1

INTERLAMINAR PEELING IN CONCRETE BEAMS

REPAIRED WITH CFRP

Pedro Alexandre Simões Martins Natário

MSc dissertation in Civil Engineering

Civil Engineering and Architecture Department, Instuto Superior Técnico, Technical University of Lisbon

Department of Civil Engineering, Technical University of Denmark

October, 2007

ABSTRACT

The strengthening of beams by external bonding of fibre reinforced polymers to the tensile surface has

proven to be a successful technique. However, the existing design guidelines are still insufficient and

incomplete, in regard to debonding failure modes.

This work presents an experimental program carried out on full scale reinforced concrete beams,

strengthened by external bonding of Carbon Fibre Reinforced Polymer (CFRP) laminates to their soffit.

More specifically, the aim of this study consisted on the observation of crack formation and debonding

initiation and propagation, by means of utilizing a digital optical measurement system to document the

strain field near the midspan region. In order to investigate the influence of cracking prior to

strengthening, two specimens were damaged before the laminates were bonded, while two other

specimens were strengthened without any previous damage.

The beam specimens failed by concrete cover separation. The results indicate that strengthening

successfuly increases the load capacity at the cost of ductility. Furthermore, the load capacity

predictions, based on the existing guidelines, proved to be inaccurate, seen as the considered failure

mode is erroneous and the ultimate load was not reached.

Key Words: Carbon fibre reinforcement polymer; Strengthening; Debonding; Digital optical

measurement systems

2

1. INTRODUCTION

Over the past few decades, the civil engineering community has shown an increasing attention

towards strengthening, repairing and retrofitting of existing structures. The overall ageing of the

existing structures, accidents with natural or human cause, exposition to extreme environmental

conditions and the misregard for maintenance have created the need for structural rehabilitation.

Other structures may require retrofitting due to increasing load demands. Furtherrmore, errors during

project or construction may cause structural problems, requiring a swift intervention.

Even though several alternatives exist, external plate bonding has proved to be a successful

intervention, while presenting several advatanges, such as being less intrusive and costly. The first

field applications utilized steel laminates bonded or bolted to the structural member, which eventually

were substituted by fibre reinforcement polymers due to their high corrosion resistance, high strength

and low weight, which eases handling and transport.

Several failure modes have been documented, when studying the behaviour of reinforced concrete

beams strengthened by external plate bonding. In this context, debonding failure modes have been

investigated as they prevent the member from achieving its ultimate capacity. These failure modes

involve loss of composite action between the laminate and the concrete and have been classified in

two main types: plate end debonding and intermediate crack debonding.

The first debonding type is defined by a failure occuring near the laminate cut-off edge, due to high

normal and shear stress concentration in this region. Once the concrete presents normal stresses due

to bending deformation, the adhesive needs to bring the stresses in the laminate to levels similar to

those on the concrete. This is achieved through shear stresses in the adhesive, which will then

propagate to the weakened concrete cover, which will fail once the concrete strength is exceeded.

The observed end peeling modes are either concrete cover separation, where the concrete fails near

the embedded steel reinforcement, and interfacial debonding, when the failure occurs in a thin

concrete layer near the adhesive.

As for the intermediate crack debonding, it occurs in the midspan region, when a critical flexural or

shear-flexural crack causes a vertical displacement of the concrete cover and adhesive, resulting in a

horizontal crack that propagates towards the laminate end. This failure was investigated by Sebastian

(2001), who utilized a crack inducer to observe initiation and propagation of the debonding cracks.

There exists an extensive experimental background on full scale specimens, which has been used to

develop analytical and numerical models focusing at the calculation of interfacial stresses and bond

strength. These would prove of valuable interest when creating a complete guideline, regarding the

strengthening of stuctural members in bending with external plate bonding. The first experimental

works aimed at the observation of debonding failures modes. Arduini and Nanni (1997) tried to

establish the differences between strengthening of virgin beams and precracked beams. Other

experimental works aimed at finding the parameters involved in these failure modes, such as the

works by Tumialan et al. (1999), Fanning and Kelly (2001), Rahimi and Hutchinson (2001).

3

Lately, the attention has shifted to identifying the parameters responsible for the initiation and

propagation of debonding cracks. This has been investigated by authors, such as Wu and Yin (2003).

Coronado and Lopez (2006) and Niu et al. (2006) agree that the fracture energy on the concrete-

adhesive interface is an important conditioning parameter.

Other authors have tried to utilize experimental background to develop strength models intended for

design purposes. Teng and Yao (2005) developed predictive strength models, which were validated

through comparison with available test results. Casas and Pascual (2007) presented an equation for

the computation of the ultimate force per width unit, in order to prevent debonding failure, for members

in bending.

The mechanisms behind debonding have also been subject of extensive investigation. It has been

concluded that the interfacial strength between concrete and laminate plays a major role in debonding.

Wang (2007) classified the bond strength models according to the considered approach: strength of

materials method; linear elastic fracture mechanisms and nonlinear fracture mechanisms.

The first method is rather simple and provided important insight in the discussed failure modes. It was

used by Roberts (1989), Roberts and Haji-Kazemi (1989), Malek et al. (1998) and Smith and Teng

(2001), as referenced by Wang (2007), to obtain interfacial stress solutions.

The linear elastic theory was used by Täljsten (1997) to obtain the critical stress levels at the plate

short edge, when bonded to the tensile surface of reinforced concrete beams subjected to a point

load. Rabinovich and Frostig (2000) used an high order analysis, which allowed the consideration of

non-null shear stresses at the end of the adhesive layer, although this solution was later criticized by

Shen et al. (2001). Other models such as the one presented by Ascione and Feo (2000) aimed at the

prediction of both shear and normal stresses at the laminate cut-off edge.

On a different approach, numerical models have been created, based on linear elastic fracture

mechanics, such as the one by Yang et al. (2003). Yang and Ye (2005) presented simplified solutions,

suitable for computational use, based on a rigorous elastic formulation for the interfacial stresses.

Non-linear fracture mechanics has been used by an increasing number of authors to describe the

stress-slip relations between the concrete-laminate interface, such as Ahmad et al. (2006). These

relations have received experimental support by means of simple shear pulloff tests, which have been

used to model the intermediate crack debonding initiation (mode II). An extensive review on these

tests and slip-bond relations is presented by Lu et al. (2005). The latter approach has also been

considered in several numerical solutions. Ferracuti et al. (2006) or Niu and Wu (2006) display two

different finite element models, based on non-linear fracture mechanics.

Further studies have concluded that the stress-slip relations depend on the shear stress responsible

for debonding initiation, the residual stress right after debonding occurs and a parameter that takes

into consideration the shear stress reduction with sliding (Leung and Tung, 2006), usually referenced

to as softening due to an interlocking effect in the crack (Pan and Leung, 2007).

Several studies have pointed out that intermediate crack debonding might be due to a critical diagonal

crack, rather than a vertical flexural crack. In this field, Karbhari and Engineer (1996) developed an

interesting test, where the peeling angle from the concrete was varied.

4

The latter afirmation has been confirmed by several studies, such as Chen and Teng (2001) and

Camata et al. (2004). As a result, the simple shear tests have been criticized and several other

experimental set-ups have been proposed. Lee (2003), referenced by Camata et al. (2004), presented

a shear/normal bond test, while Pan and Leung (2007) utilized an experimental set-up that induced

vertical displacements in the laminate.

The influence of multiple cracks has also been subject of investigation and several authors have

explicitely stressed the influence of this parameter on their models. Teng et al. (2003) highlighted that

multiple cracks are one of the main differences between intermediate crack debonding and the simple

shear tests. Chen et al. (2007) specifically studied the effect of multiple cracks by means of a modified

shear test, where the laminate is pulled from both sides, thus representing two consecutive cracks.

Investigation has also been carried out in the field of external anchorage systems, aimed at preventing

debonding failure modes. Several authors have concluded that U-wrappings near the laminate end are

rather innefective to prevent midspan debonding (Leung, 2006; Pham and Al-Mahaidi, 2006). The

correct placing of sheet wraps was investigated by some authors, such as Pimanmas and

Pornpongsaroj (2004), Pham and Al-Mahaidi (2006) and Xiong et al. (2007).

Even though several research programs have been conducted, the design guidelines are still

incomplete in terms of prediction and prevention of debonding failure modes. Furthermore, it has been

evidenced that these failure mechanisms occur before the specimens reach their ultimate load

capacity. This experimental work aims at a better understanding of the fracture mechanisms

responsible for midspan debonding. In order to accomplish this, a digital optical measurement system

was used to obtain detailed images of crack formation and debonding initiation and propagation near

midspan. The specimen damage level was varied by means of testing precracked and repaired

specimens and beams strengthened without any previous damage.

2. EXPERIMENTAL WORK

2.1. SPECIMENS

The experimental tests were conducted on five reinforced concrete beam specimens. The first beam

was tested without any external strengthening system and was utilized as reference for the remaining

results. Two beams were precracked by means of applying a load of 10 kN in the same set-up utilized

to test them after they were repaired. The remaining two specimens were strengthened and tested

without any kind of previous damage.

The beams were 120 mm wide, 170 mm high and 2000 mm long. The internal reinforcement consisted

of two ø10 steel reinforcement bars, both in the tensile and in the compressive zone. The concrete

cover was 30 mm on the beam under side and 25 mm on the upper side.

The shear reinforcement is as seen in Figure 1 and is based in ø6 steel stirrups. Their spacing was

prescribed in order to prevent traditional shear failure.

5

Figure 1– Geometry and static system of a strengthened beam

In regard to the external reinforcement, both CFRP sheets and laminates were utilized. A 300 mm

wide sheet was applied from midspan towards the right side of the specimen. It was 300 mm wide and

it was placed to prevent failure from occurring in that side of the beam.

The specimens were strengthened in bending by means of bonding two 5 mm wide laminates to their

under side. This solution was thought to reduce the ammount of external reinforcement, which eased

the application. Furthermore, composite, adhesive and concrete were needed in the same plane.

The strenghtening procedure consisted of a series of steps that ensured the correct application.

Firstly, the bond surfaces were sandblasted, leaving the coarse aggregate exposed. These surfaces

were then carefully cleaned and a primer was applied. Finally, the adhesive was placed with a

thickness of approximately 1 mm.

2.2. TEST SET-UP

The same set-up was utilized for all the specimens. The beams were subjected to a 3-point bending

test and the load was applied at midspan with displacement control at a rate of 0,7 mm/min.

The testing equipment was a Instron universal test machine with a 5 MN capacity.

2.3. INSTRUMENTATION

2.3.1. DEFLECTION AND SUPPORT SETTLEMENT MEASUREMENT

Four Linear Variable Differential Transformers (LVDTs) were installed. Their location was conditioned

by the testing equipment and, as such, two LVDTs were placed 250 mm away from midspan for

practical reasons. Two other LVDTs were placed over the supports to evaluate the possible

settlements.

6

2.3.2. STRAIN MEASUREMENT

Strain gauges were installed in both the tensile and the compressive steel reinforcement in two

prescribed sections, as seen in Figure 2. Being the laminate behaviour one of the key aspects to be

observed during the tests, ten strain gauges were placed from midspan, spaced 40 mm apart. Another

gauge was placed the laminate, aligned with section II.

section II

40

section I

1800

beam's underside plane view

[mm]

50

5020

beam's lateral view250 250 50

LVDT LVDT LVDT

50

LVDT

Figure 2 – Location of LVDTs and strain gauges in CFRP laminate and steel reinforcement

2.3.3. OPTICAL MEASUREMENT EQUIPMENT

The utilization of digital optical measurement systems in the study of crack formation and debonding

initiation and propagation is yet to be reported in full scale tests. Ahmad et al. (2006) utilized a similar

system to observe the strain field on a shear test. Pease et al. (2006) used photogrammetry to

observe cracking on concrete and concrete-steel interfaces. It is our belief that this equipment

presents numerous advantages when observing crack phenomena and documenting the strain field.

The equipment was supplied by GOM mbH and consists on two digital cameras, which capture photos

of the surface on predefined time intervals. They are stored in a computer for later analysis.

The equipment was placed from midspan towards the left. The measurement area required surface

preparation. An adequate contrast on the greyscale had to be obtained to ensure the correct system

performance. The area was smoothened and sprayed with white and black paint in a spatter pattern.

During the tests, the measurement surface was illuminated with two regular white light sources, in

order to achieve constant light intensity.

(a) (b)

Figure 3 – (a) Cameras, measurement surface and load cell (b) Experimental set-up

7

2.4. MATERIAL PROPERTIES

2.4.1. REINFORCING STEEL

Accordingly to the European Standard prEN 1992-1-1 (Eurocode 2) and European Standard EN

10080, simple tensile strength tests were performed on six ø10 samples and four ø6 samples. The

considered tensile strength and yield strain were the mean value of the results obtained. As such, the

longitudinal rebars were considered to have a mean yield strength, ymf , of 676 MPa and a yield strain.

ymε , of 3,7‰. The values for the stirrup steel were 605 MPa and 3,4‰, respectively. The elastic

modulus was taken as 200 GPa, as prescribed by Eurocode 2.

2.4.2. CONCRETE

The concrete properties were determined by means of testing twelve cylinder samples with diamater

100 mm and height 200 mm, in accordance with the European Standard EN 12390-1. Seen as the

concrete beams and the cylinders were cast in December 2006 and the experimental program and

cylinder tests were performed at different moments during 2007, the expressions presented in

Eurocode 2 were considered to calculate the concrete properties in the different stages of the

experimental work. The results are those presented in Table 1. The concrete ultimate strain was also

evaluated and taken as the mean value of the experimental results. A value of 3,9‰ was considered.

Table 1 – Concrete mechanical properties at different stages of the experimental work

Concrete

t (days) Comment fcm

[MPa] fck

[MPa] βcc[ ]

Ecm [GPa]

fctm [MPa]

28 C40/50 (Eurocode 2 - Table 3.1) 48,0 40 1,000 35,0 3,5

150 reference beam; precracking 53,8 40 1,120 36,4 3,8

180 repaired and strengthened beams 54,2 40 1,129 36,5 3,8

210 cylinder tests 54,5 40 1,135 36,6 3,8

2.4.3. EPOXY ADHESIVE AND CFRP PLATES

The adhesive and plates were supplied by Sto Scandinavia AB. The epoxy adhesive utilized was the

StoPox Sk 41, and the laminate was the StoFRP E 50 C. The mechanical properties were provided by

the manufacturer and are resumed in Table 2.

Table 2 – Adhesive and CFRP properties

Adhesive Pox SK 41 CFRP Plate E 50 C

Compressive Strength [MPa] 100 Tensile Strength [MPa] 2000

Flexural Strength [MPa] 30 Young’s Modulus [GPa] 155

Young’s Modulus [GPa] 11 Thickness [mm] 1.4

Shear Modulus [GPa] 4,23 Failure Strain [‰] ~15

8

3. ANALYTICAL WORK

The analytical calculations were performed to provide a theoretical background for comparison with

the experimental results. In order to obtain results comparable with the experimental data, current

design guidelines were utilized with the actual material properties, instead of design values.

First of all, the cracking load and crack opening was evaluated. The theoretical load responsible for

crack initiation was slight over 4,5 kN, while the crack opening due to the precracking stage on beams

2 and 3 resulted in a value of 0,14 mm.

Further on, the load bearing capacity of the specimens was considered. The design guidelines

indicate that the strain field should be documented and inserted in the strengthening calculations.

However, this was neglected both due to the fact that the strains due to self weight were insignificant

when compared with the deformations due to the point load. Furthermore, the laminates were bonded

with the under side facing upwards, which means that the strain field was opposite to that caused by

the point load. It was considered that precracking had no influence, in terms of the calculations.

In order to obtain the section resistant moment, four failure modes were considered: specimen failure

by tensile failure of the laminate with (mode 1) our without (3) compressive steel reinforcement

yielding and concrete crushing with (2) or without (4) compressive steel reinforcement yielding. It was

considered that the tensile reinforcement yielded in any of the latter failure modes. The assumptions

were then compared with the sectional strain distribution. Table 3 resumes the calculations. It is

observed that the only strain field compatible with the yielding and failure assumptions is mode 4. This

is indeed as expected, as the composite reinforcement was exaggerated. The ultimate moments gives

a load capacity of approximately 103 kN.

Table 3 – Ultimate moment and strain distribution for different failure modes

Beam specimens

Failure Mode Mode 1 Mode 2 Mode 3 Mode 4

x [m] 0,0538 0,0451 0,0548 0,0438

Mr [MPa] 52722,45 46792,65 52656,21 49205,59

εc [‰] 6,95 3,90 7,13 3,90

εsc [‰] 3,07 1,30 3,22 1,23

εst [‰] 10,48 7,78 10,44 8,13

εf [‰] 15,00 10,81 15,00 11,25

Following, a bond strength model was applied to our case study to obtain a reduced laminate strength,

which explicitely takes debonding into consideration. The model proposed by Chen and Teng (2003)

was considered. The latter authors propose a solution for the computation of the bond strenght, i.e.

the stress responsible for debonding, occuring in the concrete-laminate interface. The obtained value

was 345,7 MPa, which was used to replace the ultimate composite strength.

When considered the reduced laminate strength, it is obvious that the differences occur in the failure

modes where the laminate failure is assumed (Table 4). As such, the new failure strain was calculated

and the failure mode analyzed. Theoretically, the laminate fails without compressive reinforcement

yielding (mode 3). The load capacity is thus reduced to approximately 28,5 kN.

9

Table 4 – Ultimate moment and strain distribution considering a reduced laminate strength

Beam specimens without precracking

Failure Mode Mode 2 Mode 3 Mode 4

x [m] 0,0451 0,0190 0,0438

Mr [MPa] 46792,65 13772,72 49205,59

εc [‰] 3,90 0,24 3,90

εsc [‰] 1,30 -0,14 1,23

εst [‰] 7,78 1,48 8,13

εf [‰] 10,81 1,92 11,25

Finally, the shear stress distribution in the adhesive layer was investigated. In order to accomplish this,

the model proposed by Smith and Teng (2001) was considered. It was assumed that the stresses are

constant across the adhesive thickness. The model is based on a linear elastic behaviour between the

concrete and the laminate. The theoretical results show that the shear stresses have a peak near the

plate’s cut-off edge and tend to decrease rapidly as the distance from the cut-off edge increases.

(a) (b)

Figure 4 – Shear stress distribution in the adhesive layer due to: (a) self weight and (b) point load

4. RESULTS AND DISCUSSION

Firstly, the load capacity of the tested specimens is presented. As seen in Figure 5, the strengthened

specimens improve their load carrying capacity to about 250% of that of the reference beam.

Furthermore, it is seen that strengthening increases the overall member rigidity significantly. The latter

is documented by the rigidity coefficient calculation presented in Table 5. It is seen that this parameter

reaches an 80% to 90% increase with strengthening.

When compared with the theoretical load carrying capacity prediction, it is seen that the value is far

from being achieved (103 kN). This supports the statement that the design guidelines are not

sufficiently complete, seen as they neglect debonding failure modes, which effectively prevent the

strengthened members from achieving their ultimate capacity. The utilized materials are thus not used

to their maximum capacity.

10

Some authors, such as Chen and Teng (2003) have presented bond strength models that may be

used to calculate a reduced laminate strength, taking into account the stress level that initiates

intermediate debonding. The model presented by the latter authors was utilized to the particular case

studied in this work, and a load capacity of 28,5 kN was achieved, as presented in Figure 5.

The observed failure mode was concrete cover separation, which started in laminate’s edge in the

side of the beam where a CFRP sheet was wrapped. The confined concrete in the wrapped section

increased the rigidity of this region, thus inducing the observed failure mode.

Table 5 – Failure mode, ultimate load capacity and rigidity coefficient of the tested specimens

Specimen Beam 1 Beam 2 Beam 3 Beam 4 Beam 5

Failure Mode Steel yield Concrete cover

separation

Concrete cover

separation

Concrete cover

separation

Concrete cover

separation

Ultimate load [kN] 27,6 71,4 70,5 73,1 72,9

Rigidity Coefficient

[kN/mm] 2,7 5,0 5,2 4,9 4,8

0 5 10 15 20 25 30Deflection [mm]

0

10

20

30

40

50

60

70

80

Load [kN]

Reference

Beam 2 (rep.)

Beam 3 (rep.)

Beam 4 (str.)

Beam 5 (str.)

Figure 5 – Evaluation of the load capacity and deflection

Section I Section II

0 2000 4000 6000 8000

Strain [µm/m]

0

10

20

30

40

50

60

70

80

Load [kN]

Ref.

B.2 (rep.)

B.3 (rep.)

B.4 (str.)

B.5 (str.)

0 1000 2000 3000 4000

Strain [µm/m]

0

10

20

30

40

50

60

70

80

Load [kN]

Ref.

B.2 (rep.)

B.3 (rep.)

B.4 (str.)

B.5 (str.)

Figure 6 - Evaluation of the strains in the tensile reinforcement in both control sections

11

The strains in the steel reinforcement were also evaluated and are presented in Figure 6. It is seen

that the external CFRP laminate bonding drastically decreases the ductility of the strengthened

member. Furthermore, it is seen that the repaired beams failed in a purely brittle manner, seen as

tensile steel yielding was not achieved. As for the beams strengthened without previous damage, steel

yielding is reached shortly before the specimens failed.

The optical measurement system proved to be crucial in terms of observation of crack and debonding

development. Debonding will most likely occur at the bottom of a shear-flexural diagonal crack. As

such, the cracks closer to midspan, which are sub-vertical (flexural cracks), are less likely to initiate

debonding. Figure 7 presents the strain field in the measurement area of the repaired Beam 2

specimen. It is seen that, in this particular case, the sub-vertical crack closer to midspan does initiate

debonding.

Figure 7 – (a) Strain field in Beam 2 when the specimen failed; (b) Crack and debonding width

evolution

Figure 8 presents the final stage of the strain field documented with Aramis. It is seen that debonding

occurs in all the specimens, regardless of their previous damage condition. Furthermore, it may be

observed that debonding more likely occurs from shear-flexural diagonal cracks than from flexural

cracks.

(b)

12

Beam 2 Beam 3 Beam 4 Beam 5

Figure 8 – Strain field of the measurement surface of each specimen at failure

The following analysis refers to Beam 4. As seen in Figure 9, the first cracks are rather vertical and are

thus flexural. At later stages, small diagonal cracks start, close to the first ones, and propagate until

they intersect. The new diagonal crack will be crucial for debonding initiation. Another debonding crack

also starts at the bottom of a flexural crack, but their dimension is inferior. It is seen that the flexural

crack closer to midspan does not seem to initiate debonding.

Figure 9 - Crack and debonding evolution in the measurement surface (Beam 4)

It would have been interesting to observe failure in the measurement surface, in order to provide

further insight on the possibility that debonding might occur in a specific zone. All the specimens

present a very important diagonal crack, which assumes a large dimension at the later load stages.

This is clearly identified in Beam 4 (Figure 10), where debonding caused disruption of the optical

measurements.

13

Figure 10 – Strain field distribution in Beam 4 measurement surface, when the specimen failed

Another important observation consists on the fact that debonding is actually the horizontal

development of the main flexural or shear-flexural crack, along the concrete-laminate interface. In fact,

adhesive shear stresses have little influence on debonding initiation, seen as they assume

insignificant values as we move towards midspan. However, as evidenced in Figure 4, the risk of

debonding occuring near the laminate’s cut-off edge is significant.

In order to evaluate the CFRP strain distribution, strain gauges were installed in the laminates. The

obtained results are rather discontinuous, which is probably due to cracking phenomena in the

concrete cover. As such, the results do not appear to be accurate enough to calculate reliable values

for the adhesive shear stresses.

-400 -300 -200 -100 0Distance [mm]

3600

4000

4400

4800

5200

5600

6000

Strain [µm/m

]

B.2 (rep.)

B.3 (rep.)

B.4 (str.)

B.5 (str.)

70 kN

Figure 11 - Evaluation of the CFRP strain distribution at 70 kN

14

5. CONCLUSIONS

During the development of this experimental work, a number of important conclusions were drawn.

They are presented as follows.

1. The mechanical behaviour is successfuly improved by means of externally CFRP laminate

bonding. Furthermore, it is seen that there is little difference between the precracked

specimens and those strengthened prior to any previous damage.

2. The load carrying capacity of the structural members is greatly improved with this

strengthening technique. The experimental results showed a load capacity increase of over

150%. The rigidity of the specimens was also increased.

3. In regard to ductility, it is seen that the strengthened members see this parameter drastically

reduced. In fact, the repaired beams failed before the tensile steel reinforcement achieved

yielding. The remaining strengthened beams failed shortly after yielding. Debonding consists

of a brittle failure mode, which prevents the strengthened member from achieving its ultimate

capacity

4. The existing design guidelines are not yet sufficiently complete, as debonding failure is not

considered. The classical failure modes used to calculate the ultimate capacity lead to values

that largely surpass the experimental observations. The utilization of a reduced laminate

strength, such as the one obtained through the model presented by Chen and Teng (2003),

leads to very conservative results. However, this presents a viable option, provided that a

more accurrate model is considered.

5. The digital optical measurement system provided important insight on the crack and

debonding initiation and propagation. The observations evidenced that debonding will most

likely occur from a diagonal shear-flexural crack, than from a sub-vertical flexural crack.

6. Midspan debonding is a failure mechanism that depends on the crack development near

midspan. The shear stress concentration in the adhesive near midspan is insignificant and will

not initiate debonding.

REFERENCES

[1] Ali-Ahmad M, Subramaniam K, Ghosn M. Experimental Investigation and fracture analysis of

debonding between concrete and FRP sheets.Journal of Engineering Mechanics, ASCE,

2006;132 (9): 914-923.

[2] Arduini, M., Nanni, A.. Behavior of precracked RC beams strengthened with carbon FRP sheets.

Journal of Composites for Construction, ASCE, 1997;1 (2): 63–70.

[3] Ascione L, Feo L. Modeling of composite/concrete interface of RC beams strengthened with

composite laminates. Composites. Part B: engineering [online] 2000;31:535-540. [Available at

www.elsevier.com/locate/compositesb, accessed 22 June 2007].

15

[4] Camata G, Spacone E, Al-Mahaidi R et al. Analysis of test specimens for cohesive near-bond

failure of fiber-reinforced polymer-plated concrete. Journal of Composites for Construction, ASCE,

2004 Nov/Dec: 528-538.

[5] Casas JR, Pascual J. Debonding of FRP in bending: Simplified model and experimental

validation. Construction and Building Materials [online] 2007; 21:1940-1949. [cited 10 Aug 2007,

Available at www.elsevier.com/locate/conbuildmat, accessed 22 June 2007].

[6] Chen JF, Teng J.G. Anchorage strength models for FRP and steel plates bonded to

concrete.Journal of Structural Engineering. ASCE, 2001;127(7):784-791.

[7] Chen JF, Teng JG Shear capacity of FRP-srtrengthened RC beams: FRP debonding.

Construction and Building Materials; 2003;17:27-41. [Available at

www.elsevier.com/locate/conbuildmat, accessed 23 June 2007].

[8] Chen JF, Yuan H, Teng JG. Debonding failure along a softning FRP-to-concrete interface

between two adjacent cracks in concrete members. Engineering Structures [online]; 2007;29:259-

270. [ Available at www.elsevier.com/locate/engstruct, accessed 3 July 2007].

[9] Coronado AC, Lopez AM. Sensitivity analysis of reinforced concrete beams strenghetened with

FRP laminates. Cement and Concrete Composites, Elsevier, 2006; 28(1):102-114.

[10] European Committee for Standardization , Ensaios do betão endurecido – Parte 1: Forma,

dimensões e outros requisitos para o ensaio de provetes e para os moldes, Ref. No. NP EN

12390-1:2004

[11] European Committee for Standardization, Ensaios do betão endurecido. Parte 3: Resistência à

compressão dos provetes de ensaio, Ref. No. NP EN 12390-3:2003

[12] European Committee for Standardization, Eurocode 2: Design of concrete structures - Part 1-1:

General rules and rules for buildings, Ref. No. prEN 1992-1-1:2003 E

[13] European Committee for Standardization, Materiais metálicos. Ensaio de tracção – Parte 1:

Método de ensaio à temperatura ambiente, Ref. No. NP EN 10002-1:2006

[14] Fanning PJ, Kelly O. Ultimate response of RC beams strengthened with CFRP plates. Journal of

Composites for Construction, ASCE, May 2001; 5(2): 122-127.

[15] Ferracuti b, Savoia M, Mazzotti c. A numerical model for FRP-concrete delamination. Composites.

Part B:engineering 2006;37:356-364. [Available at www.elsevier.com/locate/compositesb ,

accessed 2 July 2007].

[16] Karbhari VM, Engineer M. Investigation of bond between concrete and composites: use of a peel test.

Journal of Reinforced Plastics and Composites 1996;15:208-227. [Available at

http://jrp.sagepub.com/cgi/content/abstract/15/2/208, accessed 1 July 2007].

[17] Leung CKY, FRP debonding from a concrete substracte: some recente findings against

conventional belief. Cement & Concrete Composites 2006; 28: 742-748.

[18] Leung CKY, Tung WK. Three-parameter model for debonding of FRP plate from concrete

substracte.[online]), Journal of Engineering Mechanics, ASCE, 2006 ; 132(5):509-518.

[19] Lu XZ, Teng JG, Jiang JJ. Bond-slip models for FRP sheets/plates bonded to concrete.

Engineering Structures [online]; 2005;27:920-937. [Available at

www.elsevier.com/locate/engstruct, accessed 2 September]

[20] Malek AM, Saadatmanesh H, Ehsani MR. Prediction of failure load of R/C beams strengthened

with FRP plate due to stress concentration at the plate end. ACI Struct J 1998;95(1):142–52.

16

[21] Niu H, Wu Z. Effects of FRP-concrete interface bond properties on the performance of RC beams

strengthened in flexure with externally bonded FRP sheets. Journal of Materials in Civil

Engineering. ASCE., September/October 2006: 723-731.

[22] Pan J, Leung CKY. Debonding along the FRP-concrete interface under combined pulling/peeling

effects. Engineering Fracture Mechanisms, 2007; 74: 132-150. [Available at:

www.elsevier.com/locate/engfracmech, accessed 22 June 2007].

[23] Pease BJ, Geiker MR, Stang H, Weii WJ. Photogrammetric assessment of flexure induced

cracking in reinforced concrete beams under service loads. Department of Civil Engineering,

Tecnical University of Denmark, presented for publication in RILEM Journal: Materials ans

Structures, revised edition 19 Jan 2006: p.11

[24] Pham HB, Al-Mahaidi R. Prediction models for debonding failure loads of carbon fiber reinforced

polymer retrofitted reinforced concrete beams, ASCE, Journal of Composites for Construction,

2006; 10(1): 248-59.

[25] Pinanmas A, Pornpongsaroj P. Peeling behaviour of reinforced concrete beams strengthened with

CFRP plates under various end restraint conditions. Magazine of Concrete Research; 2004;

56(2):73-81.

[26] Rabinovich O, Frostig Y. Closed-form high-order analysis of RC beams strengthened with FRP

strips. Journal of Composites for Construction.ASCE, May 2000: 65-74.

[27] Rahimi H, Hutchinson A. Concrete beams strengthened with externally bonded FRP

plates.[online] ASCE, Journal of Composites for Construction; 2001; 5(1):44-56.

[28] Roberts TM, Haji-Kazemi H. Theoretical study of the behaviour of reinforced concrete beams

strengthened by externally bonded steel plates. Proceedings of the Institution of Civil Engineers

1989;87(Part 2):39–55.

[29] Roberts TM. Approximate analysis of shear and normal stress concentrations in the adhesive

layer of plated RC beams. The Structural Engineer 1989;67(12):229–33.

[30] Rusinowski P, Täljsten B. Interlaminar peeling in concrete beams strengthened with CFRP plates

– experimental study. Asia Pacific Conference on FRP in Structures (APFIS 2007), Internation

Institute for FRP in Construction, 2007

[31] Sebastian WM. Significance of midspan debonding failure in FRP-plated concrete beams. Journal

of Structural Engineering,EBSCO, July 2001:792-798.

[32] Shen H-S, Teng JG, Yang J. Interfacial stresses in beams and slabs bonded with thin plate.

Journal of Engineering Mechanics, ASCE, 2001;127(4): 399-406.

[33] Smith ST, Teng JG. Interfacial stresses in plated beams. Engineering structures [online] 2001,23:

857-871.[ Available at www.elsevier.com/locate/engstruct, accessed 3 September 2007]

[34] Täljsten B. FRP strengthening of existing concrete structures. Design guideline. Sweden: Lulea

University of Technology. Division of Structural Engineering. Department of Civil Engineering;

2004. ISBN 91-89580-03-6

[35] Täljsten, B. Strengthening of beams by plate bonding, Journal of Materials in Civil Engineering,

1997; 9(4): 206-212.

[36] Teng JG, Yao J. Plate end debonding failures of FRP- or steel-plated RC beams: a new strength

model. In: Chen and Teng (eds). Proceedings of the International Symposium on Bond Behaviour

of FRP in Structures (BBFS 2005). International Institute for FRP in Construction; 2005:283-290.

17

[37] Teng, J.G., Smith, S.T., Yao, J. and Chen, J.F. (2003). “Intermediate crack-induced debonding in

RC beams and slabs”, Construction and Building Materials, 17(2003), 447-462.

[38] Tumialan G, Serra P, Nanni A, Belarbi A. Concrete cover delamination in RC beams strenghened

with FRP sheets. SP-188, American Concrete Institute, Proc., 4th International Symposium on

FRP for reinforcement of concrete structures (FRPRCS4), Baltimore, MD, 1999 Nov: 725-735.

[39] Wang J. Cohesive zone model of FRP-concrete interface debonding under mixed-mode loading.

International Journal of Solids and Structures 2007; 44: 6551-6568.

[40] Wu Z, Yin J. Fracturing behaviors of FRP-strengthened concrete structures. Engineering Fracture

Mechanics, 2003; 70: 1339-1355. [Available at: www.elsevier.com/locate/engfracmech, accessed

2 July 20007].

[41] Xiong GJ, Jiang X, Liu JW, Chen L. A way for preventing tension delamination of concrete cover

in midspan of FRP strengthened beams. Construction and Building Materials [online]; 2007; 21:

402-408. [cited 30 Aug 2007]. Available at: www.elsevier.com/locate/conbuildmat, accessed 2

September 2007].

[42] Yang ZJ, Chen JF, Proverbs. Finite element modelling of concrete cover separation failure in FRP

plated RC beams. Construction and Building Materials, 2003; 17:3-13. [Available at

www.elsevier.com/locate/conbuildmat, accessed 2 September 2007].

[43] Ye J, Yang J. An investigation on the stress transfer in concrete beams bonded with FRP plates.

In: Chen and Teng (eds). Proceedings of the International Symposium on Bond Behaviour of FRP

in Structures (BBFS 2005). International Institute for FRP in Construction; 2005:183-188.