Reinforced concrete beams strengthened with CFRP laminates ... · concrete beams with the purpose of determining the effect of crack repair on reinforced concrete beams strengthened

Reinforced concrete beams strengthened with CFRP

laminates: an experimental study on the effect of crack

repair

Pedro Colaço Franjoso da Silva Duarte

Extended Abstract

Jury

President: Professor Doutor Jorge Manuel Calico Lopes de Brito

Supervisors: Professor Doutor João Paulo Janeiro Gomes Ferreira

Professor Doutor João Pedro Ramôa Ribeiro Correia

Examiner: Professor Doutor Paulo Miguel de Macedo França

December 2011

1

1. Introduction

Reinforced concrete can suffer several types of damages that may compromise the durability of

a given structure. Along with excessive deflections and reinforcement corrosion, one of the most

common damages in reinforced concrete is cracking [1]. Structural overloads or the occurrence

of other causes of damage may lead to crack widths higher than those considered in the design

stages. If the concrete cracking is caused by a structural overload it will be necessary to apply a

strengthening system that enables the structure to sustain the new loads.

Based on the recommendation that a crack repair must take place before applying the

strengthening system, the main objective of this paper is to evaluate this recommendation by

conducting a series of experimental tests performed on reinforced concrete beams. A total of six

T-shaped beams were tested, as follows: (i) two reference beams, (ii) two cracked and

strengthened beams and (iii) two cracked, repaired and strengthened beams. The repair and

strengthening techniques used in this experimental campaign consisted of the crack repair by

epoxy injection and the application of CFRP laminates by the externally bonding reinforcement

(EBR) technique, respectively.

Given the techniques used in the experimental campaign, this paper presents a review on crack

injection, which is the most used technique for concrete crack repair [2], based on the types of

resins available, repair procedures and the mechanical behaviour of concrete beams repaired

with resin injection reported in previous studies.

The use of carbon fibre reinforced polymers as a strengthening material is also presented in this

paper. The application of the various forms that this composite material can present as a

strengthening system has been gaining more acceptance, since it is considered a very simple,

convenient and effective way to enhance the mechanical properties of reinforced concrete.

2. State of the Art

2.1 Crack repair by resin injection

There are two main types of resins for injection: epoxy and polyurethane resins. Epoxy resins’

high mechanical properties and chemical resistance make this material more suitable for

structural repair of cracked concrete, whereas polyurethane resins’ impermeability and high

adherence levels in wet conditions makes it more suitable for waterproofing [3]. In order to

determine the most adequate type of resin and its properties such as its viscosity and pot-life it

is necessary to analyze the crack conditions and characteristics namely, the crack width, extent,

activity and moisture content.

There are several crack injection procedures, the majority of them starting off by sealing the

cracks to prevent resin leaking and improve its penetration. The injection can be made with high

2

and low pressure. Although high pressure injection ensures a better resin penetration for lower

crack widths, it may also cause additional stresses to develop in the concrete. On the other

hand, low pressure injection often requires a resin with a longer pot life. For the crack injection

there are two main types of injectors: adherent injectors or packers. The adherent injectors are

applied directly on the surface of cracks while the packers are placed inside previously drilled

holes, next to the crack, intersecting the crack plane, and then tightened to ensure a mechanical

grasp. The injection proceeds until the resin exits from the next hole or injector or when a

determined pressure is achieved [4].

The efficiency of crack repair with epoxy resins has been tested in previous works such as the

one developed by Issa and Debs [5] where a series of concrete cubes, 15 cm side, were tested

in order to measure their compressive strengths. Of a total of 15 cubes, 6 cubes included cracks

without repair, 6 cubes included cracks repaired with gravity filled epoxy and 3 cubes had no

cracks. The cracks caused a reduction in compressive strength up to 41% whereas the epoxy

system restored the compressive strength by decreasing the reduction down to 8%. Another

important work was carried out by Chung [6], where 3 concrete beams, approximately 3 m long,

were subject to bending tests. The beams were loaded until failure and subsequently repaired

with epoxy injection. The load-deflection curves of the tests on the original and repaired beams

are presented in Figure 1. The results obtained in these tests show that the behaviour of the

repaired beams was similar to that of the original beams and, therefore, this repair procedure is

capable of restoring the integrity of the beams.

Figure 1 - Load-deflection curves of the original and repaired beams [6]

2.2 Strengthening with CFRP laminates

Carbon fibre reinforced polymers have high mechanical properties, are immune to corrosion

and have a good resistance against chemical agents [7]. Besides their application on concrete

structures being very simple and efficient, this composite material has been receiving more

acceptance in strengthening operations. The CFRP strengthening systems can be presented as

sheets, fabrics, wraps, strips, reinforcing bars and profiles like laminates. The type of CFRP

used depends on the type of structural member that needs to be strengthened [8]

3

CFRP laminates are frequently used for bending and shear strengthening of reinforced concrete

beams and their bonding is guaranteed by using epoxy based resins. In order to ensure good

laminate bonding, the concrete surface must be prepared through the action of water jets or

needle scalers. Since CFRP does not yield, there is a differential behaviour between CFRP and

reinforced concrete that does not allow taking total advantage of the CFRP’s mechanical

potential. In fact, a loss of bonding of the strengthening system before the failure of the

individual elements is commonly observed. To prevent this failure mode, the bond stresses,

namely at the anchorages, must be controlled in the design stages [9].

3. Experimental Campaign

As previously mentioned, the experimental campaign consisted of a series of tests on reinforced

concrete beams with the purpose of determining the effect of crack repair on reinforced

concrete beams strengthened with CFRP laminates. These tests took place at the Laboratory of

Structures and Strength of Materials (LERM) of Instituto Superior Técnico (IST) where a total of

six T-shaped beams were tested and three different types of treatments were applied. Two

reference beams were tested up to failure in order to study their mechanical behaviour and the

crack development. The other four beams were initially loaded in order to induce concrete

cracking concrete and, before the bonding of the CFRP laminates, two of the beams had their

cracks repaired through epoxy resin injection. After strengthening, these four beams were

loaded until failure. Concrete cubes and cylinders and steel rebars were tested for quality

control of the materials.

Similarly to most reinforced concrete structures, the production of the reinforced beams was

carried out with plywood formwork and applying the necessary vibration for the concrete

compaction. The production of the beams used standard Portland cement concrete C20/25

strength class, while the reinforcement steel bars’ class was A500 NR. The beams were 3,30 m

long and the beam’s cross section and reinforcement are presented in Figures 2 and 3,

respectively. The concrete cover was 2,0 cm thick.

The beams were subjected to a bending test with a monotonic loading, with the load application

and support conditions being presented in Figure 4.

Figure 2 - Geometry of the beams' section Figure 3 - Reinforcement detailing

4

Figure 4 - Test setup

During the experimental tests, applied load, beam deflections and reinforcement strains were

measured. The beams deflections were measured with displacement transducers placed at mid

span and at the loaded sections. As for the reinforcement strains, four strain gauges were

placed at the longitudinal reinforcement of each beam, also at mid span, prior to the concrete

casting, whose disposal is presented in Figure 5 and 6.

Figure 5 - Strain gauges disposal Figure 6 – Application of the strain gauges to the longitudinal reinforcement

As previously mentioned, the crack repair was performed on two of the beams while the

cracking load was being applied. For this procedure, Sikadur®52-Injection epoxy based resin,

from SIKA was used. Based on the results of the reference beams, the cracking load was

defined in order to meet the crack width criteria without causing the steel reinforcement to yield.

The crack widths were measured at different levels of loading with a crack measuring

microscope on the crack closest to mid span.

The crack injection was performed with the use of packer injectors so the repair procedure was

initiated by drilling holes in the concrete that would intersect the crack plane. Two holes were

drilled per crack plane, one on each side of the beam’s web. Afterwards, the two resin

components were mixed, with the manufacturer’s recommended ratio, until the mixture

5

presented a homogeneous appearence, presented in Figure 7. The equipment used for the

injection consisted of a monocomponent manual pump (Figure 8), thus the need to mix the

components before. The pump was cleaned by pumping acetone through the hose before the

initial injection and every time a new mixture was made.

Figure 7 - Component mixing Figure 8 - Manual pump

A packer was then introduced into the first hole and tightened to ensure the mechanical grasp.

By connecting the pump’s hose to the packer, the resin injection started and the resin dripping

through the crack was almost immediately visible, as seen in Figure 9. Given the present

conditions, the sealing of the cracks was considered not practical. It was noted that the resin

was not penetrating the lower parts of the beam and so, new holes were drilled on the bottom

surface of the beam. The pressure applied in the injection was approximately 2 to 3 bar (low

values). Moreover, when injecting the cracks, the flowing of resin through new cracks which, in

most cases, intersected the drill holes, visible in Figure 10, suggests that these cracks were

caused by the stresses induced in the drilling operation.

Figure 9 - Resin injection Figure 10 - Example of new crack formation

This injection operation was repeated on the remaining holes and, in each hole, the injection

was considered complete when the resin covered most of the crack surface or when the resin

stopped reaching new areas after several attempts.

The beams were then left with the cracking load applied for about 24 hours, a period of time

considered to be sufficiently long for the resin to attain enough strength. Before unloading the

6

beams, the exterior part of the packers was removed, leaving the rubber part inside. With the

beams unloaded it was necessary to repair the holes left by the injectors. This repair was

accomplished with the use of the Emaco S88 thixotropic mortar, from BASF. As most mortars,

there was a mixture with water, and before its application, the concrete holes were saturated in

order to prevent the mortar’s water absorption by the concrete.

As for the strengthening of the beams, S&P Laminates CFK 150/2000 were used, with a cross-

sectional area of 80 x 1,4 mm2 and a length of 2,5 m per beam. The laminates were bonded to

the concrete surface with the S&P Resin 220 epoxy resin-based adhesive and no anchorage

system was placed.

With the beams unloaded and on an upside-down position, the laminate area was outlined on

the concrete surface and needle scalers were used on the same area to remove the superficial

layer of the concrete and leave the aggregates exposed. The laminate was provided in a roll

about 10 m long, presented in Figure 11, whereby a small electric saw was used to cut the

laminate roll according to the desired dimensions. Before the bonding of the laminate, the

surface of the laminate was cleaned with acetone to remove any possible particles on the

surface and to ensure the best possible bond. Similarly to the injection resin, the bonding resin

was provided in two components whose mixture was carried out until it presented an

homogeneous aspect. The thickness of the epoxy resin is important to the performance of the

strengthening system, since a very low thickness may compromise the bonding of the laminate,

while a high thickness may influence the stress distribution. Therefore, and according to the

manufacturer’s recommendations, a 3 mm thickness of epoxy resin was applied, in which a

layer of 1 mm was applied on the concrete surface while the other 2 mm were applied on the

laminate surface [10]. Finally, the laminates were positioned on the concrete surface, as seen in

Figure 12, held by its extremities and with no additional pressure being applied on its surface.

Figure 11 - CFRP laminate roll Figure 12 - Placement of the CFRP laminates on the concrete surface

7

4. Results and discussion

In order to distinguish the different types of beams tested in the experimental tests, the following

identification was made:

Beams V1 and V2 – reference beams;

Beams V3 and V4 – simply strengthened beams;

Beams V5 and V6 – repaired and strengthened beams;

During the experimental tests, measurements were made on the applied load, beam’s

deflection, reinforcement strain and crack width, as well as the marking of the crack

development.

4.1 Material characterization tests

The concrete cubes were subjected to compressive testing, while the cylinders were subjected

to diametral compression. The results of these tests show that the concrete actually belonged to

the C25/30 strength class: at the age of 28 days, the average values of compressive strength

and splitting tensile strength were 31,0 MPa and 2,6 MPa, respectively.

The steel bar specimens, with 8 and 12 mm of diameter, were subjected to tension tests and

their behaviour confirmed the correct consideration of their strength class: the average values of

yielding and failure stresses for de 8 mm steel bars were 583 MPa and 691 MPa, respectively,

while the average values of yielding and failure stresses for the 12 mm steel bars were 535

MPa and 646 MPa, respectively.

4.2 Reference beams

The reference beams were tested monotonically up to failure. Figure 13 shows the development

of beam deflections at mid span, and the expected deflections determined through calculation

based on the beam’s characteristics and the properties of the materials used.

Figure 13 – Measured deflections at mid span for reference beams

0

20

40

60

80

100

120

140

0 10 20 30 40 50 60 70 80 90 100

Lo

ad

[kN

]

Deflection [mm]

V1

V2

Theory

8

The reference beams’ deflection results indicate a very similar behaviour to what was expected,

with similar values to the ones determined analytically. The results also show the expected

three behavioural stages of reinforced concrete beams. The first one, up to about 15 kN, where

the beam presents an elastic-uncracked behaviour. When the cracking occurs there is a visible

loss of stiffness of the beam, in which the beam assumes an elastic-cracked behaviour. The

plastic stage starts at approximately 100 kN and 110 kN for beams V1 and V2, respectively,

caused by yielding of the reinforcement. Based on the values of the beams’ deflection, the most

relevant values obtained in the tests performed on beams V1 and V2 are presented in Table 1.

Table 1 - Main values obtained in the beams' V1 and V2 tests

Stiffness (I) Stiffness (II) Yielding load Ultimate load

V1 42,3 kN/mm 7,8 kN/mm 107,4 kN 125,0 kN

V2 41,5 kN/mm 6,4 kN/mm 100,3 kN 111,9 kN

Average 41,9 kN/mm 7,1 kN/mm 103,9 kN 118,5 kN

The strains developed in the upper and bottom rebars are presented in Figures 14 and 15,

respectively, along with the theoretical strains determined analytically.

Figure 14 - Measured strains at the upper rebars of the reference beams

Figure 15 - Measured strains at the bottom rebars of the reference beams

0

20

40

60

80

100

120

-250 -200 -150 -100 -50 0 50

Lo

ad

[kN

]

Strain [μm/m]

V1-A

V1-B

V2-A

V2-B

Theory

0

20

40

60

80

100

120

0 1000 2000 3000 4000 5000 6000 7000 8000 9000

Lo

ad

[kN

]

Strain [μm/m]

V1-C

V1-D

V2-C

V2-D

Theory

9

As expected, the strains developed in the bottom rebars are significantly higher than those

developed on the upper rebars, mainly due to the fact that the neutral axis of the beam is

located near the upper rebars. Considering the theoretical values of the reinforcement strains,

these are very similar to the measured strain values on the bottom rebars, whereas on the

upper rebars, there is a significant difference. This fact suggests that the neutral axis position

was actually slightly lower than the calculated beams centre. It is also worth mentioning that,

similarly to the deflection curves, the three behavioral stages of the beams are also presented in

the bottom rebars strain curves, differing in the concrete cracking point, where the concrete

stresses are transmitted to the reinforcements, causing a sudden increase in the reinforcement

strains.

Figure 16 shows the moment-curvature relation. The curvature values were determined based

on the measured values of the reinforcement strains.

Figure 16 - Moment-curvature relation of the reference beams’ tests

Considering the loading and support conditions of these tests, the existing bending moment is

equal to 0,5P [kNm] with P [kN]. Figure 16 displays a high similarity to the bottom rebars strain

curves, with the three behavioural stages for bending actions being also visible, though

presenting a larger difference between measured and theoretical values. The theoretical values

differ slightly, mainly after the cracking point, due to the non consideration, in the theoretical

values, of the concrete influence before attaining the fully cracked stage and the possible effect

of the loading and unloading actions

The crack width measurement made throughout the loading stage showed that, in order to

execute crack repair, due to the resin appliance restrictions, the cracking load would have to be

around 90 kN, for which the crack width values were close to 0,3 mm. As for the observed

failure, both beams failed due to concrete crushing as it is seen in Figures 17 an 18.

0

10

20

30

40

50

60

0 5000 10000 15000 20000 25000 30000

Mo

me

nt [k

Nm

]

Curvature [x10-6 m-1]

V1

V2

Theory

10

Figure 17 - Concrete crushing of beam V1 Figure 18 - Concrete crushing of beam V2

4.3 Strengthened beams

As mentioned before, beams V3 and V4, although being only strengthened, were also initially

loaded with 90 kN in order to cause a crack pattern similar to that of beams V5 and V6. Unlike

beams V3 and V4, beams V5 and V6 presented some residual deflections after their unloading

due to presence of the epoxy resin inside the cracks. With the strengthening system applied,

the remaining beams were then loaded until failure. The deflections measured at mid span of all

the strengthened beams can be observed in Figure 19.

Figure 19 - Measured deflections at mid span for the strengthened beams

There are some noteworthy differences in the load-deflection behaviour of the two types of

beams. The load-deflection behaviour of beams V5 and V6, presents a less constant stiffness

comparing to beams V3 and V4. The repaired beams exhibit higher stiffness for initial loads,

though gradually decreasing throughout the test, whereas the unrepaired beams only show a

slight stiffness loss near 140 kN. This led to the conclusion that the deflections observed in the

brink of collapse were very similar to the ones observed in beams V3 and V4. Considering the

residual deflections of these beams, in the end of the tests, the beams’ deflections were even

higher for the repaired beams. In terms of ultimate strength, there seemed to be a slight

strength increase for the repaired beams, although it was only considerable in beam V5.

0

20

40

60

80

100

120

140

160

180

0 5 10 15 20 25 30 35 40 45

Lo

ad

[kN

]

Deflection [mm]

V3

V4

V5

V6

11

The main values of the tests performed on the unrepaired and repaired beams are presented in

Table 2 and Table 3, respectively, where the stiffness values only considered the initial loading

of the beams and the comparisons, are based on the average values obtained on the different

types of beams. Graphs 1 and 2 display the values these tests’ values for a better analysis of

the stiffness and ultimate strength increase.

Table 2 - Main values obtained in the beams' V3 and V4 tests

Stiffness Ultimate load

V3 12,2 kN/mm 155,5 kN

V4 12,3 kN/mm 162,2 kN

Average 12,3 kN/mm 158,9 kN

Increase compared to beams V1 and V2 73,2 % 34,1 %

Graph 1 – Stiffness and ultimate strength comparison for standard beams and unrepaired beams

Table 3 - Main values obtained in the beams' V5 and V6 tests

Stiffness Ultimate load

V5 17,1 kN/mm 170,6 kN

V6 18,0 kN/mm 163,5 kN

Average 17,6 kN/mm 167,1 kN

Increase compared to beams V1 and V2 148,9 % 41,0 %

Increase compared to beams V3 and V4 43,4 % 5,1 %

Graph 2 - Stiffness and ultimate strength comparison for standard beams, unrepaired beams, and repaired beams

0

5

10

15

Stiffness [kN/mm]

0

50

100

150

200

Ultimate strength [kN]

V1

V2

V3

V4

0

5

10

15

20

Stiffness [kN/mm]

0

50

100

150

200

Ultimate strength [kN]

V1

V2

V3

V4

V5

V6

12

The strains developed in the upper and bottom rebars are presented in Figures 20 to 23.

Considering initially the measured strains of the upper rebars it is visible that the strain values

obtained in beam V5 are much higher than those measured in the other beams, which suggests

the influence of small variations of the neutral axis position. The measured values of the strain

gauge D (at Figure 5) in beam V3 is not represented due to the erratic values and behaviour

obtained.

Figure 20 - Measured strains in the upper rebars of beams V3 and V4

Figure 21 - Measured strains in the upper rebars of beams V5 and V6

Regarding the strains developed in the bottom rebars it is possible to claim, discarding the

values of V6-C which showed some erratic behaviour in the initial loading stages, that the

strains developed in the repaired beams are lower than the ones developed in the unrepaired

beams and have a less constant stiffness. The lower strains presented in the repaired beams

can be explained by the presence of epoxy resin in the cracks and the subsequently existence

of the residual deflections when the strengthening system was applied. The residual deflections

imply the existence of residual strains in the reinforcements that are not taken into account in

Figure 23. This explanation may also apply to the lower loads for which the steel yielding occurs

in the repaired beams.

0

20

40

60

80

100

120

140

160

180

-400 -300 -200 -100 0

Load [kN

]

Strain [μm/m]

V3-B V4-A V4-B

0

20

40

60

80

100

120

140

160

180

-800 -600 -400 -200 0 Load [kN

]

Strain [μm/m]

V5-A V5-B V6-A V6-B

Figure 22 - Measured strains in the bottom rebars of beams V3 and V4

Figure 23 - Measured strains in the bottom rebars of beams V5 and V6

0

20

40

60

80

100

120

140

160

180

0 2000 4000 6000

Lo

ad [

kN

]

Strain [μm/m]

V3-C V3-D V4-C V4-D

0

20

40

60

80

100

120

140

160

180

0 2000 4000 6000

Load [kN

]

Strain [μm/m]

V5-C V5-D V6-C V6-D

13

The moment-curvature relation of the strengthened beams is presented in Figure 24.

Figure 24 - Moment-curvature relation of the strengthened beams’ tests

As seen in the reinforcement strains, the curvature developed in a more linear manner in the

unrepaired beams. Considering that the values of beam V4 are affected by the results of the

strain gauge C, it is fair to claim that Figure 24 confirms the higher initial stiffness of the repaired

beams. It is also worth mentioning that the large value difference displayed in the upper

reinforcement strains of the repaired beams had only a small influence on the curvature, mainly

in the initial loading stage.

The failure of all the strengthened beams was caused by the loss of bonding of the CFRP

laminate due to the concrete pullout at the anchorage zone. This loss of bonding is presented in

Figures 25 and 26 for beams V3 and V5, respectively.

Figure 25 – Beam failure by loss of bonding of the CFRP laminate in beam V3

Figure 26 - Beam failure by loss of bonding of the CFRP laminate in beam V5

0

10

20

30

40

50

60

70

80

90

0 5000 10000 15000 20000 25000 30000

Mo

me

nt [k

Nm

]

Curvature [x10-6 m-1]

V3

V4

V5

V6

14

As for the measured crack widths, the registered results on the strengthened beams’ tests are

presented in Figure 27, along with the crack width results from reference beam V2.

Figure 27 - Crack width measurements

The repaired beams present higher crack width values for initial loads that can be explained by

the presence of epoxy resin inside the cracks and the subsequent residual deflections. The

crack repair appears to have no influence in the crack width development as the crack

behaviour of the strengthened beams is very similar. However, comparing the crack width

values of the strengthened beams to the ones on the reference beams, it is noticeable the effect

of the strengthening system as the crack width growth of the strengthened beams is

considerably lower.

Regarding the crack pattern of the strengthened beams, in the unrepaired beams a

considerable amount of new cracks was noticed arising in the lower region of the beams’ web

for high loads, reducing the crack spacing, while in the repaired beams the number of new

cracks in this region was considerably lower. This fact is illustrated in Figures 28 and 29 in

which the crack patterns of the beams V3 and V5 are presented, respectively.

Figure 28 - Crack pattern observed in beam V3

Figure 29 - Crack pattern observed in beam V5

0

20

40

60

80

100

120

0 0,1 0,2 0,3 0,4 0,5

Lo

ad

[kN

]

Crack width[mm]

V2

V3

V4

V5

V6

15

5. Conclusions

The main objective of this work was to determine the contribution of the crack repair through

epoxy injection on the mechanical behaviour of reinforced concrete beams strengthened with

CFRP laminates.

Analyzing the experimental results, the added value of the crack repair was not sufficiently clear

compared to the simple strengthening of the reinforced concrete beams without any other

treatment. In terms of ultimate strength, both repaired beams failed at higher loads, although

only one of them showed a failure load value that stood out and was considered to be

significantly higher. Therefore, there are still some doubts on the effect of the crack repair on

this strength gain. As for the beams deflections, the behaviour between the two types of beams

was somewhat more different. In the repaired beams, a considerable increase of the initial

stiffness was observed, although decreasing more significantly for higher loads, resulting in

similar deflections near the occurrence of the beam’s failure. The crack repair had also the

disadvantage of causing some residual deflections, which, considering the experimental results,

indicate that the total deflection on the repaired beams at collapse was higher. On the other

hand, for the repaired beams, a fewer number of new cracks occurred, especially in the lower

region of the beams.

Given the fact that the crack repair with epoxy injection also serves the purpose of protection of

the reinforcement against corrosion the crack repair is still recommended in order to ensure the

durability of the strengthened structure.

6. References

[1] Helene, P., “Manual for Repair, Strengthening and Protection of Concrete Structures,

PINI, 1992. (In Portuguese)

[2] Chastre Rodrigues, C., “Cyclic Action Behaviour of Reinforced Concrete Pillars

Strengthened with Composite Materials”, PhD thesis in Civil Engineering, New

University of Lisbon, 2005. (In Portuguese)

[3] Smoak, W., “Guide to Concrete Repair”, United States Department of the Interior,

Bureau of Reclamation, 2002.

[4] SIKA Portugal, S.A., “Sika®’s Injection Systems – “Waterproofing in concrete

structures”, Technical Sheet nº037, 2009. (In Portuguese)

[5] Issa, C., Debs, P., “Experimental study of epoxy repairing of cracks in concrete”,

Construction and Building Materials, No. 21, pp 157- 163, 2007.

[6] Chung, H. W., “Epoxy-Repaired Reinforced Concrete Beams”, ACI Journal,

Proceedings, No. 5, pp 233 – 234, 1975.

16

[7] Barros, J., “Composite Materials in Structural Retrofitting”, Minho University, 2004. (In

Portuguese)

[8] Bank, L.C., “Composites for construction: Structural design with FRP materials”, Wiley,

Hoboken, NJ, 2006.

[9] FIB – Bulletin 14, “Externally bonded FRP reinforcement for RC structures”, Federation

International du Beton, 2001.

[10] S&P – Clever Reinforcement Company AG, “Guide for the application of S&P FRP

Systems”, Document available at http://www.reinforcement.ch/fileadmin/redakteur/pdf-

en/frp/application_guide_line/Application_Guide_Line.pdf, October 2006.