Download as pdf or txt
Download as pdf or txt
You are on page 1of 20

SP-327—41

Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete


Sándor Sólyom, Matteo Di Benedetti and György L. Balázs

Synopsis: This paper presents an extensive experimental study of the effect that the surface profile has on the bond
behavior of FRP bars in concrete. Studied parameters include: concrete strength (~35 MPa (5075 psi) and ~66 MPa
(9570 psi)), test type (pull-out, direct tension pull-out and beam pull-out) and surface profile of the FRP bar (helically
wrapped, indented, two types of sand coated and two types of helically wrapped and sand coated. For comparison,
ribbed steel bars were also used). It was found that the bond strength can vary considerably when different finishing
of the same surface profile type are used (e.g., fine and rough sand coating) and that the concrete strength does influence
the bond strength even if it is higher than the limit (~30 MPa (4350 psi)) stated in literature. Furthermore, the bond
strength results of all FRP bars were consistently higher than those of steel bars. The highest slip value to reach the
bond strength was observed for helically wrapped FRP bars, and the lowest for sand coated bars. Finally, the test setup
was found to affect the bond strength, while no significant effect of bar diameter was observed in the results.

Keywords: bar diameter; bond; bond strength; concrete strength; FRP; slip; surface profile; test type.

41.1
Sólyom et al.

Sándor Sólyom, is a Marie Curie Fellow and PhD student at Budapest University of Technology and Economics in
Hungary. His research interest includes the application of FRP materials for concrete structures and bond behavior of
FRP bars. Member of the fib T5.1 (FRP reinforcement for concrete structure) and fib YMG (Young Members Group)
Representative board.

Matteo Di Benedetti, is a University Teacher with more than 7 years of research experience focusing on experimental
mechanics, structural health monitoring and prognosis, non-destructive evaluation techniques and sustainable materials
for repair and rehabilitation. His publications track-record includes more than 20 internationally recognized journal
papers and conference proceedings. Dr Di Benedetti is a member of fib TG5.1 and TG2.5 as well as of ACI Committees
228 and 437.

György L. Balázs is professor at the Budapest University of Technology and Economics in Hungary. His main fields
of activities are: FRC, FRP, EBR, NSM, Durability, Service Life. Fire resistance and fire design. Bond and cracking.
HPC/UHPC. Sustainability. President of fib for the period of 2011 and 2012. Served as Immediate Past President of
fib for 2013 and 2014, continues then as Honorary President. Participating in fib Presidium meetings by invitation.
Activities in ACI and RILEM.

INTRODUCTION

The use of Fibre Reinforced Polymer (FRP) bars as internal reinforcements has significantly increased over the last
decades in civil engineering applications (Nanni et al. 2016). However, current design codes, usually adapted versions
of those developed for steel reinforcement, do not always provide adequate provisions (Nanni et al. 1995; Bakis et al.
1998; Focacci et al. 2000; Muñoz 2010; Yan et al. 2016), highlighting the need for more detailed investigations. For
example, despite the fact that numerous studies can already be found in the literature, the bond behavior of FRP bars
in concrete remains not fully understood. In particular, while several researchers investigated the bond strength of FRP
bars in concrete studying the influence of various parameters including, among others, concrete cover, concrete
strength, bar diameter, embedment length and surface treatment, a limited number of studies were carried out to define
the parameters used to standardize bond strength calculations. An accurate definition of standard parameters to estimate
bond strength is essential to respond to the improved FRP products constantly introduced in the market.

One of the bond-affecting parameters that still calls for further study is the effect of the surface characteristics of FRP
bars on their bond behavior in concrete. Even though there are numerous research studies (AL-Mahmoud et al. 2007;
Lee et al. 2008; Baena et al. 2009; Hao et al. 2009; Muñoz 2010; Wang & Belarbi 2010; Sayed Ahmad et al. 2011;
Arias et al. 2012; Vint 2012; Tasevski 2013; Mazaheripour et al. 2013; Lin & Zhang 2013; Pour et al. 2016; Yan et al.
2016) indicating that bond behavior of FRP bars can be considerably affected by the type of surface profile, it is not
taken into consideration in some standards and guidelines, such as ACI 440.1R (ACI Committee 440 2015) and JSCE
(Japanese Society of Civil Engineers (JSCE) 1997). Conversely, bond stress calculation according to CSA-S806-12
(CSA-S806-12 2012) takes into account different factors (Eq. 1), including the surface profile (k5).

𝑑𝑑𝑐𝑐𝑐𝑐 �𝑓𝑓𝑐𝑐′
𝜏𝜏𝑏𝑏,𝑚𝑚𝑚𝑚𝑚𝑚 = 1.15(𝑘𝑘 (1)
1 𝑘𝑘2 𝑘𝑘3 𝑘𝑘4 𝑘𝑘5 )𝜋𝜋𝑑𝑑𝑏𝑏

Where:
dcs - the smaller of the distance from the closest concrete surface to the center of the bar being developed; or two-thirds
of the center-to-center spacing of the bars being developed,
f’c - compressive strength of concrete,
k1 - bar location factor,
k2 - concrete density factor,
k3 - bar size factor,
k4 - bar fibre factor,
k5 - bar surface profile factor,
db - nominal diameter of FRP.

With regards to the surface profile parameter, the proposed values are 1.0 for surface-roughened or sand-coated
surfaces as well as for braided surfaces, 1.05 for spiral pattern surfaces and for ribbed surfaces and 1.80 for indented

41.2
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

surfaces. This can translate, that all the different surface profile types give similar bond strength results except the
indented ones, which give 80% lower bond strength values. However, based on the authors’ previous experimental
studies (Sólyom & Balázs 2016; Sólyom & Balázs 2017) and on data available in literature (i.e.: (AL-Mahmoud et al.
2007; Baena et al. 2009), the values proposed by CSA-S806-12 (CSA-S806-12 2012) for the surface factor do not
seem to be sufficiently accurate. Alternatively, CAN/CSA-S6-06 (CAN/CSA-S6-06 2006) recommends that the factor
which takes into account the effect of bar surface (k4- symbol of surface factor in CAN/CSA-S6-06 standard), to be
defined as the ratio between the bond strength of the FRP bar and that of a steel deformed bar with the same cross
sectional area as the FRP bar, but not greater than 1.0. In absence of experimental data, CAN/CSA-S6-06 recommends
that k4 is taken equal to 0.8.

This study has been directed to verify and if needed propose new values or new methods to define the surface profile
parameters, based on an extensive experimental study.

An experimental plan involving the study of various parameters (i.e.: FRP surface profile, concrete strength, test setup
and specimen type) was designed and performed. In order to cover a wide range of commercially available bars, FRP
bars with different profiles were chosen, including two types of sand coated [SC] surfaces, one type of helically
wrapped [HW] surface and two types of helically wrapped and sand coated [HW+SC] surfaces. Concrete strength
range was chosen wide enough to account for the variability in the concrete strength typically used for structures and,
furthermore, to study how FRP bars with different surface profiles behave in concretes with different strengths. The
lower boundary of this range was chosen according to available literature (i.e., (Baena et al. 2009)) stating that bond
failure always occurs in concrete if the strength is lower than approximately 30 MPa (4350 psi). Finally, different test
setups were used to be able to draw more general conclusions about the effect of FRP surface profile on bond behavior.

EXPERIMENTAL WORK

Materials
Concrete—specimens with two different concrete compositions were prepared in the laboratory (Table 1).
The coarse aggregate size was limited to 16 mm (0.63 in) to ease the placement of concrete inside the formwork. The
concrete was mixed in a 0.08 m3 (0.10 yd3) net capacity pan type mixer (with activator). The concrete compressive
tests were carried out on three cubic specimens (side length of 150 mm (5.90 in)) for each individual mixture
(EN12390-3:2009), while the tensile splitting tests were carried out on cylinders (150 mm (5.90 in) in diameter and
300 mm (11.80 in) in height) (EN12390-6:2010).The concrete average properties are summarized in Table 2. Glenium
C330 plasticizer admixture was used to set the consistency of concrete to flow class F4 (EN 12350-5:2009).

Table 1—Concrete mix design (kg/m3) (1 mm = 0.04 in; 1 kg/m3 = 0.062 lb/ft3)
Symbol Cement Water Sand Aggregate Concrete
(CEM II/B-S 42.5) (0-4 mm) consistency
(4-8 mm) (8-16 mm)
*
C1 300 195 824 366 641
F4
C2 400 160 824 366 641
*Flow table test (EN 12350-5:2009 2009)

Table 2—Concrete strength (1 MPa = 145.03 psi)


Symbol Concrete strength
Compressive (MPa) Tensile splitting (MPa)
Average Standard deviation Average Standard deviation
C1 35.33 2.38 2.85 0.15
C2 66.10 0.83 - -

41.3
Sólyom et al.

FRP bars—various types of FRP bars (Fig. 1) were used to study the effect that the surface profile has on the
bond behavior of FRP RC. Additionally, bond tests were also performed on steel bars (Fig. 5) for comparison.

Sand coated Basalt FRP (BFRP) with different diameters (6 to 12 mm) (#2 to #4) and sand roughness were used. Lake
sand with maximum grain size of 0.4-0.8 mm (0.015-0.031 in) and 0.5-1.2 mm (0.019-0.047 in) were employed for
“fine” and “rough” coating, respectively (Fig. 2). The tensile strength ranged from 1350 to 1565 MPa (196 to 227 ksi)
while the tensile modulus of elasticity from 51.8 to 56.0 GPa (7500 to 8100 ksi), for the small and large diameter
respectively.

Sand coated Glass FRP (GFRP) bars possess similar surfaces as the BFRP bars; their properties are 1308-1415 MPa
(190-205 ksi) tensile strength and 53.0-54.7 GPa (7700-7900 ksi) tensile modulus of elasticity, depending on bar
diameter.
GFRP bars with helically wrapped and sand coated (HW+SC) surface were used from two different producers (Fig.
3), with a tensile strength of 758-896 MPa (110-130 ksi) and 1003-1085 MPa (145-157 ksi) and a tensile modulus of
elasticity of 46.0 GPa (6700 ksi) and 47.0-49.2 GPa (6800-7100 ksi) for the type of HW+SC bar visible in Fig. 3-left
and right, respectively.

GFRP bars with helically wrapped surface (Fig. 4 right) had a tensile strength ranging from 1350 to 1500 MPa (196 to
218 ksi) and a tensile modulus of elasticity from 55.0 to 58.0 GPa (8000 to 8400 ksi) , depending on the bar diameter
(6-10 mm, 0.24-0.40 in).

GFRP bars with 8 and 12 mm (0.32 in and #4) diameter and indented surface were used for the current experimental
work. Manufacturer’s properties of the bars include a characteristic tensile strength over 1000 MPa (145 ksi) and a
modulus of elasticity of 60 GPa (8700 ksi). All values are reported according to the producer’s information.

Fig. 1—Different FRP bars used for presented study. HW - helically wrapped, SC - sand coated, “fine” and “rough”
refers to the roughness (granulometry) of the sand

41.4
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

Fig. 2—Close-up of sand coated FRP bars. Left: GFRP bars (Ø6 (#2)), right: BFRP bars (Ø12 (#4)). In both pictures
the left FRP bar has the “fine” while the right bar has “rough” sanding (Note: the two pictures are not on the same
scale)

Fig. 3—Close-up of GFRP HW+SC bars, left: 6 mm (#2); right: #2 (6.4 mm). (Note: bars are at the same scale)

Fig. 4—Close-up of GFRP bars (Ø8 (0.32 in)) with indented (left) and helically wrapped (right) surfaces. (Note: bars
are at the same scale)

41.5
Sólyom et al.

Fig. 5—Close-up of steel reinforcing bar (Ø8 (0.32 in))

Test specimens
When the authors designed the experimental plan and consequently the size of the specimens, the RILEM
recommendations (Anon 1982; Anon 1983) were taken into consideration, with some adjustments.

All specimens were cast with concrete prepared in the laboratory. Following the pouring, the specimens were left in
the molds under laboratory conditions for one day, demolded, marked and placed under water for 6 days. After this
period, the concrete specimens were taken out of water and kept in laboratory conditions until testing (mixed curing
was applied). Concrete was poured in two (pull-out, direct tension and concrete compressive test specimens) or three
layers (beam pull-out and cylinder specimens), and each layer was compacted using a vibrating table.

Pull-out test specimens—150 mm (5.90 in) cubic molds were used to manufacture the pull-out specimens.
Each specimen consisted of a concrete cube with a single rebar embedded vertically in the center of the cube. The bond
length of the rebar was set to 5Ø (Ø - bar diameter), which was achieved by wrapping a part of the FRP bars with soft
insulating material to prevent bonding between the bar and the concrete. During testing, the unbonded portion of the
FRP lays within the portion of the concrete block always reacted against the metallic plate (Fig. 6 and Fig. 7).

Fig. 6—Schematic representation of pull-out test specimen, dimensions are in mm (1 mm=0.04 in)

41.6
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

Fig. 7—Photo of pull-out test setup and test specimen

Direct tension pull-out test specimens—each specimen consisted of a concrete prism (cross section of
100x100 mm (3.94x3.94 in) and length of 200 mm (7.87 in)) with two bars embedded in the center of the prism and
protruding from two opposite sides. One of those bars was an FRP bar with an embedded length equal to 100 mm
(3.94 in), 5Ø of which was the bond length. A schematic representation of the specimen is visible in Fig. 8, along with
a photo in Fig. 9 where the test setup and the specimen are shown.

Fig. 8—Schematic representation of direct tension pull-out test specimen, dimensions are in mm (1 mm=0.04 in)

41.7
Sólyom et al.

Fig. 9—Photo of direct tension pull-out test setup and test specimen

Beam pull-out test specimens—similar to those recommended by RILEM (Anon 1982) were adopted in this
study for determining bond behavior of FRP bars. Specimens are 600 mm (23.62 in) long with a cross section of
150x150 mm (5.90x5.90 in). They comprise two prismatic concrete blocks, connected by an FRP bar as flexural
reinforcement in the tension zone and by a steel hinge at the top. Concrete cover is 20 mm (0.79 in) for all specimens.
Owing to the relatively short bond length of 5Ø consistent with bond lengths in the pull-out tests, trial tests were
performed to verify whether stirrups were required. The outcomes consistently provided a pull-out failure with no
visible splitting cracks, confirming that stirrups were unnecessary. FRP bars are provided with unbonded length, using
the same technique as described for the pull-out tests, to provide the necessary (5Ø) bond length. A photo of the beam
pull-out test setup is visible in Fig. 10.

Fig. 10—Photo of beam pull-out test setup

41.8
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

Test setup and procedure


Pull-out (P-O) test— is the most frequently chosen method to compare the bond behavior of different FRP
reinforcing bars in various concrete compositions. This test is a powerful tool to study the effect of different parameters
on bond strength, owing to its simplicity and ease of application. One part of the current experimental work is based
on pull-out tests. The experimental test setup is shown in Fig. 7. The pull-out tests were performed using a servo-
hydraulic testing machine with a capacity of 600 kN (135 000 lb). Displacement control (1 mm/min - 0.04 in/min) was
selected to capture the post-peak bond behavior. The concrete pull-out specimens were placed into a metal frame and
the FRP bars were gripped by the test machine. This gripped part of the FRP bar is considered as the loaded end of the
test specimen. The relative displacement between the FRP bar and concrete was measured with three Linear Variable
Differential Transducers (LVDT). At the other end, usually referred to as the free end, the slip was measured by one
LVDT. An automatic data acquisition system was used to record the data transmitted by LVDTs. Four nominally
identical specimens for each configuration were tested.

Direct tension (DT) pull-out test—researchers agree (Tastani & Pantazopoulou 2007; Ayudhya & Ungkoon
2010; Muñoz 2010) that the stress conditions developed in concrete during pull-out tests can substantially differ from
those developed in RC members and, in general, bond strength is higher in the case of pull-out tests than in RC
members in most practical conditions. However, due to various advantageous properties of pull-out tests, they remain
the most common test method. Based on some considerations found in the literature (Tastani & Pantazopoulou 2007;
Ayudhya & Ungkoon 2010), the authors decided to opt for a less conventional direct tension test setup (Fig. 8 and Fig.
9). The test procedure and instrumentation were similar to that presented in the case of pull-out tests. The main
differences include that the measurement of the free end slip is no longer possible and that the concrete is in tension
during the whole test, since no compression stresses are generated because of the reaction against the metal plate. It is
anticipated that this test setup will yield lower bond strength values, but closer to what can appear during practical
cases. Hence, this test setup possesses both the advantageous properties of traditional pull-out and beam pull-out tests.

Beam (bending) pull-out (BP-O) test—researchers claim that the most realistic bond strength values can be
reached when the beam pull-out test is performed (Lin & Zhang 2013; Tighiouart et al. 1998; Gudonis et al. 2014;
Pecce et al. 2001; Mazaheripour et al. 2013); however, due to its complexity it is less frequently implemented. BP-O
tests were thus performed to study how the bond behavior varies for different surface profiles when changing the test
setup, and thus the stress state in the concrete. Fig. 10 shows a photo of the test setup configuration. To measure the
slip at the loaded and free ends of the FRP bar, four LVDTs were used. Two LVDTs were fixed to the free extremities
of the bar. These measured the relative displacement between the free ends of the bar and the concrete surfaces.
Simultaneously, two LVDTs were fixed to the bar, close to the vertical axis of the beam, measuring the relative
displacement and thus the elongation of the FRP bar. These measurements will be critical for the calculations
performed for the bond stress-slip diagrams.

Two additional LVDTs have been used for the measurement of the mid-span (i.e., the innermost point of the left
concrete prism) vertical and horizontal displacement. During the first series of BP-O tests, strain gauges were also
used. One strain gauge was installed on the FRP bar, coinciding with its longitudinal axis in order to record the strains
in the bar during the test.

A load cell with a capacity of 200 kN (45 000 lb) was utilized to register the applied load. Using the registered load
values, the pull-out force was calculated. The force transferred in the bar was found by equilibrium between the applied
load and the corresponding reactions to the specimen. The effect of the change of the moment arm (i.e., distance
between hinge connection point and the center of the bar) at mid-span was neglected due to its low influence on the
results (Mazaheripour et al. 2013). Pull-out forces could be also calculated with the help of the strain gauge
measurements considering the modulus of elasticity and nominal cross sectional area of the bar.

The tests were carried out using the same testing machine as for the DT and P-O tests. Force controlled loading was
chosen in accordance with RILEM recommendations (Anon 1982). The load was applied in consecutive increments
corresponding to stresses in the bar successively equal to 80, 160, 240, 320 MPa (11.6, 23.2, 34.8, 46.4 ksi).
Thirty seconds were allowed for the load to increase between two consecutive levels. Once the target was reached, the
load was kept constant for two minutes.

41.9
Sólyom et al.

RESULTS AND DISCUSSION

Test results

Experimental results are summarized in Table 3 to Table 5, including the average bond strength, the loaded and free
end slip values (when maximum bond stress was reached) and bond stiffness. In particular, the bond strength (τb,max,)
is defined considering uniform bond stress distribution along the bond length, calculated by dividing the load by the
shear surface (Eq. 2).

𝐹𝐹
𝑢𝑢𝑢𝑢𝑢𝑢
𝜏𝜏𝑏𝑏,𝑚𝑚𝑚𝑚𝑚𝑚 = 𝜋𝜋∗∅∗𝑙𝑙 (2)
𝑏𝑏

where Fult is the maximum load (N); Ø is the FRP bar nominal diameter (mm); and lb is the bond length (equal to 5Ø).
The slip values at τb,max represent the slip when the bond stress is maximum. Finally, the bond stiffness is defined by
calculating the inclination of the ascending branch of the bond stress-slip diagram, which was assumed to be linear.

Definition of Fult is straightforward in the case of P-O and DT tests, since it is directly measured and registered during
the tests. However in the case of BP-O test, the force in the bar cannot be directly measured. The following equation
(Eq. 3) was employed for calculation of the maximum force in the bar during BP-O test:

𝑃𝑃𝑢𝑢𝑢𝑢𝑢𝑢
𝐹𝐹𝑢𝑢𝑢𝑢𝑢𝑢 = (𝑎𝑎1 − 𝑎𝑎2 ) (3)
2𝑑𝑑𝑎𝑎

where Pult is the maximum load measured during BP-O tests (N); da is the distance between the center of the hinge and
the longitudinal axis of the FRP bar (mm); a1 is the distance between the beam support to the vertical axis of the beam
(mm); a2 is the distance between the loading point of the beam to the vertical axis of the beam (mm). Due to the test
setup and specimen design, da varied with the FRP bar diameter; however a1 and a2 had constant values of 275 and
75 mm (10.82 and 2.95 in), respectively.

Effect of surface characteristics on bond behavior


Bond stress-slip curves recorded during the P-O and DT tests are shown in Fig. 11 for SC GFRP “fine” and “rough”
bars, for HW+SC GFRP bars and for steel bars. For visual clarity, the results of BP-O tests are not shown at this stage.

The bond failure always occurred by pulling out of the FRP bars; no splitting, bar rupture or anchorage failures were
recorded. In general, the bond stress-slip responses are characterized by a linear part of the ascending branch, when
the bond surface is still not totally damaged, followed by a nonlinear section up to peak bond stress (bond strength)
due to the accumulated damage on the bar surface and in the concrete surrounding the bar. The slope of the post peak
phase (descending branch) varies depending on the bar type (surface and fibre) until the bond stress remains sufficiently
constant (i.e., residual bond strength) due to the friction between the FRP bar and the concrete. As expected, the graphs
show that both the bond strength values and the corresponding slips depend on the surface profile of the bars.

It can also be noticed that, with regards to SC GFRP bars (Fig. 11, a and b), higher bond strength values are reached
by the “rough” sanding, both in P-O and DT tests. Furthermore, in the post-peak phase, SC GFRP “fine” bars have a
gradual bond stress reduction and the bond failure is always ductile, reaching high slip values. Conversely, for the
same phase, a sudden reduction is visible in the case of SC GFRP “rough” bars, which is followed by either a ductile
or brittle failure. The outcomes of HW+SC GFRP bars (type 1) are similar to those observed for GFRP SC “fine” bars
when considering the bond strength and the shape of bond stress-slip curves. This can be explained by the relatively
small diameter of the wire used to helically wrap the GFRP bar (Fig. 3 left). The diameter of the wire is similar in size
to the thickness of the “fine” sand coating, yielding comparable surface profile performances. For comparison, bond
stress-slip curves for steel bars are also presented (Fig. 11 d).

It is also worth mentioning that all the studied FRP bars presented higher bond strength values than the steel bars,
with the only exception the DT test result of GFRP SC “fine” bars.

41.10
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

Table 3—Pull-out bond test results of different FRP bars (diameter of 6 mm (#2), C1 concrete mix). (1 mm =
0.04 in; 1 MPa = 145.03 psi)
Bond strength Loaded end slip Free end slip
Sur- Bar Test τb,max τb,maxa St.D. C.O.V. Slipb Slipa,b St. C.O.V. Slipb Slipa,b St. C.O.V. Bond
face No. Dev Dev stiffness
[MPa] [MPa] [MPa] [%] [mm] [mm] [mm] [%] [mm] [mm] [mm] [%] [N/mm3]
1 14.23 0.270 0.178
2 11.06 0.311 0.161
BFRP
Sand coated ("fine")

12.91 1.43 11.09 0.291 0.047 16.22 0.155 0.025 16.28 44.33
3 12.52 0.238 0.119
4 13.83 0.346 0.162
1 16.22 0.317 0.258
2 15.23 0.365 0.180
GFRP

13.97 2.67 19.09 0.371 0.044 11.93 0.198 0.058 29.42 37.70
3 14.27 0.425 0.228
4 10.15 0.374 0.125
1 NA NA NA
Sand coated ("rough")

2 21.35 0.826 0.283


BFRP

20.71 0.60 2.91 0.812 0.060 7.34 0.301 0.036 11.83 25.51
3 20.16 0.746 0.278
4 20.62 0.863 0.342
1 21.28 0.335 0.248
2 21.28 0.594 0.217
GFRP

18.78 2.89 15.38 0.489 0.152 30.99 0.243 0.045 18.62 38.40
3 16.29 0.642 0.304
4 16.26 0.386 0.202
1 19.45 3.085 1.638
2 22.90 3.868 2.575
GFRP
HW

17.90 4.08 22.80 2.765 1.066 38.57 1.695 0.873 51.49 6.47
3 15.38 1.318 0.519
4 13.87 2.788 2.049
1 15.54 0.671 0.168
2 15.35 0.485 0.229
GFRP

16.22 0.90 5.53 0.561 0.084 14.91 0.163 0.113 69.37 28.92
3 17.01 0.579 0.252
HW+SC

4 16.97 0.508 0.002


1 14.02 0.191 0.159
2 14.23 0.296 0.156
GFRP

13.58 1.51 11.13 0.273 0.061 22.38 0.153 0.006 3.74 49.68
3 14.73 0.336 0.151
4 11.36 0.271 0.146
1 12.42 1.165 1.009
Ribbed

2 13.34 0.910 0.796


Steel

13.06 0.56 4.28 0.977 0.164 16.83 0.873 0.118 13.56% 13.36
3 NA NA NA
4 13.43 0.857 0.813
Notes: a
- Average value of similar specimens;
b
- slip corresponding to bond strength;
NA - Not Available (error during data acquisition);
HW - Helically wrapped;
HW+SC - Helically wrapped and sand coated.

41.11
Sólyom et al.

Table 4—Direct tension pull-out bond test results of different FRP bars (diameter of 6 mm (#2), C1 concrete mix).
(1 mm = 0.04 in; 1 MPa = 145.03 psi)

Bond strength Loaded end slip


Surface Bar Test τb,max τb,maxa St.D. C.O.V. Slipb Slipa,b St.D. C.O.V. Bond
No. stiffness
[MPa] [MPa] [MPa] [%] [mm] [mm] [mm] [%] [N/mm3]
("fine")

1 NA NA
coated

BFRP GFRP
Sand

7.20 NA NA 0.509 NA NA 14.14


2 7.20 0.509
1 10.73 0.526
13.25 3.56 26.88 0.615 0.126 20.50 21.55
Sand coated
("rough")

2 15.77 0.704
1 11.91 0.602
GFRP

13.17 1.79 13.59 0.576 0.037 6.50 22.87


2 14.44 0.549
1 12.19 1.781
GFRP
HW

15.61 4.84 31.00 1.378 0.570 41.36 11.32


2 19.03 0.975
1 12.02 0.426
Steel GFRP GFRP

11.94 0.12 0.97 0.406 0.027 6.74 29.39


HW+SC

2 11.86 0.387
1 11.65 0.326
10.22 2.03 19.83 0.681 0.502 73.75 15.01
2 8.79 1.036
Ribbed

1 12.28 0.563
9.98 3.25 32.53 0.846 0.400 47.31 11.80
2 7.69 1.129

Notes: a
- Average value of similar specimens;
b
- Slip corresponding to bond strength;
NA - Not Available (error during data acquisition)
HW - Helically wrapped;
HW+SC - Helically wrapped and sand coated.

41.12
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

Table 5—Beam pull-out bond test results of different FRP bars (diameter of 6 mm (#2), C1 concrete mix). (1 mm = 0.04 in; 1 MPa = 145.03 psi)

Left Right
Bond strength Loaded end slip Free end slip Loaded end slip Free end slip
Sur-face Bar Test No. τb,max τb,maxa St.D. C.O.V. Slipb Slipa,b St.D. C.O.V. Slipb Slipa,b St.D. C.O.V. Slipb Slipa,b St.D. C.O.V. Slipb Slipa,b St.D. C.O.V. Bond
stiffness
[MPa] [MPa] [MPa] [%] [mm] [mm] [mm] [%] [mm] [mm] [mm] [%] [mm] [mm] [mm] [%] [mm] [mm] [mm] [%] [N/mm3]
Sand coated ("rough") SC ("fine")

1 13.27 0.442 0.171 0.382 0.026


GFRP

15.19 2.72 17.91 0.934 0.696 74.50 0.230 0.083 36.05 0.648 0.376 58.05 0.039 0.018 45.92 16.27
2 17.12 1.426 0.288 0.914 0.051

1 19.01 0.580 0.386 0.284 0.062


BFRP

18.35 0.93 5.08 0.771 0.271 35.11 0.249 0.194 77.81 0.640 0.503 78.69 0.211 0.210 99.77 23.79
2 17.69 0.963 0.112 0.996 0.359

1 17.02 0.288 0.151 0.124 0.046


GFRP

16.92 0.14 0.84 0.622 0.471 75.78 0.182 0.044 24.09 0.245 0.170 69.42 0.073 0.038 52.31 27.23
2 16.82 0.955 0.213 0.365 0.100

1 13.21 0.954 0.260 0.908 0.225


GFRP
HW

13.26 0.06 0.46 1.327 0.528 39.79 0.318 0.082 25.79 1.724 1.154 66.96 0.740 0.728 98.42 9.99
2 13.30 1.700 0.376 2.540 1.255

1 NA NA NA NA NA
GFRP

13.20 NA NA 0.828 NA NA 0.086 NA NA 0.849 NA NA 0.131 NA NA 15.94


2 13.20 0.828 0.086 0.849 0.131
HW+SC

1 16.66 0.181 0.237 0.194 0.044


GFRP

15.95 1.00 6.27 0.320 0.197 61.46 0.161 0.108 67.41 0.186 0.012 6.34 0.045 0.001 3.14 49.80
2 15.24 0.459 0.084 0.177 0.046

1 9.53 1.037 0.696 1.012 0.738


Ribbed

Steel

10.01 0.68 6.81 0.974 0.090 9.19 0.589 0.152 25.83 0.966 0.065 6.70 0.606 0.187 30.95 10.28
2 10.49 0.910 0.481 0.920 0.473

Notes: a
- Average value of similar specimens;
b
- Slip corresponding to bond strength;
NA - Not Available (error during data acquisition)
HW - Helically wrapped;
HW+SC - Helically wrapped and sand coated.

41.13
Sólyom et al.

P-O_1
P-O_1
20 20
P-O_2
P-O_2
P-O_3 P-O_3

Bond stress [MPa]


Bond stress [MPa]

15 15
DT_2 P-O_4

DT_1
10 10 DT_2

5 5

0 0
0 5 10 0 5 10
Loaded end slip [mm] Loaded end slip [mm]
(a) (b)
P-O_1 P-O_1
20 20
P-O_2 P-O_2
P-O_3 P-O_4
Bond stress [MPa]

Bond stress [MPa]

15 P-O_4 15
DT_1
DT_1 DT_2
10 DT_2 10

5 5

0 0
0 5 10 0 5 10
Loaded end slip [mm] Loaded end slip [mm]
(c) (d)
Fig. 11—Bond stress-slip diagrams for bars (Ø6) with different surfaces: sand coated GFRP fine (a) and “rough” (b),
helically wrapped and sand coated (c) and steel (d). (1 mm = 0.04 in; 1 MPa = 145.03 psi)

To isolate the effect of different surface types, the bond behavior of FRP bars is investigated considering the results of
the same test setup (P-O) for all the bars used in this experimental study. In particular, the bond stress- slip curve of
one bar for each surface profile is plotted in Fig. 12. FRP bars show higher bond strength values than steel bars, which,
in turn, tend to show relatively higher residual bond strength values. In addition, GFRP HW bars, despite having the
lowest bond stiffness due to the large slip recorded before reaching the bond stress peak, are characterized by the
highest residual bond strength.

41.14
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

BFRP_fine
20 GFRP_fine
BFRP_rough
GFRP_rough
15 GFRP_HW
Bond stress [MPa]

GFRP_HW+SC_1
GFRP_HW+SC_2
10 Steel

0
0 1 2 3 5 4
6 7 8 9 10 11 12
Loaded end slip [mm]
Fig. 12—Bond stress-slip (P-O test) diagrams for bars (Ø6) with different surfaces, only one representative curve is
showed for each bar. (1 mm = 0.04 in; 1 MPa = 145.03 psi)

The histogram in Fig. 13 summarizes the bond strength (vertical axis) recorded for 6 mm (#2) diameter bars and
concrete composition C1. In particular, for the different surface profiles (horizontal axis), all three test methods (P-O,
DT and BP-O) are represented with bins of different colors. The main observation confirms the theory of the authors
that the bond strength values can significantly vary even when the same surface category is used (such as SC or
HW+SC). This behavior is not accounted for in some standards when the factors describing the effect of surface type
on bond strength are recommended. For example, when SC surface finishing is considered, P-O test results show 18.78
and 20.71 MPa for “rough” GFRP and BFRP bars, while only 13.97 and 12.91 MPa for “fine” bars, while, in the case
of HW+SC bars, the difference is lower, being 16.22 MPa for type 1 and 13.58 MPa for type 2. This variability in bond
strength results within the same surface categories cannot be accounted by only one parameter underlining that the
method defined by CAN/CSA-S6-06 2006 to determine the surface factor for bond strength seems better able to capture
the effect of various surface types. Furthermore the effect of test types is also visible in case of studied FRP bars, and
it is discussed later.

20.71 Pull-out DT pull-out Beam pull-out


18.35 18.78
17.90
16.92
15.61 16.22 15.95
15.19
13.97 13.58
12.91 13.25 13.17 13.26 13.20 13.06
11.94
10.22 10.01

7.20
5.64

SC "fine" B SC "fine" G SC "rough" B SC "rough" G HW HW+SC_1 HW+SC_2 Steel


Fig. 13—Summary of bond strength results (in MPa) for Ø6 bars in C1 concrete mix with different surfaces and test
types. (More information about test results and notation can be found in Table 3 to Table 5). (1 MPa = 145.03 psi)

41.15
Sólyom et al.

20.71 Ø6 Ø12 25.17 26.88 C1 C2


21.86
16.38 20.10
15.17
12.91 12.87 13.58 13.04 16.38 15.17
12.87 13.04

SC "fine" B SC "rough" B HW+SC_2 Indented SC "fine" B SC "rough" B HW+SC_2 Indented


(a) (b)
Fig. 14—Summary of bond strength results (in MPa) of FRP bars with diameters of Ø6 (#2) and Ø12 (#4) (a) and
concrete composition of C1 and C2 (b). (More information about test results and notation can be found in Table 3 to
Table 5). (1 MPa = 145.03 psi)

The same observation can be drawn from Fig. 14 (a). If the SC “fine” is considered as the reference surface profile
(having k5=1, according to CSA-S806 (CSA-S806-12 2012)), the indented bars should also have a value for the k5
factor around 1 and not 1.8 as proposed in (CSA-S806-12 2012), however, if SC “rough” bars are considered as the
reference surface profile (having k5=1), the value proposed by CSA-S806 (CSA-S806-12 2012) for indented bars
(having k5=1.8) is more accurate, but still conservative. This highlights that defining values for k5 is not
straightforward, because of the considerable bond strength difference even within the same surface profile category
(SC). Furthermore, it can be observed in the same figure (Fig. 14 left) that the effect of FRP diameter on bond strength
change is negligible in the case of the studied bars. Additionally, while numerous studies (i.e. Lee et al. 2008; Baena
et al. 2009; Muñoz 2010; Yan et al. 2016) show that for concrete strength higher than 30 MPa the bond failure happens
due to the FRP surface, resulting in a negligible effect of concrete strength on the bond strength values, the results of
the presented experimental study show a significant increase (varying between 44% and 95%) in bond strength with
concrete strength increase (Fig. 14 right).

It is know from the literature (Muñoz 2010) that in the case of steel bars, bond failure occurs due to concrete crushing
in front of the bar deformations, and is similar both in normal and high strength concrete. However, the bond failure
of FRP bars can be largely influenced by the surface profile because bond failure always take place at the surface with
the lowest shear strength. In order to observe the surface where bond failure occurred, specimens were split after testing
and analyzed. Fig. 15 shows damage associated with the failure of the specimens in concrete mixture C2. When BFRP
bars with “fine” sand coated surface were considered, inter-laminar shear failure happened on the surface between the
resin and the bar fibers (resin reach layer, Fig. 15 a). However, in the case of BFRP SC “rough” bars, bond failure
occurred partly on the surface between the concrete and the resin and partly in the resin rich layer (Fig. 15 b). Finally,
in the case of HW+SC bars, both the sand coating and wrapping were sheared off (Fig. 15 c), as similarly happened to
the helical wrap of HW bars (Fig. 15).

The possible effect of different environmental conditions on the bond behavior of FRP bars with different surfaces was
not studied in this work. Additional research is necessary to be performed to study the validity of the above presented
observations on bars that have been previously exposed to aggressive environments to assess the long term
performance of bond.

Effect of test setup on bond behavior


In Fig. 13, the bond strength results are plotted for 6 mm (#2) bars and concrete composition C1. In addition to
analyze the effect of FRP bar surface on bond strength, the figure shows that FRP bars with different surface
characteristics are affected by changes in the test setup. This can be particularly important when the test method is
chosen to define surface factors (CAN/CSA-S6-06 2006). It can be noticed that the bond strength is affected by the
test setup, since, while using nominally identical materials (FRP bar, concrete), the P-O test setup consistently
returns the highest bond strength values, owing to the higher compression stress that is generated in the concrete.
Conversely, the lowest bond strength results are achieved by DT test, where the concrete is only in tension. Bond
strength results of BP-O tests are usually intermediate between the results obtained by the other two test methods. In
particular, it was found that the ratio between bond strength of P-O and DT tests decreases as the concrete strength

41.16
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

increases and that this ratio is lower for SC bars than for HW and HW+SC bars. Finally, the authors could not
observe any influence of the diameter of the bars on this bond strength ratio.

(a)

(b)

(c)

(d)

Fig. 15—Bond failure surface after testing (C2 concrete mix): (a) BFRP SC “fine” Ø6 (#2); (b) BFRP SC “rough”
Ø6 (#2); (c) GFRP HW+SC Ø8 (0.32 in); (d) GFRP HW Ø6 (#2);

41.17
Sólyom et al.

CONCLUSIONS

In this paper an experimental study is presented that investigates the effect of surface profile of FRP bars on the bond
behavior in concrete. Studied parameters included the concrete strength and the type of test method. FRP bars with
different surface profiles (two types of sand coated surfaces, helically wrapped surface, indented surface and two types
of helically wrapped and sand coated surfaces) were used to cover a wide range of commercially available FRP bars.
For comparison, ribbed steel bars were also tested. A suitable range of concrete strength (~35 MPa (5075 psi) and
~66 MPa (9570 psi)) was chosen to cover the strength range typical for concrete structures. Three different types of
test specimens: pull-out, direct tension pull-out and beam pull-out were used to observe whether the effect of surface
profile on bond behavior is dependent on the test method.

Results highlight that bond strength values within the same surface profile — such as sand coated (SC) or helically
wrapped and sand coated (HW+SC) — can significantly vary. This observation is generally not taken into
consideration in standards and guidelines therefore, it is not reflected in the selection of design factors describing the
effect of the surface type on the bond strength. The definition of such factors relying solely on the type of surface
profile might result in inaccurate values. Subgroups within the same surface types are necessary, since bond strength
was considerably different in the case of two surface finishes (“fine” and “rough”) of the same profile (sand coated).
The ratios among the bond strength values of FRP bars with different surface profiles vary with diameter and concrete
strength, in addition to the change within the same surface profile, making it difficult to propose values of surface
factors (k5). Alternatively, this factor could be defined more accurately by experiments. However, a standardized test
method should be specified. The presented direct tension pull-out test method could be used, since it encompasses both
the advantages of the pull-out test (simplicity, ease of application) and the beam pull-out test (tension stresses in
concrete).

In the case of 6 mm (#2) bars, the highest bond strength results were achieved with SC “rough” profiles, followed by
the HW and HW+SC profiles. SC “fine” and indented surface profiles show only slightly higher bond strength results
than the steel bars. SC “rough” surface profile always led to the highest bond strength results. The difference in SC
“rough” and “fine” bond strength results can be explained by the difference in the sand grain size used for sand coating
(larger grain size for “rough” profile). Bond failure of SC “fine” bars always happened in a ductile manner allowing
gradual bond strength decrease after the bond strength was reached, while in the case of SC “rough” bars both ductile
and brittle failure modes were observed. Future work should aim to define the optimum grain size to reach maximum
bond strength with ductile bond failure.

The highest slip value recorded at the peak of bond stress was observed for HW FRP bars. Bond stiffness (calculated
assuming a linear ascending bond stress-slip curve), which is particularly important in SLS, was higher for all the
studied FRP bars than that of steel, except for HW FRP bars.

In addition, it was found that contrary to the available results in the literature, concrete strength does influence the
bond strength even if it is higher than the limit stated in the literature (~30 MPa (4350 psi)). Results of this experimental
work indicate that the inter-laminar shear strength of the bars in this study are higher thanthat from the related literature.
The test method did have a tangible effect on the bond strength values the pull-out test provided the highest bond
strength results, while the direct tension pull-out test provided the lowest. Finally, no significant effect of bar diameter
was observed in this study on bond characteristics within the diameter range of 6 to 12 mm.

ACKNOWLEDGEMENTS

Authors gratefully acknowledge the financial support of the European Union by Marie Curie ITN: European Network
for Durable Reinforcement and Rehabilitation Solutions (endure), Grant: PITN-GA-2013-607851. FRP bars were
provided by several producers, special thanks to: Dr André Weber, Boris Miller, Doug Gremel, Bert Kriekemans, Dr
Jan Prokes and Dr Péter Molnár.

41.18
Effect of Surface Characteristics of FRP Bars on Bond Behavior in Concrete

REFERENCES

ACI Committee 440, 2015. Guide for the Design and Construction of Structural Concrete Reinforced with FRP Bars
(ACI 440.1R-15), Farmington Hills, MI.
AL-Mahmoud, F. et al., 2007. Effect of surface pre-conditioning on bond of carbon fibre reinforced polymer rods to
concrete. Cement and Concrete Composites, 29(9), pp.677–689.
Anon, 1982. RILEM RC 5 Bond test for reinforcement steel. 1. Beam test.
Anon, 1983. RILEM RC 6 Bond test for reinforcement steel. 2. Pull-out test.
Arias, J.P.M., Vazquez, A. & Escobar, M.M., 2012. Use of sand coating to improve bonding between GFRP bars and
concrete. Journal of Composite Materials, 46(February), pp.2271–2278.
Ayudhya, B.I.N. & Ungkoon, Y., 2010. Bond Strength of Fiber Reinforced Polymer ( FRP ) Bars in Autoclaved
Aerated Concrete ( AAC ). In The 5th International Conference on FRP Composites in Civil Engineering
September 27-29, 2010, Beijing, China. Bejing, China, pp. 1–4.
Baena, M. et al., 2009. Experimental study of bond behaviour between concrete and FRP bars using a pull-out test.
Composites Part B: Engineering, 40(8), pp.784–797. Available at:
http://dx.doi.org/10.1016/j.compositesb.2009.07.003.
Bakis, C.E. et al., 1998. Analysis of bonding mechanisms of smooth and lugged frp rods embedded in concrete.
Composites Science and Technology, 58(8), pp.1307–1319.
CAN/CSA-S6-06, 2006. Canadian Highway Bridge Design Code, Mississauga, Ontario, Canada.
CSA-S806-12, 2012. Design and construction of building structures with fibre-reinforced polymers, Mississauga,
Ontario, Canada: Canadian Standards Association.
EN12390-3:2009, 2009. Testing hardened concrete. Part 3: Compressive strength of test specimens.
EN12390-6:2010, 2010. Testing hardened concrete. Part 6: Tensile splitting strength of test specimens.
EN 12350-5:2009, 2009. Testing fresh concrete: Flow table test.
Focacci, F., Nanni, A. & Bakis, C.E., 2000. Local bond-slip relationship for FRP reinforcement in concrete. Journal
of Composites for Construction, (February), pp.24–31.
Gudonis, E. et al., 2014. Mechanical Properties of the Bond Between GFRP Reinforcing Bars and Concrete. Mechanics
of Composite Materials, 50(4), pp.457–466. Available at: http://link.springer.com/10.1007/s11029-014-9432-0.
Hao, Q. et al., 2009. Bond strength of glass fiber reinforced polymer ribbed rebars in normal strength concrete.
Construction and Building Materials, 23(2), pp.865–871. Available at:
http://dx.doi.org/10.1016/j.conbuildmat.2008.04.011.
Japanese Society of Civil Engineers (JSCE), 1997. Recommendations for design and construction for concrete
structures using continuous fibre reinforcing materials. Concrete Engineering Series, (23).
Lee, J.Y. et al., 2008. Interfacial bond strength of glass fiber reinforced polymer bars in high-strength concrete.
Composites Part B: Engineering, 39(2), pp.258–270.
Lin, X. & Zhang, Y.X., 2013. Bond-slip behaviour of FRP-reinforced concrete beams. Construction and Building
Materials, 44, pp.110–117. Available at: http://dx.doi.org/10.1016/j.conbuildmat.2013.03.023.
Mazaheripour, H. et al., 2013. Experimental study on bond performance of GFRP bars in self-compacting steel fiber
reinforced concrete. Composite Structures, 95, pp.202–212. Available at:
http://dx.doi.org/10.1016/j.compstruct.2012.07.009.
Muñoz, M.B., 2010. Study of bond behaviour between FRP reinforcement and concrete. University of Girona.
Available at: http://www.tdx.cat/handle/10803/7771.

41.19
Sólyom et al.

Nanni, A. et al., 1995. Bond of FRP reinforcement to concrete - Experimental results. In L. Taerwe, ed. Non-metalic
(FRP) reinforcement for concrete structurestalic (FRP) reinforcement for concrete structures. London: E & FN
Spon, 2-6 Boundaty Row, pp. 135–145.
Nanni, A. et al., 2016. Concrete and Composites Pedestrian Bridge. Innovative use of various types of FRP
reinforcement at the University of Miami. , (November), pp.57–63.
Pecce, M. et al., 2001. Experimental and analytical evaluation of bond properties of GFRP bars. Journal of Materials
in Civil Engineering, (July/August), pp.282–290.
Pour, S.M., Alam, M.S. & Milani, A.S., 2016. Improved Bond Equations for Fiber-Reinforced Polymer Bars in
Concrete. materials, 9(737), pp.1–14.
Sayed Ahmad, F., Foret, G. & Le Roy, R., 2011. Bond between carbon fibre-reinforced polymer (CFRP) bars and ultra
high performance fibre reinforced concrete (UHPFRC): Experimental study. Construction and Building
Materials, 25(2), pp.479–485. Available at: http://dx.doi.org/10.1016/j.conbuildmat.2010.02.006.
Sólyom, S. & Balázs, G.L., 2016. Influence of FRC on bond characteristics of FRP reinforcement. In 11th fib
International PhD Symposium in Civil Engineering. Tokyo, J, pp. 271–278.
Sólyom, S. & Balázs, G.L., 2017. Non-metallic reinforcements with different moduli of elasticity and surfaces for
concrete structures. In CCC2017 - The 12th Central European Congress on Concrete Engineering. Tokaj, pp.
1–8.
Tasevski, D., 2013. Effect of surface treatment on the bond behavior of Fiber Reinforced Polymer (FRP) reinforcement
and concrete. Lafarge Research Center / Laboratoire Navier / ENPC.
Tastani, S.P. & Pantazopoulou, S.J., 2007. Design Values of Bond – Slip Law for Frp Bars. In FRPRCS-8. Patras, pp.
1–10.
Tighiouart, B., Benmokrane, B. & Gao, D., 1998. Investigation of bond in concrete member with fibre reinforced
polymer (FRP) bars. Construction and Building Materials, 12(8), pp.453–462.
Vint, L.M., 2012. Investigation of Bond Properties of Glass Fibre Reinforced Polymer (GFRP) Bars in Concrete under
Direct Tension. University of Toronto. Available at: https://tspace.library.utoronto.ca/handle/1807/33573.
Wang, H. & Belarbi, A., 2010. Static and Fatigue Bond Characteristics of FRP Rebars Embedded in Fiber-reinforced
Concrete. Journal of Composite Materials, 44(13), pp.1605–1622. Available at:
http://jcm.sagepub.com/cgi/doi/10.1177/0021998309355845.
Yan, F., Lin, Z. & Yang, M., 2016. Bond mechanism and bond strength of GFRP bars to concrete: A review.
Composites Part B: Engineering, 98, pp.56–69. Available at:
http://linkinghub.elsevier.com/retrieve/pii/S1359836816305200.

41.20

You might also like