Advances in Engineering Software 67 (2014) 47–56

Contents lists available at ScienceDirect

Advances in Engineering Software

journal homepage: www.elsevier.com/locate/advengsoft

Numerical simulation of the shear behaviour of reinforced concrete

rectangular beam specimens with or without FRP-strip shear
reinforcement
G.C. Manos ⇑, M. Theofanous, K. Katakalos
Department of Civil Engineering, Aristotle University of Thessaloniki, Thessaloniki, Greece

a r t i c l e i n f o a b s t r a c t

To the memory of Professor Sir Gordon The successful validation of a numerical model is presented that can realistically approximate the shear
Higginson, Founder of the Dept. of behaviour of reinforced concrete (R/C) rectangular beams strengthened against shear with externally
Engineering Science Durham University, applied open hoop fibre reinforcing polymer (FRP) strips. For this purpose, the measured load-deforma-
Vice Chancellor of Southampton University, tion response of ten (10) full-scale R/C beam specimens is utilised. These specimens were loaded mono-
U.K.
tonically in a four-point bending arrangement up to failure. Open hoop FRP strip shear reinforcement was
applied externally to upgrade the shear capacity of eight (8) R/C beam specimens. Four of these speci-
mens had these FRP strips without anchorage, whereas for the other four the FRP strips were attached
Keywords:
together with novel anchoring devices. This successful numerical simulation predicts with a very good
Shear strengthening degree of approximation the observed load-deformation behaviour and the ultimate shear capacity of
Reinforced concrete beams all these specimens as well as the observed modes of failure including diagonal concrete cracking, deb-
Fibre reinforcing plastics onding of the FRP strips in the case of no anchoring, or the plastification of parts of the anchoring devices
Numerical simulation plus the adjacent crushing of the concrete.
Experimental behaviour Ó 2013 Elsevier Ltd. All rights reserved.
FRP strip anchoring

1. Introduction sufficient ductility [1,14,32]. This ductile behaviour is also desir-

Reinforced Concrete (R/C) structures represent a substantial
portion of the current infrastructure and building stock in most
countries worldwide. A large percentage of these were built several
decades ago and in some cases require retrofitting or structural
upgrading in order to conform to more stringent design regula-
tions, to accommodate additional live loads due to possible
changes in the building’s function or to make up for the loss of
strength brought about by environmental parameters that speed
up structural deterioration. Moreover, the rapid evolution of seis-
mic codes over the past two decades, which utilises the knowledge
gained following devastating earthquakes, necessitates the struc-
tural upgrading of most concrete structures that were built accord-
ing to older design regulations, following a different design
philosophy, in which the role of ductility was not central to the de-
sign process. On the contrary, all current seismic design codes pay
particular attention to enforce ductile behaviour on the various R/C
structural elements, which is accomplished by prohibiting shear
types of failure and favouring instead flexural types of failure with
⇑ Corresponding author. Address: Aristotle University of Thessaloniki, School of
Engineering, Department of Civil Engineering, Division of Structural Engineering,
Laboratory of Strength of Materials and Structures, Earthquake Simulator Facility,
Greece. Tel.: +30 2310 995653; fax: +30 2310 995769.
E-mail address: gcmanos@civil.auth.gr (G.C. Manos).
able when one is faced with the task of designing a retrofitting
scheme that aims at upgrading the seismic performance of R/C
structures designed by older design codes [30]. This paper deals
with the objective of upgrading the shear capacity of R/C beams,
that are poorly reinforced against shear, as part of such a retrofit-
ting scheme [25,26,31].
Among several upgrading techniques available, the utilisation
of FRP sheets or plates that can be attached externally to existing
structural components with the aid of resin has attracted consider-
able attention from both academics and practising engineers
[2,15,6,28]. Various researchers have derived models for the pre-
diction of the bond-slip response and required anchor lengths of
FRP sheets attached to concrete [36,37,35,43,22,42,41]. Depending
on the application, fibre reinforcing polymer (FRP) sheets can be
employed to increase the flexural strength, the shear strength
[38], the ductility and the serviceability performance of R/C struc-
tures, utilising these FRP sheets as additional external reinforcement.
Both carbon-CFRP or steel-SFRP sheets possess very high
strength in the orientation of their fibres; however only a small
part of the available strength can be exploited in practice, due to
the onset of debonding at relatively small strains, which can initi-
ate either at the ends of the FRP sheets, due to insufficient anchor-
ing or within the mid-parts of the FRP sheets, when anchoring is
provided at the FRP-strip ends, following the formation of cracks

0965-9978/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.advengsoft.2013.08.001
48 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56
in the concrete [5,7,10,11,39]. Debonding occurs either within the
resin layer itself that attaches the FRP sheet to the concrete surface
or, more frequently, below that resin layer within the adjacent re-
sin rich concrete substrate, the strength of which depends on the
concrete tensile strength and surface preparation. Hence, central
to the successful application of FRPs on existing structures is the
preparation of the concrete substrate, which often includes the
partial removal of damaged concrete cover and sandblasting so
that such FRP debonding is postponed. Alternative techniques aim-
ing at mechanically strengthening the substrate have also been
proposed [20]. The quality of the concrete substrate linked with
surface preparation strongly depends on the working conditions
and craftsmanship and hence there is an inherent high uncertainty
in each application employing FRPs as a means of structural
upgrading, which often leads to overly conservative design. There-
fore, the development of innovative means for effective anchorage
of the FRP sheets/plates in the concrete volume, which ensures the
satisfactory transfer of forces from the FRP sheet to the concrete
part of the structural member and allows for a more efficient util-
isation of the FRP strength, is warranted [24–28].
In the present paper, a novel anchorage device [24] for FRP
sheets employed as external shear reinforcement, which was re-
cently developed at the Laboratory of Strength of Materials and
Structures of Aristotle University of Thessaloniki is studied numer-
ically. Its effectiveness has been experimentally verified
[21,25,26,31] The aim of the present study is to replicate numeri-
cally the experimentally obtained results and to identify the key
parameters controlling the structural response of reinforced con-
crete beams strengthened against shear by external FRP reinforce-
ment, that will allow the optimisation of the anchorage device and
hence the more effective utilisation of the FRP sheets, thus leading
to more efficient design.
2. Experimental Investigation
Several theoretical and experimental studies [9,44,4] have been
carried out to analyze the phenomenon of the shear failure of rein-
forced concrete (RC) beams. Russo et al. [33] presents a compre-
hensive review of various proposed methods for predicting the
shear behaviour of reinforced concrete beams without any trans-
verse reinforcement. In the design of R/C beams, this is usually con-
sidered as an initial condition that is supplemented by the strength
provided by the transverse reinforcement in order to predict the
total shear capacity of such a structural element. Following this
rationale a comprehensive experimental program has been de-
signed and carried out at the Laboratory of Experimental Structural
Mechanics at the Aristotle University of Thessaloniki aimed to as-
sess the structural efficiency and effectiveness of shear strength
provided to R/C beams by externally attached open hoop FRP trans-
verse shear reinforcement. Initially, a number of R/C beam speci-
mens with only longitudinal reinforcement were tested. Identical
R/C beam specimens were then tested which were provided with
open hoop FRP strips external shear reinforcement in order to
study in this way the additional shear strength that can be attained
by such a strengthening scheme. Many researchers examined the
effectiveness of this type of shear reinforcement in the form of
closed hoop FRP strips attached externally on rectangular R/C
beam specimens [5,7,10,11]. However, although such closed hoop
FRP strips are very effective they cannot be applied in practice as
most R/C beams are accompanied by an R/C slab that makes such
a closed hoop application almost impossible. Consequently, the full
experimental program conducted at Aristotle University included
either rectangular R/C beams or R/C T-beams being strengthened
by open hoop FRP strips, as will be described in summary below.
This open hoop FRP strip shear reinforcement was examined in
two distinct forms. First, these open hoop FRP strips were simply
attached on the external surface of the R/C specimens. This was
done following the proper construction technique and using the
proper resin recommended by current practice. In this case, as
has been established by previous research [5,7,10,11], as well as
by the current experimental sequence [25,26,28], the prevailing
mode of failure is the debonding of the FRP strips which renders
this type of shear reinforcement ineffective. This is because the
FRP strips’ debonding occurs long before the tensile strength
capacity of such FRP strips is reached. In order to prohibit such a
premature debonding type of failure a novel anchorage device
[24–27] is introduced together with the open hoop FRP strips that
can be attached relatively easily at R/C T-beams, which is the pri-
mary practical application. In this Section, key experimental results
dealing with the upgrading of the shear capacity of R/C rectangular
beams are briefly presented. The observed performance from this
experimental sequence will be utilised in the subsequent numeri-
cal simulation. More information on the experimental sequence
and results is included in the work published so far [25–27,31].
2.1. Beam tests
A total of ten rectangular beam specimens were subjected to
four point bending as part of the present study to assess the effec-
tiveness of the proposed anchorage device. The mean dimensions
and reinforcement details of these specimens are shown in Fig. 1,
whilst the employed test arrangement and instrumentation is de-
picted in Fig. 2. Their cross-section was 120 mm wide and 360 mm
deep (Fig. 1). Their overall span was 2700 mm whereas the shear
span was equal to 900 mm (1/3 of the total length). These rectan-
gular section beam specimens were simply supported and loaded
up to failure.
Initially, two concrete beam specimens both without any FRPs,
namely CRBs with Ø8/250 steel closed stirrups as shear reinforce-
ment, and CRB without shear reinforcement, were tested to obtain
the basic structural response. These two tests were utilised in the
present numerical study to calibrate the material parameters of
the concrete material model. Subsequently, four rectangular con-
crete beam specimens without any steel stirrups were reinforced
against shear by externally attaching U-shape open-hoop FRP strips
with the aid of resin (Fig. 3a); these specimens were tested to assess
the effectiveness of such a conventional FRP strengthening tech-
nique without the application of additional anchorage. Two types
of FRPs, namely CFRPs and SFRPs and two spacings (150 mm and
200 mm) were considered. Finally, another four specimens with
the same open-hoop FRP strips arrangement as before but with
additional anchorage were tested. This anchorage was provided
by the novel anchorage device [24] depicted in some detail in
Fig. 3a and b. In this anchoring device a steel rod is used to wrap
around it the FRP strip; this steel rod is secured by a steel plate of
equal length which is bolted by two anchor bolts as shown in Fig. 3b.
Table 1 lists all the particular aspects that characterise each one
of these specimens. Instrumentation was provided to monitor the
applied load as well as the deformation of these specimens. The
cross-sections and conventional steel reinforcement details were
identical for all specimens (Fig. 1).
2.2. Material tests
A series of standard tests have been carried out in order to
establish the basic material properties for these beam specimens;
e.g. basic material properties for the concrete, the reinforcing bars,
the FRP strips as well as the steel plates and steel bolts employed in
the anchoring device. These experimentally derived material prop-
erties were utilised in the subsequent numerical study. The mea-
sured mean concrete cylindrical strength for each specimen is
 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56 49

3+3 rebars (top+bottom) Section A-A

(d=20mm)

3+3 rebars
(d=20mm)
transverse rebars (stirrups)
3+3 rebars (top+bottom)
(d=8mm/250mm)
(d=20mm)
A

Fig. 1. Main dimensions and reinforcement details of the tested specimens.

Pshear Pshear

Strain gauge
LVDT

LVDT

FRP strips

Fig. 2. Employed loading arrangement and instrumentation.

listed in column 3 of Table 1. The longitudinal and transverse rein-

a Detail of Patented Anchorage forcement were grade S500, whilst the steel plates were grade
Steel rod used to wrap arround the S235. The bolts utilised in the anchoring device were manufac-
FRP strip tured by HILTI and are designated as HILTI HUS herein. The full
Simply attached
FRP strip stress–strain response of all steel parts was experimentally deter-
mined and utilised in the numerical studies. The experimentally
determined Young’s modulus E, yield strength fy and ultimate
strength fu are reported in Table 2.
Anchor Finally, unidirectional tension tests were conducted on resin
Bolt rich CFRP and SFRP sheets up to failure to obtain the Young’s mod-
ulus (E1) and strength (fu1) in their fibre orientation and the Pois-
son’s ratio (m12 = (E1/E2)m21). The remaining two material
Anchored FRP strip properties pertaining to the orthotropic nature of the material un-
attached on the der the plane stress assumption, namely the transverse Young’s
concrete surface modulus E2 and the shear modulus G12, were derived by applying
R/C Beams the conventional rule of mixture for composites [18], taking into
account the manufacturer’s specifications for the resin (citation
b Detail of Patented Anchorage SIKA) and the FRP (citation SIKA) sheets and the experimentally
determined volume fraction of resin and FRP sheets.
Steel rod used to wrap arround the FRP strip
Both the experimentally and the analytically determined mate-
rial properties are reported in Table 3. It should be noted that the
analytical values obtained by applying the rule of mixture for E1
and v12 compare very well (discrepancy within 5%) with the exper-
Anchor imental ones.
Anchor
Bolts FRP Bolt
Steel
strip plate 2.3. Summary of observed behaviour – modes of failure
Anchored FRP strip
attached on the The key findings of all beam tests are summarised in Table 1,
concrete surface where the ultimate shear force (expVmax) resisted by these speci-
R/C Beam
mens is listed in column 7 of this table. This shear force value
Fig. 3. Cross section of the tested rectangular beam specimens with the reinforce- equals half of the totally applied external load. As can be seen,
ment details and the proposed anchoring device (dimensions in mm). the shear strength of the control rectangular beam (CRB) without
50 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56

Table 1
Summary of beam tests.

Designation Shear reinforcement Mean Concrete Width of each FRP Spacing of FRP Anchorage Measured Ultimate shear Failure mode
Cylindrical Strength strip (mm) strips (mm) capacity expVmax (kN)
(MPa)
(1) (2) (3) (4) (5) (6) (7) (8)
CRB None 22.40 – – None 39.4 Shear
cracking
CRBs Steel Stirrups 25.92 – – ’’ 90.9 Yielding of
d = 8 mm / 250 mm stirrups
RB200C Open hoop CFRP 22.40 100 200 ’’ 97.8 Debonding
strips (1 layer)
RB200Ca Open hoop CFRP 22.40 100 200 Patented 115.1 Premature
strips (1 layer) FRP rupture
RB200S Open hoop SFRP 25.92 100 200 None 94 Debonding
strips (1 layer)
RB200Sa Open hoop SFRP 25.92 100 200 Patented 122 Anchorage
strips (1 layer) failure
RB150C Open hoop CFRP 22.40 100 150 None 101 Debonding
strips (1 layer)
RB150Ca Open hoop CFRP 22.40 100 150 Patented 123 FRP rupture
strips (1 layer)
RB150S Open hoop SFRP 22.40 100 150 None 103 Debonding
strips (1 layer)
RB150Sa Open hoop SFRP 22.40 100 150 Patented 119 Anchorage
strips (1 layer) failure

Nominal thickness of 1 layer: CFRP 0.131 mm, SFRP 0.1184 mm.

Table 2
Material properties of employed steel plates, rebars and bolts.

Material property Longitudinal rebars of d = 20 mm diameter Transverse rebars (stirrups) d = 8 mm Steel plates of anchoring device Anchor bolts Hilti HUS
E (MPa) 202,700 203,600 210,000 210,000
fy (MPa) 527 570 255 800
fu (MPa) 645 710 390 917

7.6% when using CFRP strips without anchoring devices and 34.2%
Table 3
when using SFRP strips with anchors. The beneficial effect of
Material properties of CFRP and SFRP.
anchoring the U-shape open-hoop FRP strips in order to inhibit
Material property E1 (MPa) m12 fu1 (MPa) E2 (MPa) G12 (MPa) the debonding mode of failure is thus demonstrated. The subse-
CFRP 41,784 0.305 376 4476 1935 quent numerical study will focus on these important features of
SFRP 36,350 0.312 330 4400 1900 the observed behaviour, that is the prevailing mode of failure
and the measured ultimate shear capacity (expVmax).

any stirrups or CFRP/SFRP strips was recorded equal to 39.4 kN,
whereas the shear capacity of specimen CRBs with stirrups Ø8/
250 mm as transverse reinforcement was increased to 90.9 kN.
From the results shown in Table 1, it is evident that the maximum
recorded shear capacity for all strengthened rectangular beams
with CFRP/SFRP strips was significantly greater than the shear
capacity of either the CRB or the CRBs specimens. The observed
mode of failure is listed in column 8 of Table 1. As can be seen,
when FRP strips are without anchors the debonding mode of fail-
ure of the FRP strips prevails. This debonding mode of failure limits
the observed shear capacity increase to 138% (nearly 2.4 times lar-
ger than the initial value of 39.4 kN), when the shear capacity of
the strengthened specimens is compared to that of specimen CRB
(without any shear reinforcement). When anchoring devices were
utilised for prohibiting this debonding mode of failure they re-
sulted in shear capacity increase of up to 210% (3.1 times larger
than the initial value of 39.4 kN). When the shear capacity of the
strengthened specimens is compared to that of the virgin specimen
with nominal transverse reinforcement in the form of steel stirrups
(CRBs), the resulting maximum shear capacity increase is equal to
3. Numerical analysis
The general purpose FE software ABAQUS [19] was employed to
generate FE models to simulate numerically the structural re-
sponse of the previously described concrete beams strengthened
with externally attached FRP sheets. The generated models were
validated against all respective experimental results. The aim of
this part of the study is to generate reliable FE models that can
be utilised to enhance the understanding of the fundamental struc-
tural response of the FRP shear strengthening scheme of R/C beams
with and without anchorage devices and hence optimise the
anchorage device design.
3.1. Modelling assumptions
The mean measured specimen geometries were utilised to
simulate the test specimens. The full 3D geometry of the concrete
volume and the steel plates employed for the anchorage of the FRP
was modelled with 3-D brick-type F.E. elements, whilst the FRP
sheets were idealised as planar F.E. elements and their actual
 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56
thickness (0.75 mm) was utilised as a section property. The bolts
were assumed cylindrical with a diameter equal to the equivalent
diameter corresponding to the net bolt area (i.e. the threaded part
of the bolts was not explicitly modelled). A further idealisation per-
tinent to the bolt assembly simulation is the assumption that the
bolt hole diameter of the steel plate is equal to the assumed bolt
diameter (i.e. the clearance has not been accounted for).
In order to reduce the computational cost, the symmetry of the
structure with respect to the x–y and z–y mid-planes was
exploited, by numerically simulating one quarter of the physical
model and applying the appropriate boundary conditions at each
of the two planes of symmetry as shown in Fig. 4. The load was ap-
plied incrementally as prescribed displacement on a steel plate
(SP1) at the point of load application, which was applied iteratively
utilising a smooth amplitude curve available in ABAQUS. Restraint
of the structure in the direction of the prescribed displacement was
provided by fixing the uy at the supporting steel plate (SP2). All the
restraints introduced to the model are shown in Fig. 4. The sum of
the reaction forces in the y direction at SP2 yields reaction Ry that
corresponds to half of the shear force that is exerted to the model
structure (e.g. expVmax Table 1).
The contact between the various parts of the model (steel bolts,
steel plates and concrete volume) was explicitly modelled and a
friction coefficient equal to 0.3 was assumed for tangential contact
behaviour except for the contact between the concrete beam and
the loading/reaction plates which was assumed frictionless. Varia-
tion of the friction coefficient applied to the bolt-concrete interface
between the idealised frictionless and rough contact extreme cases
was attempted, but no significant effect on the overall solution was
observed. Moreover, in order to reduce computational cost, the ac-
tual detail of the cylindrical rod of the anchoring device through
which load is transferred from the FRP strips to the concrete beam
through the steel plate was somewhat simplified (see Figs. 3b and
5); towards this end, the displacements of the FRP’s edge were con-
strained to be equal to the respective displacements of the steel
plate’s mid plane for the numerical representation of the anchoring
device. The interface between the FRP and the concrete was simu-
lated as a cohesive zone endowed with a suitable traction separa-
tion response, the properties of which are discussed below.
3.2. Material modelling
The auxiliary test results reported previously were utilised
herein to define the material response. The SFRP/CFRP sheets were
assumed orthotropic under plane stress conditions and the mate-
rial properties reported in Table 3 were adopted. Moreover, due
to some specimens failing by FRP rupture, damage and failure of
the FRPs was explicitly modelled by adopting the Hashin damage
Fig. 4. Modelled geometry and applied boundary conditions.
51
model [17] incorporated in ABAQUS. Only tensile failure in the
principal direction was considered and the ultimate stress reported
in Table 3 was adopted, whilst other failure modes were consid-
ered irrelevant to the present study. A fracture energy equal to
0.01 was assumed to simulate damage evolution, which was rapid,
due to the severe stress concentration following the onset of fail-
ure. Steel plates, longitudinal reinforcing bars, steel stirrups and
steel bolts and plates were assumed elastic-isotropic hardening
and the experimentally obtained stress–strain curves for the steel
used for these parts were converted into the true stress-logarith-
mic plastic strain format according to Eqs. (1) and (2) and utilised
to define the material response.
rtrue ¼ rnom ð1 þ enom Þ ð1Þ
rtrue
eplln ¼ lnð1 þ enom Þ  ð2Þ
E
The damaged plasticity model for concrete available in the ABA-
QUS material library was adopted to model concrete response,
since it has been shown to perform satisfactorily in similar applica-
tions [23]. All material parameters were initially calibrated to
accurately simulate the response of the two control beam speci-
mens CRB and CRBs. The values adopted for the relevant material
parameters, namely the angle of dilation h, the eccentricity, the ra-
tio of equibiaxial to uniaxial compressive stress fb0/fc0, the ratio of
the second stress invariant on the tensile meridian to that on the
compressive meridian at initial yield Kc and the viscosity parame-
ter were 40, 0.1, 1.16, 0.666 and 0 respectively and were all within
the range encountered in literature [23]. In accordance with similar
studies [29], the stress strain response of concrete in compression
was derived according to Saenz [34], whilst the tension stiffening
response was defined in terms of tensile stress and axial deforma-
tion according to Cornelissen et al. [12], assuming a fracture energy
equal to 0.06. The Young’s modulus E and tensile strength fct were
determined according to ACI [1] as a function of the compressive
cylindrical strength, whilst Poisson’s ratio was assumed equal to
0.2. The comparison between this assumed in the numerical simu-
lation stress–strain concrete behaviour and the one measured by
testing concrete cylindrical specimens taken during the construc-
tion of the beam specimens is depicted in Fig. 6. It must be borne
in mind that these specimens were simply loaded in compression
in such a way that the descending branch of the stress–strain curve
could not be obtained.
The resin rich layer, within which the debonding of the FRP strip
from the concrete volume occurs, was modelled as a cohesive zone
[3] endowed with a traction-seperation response. The initial elastic
stiffness for all three modes of debonding was set equal to 106 N/
mm2, which is within the range of values proposed by Turon et al.
[40]. The quadratic stress-based damage initiation criterion
52 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56

Fig. 5. Finite element discretisation of the rectangular beam specimens and close-ups of the discretization of the anchoring device.

in Fig. 5. Due to severe convergence problems typically associated

with strongly nonlinear response and with materials exhibiting
non-monotonic stress–strain response [13], such as concrete and
cohesive elements, the explicit dynamics solver ABAQUS/EXPLICIT
was employed to perform the nonlinear analyses. Quasi-static re-
sponse was achieved by specifying a slow displacement rate and
checking that the kinetic energy was smaller than 2% of the internal
energy for the greatest part of the analysis.

4. Results and discussion

The F.E. numerical predictions regarding the ultimate shear

Fig. 6. Measured and assumed concrete stress-strain behaviour in unconfined uni- capacity (numVmax) and failure modes with the ones observed dur-
axial compression.
ing testing (see Table 1) are summarised in Table 4.
The ratio of numerically predicted over measured ultimate
shear capacity is listed in column 3 of Table 4 with values very
available in ABAQUS was adopted, with the stress limit in all three
close to 1.0, which signifies a very good agreement of the
principal directions being equal to the respective tensile concrete
strength. Damage evolution was defined in terms of fracture en-
ergy with linear softening, whilst mixed mode behaviour was ac- Table 4
counted for according to the model proposed by Benzeggagh and Summary of Numerical results.
Kenane [8], with a power coefficient equal to 1.45, in accordance num
Designation Numerical Vmax/expVmax Numerical Observed
with the proposals of Obaidat et al. [29]. Best results were obtained ultimate shear failure failure
for a fracture energy in both shearing and peeling modes equal to capacity mode mode
num
0.25 (GII = GIII = 0.25 N/mm2), whilst the fracture energy GI associ- Vmax (kN)
(1) (2) (3) (4) (5)
ated with the opening mode was set equal to 0.025. It should be
noted that overall response is relatively insensitive to the assumed CRB 39.44 1.00 Shear Shear
elastic stiffness and limiting stress adopted, but these properties cracking cracking

have a significant effect on the numerical cohesive zone length CRBs 87.89 0.97 Yielding of Yielding of
[16,40], which defines the minimum required mesh size. stirrups stirrups
RB200C 97.99 1.00 Debonding Debonding

3.3. Non-linear analysis and discretisation RB200Ca 118.79 1.03 Anchorage Premature
failure FRP
rupture
Linear 8-noded brick elements were adopted for the discretisa-
RB200S 87.54 0.93 Debonding Debonding
tion of the concrete beam, steel bolts, steel plates and loading
plates, whilst the longitudinal reinforcing bars and the steel stirrups RB200Sa 112.47 0.92 Anchorage Anchorage
failure failure
were discretised with linear truss elements embedded in the con-
crete region (i.e. no relative displacement between reinforcement RB150C 99.65 0.99 Debonding Debonding

and concrete was allowed for). Linear 4-noded shell elements were RB150Ca 120.58 0.98 FRP FRP
used to discretise the FRP sheets, the degrees of freedom (DOF) of rupture rupture

which were tied to the respective DOFs of the underlying 8-noded RB150S 96.95 0.94 Debonding Debonding
3D cohesive elements. The mesh size was dictated by the attempt RB150Sa 123.52 1.04 Anchorage Anchorage
to minimise computational time whilst maintaining accuracy and, failure failure
upon extensive mesh convergence studies, a uniform mesh size of Mean value
15 mm was adopted for the concrete beam, cohesive zone and 0.98
FRP sheets, whilst smaller mesh sizes were adopted for the steel Coef. of
variation 0.05
plates, bolts and concrete regions surrounding the bolts as shown
 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56 53

numerically predicted ultimate shear capacity values with the
respective experimental ones included in Table 1. Moreover, the
numerically predicted modes of failure are also in very good agree-
ment with the ones observed during testing, with the exception of
specimen RB200Ca. This discrepancy can be attributed to the pre-
mature failure of the CFRP strip due to the non-ideal preparation of
the concrete surface at the bottom fibre of this specimen prior to
gluing this CFRP strip. Detailed discussion for each of the three
types of rectangular beams considered in the present study
follows.
4.1. Specimens of rectangular beams without FRP sheets
The observed behaviour of the two control beam specimens,
which were initially tested to obtain the basic structural response
of the reinforced concrete rectangular beam specimens without
the added complexity of the attached FRP sheets, was, in general,
well-predicted by the FE numerical models, as can be seen in
Fig. 7, where the experimental and numerical load–displacement
response is depicted. The load V refers to the applied shear force
and equals the half of the totally applied external load (Fig. 2),
whilst the d is the corresponding average displacement at the
points of load application. In Fig. 8 the experimental failure mode,
in the form of the dominant shear crack, and the numerical failure
mode of specimen CRB are compared. Specimen CRBs displays a
similar failure mode at a much higher load, with a major diagonal
crack forming upon the yielding of the stirrups. It should be noted
that in all these figures depicting numerical failure modes, the
maximum principal plastic strain of concrete is depicted as a
means to visualise concrete cracking since no discrete cracks form
during the adopted numerical analysis.
4.2. FRP-strengthened beams without anchorage
The experimental and numerical load–displacement response
of the concrete beam specimens strengthened with FRP sheets
without the provision of additional anchoring devices is depicted
in Fig. 9 for specimens RB200C and RB150S, respectively; as can
be seen, an excellent agreement between the experimental and
the numerical results is achieved. All these FE numerical simula-
tions of the observed behaviour exhibited almost identical struc-
tural response, regardless of the employed FRP material (i.e. CFRP
or SFRP) and the spacing considered. The beam specimens behaved
elastically prior to the formation of the initial flexural cracks,
whereupon the flexural reinforcement prevented the opening of
the cracks. Subsequently, a diagonal crack started to form beneath
the point of application of the load, which was interrupted by the
presence of the FRP sheets which were stressed and prevented the
opening of the diagonal crack. Finally, the debonding of the FRP
sheets commenced and propagated rapidly, leading to the

Fig. 7. Experimental and numerical load-displacement curves for specimens CRB and CRBs.

Fig. 8. Experimental and numerical failure modes for specimen CRB.

Fig. 9. Experimental and numerical load-displacement curves for specimens RB200C and RB150S.
54 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56

formation of critical diagonal cracks and thus to failure. This ob-

Fig. 10. Flexural cracks for specimen RB200S at d = 1.1 mm.
served behaviour was also accurately reproduced by the previously
described numerical simulations.
The debonding occurred within a thin resin rich layer of con-
crete, the properties of which were assumed to be incorporated
in the cohesive zone material properties. Typical numerical flexural
cracks and the corresponding axial stresses in the reinforcing bars
are depicted in Fig. 10, whilst the crack pattern and the FRP stres-
ses in their principal direction at ultimate load are shown in
Fig. 11. The numerical failure mode is compared with the corre-
sponding experimental one for specimen RB200S in Fig. 12.

Fig. 11. Crack pattern and FRP stresses for specimen RB200S at ultimate load.

Fig. 12. Experimental and numerical failure modes for specimen RB200C.

Fig. 13. Experimental and numerical load-displacement curves for specimen RB200Sa and detail of localised deformation at bolt hole.

Fig. 14. Experimental and numerical cracking patterns for specimen RB200Ca.
 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56
4.3. FRP-strengthened beams with anchorage device
As described in Sections 4.1 and 4.2 all FRP-strengthened spec-
imens to which no additional anchorage was provided failed
abruptly, due to debonding of the FRP sheets. The provision of
additional anchorage by means of the device shown in Fig. 3b al-
lows for additional forces to be carried by the FRPs once debonding
has occurred; these forces are transferred by the anchoring device
finally through the bolts into the concrete thus resulting in an in-
creased ultimate shear capacity and thus in a more efficient utilisa-
tion of the FRP material. The anchorage zone is stressed from the
onset of loading but deformations and stresses remain small as
long as the resin remains effective. Upon debonding, the anchorage
zone experiences high localised stresses. The failure is no longer
due to FRP debonding but rather localised above the anchorage
zone, due to the high stresses transmitted by the bolts in the vicin-
ity of the bolt holes. The failure mode is complex and involves a
combination of local concrete cracking and crushing near the bolt
holes, flexure of the compressive reinforcement due to the forces
exerted by the bolts and possibly pullout of the anchoring bolts.
In some cases, FRP rupture may occur prior to the failure of the
anchorage device or anchorage zone. The generated FE models
were in general able to capture both the FRP rupture and localised
cracking above the anchorage zone. Fig. 13 depicts the experimen-
tal and numerical V-d response for specimen RB200Sa and the cor-
responding numerical failure mode which is principally extensive
cracking and crushing of the concrete in the vicinity of the bolt
holes. This was also observed during testing that led to the pullout
of the anchor bolts. Fig. 14 displays the experimental and numeri-
cal failure modes for RB200Ca, which involves localised cracking in
the vicinity of the anchorage zone which evolves into critical diag-
onal cracks.
5. Conclusions
1. The successful numerical simulation of the application of a
novel anchoring device together with open hoop FRP strips,
which enhances the efficiency of these FRP sheets applied as
shear reinforcement for reinforced concrete rectangular beams,
was demonstrated by this study.
2. The validation of the numerical simulation is based on the mea-
sured behaviour of two (2) full-scale reinforced concrete rectan-
gular beams with or without nominal shear reinforcement as
well as with eight (8) full-scale reinforced concrete rectangular
beams that employed open hoop FRP strips, with or without
novel anchoring devices, as a means of upgrading their shear
capacity.
3. The success of the deployed numerical simulation is fully dem-
onstrated by comparing first the observed and numerically pre-
dicted ultimate shear capacity for all these specimens. The
achieved accuracy of the predicted ultimate shear capacity is
excellent.
4. The success of the deployed numerical simulation is fully dem-
onstrated by then comparing the observed and numerically pre-
dicted mode of failure for each one of the ten (10) tested
specimens. It is shown that the deployed numerical simulations
have been articulated with all the significant features, to be able
to portray in a very realistic manner complex non-linear mech-
anisms such as the cracking and crushing of concrete, the deb-
onding and fracture of the FRP strips and the yielding and
fracture of the steel parts of the novel anchoring devices.
5. The success of the deployed numerical simulation also extends
to the shape of the load–deflection curves when they are com-
pared with the ones obtained during testing. Within the scope
of the present paper, the studied beam specimens were sub-
jected to static monotonic loads.
55
6. From the above, it must be seen that the accurate numerical
prediction of a relatively complex behaviour is the result of all
these intelligent features that were built into the numerical
model together with the accurate representation of all the indi-
vidual material features which were established by careful test-
ing prior to the numerical simulation. The successful FE models,
accurately capturing the basic structural response in terms of
ultimate load, failure modes and overall load–displacement
response, have been developed and discussed.
7. Based on the validated FE numerical models presented here,
further research is underway to identify the key features affect-
ing overall structural response and thus to optimise the struc-
tural arrangement of the proposed novel anchoring device in
terms of strength and ductility.
8. A further goal is to derive a generic strengthening technique
and associated design guidelines, which allow for a more effi-
cient utilisation of the FRPs’ strength and significantly enhance
the structural performance of existing structures strengthened
with FRP sheets for both static and seismic type loads.
Acknowledgements
The carbon fibres and epoxy resins were provided by Sika
Hellas. The steel fibres are not commercially available and were
provided for the present study by Bekaert Industries. Partial finan-
cial support for this investigation was provided by the Hellenic
Earthquake Planning and Protection Organization (EPPO). The
anchoring device employed in this study, is patented under Patent
No. WO2011073696, which is managed by the Research Commit-
tee of Aristotle University.
References
[1] ACI. Committee 318. Building code requirements for structural concrete and
commentary (ACI 318-99). Detroit (MI): American Concrete Institute; 1999.
[2] ACI. Guide for the design and construction of externally bonded FRP systems
for strengthening concrete structures (ACI 440.2R-08). In: ACI 440.2R-08.
American Concrete Institute: Farmington Hills; 2008. p. 45.
[3] Alfano G, Crisfield MA. Finite element interface models for the delamination
analysis of laminated composites: mechanical and computational issues. Int J
Numer Methods Eng 2001;50(7):1701–36.
[4] Angelakos D, Bentz EC, Collins MP. Effect of concrete strength and minimum
stirrups on shear strength of large members. ACI Struct J 2001;98(3):290–300.
[5] Arduini M, Di Tommaso A, Nanni A. Brittle failure in FRP plate and sheet
bonded beams. ACI Struct J 1997;94(4):363–70.
[6] Bakis et al. Fiber-reinforced polymer composites for construction – state of the
art review. J Compos Constr 2002 (May).
[7] Barton B, Wobbe E, Dharani LR, Silva P, Birman V, Nanni A, et al.
Characterization of reinforced concrete beams strengthened by steel
reinforced polymer and grout (SRP and SRG) composites. Mater Sci Eng A
2005;412(1–2):129–36.
[8] Benzeggagh ML, Kenane M. Measurements of mixed-mode delamination
fracture toughness of unidirectional glass/epoxy composites with mixed-mode
bending apparatus. Compos Sci Technol 1996;56(4):439–49.
[9] Bresler B, Scordelis AC. Shear strength of reinforced concrete beams. ACI J
1963;60(1):51–74.
[10] Casadei P, Nanni A, Alkhrdaji T, Thomas J. Performance of double-T prestressed
concrete beams strengthened with steel reinforcement polymer. Adv Struct
Eng 2005;8(4):427–41.
[11] Chen JF, Teng JG. Shear capacity of fiber-reinforced polymer-strengthened
reinforced concrete beams: fiber reinforced polymer rupture. J Struct Eng
2003;129(5):615–25 [May 1, ASCE, 5].
[12] Cornelissen HAW, Hordjik DA, Reinhardt HW. Experimental determination of
crack softening characteristics of normal weight and ligthweight concrete.
HERON 1986;31(2):45–66.
[13] Crisfield MA, Wills J. Solution strategies and softening materials. Comput
Methods Appl Mech Eng 1988;66(3):267–89.
[14] Eurocode 8. Design of structures for earthquake resistance. DRAFT No 3.
European Committee for Standardization; 2004.
[15] FIB. I.F.f.S.C., Externally bonded FRP reinforcement for RC Structures. In:
Bulletin 14, f.T.G. 9.3, editor; 2001. p. 165.
[16] Harper PW, Hallett SR. Cohesive zone length in numerical simulations of
composite delamination. Eng Fract Mech 2008;75(16):4774–92.
[17] Hashin Z. Failure criteria for unidirectional fiber composites. J Appl Mech
1980;47(2):329–34.
56 G.C. Manos et al. / Advances in Engineering Software 67 (2014) 47–56
[18] Hashin Z. Analysis of composite materials – a survey. J Appl Mech, Trans ASME
1983;50(3):481–505.
[19] Hibbitt, Karlsson, Sorensen. Inc. ABAQUS user’s manual volumes I–V and
ABAQUS CAE manual. Version 6.10.1. Pawtucket, USA; 2010.
[20] Kalfat R, Al-Mahaidi R. Investigation into bond of a new CFRP anchorage
system for concrete utilizing mechanically strengthened substrate. Compos
Struct 2010;92(11):2738–46.
[21] Katakalos K, Papakonstantinou CG, Manos GC. Comparison between carbon
and steel fiber reinforced polymer with or without anchorage. In: 6th CICE,
Rome; 2012.
[22] Lu XZ, Teng JG, Yea LP, Jiang JJ. Bond–slip models for FRP sheets/plates bonded
to concrete. Eng Struct 2005;27:920–37.
[23] Malm R. Predicting shear type crack initiation and growth in concrete with
non-linear finite element method. PhD thesis, Department of Civil and
Architectural Engineering, Royal Institute of Technology (KTH) Stockholm;
2009.
[24] Manos GC, Katakalos K, Kourtides V. Construction structure with
strengthening device and method. European Patent Office, Patent Number
WO2011073696 (A1); 2011 (23.06.11).
[25] Manos GC, Katakalos K, Papakonstantinou CG. Enhanced repair and
strengthening of reinforced concrete (R/C) T-beams utilizing external
reinforcement of fiber polymer sheets and novel anchoring devices.
15WCEE, Lisboa; 2012.
[26] Manos GC, Katakalos K, Koidis G, Papakonstantinou CG. Shear strengthening of
R/C beams with FRP strips and novel anchoring. J Civ Eng Res 2012;2(6):73–83.
http://dx.doi.org/10.5923/j.jce.20120206.04.
[27] Manos GC, Katakalos K, Kourtides V. Cyclic behaviour of a hybrid anchoring
device enhancing the flexural capacity and ductility of an R/C bridge-type pier
strengthened with CFRP sheets. J Civ Eng Res 2012.
[28] Manos GC, Katakalos K. The use of fiber reinforced plastic for the repair and
strenghtening of existing reinforced concrete structural elements damaged by
earthquakes. In: Masuelli Martin Alberto, editor. Published in the book, Fiber
Reinforced Polymers – The Technology Applied for Concrete Repair; January
23, 2013, ISBN: 978-953-51-0938-9 [chapter 3].
[29] Obaidat YT, Heyden S, Dahlblom O. The effect of CFRP and CFRP/concrete
interface models when modeling retrofitted RC beams with FEM. Compos
Struct 2010;92(6):1931–8.
[30] EPPO. Hellenic earthquake planning and protection organization. Athens:
Guidelines for Retrofitting in Reinforced Concrete Buildings; 2011.
[31] Papakonstantinou CG, Katakalos K, Manos GC. Reinforced concrete T-beams
strengthened in shear with steel fiber reinforced polymers. In: 6th CICE, Rome;
2012.
[32] Paz, Mario. International handbook of earthquake engineering: codes,
programs and examples. In: Greece by G.C. Manos, editor. Chapman and
Hall; 1994. ISBN: 0-412-98211-0 [chapter 17].
[33] Russo G, Somma G, Mitri D. Shear strength analysis and prediction for
reinforced concrete beams without stirrups. J Struct Eng 2005;66(January).
[34] Saenz LP. Discussion of ‘‘equation for the Stress–strain curve for concrete’’ by
Desayi P, Krishnan S. ACI J 1964;61:1229–35.
[35] Smith ST, Teng JG. FRP-strengthened RC beams. I: Review of debonding
strength models. Eng Struct 2002;24(4):385–95.
[36] Täljsten B. Strengthening of concrete prisms using the plate-bonding
technique. Nonlinear fracture mechanics derivations applied to plates of
steel or carbon fibre, bonded to concrete and loaded in shear. Int J Fract
1996;82(3):253–66.
[37] Täljsten B. Defining anchor lengths of steel and CFRP plates bonded to
concrete. Int J Adhes Adhes 1997;17(4):319–27.
[38] Täljsten B, Elfgren L. Strengthening concrete beams for shear using CFRP
materials: evaluation of different application methods. Compos Part B: Eng
2003;31(2):87–96.
[39] Teng JG, Smith ST, Yao J, Chen JF. Intermediate crack-induced debonding in RC
beams and slabs. Constr Build Mater 2003;17(6–7):447–62.
[40] Turon A, Dávila CG, Camanho PP, Costa J. An engineering solution for mesh size
effects in the simulation of delamination using cohesive zone models. Eng
Fract Mech 2007;74(10):1665–82.
[41] Wu Y, Zhou Z, Yang Q, Chen W. On shear bond strength of FRP-concrete
structures. Eng Struct 2010;32:897–905.
[42] Yao J, Teng JG, Chen JF. Experimental study on FRP-to-concrete bonded joints.
Compos Part B: Eng 2005;36(2):99–113.
[43] Yuan H, Teng JG, Seracino R, Wu ZS, Yao J. Full-range behaviour of FRP-to-
concrete bonded joints. Eng Struct 2004;26(5):553–65.
[44] Zsutty TC. Beam shear strength prediction by analysis of existing data’’. ACI J
1968;65(11):943–51.