent elastic modulus of the FRP are used for the composite material instead of
introducing the reinforcement element to the FRP. The equivlent elastic modulus for the CFRP and
GFRP are calculated based on Equation 7 and the values are shown in Table. 4. Since the FRP is being
modeled as homogenerous materials, in order to obtain a uniform displacement at the end of the
composite, the following nodal forces, which are not the same as those used for model with
concentrated fiber, are assigned to the end of the composite (Fig. 22).

Figure 22 Percentage of nodal forces assigned to the loading end of FRP (Distributed fiber model)

obtained from the finite element analysis of models with concentrated fiber and distributed fiber,
respectively are shown in Table 8 and the corresponding load versus displacement curves of FRP
loading end is shown in Figs. 23 and 24. For the CFRP specimens, as it is shown in Table 8, the
maximum end displacement of FRP and maximum load for both models are almost the same. However,
the model with distributed fiber property showed a slightly increase in maximum displacement for the
CFRP specimen. For the GFRP specimen, it is shown that the model with distributed fiber property
showed a slightly increase in maximum displacement and also for the maximum load. Nevertheless, the
results obtained in both models only showed very little different in term of both maximum end
displacement and maximum load.



Table 8 Specimens with exponential bond-slip behaviour (concentrated versus distributed fiber)
() mm
Max. Load
(P) N
P
FE
/ P
u, ave

CFRP specimens
Concentrated fiber 1.67 34,537 0.95
Distributed fiber 1.68 34,413 0.95
GFRP specimens

0
5000
10000
15000
20000
25000
30000
35000
40000
45000
0 0.3 0.6 0.9 1.2 1.5 1.8
L
o
a
d
i
n
g

(
P
)

N
Concentrated fiber
Distributed fiber
Test range of P
u
0.3P
0.9P
1.0P
Test Average of P
u

Figure 23 Comparison of the load versus deflection of CFRP specimens with exponential bond-slip
behaviour (concentrated fiber versus distributed fiber)

0
5000
10000
15000
20000
25000
30000
0 0.5 1 1.5 2 2.5 3
L
o
a
d
i
n
g

(
P
)

N
Concentrated fiber
Distributed fiber
Test range of P
u
0.3P
0.9P
1.0P
Test Average of P
u

Figure 24 Comparison of the load versus deflection of GFRP specimens with exponential bond-slip

Comparison of the finite element strain results of the models with exponential bond-slip behaviour
(concentrated versus distributed fiber) were made with the test results as shown in Figs. 25 and 26 for
CFRP and GFRP specimens, respectively. Similarly, the strain results were compared in three different
load levels (30%, 90% and 100% of maximum loading level) and it is shown that for both CFRP and
GFRP specimens, the finite element strain results are almost the same for models with concentrated
fiber and distributed fiber, respectively. Therefore, it is concluded that the strain results obtained from
both type of models are almost the same and they are both compared well with the test results especially
in the lower load levels. Similar results are also observed for the GFRP specimens.


0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Distributed fiber
Concentrated fiber

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Distributed fiber
Concentrated fiber


0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Distributed fiber
Concentrated fiber

Figure 25 Comparison of strain results of CFRP specimens with exponential bond-slip behaviour
(concentrated versus distributed fiber)
0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Distributed fiber
Concentrated fiber


0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Distributed fiber
Concentrated fiber

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Distributed fiber
Concentrated fiber

Figure 26 Comparison of strain results of GFRP specimens with exponential bond-slip behaviour


3.4.2 Model with bilinear bond-slip behaviour (Concentrated fiber versus
Distributed fiber)
Similar finite element analyses were also carried out for model with bilinear bond-slip behaviour.
obtained from the finite element analysis of models with concentrated fiber and distributed fiber,
respectively are shown in Table 8 and the corresponding load versus displacement curves of FRP
loading end is shown in Figs. 27 and 28. In general, the results of model with bilinear bond-slip
behaviour are in the same trend as those obtained from the model with exponential bond-slip behaviour.
For the CFRP specimens, as it is shown in Table 9, the maximum end displacement of FRP and
maximum load for both models are almost the same. However, the model with distributed fiber property
showed a slightly increase in maximum displacement and a slightly decrease in maximum load.
Nevertheless, the results obtained in both models only showed very little different in term of both
maximum end displacement and maximum load.

Table 9 Specimens with bilinear bond-slip behaviour (concentrated versus distributed fiber)
() mm
Max. Load
(P) N
P / P
u, ave

CFRP specimens
GFRP specimens


0
5000
10000
15000
20000
25000
30000
35000
40000
45000
0 0.3 0.6 0.9 1.2 1.5 1.8
L
o
a
d
i
n
g

(
P
)

N
Concentrated fiber
Distributed fiber
Test range of P
u
0.3P
0.9P
1.0P
Test Average of P
u

Figure 27 Comparison of the load versus deflection of CFRP specimens with bilinear bond-slip
0
5000
10000
15000
20000
25000
30000
0 0.5 1 1.5 2 2.5 3
L
o
a
d
i
n
g

(
P
)

N
Concentrated fiber
Distributed fiber
Test range of P
u
0.3P
0.9P
1.0P
Test Average of P
u

Figure 28 Comparison of the load versus deflection of GFRP specimens with bilinear bond-slip


The finite element strain results of the models with bilinear bond-slip behaviour (concentrated versus
distributed fiber) were compared with the test results as shown in Figs. 29 and 30 for CFRP and GFRP
specimens, respectively. Similarly, the strain results were compared in three different load levels (30%,
90% and 100% of maximum loading level) and it is shown that for both CFRP and GFRP specimens, the
finite element strain results are almost the same for models with concentrated fiber and distributed fiber,
respectively. Therefore, it is concluded that the strain results obtained from both type of models are
almost the same and they are both compared well with the test results especially in the lower load levels.
Similar results are also observed for the GFRP specimens.

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Concentrated fiber
Distributed fiber


0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Concentrated fiber
Distributed fiber

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
Concentrated fiber
Distributed fiber

Figure 29 Comparison of strain results of CFRP specimens with bilinear bond-slip behaviour

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Concentrated fiber
Distributed fiber

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Concentrated fiber
Distributed fiber


0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
Concentrated fiber
Distributed fiber

Figure 30 Comparison of strain results of GFRP specimens with bilinear bond-slip behaviour



4. COMPARISON OF FINITE ELEMENT RESUTLS WITH ANALYTICAL
SOLUTION
A full-range analytical solution is proposed by Yuan et al. (2004) for predicting the load versus deflection
of FRP bonded joint. In the full-range analytical solution, the bilinear bond-slip law which features a
linear ascending branch followed by a linear descending branch is employed. In this chapter, a brief
summary of the full-range analytical solution by Yuan et al. (2004) is presented. Then, the finite element
results which are presented in the previous chapter are compared with the prediction obtained by the
full-range analytical solution. Discussion about the comparison of the finite element results and
analytical results are presented in the following sections.
4.1 Full-range analytical solution proposed by Yuan et al. (2004)
The full-range analytical solution proposed by Yuan et al. (2004) applies to a single-lap plate bonded
joint as shown in Fig. 31. In the analytical solution, it is assumed that the width and thickness of each of
the three components (plate, adhesive layer and substrate) are constant along the length. The width and
thickness of the plate are denoted by b
p
and t
p
respectively, those of the substrate by b
s
and t
s

respectively, and the bonded length of the plate (i.e. bond length) is denoted by L. The elastic modulus
of the plate and substrate are E
p
and E
s
, respectively. In such a joint, the adhesive layer is mainly
subjected to shear deformations, so mode II interfacial fracture is the expected failure mode. A simple
mechanical model for this joint can be thus established by treating the plate and the substrate (the two
adherends) as being subject to axial deformations only while the adhesive layer can be assumed to be
subject to shear deformation only. It should be noted that in such a model, the adhesive layer represents
not only the deformation of the actual adhesive layer but also that of the materials adjacent to the
adhesive layer and is thus also referred to as the interface. Based on these assumptions, the following
fundamental equations can be found based on equilibrium considerations (Wu et al. 2002).

Figure 31 Single-lap FRP bonded joint


0
t dx
d
p
p
=

(8)

0 b t b t
s s s p p p
= +
(9)
where is the shear stress in the adhesive layer,
p
is the axial stress in the plate and
s
is the axial
stress in the substrate. The constitutive equations for the adhesive layer and the two adherends can be
expressed as

) s ( f =
(10)

dx
du
E
p
p p
=
(11)

dx
du
E
s
s s
=
(12)
The interfacial slip s is defined as the relative displacement between the two adherends, that is

s p
u u s =
(13)
After substituting Eqs. 9 to 13 into Eq. 8 and introducing the parameters of local bond strength
max
and
interfacial fracture energy G
f
yield the following equations,

0 ) s ( f
G 2
dx
s d
2
2
max
f
2
2
=

(14)
and

dx
ds
t G 2
2
p f
2
max
p
=
(15)
where

|
|
\
|
+
=
s s s
p
p p f
2
max 2
t E b
b
t E
1
G 2

(16)

Equation 14 is the governing differential equation of the bonded joint and can be solved if the local
bond-slip model relating the local interfacial shear stress to the local shear slip is defined. The interfacial
fracture energy, which is simply the area under the local bond-slip curve, is introduced because once it is
known it can be used regardless of the exact shape of the local bond-slip curve where a particular
quantity depends on the interfacial fracture energy but not on the shape of the bond-slip curve.
In the analytical solution, a bilinear bond-slip curve as shown in Fig. 32 which features a linear
ascending branch followed by a linear descending branch is used. The bond-slip model defined above,
the governing equation (Equation 14) can be solved to find the shear stress distribution along the
interface and the load-displacement response of the bonded joint. The interfacial shear stress
distribution and the propagation of debonding are shown in Fig. 33 and a typical load versus deflection
curve is shown in Fig. 34. As it is shown in Fig. 34, the load versus deflection curve is defined in four
stages: (1) elastic stage; (2) elastic-softening stage; (3) elastic-softening-debonding stage and (4)
softening-debonding stage. The corresponding equations of load versus deflection are shown in
following:

Figure 32 Local bilinear bond-slip curve

x
P
P
max
=
max
a
=
max
a = a
d
=
max
a = a
d
d a = a
u
d
=
max
a = a
u
d
<
max

Figure 33 Interfacial shear stress distribution and propagation of debonding


Figure 34 Typical full-range analytical load versus displacement curve

Elastic stage (O-A)

) L tanh(
s
b
P
1
0 1
p max
= for
0
s 0
(17)
where

|
|
\
|
+
=
s s s
p
p p 0
max
max 0
f 2 2
1
t E b
b
t E
1
s s
G 2

(18)

Elastic-softening stage (A-B)

[ ]
)
`
= ) a sin( ) a cos( ) a L ( tanh

b
P
2 2 1
1
2
2
p max
for
d
a a 0
(19)

( ) [ ]
)
`
=
0 f
f
2 2 1
1
2
0 f
s s
s
) a cos( ) a sin( ) a L ( tanh s s
for
d
a a 0
(20)
where


|
|
\
|
+
=

=
s s s
p
p p 0 f
max
max 0 f
f 2 2
2
t E b
b
t E
1
) s s ( ) s s (
G 2

(21)

|
|
\
|
=
2
1
2
d
arctan
1
a
(22)

Elastic-softening-debonding stage (B-C-D)

[ ]
)
`
= ) a sin( ) a cos( ) a d L ( tanh

b
P
d 2 d 2 d 1
1
2
2
p max

for
u
a L d 0
(23)

) d
b
P
1 ( s
p max
2
f
+ = for
u
a L d 0
(24)
where

2
u
2
a
=
(25)

Softening-debonding stage (D-E)

)) a L (
b
P
1 ( s
u
p max
2
f

+ =
(26)

4.2 Comparison of finite element results with analytical solution
The finite element results of the load versus deflection curves were compared with the analytical
solution. In the analytical solution, the FRP is considered as homogenous materials with the equivalent
elastic modulus of the composite material assigned. Therefore, the results obtained from the finite
element model with distributed fiber property were used for comparison. The material properties of brick

and FRP which were used in the finite element model with distributed fiber property are used in the
analytical solution. Although the bilinear bond-slip behaviour was used in the analytical solution, the
finite element results of both models with exponential and bilinear bond-slip behaviour, respectively,
were compared with the analytical solution. It should be noted that in the finite element model, the
deflection of the FRP is not measured directly from the end of the bonded length. It is measured from the
loading end of the bond length plus a 30 mm of unbonded length of FRP. However, the result obtained
from the analytical solution is the deflection at the end of the bonded length. Therefore, in order to
account for the deformation due to the 30 mm unbonded length, deformation of the unbonded length is
calculated based on Equation 27 and the results are added to the analytical solution.

frp frp
free , frp
free , frp
A E
PL
=
(27)
where P is the load on the FRP, L
frp, free
is the unbond length (i.e. 30 mm), E
frp
and A
frp
are the elastic
modulus and area of the FRP. According to the equations of the analytical solution, the load (P) is the
load on one FRP. Therefore, the load obtained from the equations of the analytical solution is multiplied
by 2 in order to obtain the total load for the comparison to the finite element results. Also, in the analytical
solution, the actual width of FRP and brick can be assigned. Therefore, two results of the analytical
solution, one with the width ratio equals to 50/120 and one with the width ratio equals to 1, are included
for comparison. Summary of the results obtained from the finite element analysis and the analytical
solution are shown in Table. 9 and the corresponding load versus deflection curves are shown in Figs.
35 and 36. It is shown in Table. 10 that the analytical solution compared well with the finite element
results. In general, for the maximum load, the results obtained from the analytical solution are slightly
lower than that obtained from the finite element analysis (about 3% lower for both specimens). For the
maximum deflection, the results obtained from the analytical solution are almost the same as those
obtained from the finite element analysis. Meanwhile, it is observed that the maximum loads obtained
from the analytical solution, which considered the width ratio of FRP and substrate equals to 50/120, are
slightly larger than that obtained from the analytical solution with width ratio equals to 1 (about 3% higher
for CFRP specimens and 1% higher for GFRP specimens). However, the corresponding maximum
deflections obtained from the analytical solution, which considered the width ratio equals to 50/120, are
slightly lower than that obtained from the analytical solution with width ratio equals to 1 (about 2% lower
for CFRP specimens and 1% lower for GFRP specimens).


Table 10 Comparison of finite element resutls and analytical results
Max. end disp. of FRP () mm Max. Load (P) N
CFRP specimens
FE Exp. bond-slip 1.68 34,413
FE Bilinear bond-slip 1.70 34,472
Analy. solution (width ratio = 1) 1.69 33,467
Analy. solution (width ratio = 50/120) 1.66 34,374
GFRP specimens
FE Exp. bond-slip 2.83 28,058
FE Bilinear bond-slip 2.85 28,058
Analy.solution (width ratio = 1) 2.86 27,194
Analy. solution (width ratio = 50/120) 2.83 27,547

0
5000
10000
15000
20000
25000
30000
35000
40000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Deflection (mm)
L
o
a
d

(
N
)
FE results - Exponential
FE results - Bilinear
Analytical solution
(width ratio = 1)
Analytical solution
(width ratio = 50/120)

Figure 35 Comparison of FE results and analytical solution (CFRP specimens)


0
5000
10000
15000
20000
25000
30000
0 0.5 1 1.5 2 2.5 3
Deflection (mm)
L
o
a
d

(
N
)
FE results - Exponential
FE results - Bilinear
Analytical solution
(width ratio = 1)
Analytical solution

Figure 36 Comparison of FE results and analytical solution (GFRP specimens)

4.3 Comparison of analytical solution results and finite element results with
test results
In the previous section, it is shown that the results obtained from the analytical solution compared well
with the finite element results. In Chapter 3, with the equivalent elastic modulus of CFRP and GFRP
predicted according to Equation 7, it is found that the finite element results under-predicted the
maximum load of about 5% for the CFRP specimens and over-predicted the maximum load of about 5%
for the GFRP specimens when compared to the average values of the test results. The theoretical
equivalent stiffness (E
frp
t
frp
) of CFRP and GFRP which were used in the finite element models are
43,455 N/mm and 20,260 N/mm, respectively. Although these two values are within the range of the test
results as shown in Table. 2, they are not exactly the same as the average values of the test specimens.
Therefore, finite element analysis of CFRP and GFRP specimens are carried out again with the average
test values of equivalent stiffness of the specimens (i.e. 46,260 N/mm and 18,673 N/mm, respectively)
assigned to the FRP in the finite element models. In the finite element analysis, the models with
distributed fiber property are used. Two analyzes, one with the exponential bond-slip behaviour and one
with the bilinear bond-slip behaviour, are carried out for CFRP and GFRP specimens, respectively. Also,
calculation was carried out again according to the analytical solution with the average test values of
equivalent stiffness of FRP of the specimens. Summary of the results obtained from the finite element
analysis and analytical solution for CFRP and GFRP specimens are shown in Table. 11 and the

corresponding load versus deflection curves are shown in Figs. 37 and 38. As it is shown in Table. 11
that the results obtained from both finite element analyzes and analytical solution compared very well
with the test average value of maximum load. It is also found that the results obtained from the analytical
solution gave almost the same value of the test average value of maximum load.

Table 11 Summary of results from finite element analysis, analytical solution and test (with equivalent
stiffness of FRP equals to the average value of test results)
Max. end disp.
of FRP () mm
Max. Load in
(P) N
ave , u
P
P

CFRP specimens
FE Exp. bond-slip 1.63 35,372 0.98
FE Bilinear bond-slip 1.65 35,372 0.98
Analy. solution
1.60 35,422 0.99
GFRP specimens
FE Exp. bond-slip 2.96 26,379 0.99
FE Bilinear bond-slip 2.96 26,199 0.98
Analy. solution
2.95 26,466 0.99


0
5000
10000
15000
20000
25000
30000
35000
40000
45000
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Deflection (mm)
L
o
a
d

(
N
)
Test Average, Pu
Analytical solution
FE Exponent BS
FE Bilinear BS
Test range of P
u

Figure 37 Load versus deflection results of CFRP specimens with test average FRP stiffness (E
cfrp
t
cfrp
)

0
5000
10000
15000
20000
25000
30000
0 0.5 1 1.5 2 2.5 3
Deflection (mm)
L
o
a
d

(
N
)
Test Average, Pu
Analytical solution
FE Exponential BS
FE Bilinear BS
Test range of P
u

Figure 38 Load versus deflection results of GFRP specimens with test average FRP stiffness (E
gfrp
t
gfrp
)


4.3.1 Comparison of FE strain results with test results
The strain results obtained from the finite element analysis were compared with the test results. In order
to provide a more detail comparison, the results were compared in ten different load level, (i.e. 10%,
20% up to 100% of maximum load). The corresponding results of CFRP and GFRP specimens with
exponential bond-slip and bilinear bond-slip behaviour are shown in Figs. 39 and 40, respectively. It is
shown from the figures that the finite element results of models with exponential bond-slip behaviour
were almost the same as those of the model with bilinear bond-slip behaviour. Meanwhile, the finite
element results compared well with the test results, especially when the load levels are less that 90% of
the maximum load. At the maximum load level, the finite element strain results are higher than that of the
test results, except for the results of specimen C5 which compared well with the finite element results.
Nevertheless, it is shown that the finite element results are in the same trend as the test results for both
CFRP and GFRP specimens.

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS


0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS


0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS


0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS

(d) Load level = 40% of maximum loading

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS

(e) Load level = 50% of maximum loading

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS

(f) Load level = 60% of maximum loading

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS

(g) Load level = 70% of maximum loading

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS

(h) Load level = 80% of maximum loading

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS

(i) Load level = 90% of maximum loading

0
1000
2000
3000
4000
5000
6000
7000
8000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
C1
C2
C3
C4
C5
FE Exponential BS
FE Bilinear BS

(j) Load level = 100% of maximum loading
Figure 39 Comparison of strain results of CFRP specimens (with mean value of E
cfrp
t
cfrp
assiged)

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS


0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS


0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS


0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS

(d) Load level = 40% of maximum loading

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS

(e) Load level = 50% of maximum loading

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS

(f) Load level = 60% of maximum loading

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS

(g) Load level = 70% of maximum loading

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS

(h) Load level = 80% of maximum loading

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS

(i) Load level = 90% of maximum loading

0
2500
5000
7500
10000
12500
15000
0 50 100 150 200 250
S
t
r
a
i
n

o
f

F
R
P

(

m
/
m
)
G1
G2
G3
G4
G5
FE Exponential BS
FE Bilinear BS

(j) Load level = 100% of maximum loading
Figure 40 Comparison of strain results of GFRP specimens (with mean value of E
gfrp
t
gfrp
assiged)



5. PARAMETRIC STUDY OF BOND-SLIP BEHAVIOUR OF FRP-TO-BRICK
In the previous chapter, it is shown that the results obtained from the analytical solution compared very
well with the finite element results in term of both maximum load and deflection. In order to examine the
effect of different parameters in the maximum load and deflection of FRP-to-brick joint, the analytical
solution which is proposed by Yaun et al. (2005) is applied for carrying out parametric study and the
results are discussed in this chapter.
5.1 Effect of bond length of FRP to the capacity and deflection of FRP-to-brick
joint
In the parametric study, the material properties for the brick and FRP are the same as those used in
Chapter 4, i.e. the stiffness of brick (E
brick
t
brick
) is 443,025 N/mm, the stiffness of CFRP (E
cfrp
t
cfrp
) and
GFRP (E
gfrp
t
gfrp
) is 46,260 N/mm and 18,673 N/mm, respectively. Those stiffnesses of FRP are the
average values obtained from the test. The width of FRP is 50 mm while the width of the brick is 120 mm.
The bilinear bond-slip curve and the interfacial fracture energy with the corresponding data assigned for
CFRP and GFRP specimens according to Table 3 are used in the analytical solution. With the given
material properties, geometry and bilinear bond-slip curve, the total load (which is the load applied on
the FRP multiplied by 2) versus the deflection of FRP (without considering the deflection due to the
unbonded length) are obtained for CFRP and GFRP specimens with four different bond length of FRP
(80 mm, 120 mm, 160 mm and 200 mm), respectively. Finite element analysis was also carried out to
obtain the load versus deflection curve of specimens with the mentioned four bond length. The finite
element models which were used in previous study with FRP stiffness assigned as the mean value of
the test results were used in the parametric studies. Bilinear bond-slip equations was assigned to the
interface elements according to the given bond length. Comparison of the load versus deflection curves
obtained from the finite element analysis and the analytical solution for CFRP and GFRP specimens are
shown in Figs. 41 and 42, respectively. As it is shown from the figures that, in general, the results
obtained from the analytical solution were stiffer than that obtained from the finite element analysis.
Also, the maximum deflections predicted by the analytical solution were a bit smaller than that obtained
from the finite element analysis. For the CFRP specimens, when the bond length is shorter, a higher
maximum load was obtained from the finite element analysis compared to the analytical solution. As the
bond length increases, the different between the maximum load obtained from the finite element
analysis to the analytical solution becomes less. Nevertheless, the results obtained from the analytical
solution were compared well to the finite element results in term of both maximum load and maximum
deflection and the analytical solution seems to provide conservative results.

0
5000
10000
15000
20000
25000
30000
35000
40000
0.00 0.40 0.80 1.20 1.60
Deflection (mm)
L
o
a
d

(
N
)
(Et)cfrp = 46260 N/mm
FE results
L = 80 mm
L = 120 mm
L = 160 mm
L = 200 mm
b
frp
/ b
brick
= 0.42

Figure 41 Load versus deflection curve of CFRP specimens with different bond length (Analytical
versus finite element results)
0
5000
10000
15000
20000
25000
30000
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80
Deflection (mm)
L
o
a
d

(
N
)
(Et)gfrp = 18673 N/mm
FE results
L = 80 mm
L = 120 mm
L = 160 mm
L = 200 mm
b
frp
/ b
brick
= 0.42

Figure 42 Load versus deflection curve of GFRP specimens with different bond length (Analytical
versus finite element results)

The corresponding analytical results of CFRP and GFRP specimens with different bond length are
shown in Fig. 43 (Noted that the symbol of stiffnesses of FRP is E
p
t
p
= E
frp
t
frp
and brick is E
s
t
s
= E
brick

t
brick
, respectively). As it is shown in the figure that increasing the bond length does not have significant
effect on the maximum load. It is because that the bond length of 80 mm is already longer than the
minimum bond length requirement for both specimens. The minimum required bond length can be
obtained from the following equations (Yuan et al. 2004).

) a tan(
) a tan(
ln
2
1
a L
2 2 1
2 2 1
1
e

+
+ =
(28)
where
|
|
\
|
+
=
s s s
p
p p 0
max
1
t E b
b
t E
1
s

|
|
\
|
+
=
s s s
p
p p 0 f
max
2
t E b
b
t E
1
) s s (

(
(
=
f
0 f
2
s
s s
97 . 0 arcsin
1
a

According to Equation 28, the corresponding minimum bond length for CFRP and GFRP specimens are
70.13 mm and 59.41 mm, respectively. When the bond length is less than the minimum required bond
length, the maximum load capacity of the joint can not be achieved. However, increasing the bond
length has significant effect on the maximum deflection of the joint. As it is shown in Fig. 43 that for
GFRP specimen, increasing the bond length from 80 mm to 200 mm (2.5 times) results in an increase of
maximum deflection by about 3 times. Similar behaviour is observed for the CFRP specimen which has
a higher FRP stiffness. Summary of maximum deflection versus bond length is shown in Table. 12 and
the corresponding plot is shown in Fig. 44. It is shown from the figure that the maximum deflection
increases almost linearly with respect to the bond length. However, the increasing rate is higher for
specimen with lower FRP stiffness. Therefore, in order to achieve a certain amount of maximum of
deflection, longer bond length should be applied for FRP with higher stiffness. It is also shown that
higher maximum load and smaller maximum deflection are obtained for specimen with higher FRP
stiffness. These results are consistent with the results obtained from test.



Table 12 Maximum deflection versus bond length (b
frp
/ b
brick
= 0.42)
b
frp
/ b
brick
= 0.42 Maximum Deflection (mm)
Bond length (mm) 80 120 160 200
CFRP specimen 0.44 0.75 1.06 1.38
GFRP specimen 0.84 1.41 1.97 2.54

0
5000
10000
15000
20000
25000
30000
35000
40000
0.00 0.40 0.80 1.20 1.60 2.00 2.40 2.80
Deflection (mm)
L
o
a
d

(
N
)
(Et)cfrp = 46260 N/mm
(Et)gfrp = 18673 N/mm
L = 80 mm
L = 120 mm
L = 160 mm L = 200 mm
L = 80 mm
L = 120 mm
L = 160 mm
L = 200 mm
b
frp
/ b
brick
= 0.42

Figure 43 Analytical load versus deflection curve of CFRP and GFRP specimens with different bond
length


y = 0.008x - 0.188
R
2
= 1.000
y = 0.014x - 0.291
R
2
= 1.000
0
0.5
1
1.5
2
2.5
3
50 75 100 125 150 175 200
Bond length (mm)
M
a
x
.

D
e
f
l
e
c
t
i
o
n

(
m
m
)
CFRP specimens
GFRP specimens
Linear (CFRP specimens)
Linear (GFRP specimens)
b
frp
/ b
brick
= 0.42

Figure 44 Maximum deflection versus bond length (b
frp
/ b
brick
= 0.42)

5.2 Effect of width ratio of FRP-to-brick to the capacity and deflection of
FRP-to-brick joint
In the analytical solution, the width ratio of FRP to brick is also considered. In the experimental study and
the previous analytical study, a fixed width ratio of FRP to brick of 0.42 is used. In order to examine the
effect of the width ratio of FRP-to-brick to the capacity and deflection of the FRP-to-brick joint, two more
width ratio of FRP-to-brick (i.e. b
frp
/ b
brick
= 0.042 and 1.00) were studied by applying the analytical
solution. The width ratios are achieved by keeping the width of brick as 120 mm while the width of FRP
were changed to 5 mm and 120 mm, respectively. The results of load per unit width of FRP versus the
deflection of FRP are shown in Figs. 44 and 46 for CFRP and GFRP specimens, respectively. As it is
shown from the figures that when the width ratio is close to zero (i.e. the width of the FRP is very small
compared to the width of the brick), the load capacity per unit width of FRP is the highest. When the
width ratio is closed to unity (i.e. the width of the FRP is the same as the width of the brick), the load
capacity per unit width of FRP is the lowest.


0
50
100
150
200
250
300
350
400
0.00 0.50 1.00 1.50
Deflection (mm)
L
o
a
d

p
e
r

w
i
d
t
h

(
N
/
m
m
)
L = 80 mm
L = 120 mm
L = 160 mm
L = 200 mm
E
cfrp
t
cfrp
= 46260 N/mm
b
frp
/ b
brick
= 0.042
b
frp
/ b
brick
= 0.42
b
frp
/ b
brick
= 1.00

Figure 45 Load per width of CFRP specimens with different bond length and width ratio

0
50
100
150
200
250
300
0.00 0.50 1.00 1.50 2.00 2.50
Deflection (mm)
L
o
a
d

p
e
r

w
i
d
t
h

(
N
/
m
m
)
L = 80 mm
L = 120 mm
L = 160 mm
L = 200 mm
b
frp
/ b
brick
= 0.042
b
frp
/ b
brick
= 0.42
b
frp
/ b
brick
= 1.00
E
gfrp
t
cfrp
= 18673 N/mm

Figure 46 Load per width of GFRP specimens with different bond length and width ratio


In order to show the relationship between the load capacity of the joint to both the bond length and width
ratio, a summary of the load capacity, bond length of FRP and width ratio of FRP to brick are shown in
Figs. 47 and 48 for the CFRP and GFRP specimens, respectively. In the figures, the percentage of loads
is shown in bracket for different FRP to brick width ratio. It is shown from the figures that for specimens
with higher FRP stiffness, the effect of FRP to brick width ratio to the maximum load capacity is more
pronounced. However, for the two FRP specimens in this study, the effect of width ratio to the maximum
load capacity is not significant (which is less than 5% for CFRP specimen and 2% for GFRP specimen).

340
345
350
355
360
365
60 80 100 120 140 160 180 200 220
Bond length (mm)
L
o
a
d

p
e
r

w
i
d
t
h

(
N
/
m
m
)
(100%)
(98%)
(95%)
b
frp
/ b
brick
= 0.042
b
frp
/ b
brick
= 0.42
b
frp
/ b
brick
= 1.00
E
cfrp
t
cfrp
= 46260 N/mm

Figure 47 Summary of load capacity, bond-length and width ratio of FRP of CFRP specimen


260
261
262
263
264
265
266
267
268
60 80 100 120 140 160 180 200 220
Bond length (mm)
L
o
a
d

p
e
r

w
i
d
t
h

(
N
/
m
m
)
b
frp
/ b
brick
= 0.042
(100%)
(99%)
(98%)
b
frp
/ b
brick
= 0.42
b
frp
/ b
brick
= 1.00
E
gfrp
t
cfrp
= 18673 N/mm

Figure 48 Summary of load capacity, bond-length and width ratio of FRP of GFRP specimen



6. SUMMARY AND CONCLUSION
In this report, a brief summary of the experimental work carried out by Panizza et al. (2009) about the
bond-slip behaviour of FRP-to-brick was presented. Then, finite element analyses of the test specimens
were carried out. In the finite element model, bond-slip interface elements were used according to the
exponential bond-slip curve and the bilinear bond-slip curve proposed by Panizza et al. (2009) while the
material properties for the FRP and brick were kept as elastic. Comparison of the results obtained from
different finite element models were made, they are: (1) coarse mesh versus fine mesh, (2) exponential
bond-slip versus bilinear bond-slip and (3) FRP with concentrated fibre property versus FRP with
distributed fibre property. The finite element results were also compared with the results obtained from
the analytical solution proposed by Yuan et al. (2004). As it is shown that the analytical solution provides
reasonable prediction about the maximum load and deflection, parametric studies were carried out
based on the analytical solution for different bond length and width ratio of FRP to brick.
6.1 Conclusion
Based on the results obtained from the fine mesh finite element models, it is shown that the results, in
term of maximum load, deflection and axial strain of FRP, are almost the same for models with
exponential bond-slip behaviour and models with bilinear bond-slip behaviour. Similar results were also
obtained for model with concentrated fibre property assigned to FRP and model with distributed fibre
property assigned to FRP. With the stiffness of FRP assigned as the average value of stiffness obtained
from the test (i.e. the stiffness of brick (E
brick
t
brick
) is 443,025 N/mm, the stiffness of CFRP (E
cfrp
t
cfrp
) and
GFRP (E
gfrp
t
gfrp
) is 46,260 N/mm and 18,673 N/mm, respectively), the results obtained from both finite
element methods and analytical solution compared well with the test results in term of both maximum
load and deflection. The minimum required bond length for the CFRP and GFRP specimens were
predicted based on the analytical solution. For the CFRP specimens with E
cfrp
t
cfrp
equals to 46,260
N/mm, the minimum required bond length is about 70 mm. For the GFRP specimens with E
gfrp
t
gfrp

equals to 18,673 N/mm, the minimum required bond length is about 59 mm. It is shown that longer
minimum lap length is required for specimen with higher stiffness of FRP. When the bond length is
longer than the minimum required bond length, increasing the bond length does not increase the
maximum load capacity significant, however, the maximum deflection increases with increasing bond
length. It is shown that the maximum deflection increases almost linearly with respect to the bond length.
For the current specimens studied, the width ratio of FRP to brick does not have significant affect to the
maximum load capacity and deflection, especially for specimen with lower stiffness ratio of FRP to brick.


6.2 Suggestion for further study
In current study, one layer of CFRP and GFRP were used in both experimental and numerical studies.
Therefore, the results are applicable to the FRP-to-brick specimens with the stiffness ratio of FRP to
brick same as the one used in the study. As it is shown in this study that the interfacial fracture energy,
G
f
, is different for specimens with different stiffness ratio of FRP to brick, more experimental study
should be carried out to obtained the interfacial fracture energy with different stiffness ratio of FRP to
brick. Different stiffness ratio can be achieved by either increase the number of layer of FRP or
decrease/increase the thickness of brick. Based on the results of interfacial fracture energy, calibration
of the bond-slip curve for specimens with different stiffness ratio could be carried out. Meanwhile, as it is
shown in the literatures about the bond-slip behaviour of FRP-to-concrete to the interfacial fracture
energy depends also on the compressive strength of the concrete, experimental study is suggested to
carried out for specimens of brick with different compressive strength.

REFERENCES

Aiello, M.A. and Sciolti, M.S., Bond analysis of masonry structures strengthened with CFRP sheets.
Construction and Building Materials,2005, 20: 90-100
Anzani, A., Binda, L., Cantini, L., Cardani, G., Condoleo, P., and Massetti, G. E., The Basilica of S.
Lorenzo in Cremona: Structural investigation and monitoring, The 10
th
North American Masonry
Conference, St. Louis, Missouri, June, 2007
Binda, L., Valle, S., Saisi, A, Zanzi, L. and Forde, M., Application of radar test in the survey of the load
bearing walls of the Torrazzo of Cremona, The 5
th
International Congress on Restoration of Architectural
Heritage, Firenze, September, 2000
Briccoli Bati, S., Rovero, L., Tonietti, U., Adesione frablocchi in laterizio e rinforzo in CFRP (in Italian).
Materiali ed Approcci Innovativi per il Progetto in Zona Sismica e la Mitigazione della Vulnerabilita delle
Strutture; proc. Nat. workshop ReLUIS, Salerno, Italy, 2007
Chajes, M. J., Finch, W. W., Januszka, T. F., and Thomson, T. A., Bond and force transfer of composite
material plates bonded to concrete, ACI Structures Journal, 1996, 93(2): 208-217
Dai, J. G., Sato, Y., and Ueda, T., Improving the load transfer and effective bond length for FRP
composites bonded to concrete, Proceedings of Japan concrete Institute, 2002, Vol. 24: 1423-1428
Dai, J., Ueda, T. and Sato, Y., Development of the nonlinear bond stress-slip model of fiber reinforced
plastics sheet-concrete interfaces with a simple method, Journal of Composites for Construction, ASCE,
2005, 9(1): 52-62
Diana Release 9.3, DIANA Finite Element Analysis, TNO Diana BV, 2008
Drysdale, R. G., Hamid, A. A., and Baker, L. R., Masonry structures, behaviour and design, Technical
report, Second Edition, Boulder Colorado, 1993: 888
Fernandes F., Lourenco, P. and Zanzi L., Evaluation of two novel NDT techniques: microdrilling of clay
bricks and ground penetrating radar in masonry, PhD thesis, University of Minho, Department of Civil
Engineering, Portugal, 2006
fib Bulletin 14 2001. Externally bonded FRP reinforcement for RC structures Technical Report on the
design and use of externally bonded fibre reinforced polymer reinforcement for reinforced concrete
structures, Federation Internationale du Beton.
Garrity, S. W., Retro-reinforcement a proposed repair system for masonry arch bridges, Proceddings
of the 1
st
International arch bridges conference, Bolton (UK), Thomas Telford, 1995: 557 - 566

Lorenzis, L. De., Miller, B., and Nanni, A., Bond of fiber-reinforced polymer laminates to concrete, ACI
Materials Journal, 2001, 98(3): 256-264
Lu, X. Z., Teng, J. G., Ye, L. P. and Jiang, J. J., Bond-slip models for FRP sheets/plates bonded to
concrete, Engineering Structures, 2005, 27: 920-937
Monti M, Renzelli M, Luciani P., FRP adhesion in uncracked and cracked concrete zones. In: Proc. of 6
th

international symposium on FRP reinforcement for concrete structures. Singapore: World Scientific
Publications; 2003, 183-92
Nakaba, K., Kanakubo, T., Furuta, T., and Yoshizawa, H., Bond behaviour between fiber-reinforced
polymer laminates and concrete, ACI Structures Journal, 2001, 98(3): 359-167
Neubauer U, Rostasy FS., Bond failure of concrete fiber reinforced polymer plates at inclined
cracks-experiments and fracture mechanics model. In: Proc. of 4
th
international symposium on fiber
reinforced polymer reinforcement for reinforced concrete structures, SP-188, Farmington Hills (MI): ACI;
1999, 369-82
Panizza, M, Garbin, E., Valluzzi, M. R. and Modena, C., Bond behaviour of CFRP and GFRP Laminates
on Brick Masonry, Department of Structural and Transportation Engineering, University of Padova, Italy,
2009
Sanchez, I. B., Lourenco, P., Oliverira, D. V. and Milian, A. G. A., Strengthening of arched masonry
structures with composite materials, PhD thesis, University of Minho, Department of Civil Engineering,
Portugal, 2007
Sato, Y., Asano, Y., and Ueda, T., Fundamental study on bond mechanism of carbon fiber sheet,
Concrete Library International, JSCE, 2001, No. 37: 97-115
Savioa M, Farracuti B, Mazzotti D., Non-linear bond-slip law for FRP-concrete interface. In: Proc. of 6
th

international symposium on FRP reinforcement for concrete structures. Singapore: World Scientific
Publications; 2003, 163-72
Sumon, S. K., Repair and strengthening of a damaged arch with built-in ring separation, Proceedings of
the 7
th
International conference on structural faults and repair, Edinburgh: Engineering Technics Press,
1997: 69 - 75
Vekey, R. C., Brickwork and blockwork, Construction materials, Their nature and behaviour, E & FN
Spon, ISBN 0-419-15470-1, 1998: 251-315
Wu, Z. S., Yuan, H., Niu, H., Stress transfer and fracture propagation in different kinds of adhesive joints.
Journal of Engineering Mechanics, ASCE, 2002, 128(5): 562-273
Wu, Z., and Yin, J., Numerical analysis on interfacial fracture mechanism of externally
FRP-strengthened structural members, Journal of Materials, Concrete Structures and Pavements,
JSCE, 2002, 704(55): 257-270

Yuan, H., Wu, Z., and Yoshizawa, H., Theoretical solutions on interfacial stress transfer of externally
bonded steel/composite plates, Journal of Structural Mechanics and Earthquake Engineering, JSCE,
2001, 18(1): 27-39
Yuen, H., Teng, J. G., Seracino, R., Wu, Z. S. and Yao, J., Full-range behavior of FRP-to-concrete
bonded joints, Engineering Structures, 2004, 26: 553-565

Chi Chiu Lam: Finite Element Study of Bond-Slip Behaviour of CFRP and GFRP Laminates On Brick Masonry

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chi Chiu Lam: Finite Element Study of Bond-Slip Behaviour of CFRP and GFRP Laminates On Brick Masonry

Uploaded by

Copyright:

Available Formats

C

= ) a sin( ) a cos( ) a L ( tanh

= ) a sin( ) a cos( ) a d L ( tanh

You might also like