JOURNAL OF

COMPOSITE
Article M AT E R I A L S

Journal of Composite Materials

2021, Vol. 55(21) 2939–2951
Deflection of GFRP-reinforced ! The Author(s) 2021
Article reuse guidelines:
concrete beams sagepub.com/journals-permissions
DOI: 10.1177/00219983211002922
journals.sagepub.com/home/jcm

Katarına Gajdosová , R
obert Sonnenschein and
Stanislav Blaho

Abstract
This paper presents an investigation of the performance of concrete beams reinforced with glass fiber-reinforced
polymers (GFRP) under short-term loading. A total of six specimens with rectangular cross-sections (75 mm in
height and 150 mm in width) were tested under a four-point bending test to failure. Each specimen was reinforced
with two GFRP bars with diameters of 8 mm. The results of this study demonstrated the behavior of GFRP-reinforced
concrete members and a validation of the available calculation methods for the deflection of these members and
assumed possibilities of the use of a GFRP reinforcement over the long term. The results of the study presented
show a very good agreement of an experimentally measured and theoretically calculated instantaneous deflection when
using the approaches in the European and American standards. In calculations of long-term deflections, the results are
highly inconsistent and seem to be quite overestimated in some cases. The study shows the necessity of real-time long-
term measurements to demonstrate the real deformations to be assumed during design of structures reinforced with
GFRP reinforcement.

Keywords
GFRP reinforcement, deflection, concrete beams, four-point bending test

Introduction
152
203 and 152
152 mm and with a length of
In the last few decades, progressive composite materials 2,438 mm were simply supported and subjected to a
for reinforcing and strengthening concrete structures uniform sustained loading for nearly two years. The
have been recognized. The most widely known and authors concluded that a time-dependent deflection
used are glass fiber-reinforced polymers (GFRPs) and increased with the increase in the applied load. The
carbon fiber-reinforced polymers (CFRPs), thanks to rate of this increase was higher in the initial period of
their properties, especially their resistance to corrosion, loading and tended to reduce over time. The theoretical
their high strength-to-weight ratio, and, in the case of a prediction of a deflection of concrete beams reinforced
GFRP, their relatively low cost. The short-term behav- with CFRP using the age-adjusted modulus of elastic-
ior of GFRP-reinforcement concrete members has been ity method appears to be very realistic.
demonstrated by a number of tests.1–8 Hall and Ghali10 carried out an experimental inves-
In comparison to the large database of publications tigation on the long-term deflection of concrete beams
dealing with short-term loading, the investigation of reinforced with steel and GFRP bars. The beams were
long-term properties of GFRP-reinforced concrete 280 mm wide, 180 mm high, had a span of 3,200 mm
members is insufficient. The extension of this method
of application in central Europe is confined because
Department of Concrete Structures and Bridges, Faculty of Civil
there are evident differences between the short and Engineering, Slovak University of Technology in Bratislava, Slovakia
long-term behavior of these members.
Arockiasamy et al.9 studied the long-term behavior Corresponding author:
Katarına Gajdosová, Department of Concrete Structures and Bridges,
of beams reinforced with CFRP under a uniform sus- Faculty of Civil Engineering, Slovak University of Technology in Bratislava,
tained loading. Four concrete beams reinforced with Radlinskeho 11, 810 05 Bratislava, Slovakia.
CFRP bars with rectangular cross-sections of Email: [email protected]
2940 Journal of Composite Materials 55(21)

and were reinforced with 3Ø15 mm GFRP bars. The deflection has been developed with a rational approach
beams were tested under sustained loading for approx- that incorporates a tension-stiffening model and has
imately 8 months. The variables were the level of the been adopted in Europe. The proposed equation gave
sustained loading and the reinforcement material. The an effective moment of inertia that is independent of
experimental test results of this study indicate that Ig/Icr and works well with FRP-reinforced concrete.
under similar loading conditions and the same rein- Bischoff assumptions were also verified by Park.14
forcement ratio, the GFRP-reinforced beams had In this study a total of thirty-six specimens were fabri-
long-term deflections 1.7 times greater than those of cated with the dimensions of 200 mm
300 mm

the steel-reinforced beams. 1800 mm (width
height
length). For the tensile rein-
Vijay11 dealt with the aging of the concrete members forcement, the longitudinal reinforcement consisted of
reinforced with GFRP bars. The beams were 150 mm two bars of 13 mm in diameter or three bars of 16 mm in
wide, 300 mm high, had a span of 3,050 mm and were diameter for each specimen with a single layer of rein-
reinforced with 2Ø12 mm or 2Ø16 mm GFRP bars. forcement. The specimens were placed in simulated
Four-point sustained bending loads were applied on aggressive environmental conditions in accelerated
four concrete beams up to 847 days. Although the aging for 300 days. The moment-carrying capacity
deflections are not quantified, it is concluded, that decreased, and the deflection increased as a function of
though initial deflections are high in concrete beams time when exposed to accelerated aging conditions. The
reinforced with GFRP bars as compared to those Bischoff equation provided the best predictions of load
with steel reinforcement, relative creep deflections are and deflections for the accelerated aged specimens.
less as compared to theoretical creep deflections of Miàs et al.15 developed a new method for calculating
beams with steel reinforcement. time-dependent deflections based on the application of a
Abdalla12 experimentally studied seven concrete simplified multiplicative coefficient kcreep for instanta-
beams reinforced with different types of FRP bars neous deflections. The results obtained were compared
and eight prototype one-way concrete slabs reinforced with the experimental results of the long-term deflec-
with FRP and steel reinforcements. The beams with tions of concrete beams presented by Hall and Ghali.10
cross-sections of 500
250 mm and a clear span of A good degree of accuracy in predicting long-term
2,300 mm were tested under a four-point bending test. behavior using the new method has been observed.
The length and width of the slabs were 3,500 and Miàs et al.16 experimentally studied eight GFRP-
1,000 mm, respectively, with a clear span of reinforced concrete beams tested at a service load and
3,000 mm. The thicknesses of the slabs were 150 and subsequently subjected to sustained loading for
200 mm. Based on the results of the measured deflec- 250 days. The dimensions of the beams were
tions of these 15 FRP-reinforced concrete members, the 140
190 mm and the span was 2,200 mm. The rein-
following conclusions were made: Due to the low mod- forcement consisted of two GFRP bars of 12 or
ulus of elasticity, the deflections of concrete members 16 mm in diameter. The time-dependent deflection
reinforced with FRP bars are larger than those rein- increased considerably in the initial period. At
forced with steel bars. The deflections estimated 10 days from loading, the total deflections increased
according to the ISIS Canada approach were in good around 50–60% of the total increase, while 90% of
agreement with the experimental results. the total increase was attained around 90 days after
Bischoff13 carried out an evaluation of equations loading. Two levels of the sustained load were consid-
commonly used for calculating short-term deflections ered, but no significant effect of the applied levels of
for steel and FRP-reinforced concrete members. The the sustained load was observed in the increase in total
two most often used approaches were summarized in deflections. The authors concluded that with regard to
this study. Branson’s equation is used in North the theoretical predictions, the modified Bischoff’s equa-
America and other parts of the world; it is based on tion,13 together with the modified time-dependent factor
an empirically derived effective moment of inertia to presented in ACI 440.1 R-06,17 could reasonably predict
calculate deflections. The author concluded that for the total deflections of the concrete beams reinforced
the high ratios of the uncracked-to-cracked trans- with GFRP. The use of the CSA S806-0218 time-
formed moment of inertia (Ig/Icr) typical of FRP- dependent factor in this study overestimated the exper-
reinforced concrete members, the tension stiffening is imental results in all the beams.
overestimated, thereby leading to a much stiffer Fergania et al.19 experimentally investigated the
response and underprediction of a member’s deflection. long-term tension stiffening and flexural behavior of
Modifications to Branson’s equation for FRP- GFRP-reinforced concrete members subjected to accel-
reinforced members were established to reduce the erated aging conditions. Six tension-stiffening speci-
tension-stiffening component by introducing empirical mens, square in their cross-sections (100
100 mm)
factors. An alternative expression for calculating beam and 1,100mm long with an effective bond length of
Gajdosová et al. 2941

500 mm and eight GFRP-reinforced concrete beams, 150 mm
75 mm rectangular cross-sections, that were
110 mm wide, 150 mm deep with a clear span of symmetrically reinforced with 2 / 8 mm longitudinal
1,000 mm, were exposed to different environments bars (Figure 1). The cover of the reinforcement was
and sustained stress levels for 120 and 270 days, respec- designed to be 20 mm.
tively. Subsequently, the specimens were tested to fail-
ure. The test results revealed that the stressed Materials
specimens conditioned in a wet environment experi-
For the purpose of comparing the results of the exper-
enced a reduction in their tension-stiffening response
imental and analytical modeling, the material proper-
as a result of a bond degradation. Conversely, exposure
ties of the concrete were investigated. GFRP bars with
to a moist environment without sustained loading did
helically wrapped surface treatment were used for this
not cause any noticeable degradation in bond strength.
study. The properties of the GFRP bars provided by
The results also indicate that the accelerated aging con-
the manufacturer (Table 1) were not verified because
ditions affected the overall flexural behavior and led to
good agreement with the results of previous experimen-
higher deflections. Based on the conclusions of this
tal testing carried out at our department during a
study, it can be stated that current models to predict
former investigation had already been achieved.
deflection have failed to adequately capture the com-
The same concrete mix was used for all the speci-
bined effect of a severe environment and sustained
mens. During the concreting of the specimens, cubic
loading. The predictions of the model according to
(150
150
150 mm), cylindrical (150
300 mm) and
the European standard were in a good agreement
prismatic specimens (100
100
300 mm) were also
with the experimentally measured short-term deflec-
prepared and tested according to EN 12390 standard
tions and also with the long-term deflections of the
for testing hardened concrete (compatible with ASTM)
beams that were conditioned in the open air.
just before the beams were tested (Figure 2). The results
A detailed review of the available literature illus-
are summarized in Table 2.
trates that there is a huge need for the verification
and determination of calculation approaches to capture
the actual behavior of FRP-reinforced concrete mem- Specimens
bers under a sustained loading. This paper is focused The beams were supplemented with a transversal
on an investigation of both short-term and long-term GFRP reinforcement in the prepared formwork
deflections of FRP-reinforced concrete members and (Figure 3). The biggest problem during the concreting
predictions according to the European, American, was the stabilization of the reinforcement in its position
and Canadian guidelines. The objective of this study because of its very low weight and deformability. The
is to show, that for the design of concrete structures beams were placed next to each other (2
3) to allow
reinforced with GFRP reinforcement it is necessary to
provide real-time long-term testing because of inconsis- Table 1. Properties of GFRP bars.
tent assumptions of design codes.
Value Value from
from the the previous
Experimental investigation Property manufacturer testing

Testing was performed on 6 concrete beams (N1–N6) Tensile strength (MPa) 1000 1070
reinforced with a GFRP reinforcement. The specimens Modulus of elasticity (GPa) 52 52
were designed as 1,300 mm long beams, with

Figure 1. Dimensions of the concrete beams.

2942 Journal of Composite Materials 55(21)

Figure 2. Concrete specimens for testing of material properties.

similarity of triangles in the course of relative strains

Table 2. Material properties of the concrete.
along the cross-section height, when concrete strain
Property Value was given (measured value) and the neutral axis
Cube strength (MPa) 45.7
depth and reinforcement strain were the unknown
Cylindrical strength (MPa) 38.0 values. The strains of the concrete were measured by
Static modulus of elasticity (GPa) 35.0 strain gauges. Only one strain gauge was installed in
the case of the N1–N3 beams, and two strain gauges
(next to each other) were installed on the N4–N6
the longitudinal reinforcement to be embedded contin- beams.
uously through the three beams and fixed in its position The parameters obtained during the testing were
at the ends by passing through a hole in the formwork.
recorded by computer software. The test was stopped
The stabilization through the length of the beams was
after the cross-section failure.
ensured by the transverse reinforcement bars (in thirds
of the length), which were also fixed at the ends by
passing through a hole in the formwork. Experimental results
After 7 days, the moulds were removed, and after
three months, the beams were tested. General
Items of the investigation Figure 6 shows the load–deflection curves measured by
linear variable differential transformer (LVDT) for all
All six beams were tested in the same manner, i.e. by a the beams tested. After reaching the maximum load
four-point bending test that increased the force in there was no plastic deformation observed. The load
0.5 kN steps. The span between the supports was dropped down without increase in deflection. The
1.2 m, and the two forces were acting in the 20 cm spac- results are summarized in Table 3. The strains of the
ing in the middle part of the beam (Figure 4).
concrete needed for the calculation of the curvature
The strains on the compression face of the concrete
and subsequently for the deflection are shown in
and the deflection from the increasing bending moment
Figure 7. When the two gauges were mounted on the
at the mid-span were monitored during the loading.
compression side of a cross-section, the average value
The prime objective of the experimental study was
the examination of the relationship between the deflec- was calculated and used in Figure 7. In the cases of the
tion at mid-span and the acting force. One linear var- N3 and N6 beams, the gauges peeled off during the
iable displacement transducer (LVDT) was used as it is loading, and the measurement was not finished till
considered the most accurate measurement method for the final failure of the beam.
this purpose (Figure 5). The indirect method for check- The load–deflection curves were then used for a
ing the mid-span deflection was based on the calcula- comparison of the two different measurements.
tions of the deflection from the curvature; this was The deflection for the indirect method was calculat-
determined from the measured strain in the concrete ed from a curvature following the strains:
and the calculated strain in the reinforcement. The rein-
forcement strain was calculated iteratively from three l2
wmeasurede ¼ k     ð mÞ (1)
equations: forces equilibrium, moment equilibrium and 8
Gajdosová et al. 2943

Figure 3. Stabilization of the GFRP reinforcement during the concreting.

Figure 4. Test set-up.

Figure 5. Measurement instruments.

2944 Journal of Composite Materials 55(21)

Figure 6. Load–deflection curves measured by the LVDT.

Table 3. Measured values of the maximal force and deflection. The comparison of the results from the two mea-
surement methods is shown in Figure 8. Relatively
Beam no. Maximal force (kN) Deflection (mm) good agreement was obtained: the indirectly calculated
N1 10.0 44 values from the strains measured show the deflections
N2 8.6 50 of the most stressed cross-section at the middle of the
N3 8.5 41 span of the beam in the same tendency as the LVDT
N4 9.5 41 measurements. In the final phase, the concrete surface
N5 9.0 35 in compression is completely damaged and the mea-
N6 9.4 40 sured values of strains are not accurate. The differences
Average 9.15 41.9
between the curves in Figure 8 might be the conse-
quence of the non-compliance of assumptions of per-
fect connection between the concrete and the GFRP
reinforcement used in equation (2) because of the
where: l ¼ beam span (1,200 mm), k ¼ curvature: weaker bond performance of GFRP.
The results of two measurement methods are also
ecmeasured þ eGFRPcalculated
k¼    ð1=mÞ (2) plotted in the moment-curvature relationship in
d Figure 9.

where: ec_measured ¼ strain at the compressed face of the Beam resistance

mid-span cross-section, eGFRP_calculated ¼ tensile strain
in the GFRP reinforcement (calculated according to All the beams tested in the experimental investigation
the force applied and strain measured on the com- failed in the mid-span’s cross-section (Figure 10). The
pressed face of the concrete – based on the perfect con- failure mode was specified by the crushing and spalling
nection between the concrete and the reinforcement), of the concrete. On all the beams, one large crack was
d ¼ effective depth of the cross-section. observed at the middle of the span of the beam on the
Equation (1) is based on the assumption that the tensile side of the cross-section. Two or three other
deflection curve of a simply supported beam is a circu- tensile cracks appeared at the middle part of beams
lar arch with the radius r. Curvature is calculated as with smaller widths and mutual distances of about 7
k ¼ 1/r and the maximal deflection in the middle of the to 10 cm (Figure 11). The failure of all beams was rel-
beam w can be derived from the right-angled triangle atively brittle, without any plastic deformation.
with side lengths: radius of the arch, a half-length of the The failure mode of all beams demonstrates the pre-
beam and a radius minus deflection in the middle. dicted behavior of GFRP reinforced concrete
Gajdosová et al. 2945

Figure 7. Load–concrete strain curves measured by the strain gauges.

Figure 8. Comparison of the two measurement methods (F–w).

members. The under-reinforced members (reinforced Theoretical analysis

ratio under the balanced one) fail by reinforcement
rupture. The over-reinforced members fail by concrete A theoretical analysis was carried out before testing the
crushing. The beams presented in this study where beams with the material properties measured according
designed as over-reinforced. The reinforced ratio of to a number of analytical models found from various
presented beams is 0.0064 and the value of balanced sources. The models, which mostly approximate the
reinforced ratio is 0.005 or 0.0042 according to EC220 experimental results, are further recommended for use.
or ACI21 respectively. In any case of reinforcement In the preliminary part of the analytical procedure,
ratio, the members cannot fail in ductile way. This is the resistance force was calculated based on the equi-
the biggest difference when comparing GFRP and steel librium equations for bending in the cross-section.
reinforced concrete members. According to the maximal measured strain of the
2946 Journal of Composite Materials 55(21)

Figure 9. Comparison of the two measurement methods (M–k).

Figure 10. Beams after failure.

Figure 11. Cross-section of mid-span of beams after failure.

Gajdosová et al. 2947

concrete of 0.38% and the verified dimensions of the Icr

IeScanlon ¼
 2    ðm Þ
4

 (7)
beams, the resistance force of 9.3 kN was calculated. 1  1  Ig  0:5  ME
Icr Mcr
The second part of the analytical procedure involves
the calculation and prediction of the short-term and
long-term deflections. In general, the deflection of a The American ACI 440.1 R-15 standard21 recom-
beam loaded by a four-point bending test is calculated mends applying the reduction factor bd to the
as follows: moment of inertia, which is obtained using Branson’s
equation for steel RC members:
1 ME
2 
wST ¼   3  leff  4  a2    ðmÞ (3)   "   #
24 E  Ie Mcr 3 Mcr 3
IeACI ¼  bd  Ig þ 1 þ  Icr
ME ME
where
 Ig    ðm 4 Þ (8)
ME ¼ bending moment applied
E ¼ modulus of elasticity of the concrete
Ie ¼ effective moment of inertia where bd ¼ reduction factor equals
leff ¼ effective span q 
a ¼ distance from the support in which loading force 1 f
bd ¼     ðN=AÞ (9)
F/2 is applied 5 qfb

Short-term deflection where qf ¼ the reinforcement ratio, qfb ¼ the balanced

For calculating the short-term deflection, the modulus reinforcement ratio
of elasticity is assumed by the mean value Ecm for all The short-term deflection for the abovementioned
the approaches. For the effective moment of inertia Ie, approaches is then calculated using equation (3).
there are different assumptions according to various According to the Canadian CSA-S806-12 stan-
standards and methods. dard,24 no tension stiffening is considered (the
In the European standard EC2,20 the effective moment-curvature relation is assumed to be trilinear
moment of inertia is calculated as follows: for FRP reinforced concrete members), and the short-
term deflection can be calculated using the following
1
IeEC2 ¼ f
   ðm 4 Þ (4) equation:
þ 1f
Icr Ig
0    3 1
a a
B 3   4  C
where 1 F2  l3eff B leff leff C
wSTCSA ¼  B
B     C    ðmÞ
C
Icr ¼ the moment of inertia of the transformed 24 Ecm  Icr @ Icr lg 3
A
cracked section 8 1 
Ig leff
Ig ¼ the gross moment of inertia
f ¼ the distribution coefficient equal to (10)

 2 where
Mcr
f¼1b    ðN=AÞ (5) lg ¼ the length of the uncracked part of the beam
ME
According to the EC, ACI and CSA standards in
comparison with the average deflection from an exper-
where Mcr ¼ the cracking moment imental verification, the calculated values of a deflec-
b ¼ the coefficient taking account of the effect of the tion are shown in Figure 12 and stated in Table 4. The
duration of the loading, equals 1.0 for a single short- predictions according to the EC2 and ACI standards
term loading and 0.5 for sustained or cyclic loads are in very good agreement with the measured values.
The below-mentioned Bischoff’s equation (6)13–16,22
The CSA standard overestimates the values of the
is only the adjustment of the equation for the effective
deflection. The missing consideration of tension stiff-
moment of inertia in EC2;20 it also accounts for the
ening in the CSA should be the reason.
tension-stiffening effect. Scanlon’s equation (7)16,22,23
includes b ¼ 0.5 for the cyclic loading.
Long-term deflection
Icr For calculating a long-term deflection, the approach of
IeBishoff ¼ 
 2    ðm Þ
4

(6)
the European EC2 standard differs from the other
1  1  Ig  ME
Icr Mcr
standards and methods.
2948 Journal of Composite Materials 55(21)

Figure 12. Comparison of the instantaneous mid-span deflection of the beams according to various approaches.

Table 4. Comparison of the deflections.

Calculated values

Measured value EC2 ACI CSA

Short-term deflection (mm) 41.9 37.9 37.2 51.6

Ratio between the calculated and measured value – 0.90 0.89 1.23

In the European EC2 standard,20 the modulus of wLTCSA ¼ ð1 þ nÞ  wSTCSA    ðmÞ (14)
elasticity is assumed to have the effective value Ec,eff
according to (11): where
Ecm n ¼ the time-dependent factor, including creep and
Ec;eff ¼    ðGPaÞ (11) shrinkage effects
1þu
The approach adapted was designed by Miàs15,16
where u ¼ the creep coefficient based on the good agreement of the results with
The effective moment of inertia is also calculated the experimentally measured deflections of
with the effective modulus of elasticity. According to concrete members reinforced with a GFRP reinforce-
EC2, the final value of a long-term deflection, includ- ment after almost 1 year of sustained loading. The
ing shrinkage, is calculation of the long-term deflection is based
on the short-term deflection according to the ACI
1 ME
 1
standard:
wLTEC2 ¼   3  l2eff  4  a2 þ  l2eff
24 Eceff  Ie 8
1 esh  l2eff
    ðm Þ (12) wLTMias ¼ ð1 þ kcreep Þ  wSTACI þ  ksh    ðmÞ
rcs r 8h
(15)


where
1
rcs ¼ the average curvature due to shrinkage where
r
(assuming cracks in a concrete member) esh ¼ the shrinkage strain
In the American21 and Canadian24 standards, the kcreep ¼ the coefficient for creep effects
long-term deflection is derived from the short-term
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
deflection as follows: Ef
kcreep ¼ 0:73  u   q    ðN=AÞ (16)
Ecm f
wLTACI ¼ ð1 þ 0:6  nÞ  wSTACI    ðmÞ (13)
Gajdosová et al. 2949

Figure 13. Predictions of a long-term mid-span deflection of the beams according to various approaches.

Table 5. Predictions of long-term deflections.

Calculated values

Measured value EC2 ACI CSA ACI þ Miàs

Short-term deflection (mm) 41.9 37.9 37.2 51.6 37.2

Long-term deflection (mm) – 49.9 81.8 154.8 47.7
Ratio of predicted long-term deflection to – 1.32 2.20 3.00 1.28
calculated short-term deflection
Ratio of predicted long-term deflection to – 1.19 1.95 3.69 1.14
measured short-term deflection

ksh ¼ the coefficient for shrinkage effects simplifications in the theoretical analysis. On the
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi other hand, the prediction according to the CSA over-
Ef estimates the measured deflection by 23%. This result
ksh ¼ 1 þ  q    ðN=AÞ (17)
Ecm f does not permit this approach to be considered to be
very accurate, even if the results are significantly on the
The predicted long-term deflections according to the safe side.
above-mentioned approaches are shown in Figure 13 The differences in the predictions of a deflection at
and stated in Table 5. According to the predictive cal- the end of a service life are highly inconsistent. The
culations, the deflection at the end of a beam’s service increase in the instantaneous deflection is predicted to
life would increase by 19%, 95% and 269%, consider- be 19%, 95%, and 269% according to the various
ing the EC2, ACI, and CSA standards for the beams approaches. This fact also emphasizes the need for
tested. These represent large variances and introduce a long-term tests and measurements of the use of an
high degree of uncertainty into the design and use of a FRP reinforcement in order to be able to specify the
GFRP reinforcement in concrete structures. design procedures. At present, it is only possible to
evaluate the deflections expected in comparison with
Comparison of the predictions and the experiments performed.
experimental results
Discussion of results
From the experimental investigation, the results mea-
sured of an instantaneous beam deflection are about The comparison of the experimentally measured and
10% higher than the theoretical calculations according theoretically calculated instantaneous deflections
to the EC2 and ACI. This is a product of many shows very good agreement for the GFRP-reinforced
2950 Journal of Composite Materials 55(21)

concrete beams. The approaches of the European and on the safe side, i.e., the results overestimate the
American standards show a very high degree of accu- measured deflection by 23%. The moment-
racy. Only the overestimation according to the curvature relation in the Canadian standard is
Canadian standard can lead to a very conservative assumed to be trilinear for FRP reinforced concrete
design. The difference between the European and members.
American approach on the one hand and the 3. The predicted values of a long-term deflection at the
Canadian one on the other hand is due to the consid- end of the service life of GFRP-reinforced concrete
eration of the tension stiffening in the first two stand- beams show an increase in an instantaneous deflec-
ards mentioned. tion by 19%, 95%, and 269% specifically for the
The long-term deflection after a sustained loading European, American, and Canadian standards.
has not yet been validated experimentally in this 4. According to the various standards, predictions of a
study. Therefore, the predicted values cannot be com- long-term deflection are highly inconsistent, which is
pared with the actual measurements. The experimental why it is highly recommended to perform real-time
investigation of the long-term behavior of GFRP rein- long-term testing of GFRP-reinforced concrete
forced concrete beams is in a preparatory state. To
members for the specification of more accurate
date, the results of Miàs16 are used for comparisons
and correct design procedures. To extend the find-
of predicted long-term deflections. In the study men-
ings of this study to wide variability of possible
tioned, the GFRP-reinforced beams were exposed to
cases, the test matrix should be supplemented by
sustained loads for almost 1 year, and equation (15)
was derived as the best to capture the actual behavior other parameters as concrete strength, reinforcement
of these members. The results stated in Table 5 show ratio, size effect, etc.
that the approach of the European standard most
exactly approximates the results obtained according Highlights
to Miàs from the calculations applied to the period at
the end of a beam’s service life. It follows that the pre- • There is a quite accurate prediction of the instanta-
diction of a deflection according to the European stan- neous deflection of GFRP-reinforced concrete
dard is probably the most realistic one. A more than beams in the European and American guidelines.
three times larger deflection predicted at the end of a • Approaches in the Canadian guidelines highly over-
beam’s service life according to the Canadian approach estimate the predictions of short-term and long-term
seems improbable. On the other hand, the design deflections.
according to this standard leads to very safe results; it • There is a high degree of inconsistency in the predic-
also leads to a very ineffective economic solution. It is tion of a long-term deflection according to various
very important to realize some long-term tests and standards.
measurements for the validation of the design • Real-time long-term testing of GFRP-reinforced
approaches. According to the available information concrete members is needed.
and the results of the study presented in this paper,
the approaches of the European and American stand- Contributorship
ards are the ones most recommended for the prediction Gajdosová: Conceptualization, Methodology, Validation,
of an immediate and final deflection of GFRP- Resources, Writing – Original Draft, Writing - Review &
reinforced concrete members. Editing, Visualization, Supervision, Project administration.
Sonnenschein: Investigation, Data Curation. Blaho:
Software, Formal analysis.
Conclusions
Based on the study presented herein, the following con- Declaration of Conflicting Interests
clusions have been drawn: The author(s) declared no potential conflicts of interest with
respect to the research, authorship, and/or publication of this
1. The instantaneous beam deflection calculated article.
according to the European standard, the American
standard, and the well-known Bischoff’s equation Funding
are in very good agreement (differences of about The author(s) disclosed receipt of the following financial sup-
10%) with the experimentally measured values. All port for the research, authorship, and/or publication of this
these models account for the tension-stiffening effect article: This work was supported by the Slovak Research and
in the moment-curvature relation. Development Agency under the contract No. APVV-15-0658,
2. According to the Canadian standard, the prediction by the Scientific Grant Agency VEGA under the contract No.
of an instantaneous beam deflection is significantly VEGA 1/0522/20 and by the University Science Park (USP)
Gajdosová et al. 2951

of the Slovak University of Technology in Bratislava (ITMS: 12. Abdalla HA. Evaluation of deflection in concrete mem-
26240220084). bers reinforced with fiber reinforced polymer (FRP) bars.
Compos Struct 2002; 56: 63–71.
ORCID iD 13. Bischoff PH. Reevaluation of deflection prediction for
Katarına Gajdosová https://orcid.org/0000-0002-1201- concrete beams reinforced with steel and fiber reinforced
8983 polymer bars. J Struct Eng 2005; 131: 752–767.
14. Park Y. Long-term performance of GFRP reinforced con-
crete beams and bars subjected to aggressive environments.
References PhD Dissertation. Texas A&M University, USA, 2012.
1. Nanni A. Flexural behavior and design of RC members 15. Miàs C, Torres L, Turon A, et al. A simplified method to
using FRP reinforcement. J Struct Eng ASCE 1993; 119: obtain time-dependent curvatures and deflections of con-
3344–3359. crete members reinforced with FRP bars. Compos Struct
2. Masmoudi R, Theriault M and Benmokrane B. Flexural 2010; 92: 1833–1838.
behavior of concrete beams reinforced with deformed 16. Miàs C, Torres L, Turon A, et al. Experimental study
fiber reinforced plastic reinforcing rods. ACI Struct J of immediate and time-dependent deflections of GFRP
1998; 95: 665–676. reinforced concrete beams. Compos Struct 2013; 96:
3. Pecce M, Manfredi G and Cosenza E. Experimental 279–285.
response and code models of GFRP RC beams in ulti- 17. American Concrete Institute (ACI). Guide for the design
mate and serviceability conditions. J Compos Constr and construction of structural concrete reinforced with
2000; 4: 182–190. FRP bars (ACI 440.1R-06). Farmington Hills, MI:
4. Yost JR, Gross SP and Dinehart DW. Effective moment Author, 2006.
of inertia for glass fiber-reinforced polymer-reinforced 18. Canadian Standards Association (CSA). Design and con-
concrete beams. ACI Struct J 2003; 100: 732–739.
struction of building components with fiber-reinforced pol-
5. Rasheed HA, Nayal R and Melhem H. Response predic-
ymers (CSA S806-02). Toronto, Canada: Author, 2002.
tion of concrete beams reinforced with FRP bars.
19. Fergani H, Di Benedetti M, Miàs Oller C, et al. Long-
Compos Struct 2004; 65: 193–204.
term performance of GFRP bars in concrete elements
6. Barris C, Torres L, Turon A, et al. An experimental study
under sustained load and environmental actions.
of the flexural behavior of GFRP RC beams and com-
Compos Struct 2018; 190: 20–31.
parison with prediction models. Compos Struct 2009; 91:
20. European Committee for Standardization (CEN). Design
286–295.
7. Kassem C, Farghaly AS and Benmokrane B. Evaluation of concrete structures: general rules and rules for buildings
of flexural behavior and serviceability performance of (EN1992-1-1, EC2). Brussels: Author, 2004.
concrete beams reinforced with FRP bars. J Compos 21. American Concrete Institute (ACI). Guide for the design
Constr 2011; 15: 682–695. and construction of structural concrete reinforced with
8. Ali A, Afifi M, Abdulsalam B, et al. Performance evalu- FRP bars (ACI 440.1R-15). Farmington Hills, MI:
ation of one-way concrete slabs reinforced with new Author, 2015.
developed GFRP bars. MSA 2015; 06: 420–435. 22. Bischoff PH and Scanlon A. Effective moment of inertia
9. Arockiasamy M, Chidambaram S, Amer A, et al. Time- for calculating deflections of concrete members contain-
dependent deformations of concrete beams reinforced ing steel and fiber reinforced polymer bars. ACI Struct J
with CFRP bars. Compos Part B: Eng 2000; 31: 577–592. 2007; 104: 68–75.
10. Hall T and Ghali A. Long-term deflection prediction of 23. Scanlon A and Bischoff PH. Shrinkage restraint and
concrete members reinforced with glass fiber reinforced loading history effects on deflections of flexural mem-
polymer bars. Can J Civ Eng 2000; 27: 890–898. bers. ACI Struct J 2008; 105: 498–506.
11. Vijay PV. Aging and design of concrete members rein- 24. Canadian Standards Association (CSA). Design and con-
forced with GFRP bars. Graduate Theses, Dissertations, struction of building components with fiber-reinforced pol-
and Problem Reports, 1999, 3138. ymers (CSA S806-12). Toronto, Canada: Author, 2012.