Aspects of behaviour of CFRP reinforced concrete beams in bending

Muhammad Masood Ra
*
, Ali Nadjai, Faris Ali, Didier Talamona
Fire Safety Engineering Research & Technology Centre (FireSERT), University of Ulster at Jordanstown, Shore Road, Newtownabbey BT37 0QB, UK
Received 13 February 2006; received in revised form 24 July 2006; accepted 30 August 2006
Abstract
The corrosion of steel poses a serious problem to the durability of reinforced concrete structures and bre reinforced polymer (FRP)
has emerged as a potential alternative material to the traditional steel. The results of a test series consisting of carbon FRP (CFRP) and
steel bars reinforced concrete beams are reported in this paper. The results indicated that the behaviour of CFRP and steel reinforced
beams was similar in many aspects. Both type of beams failed in their predicted modes of failure. The strength design method underes-
timated nominal moment capacity of CFRP reinforced beams. The deection of CFRP reinforced beams was satisfactory at service load
level, corresponding to theoretical load capacity. The deformability factor of CFRP reinforced beams was more than 6 indicating their
ductile nature of failure.
2006 Elsevier Ltd. All rights reserved.
Keywords: Fibre reinforced polymer; Carbon FRP; Mode of failure; Concrete beams; Reinforcement; Moment; Deection; Deformability
1. Introduction
Reinforced concrete (RC) structures have proted from
the unrivalled dominance of steel over all other reinforc-
ing materials for more than 100 years. Superior qualities
of steel, in terms of strength and compatibility with con-
crete, make steel an eective concrete reinforcement. Steel
is, however, highly susceptible to oxidation when exposed
to chlorides. Although, alkaline environment of concrete
protects steel from corrosion and makes it very durable
it is not always possible to provide an ecient protection.
Factors such as insucient concrete cover, poor design or
workmanship, poor concrete mix and aggressive environ-
ments can break down the protection layer and may lead
to corrosion of the steel rebars. These destructive environ-
ments include marine surroundings, use of deicing salts on
bridges and parking garages, and the use of salt contam-
inated aggregates in the concrete mixture. The initial signs
of distress are usually cracking and spalling of concrete,
which provides access to other environmental agents like
moisture to further intensify the oxidisation of steel. As
corrosion goes on it causes a reduction in the cross-sec-
tion of a steel bar, which leads to loss of bond between
rebar and concrete. To arrest the rusting of steel remedial
work has to be carried out in order to achieve the full
potential of the structure. This structural maintenance
incurs exorbitant costs annually. Unreliable durability of
these structures as a result of corrosion of steel is thus
a serious problem. Recent eorts and research have been
focussed towards the introduction of innovative non-
metallic materials in the construction industry. Fibre rein-
forcement polymer (FRP) materials have evolved as a
result of new developments in the elds of plastics and
bre composites.
A signicant amount of research work has been exe-
cuted to investigate various aspects of the use of FRP bars
with concrete. As a result of these eorts, the applications
of FRP bars are becoming increasingly common as a rein-
forcing material. Carbon FRP (CFRP) bars are mostly
used in prestressing applications due to their high tensile
strength, which is comparable with steel strands. This
paper presents the results of tests carried out on concrete
beams reinforced with CFRP bars. Testing of these beams
0950-0618/$ - see front matter 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2006.08.014
*
Corresponding author. Tel.: +44 28 903 68734; fax: +44 28 903 68726.
E-mail address: ra-m@ulster.ac.uk (M.M. Ra).
www.elsevier.com/locate/conbuildmat
Construction and Building Materials xxx (2006) xxxxxx
Construction
and Building
MATERIALS
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014
is a part of on-going research on the behaviour of CFRP
reinforced concrete beams at room and elevated tempera-
tures at the University of Ulster.
In the rst phase of this project a total of four specimen
beams were used. Duplicate steel and CFRP reinforced
beams were tested at room temperature. The tests results
of this series will serve as benchmark to study the perfor-
mance of similar beams under re conditions. Steel rein-
forced beams were tested as control specimens. The study
focused on the exural behaviour of these beams in terms
of stressstrain, loaddeection, modes of failure, load-car-
rying capacity, and cracking pattern. Some of the aspects
of behaviour of these beams are discussed hereafter.
Long-term behaviour and durability aspects are, however,
beyond the scope of this work.
2. Experimental programme
2.1. Test specimens
The overall length of each beam was 2000 mm and the
cross-section was 120 200 mm. Each of these beams was
reinforced with two longitudinal bars on the tension face
(CFRP bars for FRP reinforced beams and steel bars for
steel reinforced beams). The beams were cast in moulds
made of plywood stiened with aluminium angles to main-
tain the beam shape under the pressure of newly cast con-
crete. A 20 mm concrete cover was used all-around the
beam. The area and nominal yield strength of the compres-
sion steel and nominal concrete strength were kept con-
stant for all beams in this series. Mixing of concrete was
done in a rotating mixer. A vibrating table was used for
the compaction of concrete inside the mould. The sides
of the mould were stripped after 24 h of casting and beams
were covered with hessian. Intermittent curing (thrice a day
in summer) was carried out for 10 days and then the beams
were left air-drying in laboratory conditions until the day
of testing. Four 100 mm cubes were cast for each beam.
Cubes were cured by keeping them on top of their respec-
tive beam. The beams were tested as simply supported
beams over a span of 1750 mm under a four-point static
load as shown in Fig. 1. The loads were 400 mm apart giv-
ing a shear span of 675 mm.
2.2. Materials
2.2.1. Concrete
Four identical concrete mixes of 325 kg/m
3
ordinary
Portland cement, 1001 kg/m
3
of graded crushed stone,
853 kg/m
3
of sand and 216 litres/m
3
of water were used.
The maximum aggregate size was 10 mm. The exact
amount of water varied depending on the moisture con-
tents of aggregates. The slump of concrete ranged from
40 to 50 mm. Cube strength after 28 days was
49.23 MPa, this being the average strength of the four
cubes, one from each beam. This is equivalent to
43.31 MPa cylindrical strength [1]. Table 1 shows the
equivalent cylindrical strength [1] of the concrete on the
day of testing. These strengths were obtained from the
average strength of three cubes for each beam. Each beam
tested is dened by letters comparing its reinforcing mate-
rial and temperature conditions. The notation of beams is
as: the rst letter (B) stands for beam; the second letter
indicates the testing temperature as R for room tempera-
ture; the third letter represents the type of tension reinforc-
ing bar material such as S for steel and C for CFRP bars.
These notations have been consistently followed through-
out the rest of the text.
2.2.2. CFRP bars
The FRP bars consisted of 9.5 mm diameter straight
CFRP rods, as shown in Fig. 2. The bars were produced
by an American manufacturer using the pultrusion process.
Not to Scale
Dimensions in mm
6 mm stirrups
400
1750
2000
200
120
2 T8 bars
675 675
100 mm c/c Typ.
2 T10 steel /
2- 9.5 mm
CFRP bars
P
Fig. 1. Details of a typical beam.
Table 1
Strength of concrete on the day of test
Beam Compressive strength (MPa)
BRS1 46.52
BRS2 44.64
BRC1 42.55
BRC2 41.71
2 M.M. Ra et al. / Construction and Building Materials xxx (2006) xxxxxx
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014
Continuous carbon bres with a volume fraction of 60% by
volume were used. The nominal tensile strength and tensile
modulus of these bres were 4.83 GPa and 234 GPa,
respectively. The resin used to bond bres was bisphenol
epoxy vinyl ester. A textured surface was provided on the
bar through surface treatment in order to increase the bond
with the concrete. The bar had a widely spaced spiral wind-
ing imprint with textured surface in between helical rings.
The surface texture was formed with the same resin without
involving external bres. Properties of CFRP bars are
given in Table 2. The manufacturer provided results of ten-
sile tests on these bars.
2.2.3. Steel bars
Tension rebars in BRS beams consisted of 10 mm diam-
eter high strength deformed bars, as shown in Fig. 2. Mate-
rial properties of these bars are given in Table 2. These
properties were determined by the tensile tests in the labo-
ratory. The bars were chosen because their nominal cross-
sectional area was approximately equal to the CFRP bars.
It was, therefore, possible to position the steel and CFRP
bars identically in the control and in the test beams. Steel
reinforcing bars of the identical area of the CFRP bars
were not available.
Top bars were of 8 mm diameter high strength deformed
steel for all beams. Both top and bottom steel bars were
hooked at each end. The mechanical properties of top rein-
forcing bars are shown in Table 2.
The reinforcement cages (Fig. 1) were tied with iron
wire. Smooth 6 mm diameter closed rectangular stirrups
spaced at 100 mm centre to centre were chosen to comply
with the criteria of the ultimate strength design of FRP
reinforced beams given by the ACI code [2,3]. Results of
tensile tests on 6 mm bars are also included in Table 2.
2.3. Instrumentation
The instrumentation was set up to measure the deec-
tion of the beam and the deformation of the reinforcing
bar. Strain gauges were used to measure deformation and
to monitor bond of tension bars. The deection at midspan
was recorded using linear variable dierential transducers
(LVDTs). These LVDTs were placed on both sides of the
beam at the centre. Horizontal LVDTs were used at the
end of the BRC beams additionally to measure the slip of
the CFRP bars. Computer aided data acquisition systems
were used to record continuously, load, deection, slip
and strain. Thus this data could be obtained easily at any
time during each test.
2.4. Test procedure
The specimens were placed on half-round supports,
which were spaced at a distance equal to the test span of
the beam. Loading was applied in small increments,
through 38 mm diameter rollers, by means of a 200 kN
hydraulic jack. Steel plates 25 mm wide and 5 mm thick
were placed under each roller on the top of the beam in
order to avoid local crushing. Each load increment was
2.5 kN for BRS and 5 kN for BRC beams and was mea-
sured with a 200 kN load cell. All beams were tested to fail-
ure. The beams were 23 months old at the time of testing.
Immediately after the load increment, cracks were identi-
ed using a magnifying glass and marked. The ends of
the cracks were labelled with the corresponding load step.
Three minutes were allowed for the completion of the pro-
cess before the next increment of load was applied. The
monitoring of cracks continued over the entire loading
spectrum.
The operator manually controlled load, which was dis-
played on the monitor screen, and made the necessary
adjustments to keep load as constant as possible. For all
tests the load was removed after the applied load dropped
substantially below the ultimate load. A complete test took
approximately 1 h.
3. Analysis of test results
3.1. Cracking behaviour
Concrete is a weak material in tension and cracks when
subjected to high local tensile stresses. Its low tensile capac-
ity can be attributed to the high stress concentration under
load [4]. Flexural cracking in beams is not only unavoid-
able but is necessary to allow tension reinforcement to play
its part. The formation and propagation of cracks depend
on the tensile strength of the concrete. When the principal
tensile stress in concrete exceeds its tensile strength, cracks
Fig. 2. CFRP and tension steel bar.
Table 2
Mechanical properties of rebars
Bar type Nominal ultimate
strength
a
(MPa)
Ultimate
strain
a
Elastic modulus
(GPa)
CFRP 1676 0.0145 135.9
Steel 10 mm 530 0.0048 201
Steel 8 mm 566 0.0049 194
Steel 6 mm 421 0.0041 200
a
Parameters corresponding to 0.2% oset yield stress for steel bars.
M.M. Ra et al. / Construction and Building Materials xxx (2006) xxxxxx 3
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014
form perpendicular to the direction of this stress. A section
of least resistance may develop microscopic cracks as a
result of very high stresses, while the other parts may still
be subjected to low stresses. This section is likely to be
the one with many aggregate-paste interfaces [5]. A number
of factors may aect the spacing and width of cracks. How-
ever, investigators have not completely agreed on the fun-
damental factors of crack formation [5]. Extensive
cracking generally takes place in exural members like
beams and causes a change in the behaviour of these
members.
3.1.1. Cracking in BRC beams
Typical cracking behaviour of BRC beams is shown in
Figs. 3a and b. Cracking started in the constant moment
region with the cracks originating from bottom bres, as
the principal stresses were the greatest at these extreme
bres. These cracks were mainly vertical exural cracks,
which were perpendicular to the beam longitudinal axis
since the shear stresses were absent in this zone. These ini-
tial cracks traversed quite deeply into the compression
zone. The height of initial cracks was between 35 and
129 mm. This shows that immediately after cracking the
neutral axis shifted quite deeply into the compression zone.
Cracks outside this central zone originated as vertical ex-
ural cracks at a relatively lower load level, as shear stresses
were initially small. However, as the load increased shear
stresses became high and the principal stresses were gener-
ated at approximately 45. These stresses transformed the
initial exural cracks into inclined diagonal tension cracks
at higher load levels. With increasing load these inclined
cracks propagated towards the load points on top of the
beam. The crack spacing also decreased rapidly with the
increasing load. However, the spacing of cracks remained
approximately constant after an applied load of 30 kN.
As the cracking initiates on the tension side of a beam,
stresses are transferred to the reinforcement, thus relieving
the concrete, at and adjacent to these cracks. After exten-
sive cracking has taken place, the uncracked concrete
between consecutive cracks becomes small and subject to
a very high longitudinal stress gradient. These highly
stressed uncracked segments of concrete require higher
average tensile stresses to induce further cracking at this
stage, as the new cracks may be forced to pass through
the cement paste and aggregates instead of developing
along paths of least strength [5]. The strain in the FRP bars
at this stage reached an average value of 0.0034. Beyond
this load level only existing cracks grew in length with
the increase in load with almost no change in number of
cracks. Some tertiary as well as shear cracks also developed
near the ultimate capacity of the beam. Tertiary cracks
formed at the level of tension reinforcement.
3.1.2. Cracking in BRS beams
The cracking pattern of BRS1 and BRS2 is shown in
Figs. 3c and d. Similar to BRC beams cracking in BRS
beams also started with vertical exural cracks in the con-
stant moment region. However, for BRS beams very few
cracks outside the pure bending zone turned into inclined
cracks and most of these were relatively vertical as shown
in Figs. 3c and d. These beams showed a classic reinforced
concrete crack pattern involving fewer and larger cracks.
The spacing of cracks, however, decreased as the load
increased. The crack formation became stabilized once
Fig. 3. Cracking pattern of beams.
4 M.M. Ra et al. / Construction and Building Materials xxx (2006) xxxxxx
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014
again after a load of 30 kN. The average strain in steel bars
at this load level was recorded as 0.0024. Few tertiary
cracks also formed around reinforcement at higher load
levels.
3.1.3. Comparison between BRC and BRS beams
The number of cracks and the average crack spacing at
failure for all beams is shown in Table 3. Spacing was mea-
sured with a steel rule to the nearest 1 mm. It can be seen
that both BRC and BRS type of beams developed almost
same number of cracks at failure with similar average spac-
ing. The width of cracks in BRC beams was considerably
narrow. This shows a good mechanical bond between
CFRP bars and surrounding concrete.
No horizontal crack at the reinforcement level was
observed in any of the tested beams. When the load was
removed the cracks on BRC beams were observed to close.
This is due to the fact that FRP bars exhibit linearly elastic
behaviour up to failure. Since BRC beams failed by the
crushing of concrete, CFRP bars recovered most of their
deformation after the load was removed. On the contrary,
no change in the crack width was observed in BRS beams
after removal of the load as steel bars were in their irrevers-
ible strain-hardening zone once they yielded.
3.2. Modes of failure
The design of both BRS and BRC beams was based on
ACI Code [2,3]. BRC beams were designed over-reinforced
using reinforcement ratio (q) (Eq. (1)) greater than the bal-
anced reinforcement ratio (q
b
). The balanced reinforce-
ment ratio corresponds to the condition when the strain
in the extreme concrete bres in compression reaches to
its ultimate value of e
c
= 0.003 [2,3] at the same time when
the strain on the tension steel reaches to yield strain. Bal-
anced reinforcement ratio can be calculated using Eq. (2).
BRS beams were under-reinforced beams with reinforce-
ment ratio less than balanced reinforcement ratio.
q
s

A
s
bd
q
f

A
f
bd
1
q
sb
0:85b
1
f
c
f
y

600
600 f
y
q
fb
0:85b
1
f
c
f
fu

E
f
e
c
E
f
e
c
f
fu
2
where q
s
is the actual reinforcement ratio of steel, q
f
the ac-
tual reinforcement ratio of FRP, q
sb
the balanced steel
reinforcement ratio, q
fb
the balanced FRP reinforcement
ratio, A
s
the area of steel bars, A
f
the area of FRP bars,
f
c
the ultimate concrete strength (MPa), b the width of
the section, d the eective depth of the section, f
y
the yield
strength of steel (MPa), f
fu
the ultimate tensile strength of
FRP (MPa), e
c
the maximum concrete strain at extreme
compression bres, E
f
the modulus of elasticity of FRP
bars (MPa), and b
1
0:85 0:05
fc28
7
P0:65.
For design guidelines in Europe, Pilakoutas et al. [6]
proposed a minimum amount of FRP reinforcement to
achieve the failure of a section through concrete crushing.
The minimum reinforcement can be calculated from Eq.
(3). e
c
in Eq. (3) is taken as 0.0035 [6]
q
min

0:81f
c
8e
c
f
fu
f
fu
E
f
e
c
_ _ 3
Balanced and actual reinforcement ratios for both types of
beams are given in Table 4. The last column of Table 4 also
gives the minimum reinforcement ratio (q
min
) according to
Eq. (3) for BRC beams. It is evident that BRC beams ful-
lled the requirements of over-reinforced section described
by both ACI Code [3] and Eq. (3).
From Table 4 a compression failure of BRC beams and
a tension failure of BRS beams can be anticipated. Table 5
presents the observed modes of failure of all beams. The
behaviour of both BRS beams was similar, they both failed
by the crushing of concrete after the tension reinforcement
yielded and the BRC beams failed in compression as
expected. In beam BRC1 the diagonal tension crack caused
failure. Formation of diagonal tension cracks takes place in
beams in a region where large compression stresses exist in
a direction perpendicular to the maximum tensile stresses
[5]. This diagonal tension crack in BRC1 originated as a
vertical crack at a distance of approximately 120 mm from
the support. The crack, however, gradually bent over
towards the point of load application as it propagated
upward and became almost horizontal before reaching
the load point. The beam, nevertheless, kept on taking load
as this diagonal tension crack propagated and opened up.
Finally, the concrete above this crack crushed adjacent to
the load point in the shear span as shown in Fig. 3a.
BRC2 beam failed by the crushing of concrete between
point loads as shown in Fig. 3b.
Table 5 shows cracking loads of beams where it can be
seen that the cracking loads for all four beams are very
close to each other despite the two diering reinforcing
materials. Discussion on this aspect is covered in Section
3.3. The ultimate load carried by the beam is also shown
in Table 5 and it can be noticed that BRC beams carried
more than twice the load on BRS beams. This was due
Table 3
Number of cracks and average crack spacing
Beam Number of cracks Spacing of cracks (mm)
BRS1 13 96.00
BRS2 12 100.00
BRC1 18 77.00
BRC2 15 99.00
Table 4
Balanced, actual and minimum reinforcement ratio
Beam type q
b
Actual q q
min
(Eq. (3))
BRS beam 0.0277 0.0077
BRC beam 0.0032 0.0070 0.0055
M.M. Ra et al. / Construction and Building Materials xxx (2006) xxxxxx 5
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014
to strength of CFRP, which was much higher than yield
strength of steel bars (Table 2). Failure modes of beams
are presented in Fig. 3.
3.3. Cracking and ultimate moment
ACI code [2] expresses exural strength of concrete in
terms of modulus of rupture [Eq. (4)], which is the maxi-
mum tensile stress of concrete in bending [4]. Eurocode 2,
part1-1 [7] suggests the use of a mean value of axial tensile
strength of concrete, which is given by Eq. (5). To calculate
cracking moment Eq. (6) can then be used
f
r
0:62

f
c
_
4
f
ct
0:30 f
2=3
c
5
M
cr
ACI

f
r
I
g
y
b
M
cr
Eurocode

f
ct
I
g
y
b
6
where f
r
is the modulus of rupture (MPa), f
ct
the mean axial
tensile strength of concrete (MPa), M
cr
the cracking mo-
ment, I
g
the gross moment of inertia about centroidal axis
of concrete section without reinforcement, and y
b
is the dis-
tance from centroidal axis of concrete section to the ex-
treme tension bers, neglecting reinforcement.
The nominal moment capacity of under-reinforced steel
RC section can be determined using Eq. (7). This equation
is common to both ACI code [2] and Eurocode 2 [7] in the
absence of safety factors and could be arrived at using con-
ditions of compatibility of strain and equilibrium of forces.
For FRP reinforced beams failing in compression, Eq. (8)
taken from ACI code [3] can be used.
M
ns
A
s
f
y
d
A
s
f
y
1:7f
c
b
_ _
7
M
nf
A
f
f
f
d 1 0:59
A
f
f
f
f
c
bd
_ _
8
where M
ns
is the nominal moment capacity of steel rein-
forced section and M
nf
is the nominal moment capacity
of FRP reinforced section
f
f

E
f
e
c

2
4

0:85b
1
f
c
q
f
E
f
e
c

0:5E
f
e
c
_
_
_
_
6 f
fu
From ACI 3 9
It can be expected from Eq. (6) that the cracking moment
and hence cracking load is independent of the tension rein-
forcing material type. Masmoudi et al. [8] have also shown
that theoretical M
cr
is nearly unaected by the second mo-
ment of area of the reinforcement. Table 6 presents theoret-
ical and experimental cracking moment and nominal
moment capacities of all four beams. It is evident that
the tensile strength of the concrete in bending is overesti-
mated by the value of modulus of rupture, as also found
by Benmokrane et al. [9]. The theoretical M
cr
values based
on direct tensile strength of concrete adopted by Eurocode
2 [7] are quite close to the experimental values of cracking
moment. This is because the tensile strength of concrete in
bending determined from a modulus of rupture test tends
to be higher than tensile strength from a direct tension test
[10]. The slight dierence between experimental and theo-
retical M
cr
values may partly be due to the variation of
strength of concrete in larger structural element and smal-
ler test specimens (cubes/cylinders).
It can be observed from the last column of Table 6 that
Eq. (7) predicted the nominal moment reasonably well for
steel reinforced beams. However, Eq. (8) underestimated
the nominal capacity of FRP beams to about 33%. Since
both BRC beams were over-reinforced, maximum moment
capacity depended on the maximum concrete strain at fail-
ure. It thus seems likely that as a result of the connement
provided especially by the stirrups the actual strain in con-
crete exceeded maximum concrete strain of 0.003 assumed
by ACI code [3]. This has been pointed out also by Nanni
[11] and Benmokrane et al.[12].
Table 5
Load, deection and modes of failure
Beam P
cr
(kN) P
u
(kN) D at P
yield
a
(mm) D at P
u
(mm) D at P
35%
(mm) Modes of failure
BRS1 7.8 41.9 6.92 29.16 1.75 Steel yielding
BRS2 7.5 40.1 7.50 27.78 1.74 Steel yielding
BRC1 7.1 88.9 11.14 35.26 8.60 Shear compression
BRC2 7.1 86.5 10.96 35.50 8.37 Compression
a
Deection in BRC beams corresponding to yielding of steel bars in BRS beams.
Table 6
Cracking moment and moment capacity
Beam (M
cr
)
Exp.
(kN m) Theoretical M
cr
(kN m) (M
n
)
Exp.
(kN m) (M
n
)
Th.
(kN m)
Mn
Th:
Mn
Exp:
ACI Code Eurocode
BRS1 2.63 3.38 3.10 14.14 13.29 0.94
BRS2 2.53 3.31 3.02 13.53 13.36 0.99
BRC1 2.40 3.24 2.93 30.00 20.04 0.67
BRC2 2.40 3.20 2.89 29.19 19.91 0.68
6 M.M. Ra et al. / Construction and Building Materials xxx (2006) xxxxxx
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014
3.4. Loaddeection history
One of the important factors that aect the serviceability
of a RC exural member is its deection. Deection is a
function of load, span length, second moment of area
and modulus of elasticity of material. The stiness of a
member is largely dependant on both the second moment
of area and the modulus of elasticity. Table 5 presents a
comparison of deection at dierent stages of loading of
beams. Service load is considered as 35% of the ultimate
load [13].
Fig. 4 shows the experimental loaddeection histories.
The initial linear part of the graphs has a very steep slope,
which corresponds to the uncracked condition of these
beams. In this region the deection is proportional to the
applied load and the entire concrete section is considered
eective in resisting the loads. As can be seen from
Fig. 4, the behaviour of both types of beams is similar
before cracking when the beams are sti. The end point
of this linear part is an indication of the initiation of crack-
ing in the beam.
The next segment that immediately follows this linear
part provides information about the bond quality and ten-
sion stiening eects due to crack spacing. The slope of this
part is smaller than the slope of the initial linear segment.
This shows that the rate of deection per unit load is higher
after the beam has cracked, which is an indication of the
reduction in the stiness of the cracked beam. Stiness here
is dened as load per unit deection. It can be seen, how-
ever, from the widening of the gap between BRS and
BRC curves in Fig. 4 that the rate of reduction in the sti-
ness of BRC beams became higher with the increase in
load. Average dierence in the stiness of both types of
beams at the yielding of steel bars was about 38%. This
can be attributed to the low elastic modulus of CFRP
bar, which is 32% less than a steel bar. The reduced stiness
of FRP reinforced beams after cracking has also been
reported by various other researchers [9,11,1418].
The last part of the curve is an indication of possible
failure mechanism of the structure. As shown in Fig. 4,
both BRS beams showed a very ductile behaviour and both
beams failed at nearly the same load after undergoing con-
siderable deformation with very small increase in the load
once steel yielded. The ultimate load of BRS beams was
around 53% lower than BRC beams while the deection
of BRC beams at ultimate state was 25% greater than
BRS beams on the average, as can be seen in Table 5.
The ultimate load here is considered as the maximum load
carried by the beam. At failure, the deection of BRC
beams was on the average 7% less than BRS1 beam, as evi-
dent in Fig. 4. This shows that up to the yielding of steel
bars BRC beams deected more than BRS beams due to
factors related to the low elastic modulus of CFRP bars.
However, after yielding BRS beams exhibited a much fas-
ter rate of deection than BRC beams with a negligible
change in load. Span to deection ratio at failure for all
the beams is shown in Table 7. It can be seen in Fig. 4 that
some variations existed in the deection at failure for each
type of beam due to the variations in the maximum con-
crete strain. Substantial deection recovery was observed
in BRC beams after the load was taken of. This shows that
CFRP bars remained within the elastic range at failure
stage as shown in Fig. 4.
It can also be seen from Fig. 4 that load-carrying
capacity of both BRC beams dropped gradually after
crushing of concrete. This shows that despite being
over-reinforced FRP reinforced beams can have a ductile
failure mode as well as a kind of energy dissipation mech-
anism. Sudden drop in loads near the failure as visible in
loaddeection curves of BRC beams was due to the
opening of cracks.
If the deection of beams at service load, which is given
in Table 5, is compared with the deection limit of span/
360 set by ACI code [2], it appears that both BRC beams
deected more than the allowable limit. This aspect, how-
ever, does not necessarily need to control the entire design
work in every situation. The structural designers encounter
excessive deection issues very often associated with steel
reinforced beams. They are, therefore, not only aware of
the nature of this issue but are well equipped with practical
solutions if they face constraints on the design. For exam-
ple, if the situation allows an initial camber in the beam can
be provided to reduce the amount of nal deection after
the application of full service load. The design of beams
with FRP bars might make these types of solutions a reg-
ular part of the design in most of the cases. However, con-
sidering the benets oered by FRP bars this should not be
too much of a worrying factor. Nevertheless, it emphasises
the importance of a reliable method of predicting theoret-
ical deection of FRP reinforced beams.
0
10
20
30
40
50
60
70
80
90
100
0 5 10 15 20 25 30 35 40 45 50
Deflection (mm)
L
o
a
d

(
k
N
)
BRC1
BRC2
BRS1
BRS2
Fig. 4. Loaddeection history.
Table 7
Span to deection ratio at failure
Beam Span/deection
BRS1 37
BRS2 44.5
BRC1 39
BRC2 40
M.M. Ra et al. / Construction and Building Materials xxx (2006) xxxxxx 7
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014
Another aspect of this issue is that the limiting deection
of span/360 during the test reached at the applied load of
19.5 and 20.2 MPa, respectively, for BRC1 and BRC2, as
can be seen in Fig. 4. These values are 33% and 34% of the-
oretical ultimate load (Table 6) and correspond to service
load stage. Unless a dierent approach of designing FRP
reinforced concrete beams is required by code, service
deection of BRC beams corresponding to theoretical load
capacity is satisfactory.
3.5. Ductility and deformability
Ductility of a structure can be dened as its ability to
absorb energy without critical failure. Ductility generally
refers to the amount of inelastic deformation which a mate-
rial or structure experiences before complete failure. This
deformation can be measured in terms of displacement,
strain or curvature. Ductile behaviour allows a structure
to undergo large plastic deformations with little decrease
in strength and hence avoids brittle failure.
Conventional steel reinforced beams have a distinct elas-
tic and inelastic phase of deformation before and after
yielding of steel. Hence for these structures, ductility can
be dened quantitatively as the ratio of the total deforma-
tion at failure divided by the deformation at the elastic
limit.
The behaviour of FRP reinforced structures is linear up
to failure. The energy released at failure for these beams is
linear as well [19] as shown in Fig. 4. Grace et al. [20] have
suggested a ductility classication of beams based on the
ratio of inelastic energy to the total energy. They termed
this ratio as energy ratio.
Naaman and Joeong [21] proposed following equation
to compute the ductility index l
e
for beams reinforced with
FRP bars.
l
e

1
2
E
T
E
El
1
_ _
10
where E
T
is the total energy, which is area under load
deection curve and E
El
is the elastic energy, which is area
of the triangle formed at the failure load by a line whose
slope is S (Fig. 5).
Ductility indices for all beams are shown in Table 8. The
failure load has been considered as the load at failure of
each of these beams. As can be expected, it can be seen
from the column of Ratio of l
e
that BRS beams were
around 4 times more ductile than BRC beams.
However, there is an agreement between many research-
ers that FRP reinforced beams should be compared on the
basis of their deformability with steel reinforced beams
rather than ductility. Jaeger et al. [22] presented the con-
cept of deformability factor and strength factor. They
found that the product of these two factors remains
approximately between 6 and 7 for steel reinforced beams
irrespective of the reinforcement ratio. They termed this
product overall factor and proposed a value not less than
6 for all type of reinforcement.
Newhook et al. [23] proposed a method of calculating
deformability factor as given by Eq. (11)
Deformability Factor

Moment at Ultimate State Curvature at Ultimate State

Moment at Service State Curvature at Service State
11
Newhook et al. [23] proposed 0.0020 as maximum strain in
FRP reinforcement at service state and 0.0012 for steel. Ta-
ble 8 shows deformability factors for all beams according
to Eq. (11). Curvature at ultimate state in the above equa-
tion was calculated using a concrete strain of 0.0035 and
actual reinforcement strain recorded during the test. Ulti-
mate moment is taken as the maximum recorded unfac-
tored moment resisted by the beam. It can be seen that
the deformability factors for all beams are greater than 6.
These beams can, therefore, be considered safe from both
strength and deformability aspects.
4. Conclusions
The results of experimental testing on CFRP and steel
reinforced beams were presented. The behaviour of both
types of beams was found similar in many respects. The
important features of the comparison of behaviour
between BRC and BRS types of beams are listed as
follows:
P
1
P
2
S
1
S
2
L
o
a
d

Deflection
E
l
a
s
t
i
c

E
n
e
r
g
y
,

E
E
l

Inelastic Energy
S
P
u
( )
2
2 1 2 1 1
P
S P P S P
S
+
=
Fig. 5. Ductility index [21].
Table 8
Ductility index and deformability factor
Beam l
e
Ratio of l
e
Deformability
factor (Eq. (11))
BRS1 8.28 3.80 13.84
BRS2 7.26 3.33 10.10
BRC1 3.01 1.38 28.36
BRC2 2.18 1.00 30.53
8 M.M. Ra et al. / Construction and Building Materials xxx (2006) xxxxxx
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014
1. The cracking behaviour and pattern of both steel and
CFRP reinforced beams was similar. Both types of
beams developed nearly equal numbers of cracks with
an equal average crack spacing at failure.
2. BRS beams failed by steel yielding and BRC type
beams failed by concrete crushing as per their design.
3. The cracking load for all beams was nearly the same.
4. The maximum concrete strain in compression at fail-
ure exceeded 0.003 in BRC beams.
5. The ACI code equation underestimated the ultimate
capacity of BRC beams.
6. BRC beams deected more than the BRS type beams
after cracking. However, after yielding of steel the
rate of deection in BRS beams was more than
BRC beams.
7. The service deection of span/360 of BRC beams cor-
responding to theoretical ultimate load, as deter-
mined by code equations, was satisfactory.
8. The issue of deection of FRP reinforced beams can
also be addressed by practical solutions of controlling
deection at service state such as by providing an ini-
tial camber in the beam.
9. BRC beams failed in ductile fashion. However, they
were less ductile than BRS beams.
10. The deformability factor of BRC beams was greater
than 6.
11. More tests are, however, needed to investigate the
inuence of other parameters such as concrete
strength on the behaviour of CFRP bars reinforced
structures.
Acknowledgements
The authors wish to acknowledge the support provided
for this research by the School of Built Environment,
University of Ulster; Dr. Alan Leacock, Lecturer, School
of Electrical and Mechanical Engineering for tensile test-
ing of steel bars and all the laboratory technical sta
members.
References
[1] Neville MA. Properties of concrete. England: Longman Scientic &
Technical; 1981. p. 529565.
[2] American Concrete Institute. Building code requirements for struc-
tural concrete, ACI 318-95, Detroit (MI); 1995.
[3] American Concrete Institute. Guide for the design and construction
of concrete reinforced with FRP Bars, ACI 440.1R-01, Detroit (MI);
2001.
[4] Hassoun MN. Design of reinforced concrete structures. USA: PWS
publishers; 1985. p. 14182.
[5] Broms B, Raab A. The fundamental concept of the cracking
phenomenon in reinforced concrete beams. Final Progress Report
Phase I, Cornell University; 1961.
[6] Pilakoutas K, Neocleous K, Guadagnini M. Design philosophy issues
of ber reinforced polymer reinforced concrete structures. ASCE J
Compos Constr 2002;6(3):15461.
[7] Eurocode 2. Design of concrete structures, Part1-1: general rules and
rules for buildings, Brussels; 1992.
[8] Masmoudi R, Benmokrane R, Chaallal O. Cracking behaviour of
concrete beams reinforced with ber reinforced plastic rebars. Can J
Civil Eng 1996;23:11729.
[9] Benmokrane B, Chaallal O, Masmoudi R. Flexural response of
concrete beams reinforced with FRP reinforcing bars. ACI Struct J
1996;91(2):4655.
[10] Mirza SA, Hatzinikdas M, Macgregor JG. Statistical descriptions of
strength in concrete. J Struct Div Proc ASCE 1979;ST6:102137.
[11] Nanni A. Flexural behaviour and design of RC members using FRP
reinforcement. ASCE J Struct Eng 1993;119(11):334559.
[12] Benmokrane B, Masmoudi R. FRP C-bar as reinforcing rod for
concrete structures. In: El-Badry MM, editor. Proceedings of 2nd
international conference on advanced composite materials in bridges
and structures, Montreal, Que.; 1996. p. 1818.
[13] Al-Salloum AY, Alsayed HS, Almusallam HT, Amjad AM. Some
design considerations for concrete beams reinforced by GFRP bars.
In: Saadatmanesh H, Ehsani RM, editors. Proceedings of rst
international conference on composites in infrastructure, Tucson
(AZ); 1996. p. 31831.
[14] Faza SS, GangaRao HVS. Pre- and post-cracking deection behav-
iour of concrete beams reinforced with bre-reinforced plastic rebars.
In: Neale KW, Labossiere P, editors. Proceedings of 1st international
conference on advanced composite materials in bridges and structures
(ACMBS 1), CSCE, Sherbrooke, Que.; 1992. p. 15160.
[15] Saadatmanesh H, Ehsani MR. Fiber composite bar for reinforced
concrete construction. J Compos Mater 1991;25:188203.
[16] Nawy EG, Neuwerth GE. Fiberglass reinforced concrete slabs and
beams. J Struct Div 1977;103(ST2):42140.
[17] Theriault M, Benmokrane B. Eects of FRP reinforcement ratio and
concrete strength on exural behavior of concrete beams. J Compos
Construct 1998;2(1):716.
[18] Toutanji H, Deng Y. Deection and crack-width prediction of
concrete beams reinforced with glass FRP rods. Constr Build Mater
2003;17:6974.
[19] Alsyed HS, Alhozaimy MA. Ductility of concrete beams reinforced
with FRP bars and steel bers. J Compos Mater 1999;33(19):
1792806.
[20] Grace FN, Soliman KA, Abdel-Sayed G, Saleh RK. Behaviour and
ductility of simple and continuous FRP reinforced beams. ASCE J
Compos Constr 1998;2(4):18694.
[21] Naaman AE, Joeng SM. Structural ductility of concrete beams
prestressed with FRP tendons. In: Taerwe L, editor. Proceedings of
2nd International RILEM symposium (FRPRCS-2), non-metallic
(FRP) reinforcement for concrete structures, RILEM. London: E &
FN Spon; 1995. p. 37986.
[22] Jaeger GL, Mufti AA, Tadros G. The concept of the overall
performance factor in rectangular-section reinforced concrete beams.
In: Proceedings of the third international symposium on non-metallic
(FRP) reinforcement for concrete structures, vol. 2, Sapporo, Japan;
1997. p. 5518.
[23] Newhook J, Ghali A, Tadros G. Cracking and deformability of
concrete exural sections with ber reinforced polymer. ASCE J
Struct Eng 2001;128(9):1195201.
M.M. Ra et al. / Construction and Building Materials xxx (2006) xxxxxx 9
ARTICLE IN PRESS
Please cite this article in press as: Ra MM et al., Aspects of behaviour of CFRP reinforced concrete beams in bending, Constr Build-
ing Mater (2006), doi:10.1016/j.conbuildmat.2006.08.014