A large number of civil infrastructures around the world are in a state
of serious deterioration today due to durability problems. Quite a number
of these structures are also no longer considered safe due to increase
load specifications, overloading, under design of existing structures
or due to lack of quality control. These structures must either be replaced,
repaired or strengthened in order to fulfil their functions effectively.
It is becoming both environmentally and economically preferable to repair
or strengthen these structures rather than replacements, particularly
if rapid, effective and simple strengthening methods are available. Many
research works have been carried out in this field. Use of steel plate
and Carbon Fibre Reinforced Polymer (CFRP) are the most widely applied
in this field due to several advantages. Which include easy construction
work, minimum change in the overall size of the structure after plate
bonding and less disruption to traffic while strengthening work is being
carried out. In recent years, with the development of structurally effective
adhesives, the use of plating methods has increased tremendously.
The plate bonding methods, however, often poses a serious premature failure
problem due to separation of plates and rip off concrete along the tensile reinforcing
bars. Thus, other alternative methods are relatively preferred. In this regard,
many studies have been conducted to explore the failure behaviour of plate bonded
strengthened beams. Premature failure due to excessive stress concentration
can be classified into two phenomena. One is the separation that occurs within
the interface layer of the plate and the concrete and the other is the rip off
of the concrete cover along the tensile reinforcing bar due to the initiation
of shear cracking at the ends of plate. These two phenomena are closely related
to the shear and normal stress at the interface layer of plate-bonded beams
as reported by Oh et al. (2003). Saadatmanesh
and Malek (1998) pointed out that shear and normal (peeling) stress concentrations
at the cut-off point or around the flexural cracks were the main reasons for
local failures. Failure due to plate separation starts at the ends of the plates
and is referred to as shear peeling. Shear peeling according to Oehlers
and Moran (1990) is normally induced by the formation of diagonal shear
cracks which is associated with rapid separation of the plate. This debonding
mechanism is initiated by a shear crack at the plate-end which induces a horizontal
crack at the level of the tension reinforcement which propagates rapidly towards
the load point and eventually causes the separation of the plate.
In view of these, some researches were done on end anchorage to prevent the
plate end debonding of strengthened beams. Jones et al.
(1988) first studied the effects of bolt and partial L-shape end anchorage
details on the failure behaviour of strengthened beams with steel plates. They
reported that the anchorage details had some effects on the ultimate strength
and failure modes. However, they also found that anchor bolting did not prevent
the debonding of strengthened plates, but instead complete plate separation
was avoided. An increase in strengths of up to 8% over the un-plated beam was
achieved. They also experimented with glued anchor plates and found it to be
the most effective. However, sudden plate separation still occurred, producing
yielding of the tensile plates and achieving the full theoretical strength of
about 36% above the unplated beam. Similarly, Hussain et
al. (1995) and Garden (1998) reported that anchor
bolting could not totally prevent premature failure. Chahrour
and Soudki (2005) also studied the effects of end anchorage details on the
failure behaviour of CFRP strengthened beams. They used mechanical anchors consisting
of top and bottom 10 mm thick steel plates fastened together using two M12 tightened
bolts for the end anchorage of the CFRP strengthened beams. They reported that
beams strengthened with end anchors consisting partially bonded CFRP strips
gave higher strengths than the fully bonded beam with no end anchorage.
In general, the findings from these studies indicated that these anchors
increased the ultimate strength slightly but still showed sign of debonding
of the strengthened plates. Further, a design guideline for the anchorage
length of end anchors was seldom found. The main goals of the research
reported in this study are; to investigate the effect of U-shaped end
anchors on steel plate and CFRP laminate flexurally strengthened r.c.
beams and to develop a design guideline for the anchorage length of end
anchors to eliminate premature end peeling.
DESIGN OF END ANCHORAGE LENGTH
The purpose of end anchoring on plate bonded strengthened r.c. beam is to prevent
premature end peeling. Design of the anchorage length of end anchors is the
current interest of this field. In this study, a design guide line for anchorage
length of end anchors is proposed based on the fictitious shear span model of
Jansze (1998). According to his model, plate end shear
is the governing failure mechanism for end peeling which creates a fictitious
shear span (Fig. 1) on partially bonded plated beam. This
fictitious shear span can be calculated by using the Eq. 1.
Fictitious shear span (aL) = [(1(ρs)0.5
)2 ds L3/ρs]1/4
||Bar reinforcement ratio (As/bd)
||Effective depth of internal bar (mm)
||Unplated length of strengthened beam (mm)
||Fictitious shear span (mm)
However, since the shear crack at the end of the plate is the main reason
which causes the plate debonding, anchors could be provided along the
plated portion of fictitious shear span. Thus, end anchorage length can
be obtained from the Eq. 2 and it should not be more
than the effective depth (ds) of beam.
Anchorage length (x) = [aLL]
MATERIALS AND METHODS
Test specimens: A total of five identical reinforced concrete
beams were cast. Each beam was of 2300x125x250 mm (length/width/depth)
in dimensions as shown in Fig. 2.
The beams had 12 mm deformed steel bars in the tension zone. Ten millimeter
mild steel bars were used as hanger bars. Six millimeter diameter stirrups
were positioned equally along the beams length, as shown in Fig.
Major test variables: The main test variables considered in the
study were steel plate and CFRP laminate strengthened beams with or without
end anchorage. The cross section of steel plate and CFRP laminate were
2.76x73 and 1.2x80 mm, respectively. U-shaped mild steel plates of 2 mm
thickness were used for end anchoring of the steel plate and CFRP laminates.
The length and height of the end anchorage was 200 and 250 mm, respectively.
The anchorage length of end anchors is obtained from Eq.
1. These test variables are shown in Table 1.
|| Typical reinforcement details of the r.c. beams
|| Strengthening details
For all strengthened beams, the length of the bonded plate was maintained at
1900 mm, which covered almost the full-span length between the supports of the
beams (Fig. 3). The reason for the full span-length strengthening
with steel plates and CFRP laminates was to maximize the strengthening effects.
Materials: Ordinary Portland Cement was used in the concrete mix.
The maximum size of coarse aggregate was 20 mm. The mix proportion is
as shown in Table 2.
The compressive strengths of concrete were measured from three cubes
after 28 days curing in accordance with British Standard (BS 1881). The
average compressive strength of all specimens was 33 MPa. Two tensile
reinforcing deformed bars of diameter 12 mm were used. The measured yield
strength was 551 MPa and measured tensile strength was 641 MPa. Two 10
mm mild steel bars were used as hanger bars. The modulus of elasticity
of steel bar was 200 GPa. The 6 mm stirrups were positioned along the
beam as shown in Fig. 2.
|| Mix proportion of concrete
The measured yield strength
of the stirrup was 520 MPa, the tensile strength was 572 MPa and the modulus
of elasticity was 200 GPa. For strengthening, mild steel plates and Sika
CarboDur S812 laminates (CFRP) were used. The yield and tensile strengths
of the steel plates were found to be 320 and 375 MPa, respectively. The
tensile strength and modulus of elasticity of CFRP laminates were 2800
MPa and 165 GPa, respectively. U-shaped mild steel plates were used for
end anchorage of strengthened beams. The steel plates and CFRP laminates
were glued to the beam soffits before the end anchoring plates were placed
of the bottom and the two sides of the beam using epoxy Sikadur 30 adhesive.
Bonding procedure: Proper concrete surface treatment prior to
plating works has a significant effect in guarantying a perfect bonding
between the concrete and the strengthening plates. For perfect bonding,
the concrete surface was first ground using a diamond cutter to expose
the coarse aggregates. Concrete dusts were then blown out using an air
compressor. The surface of the steel plate was also sand blasted to expose
the original texture of steel and to eliminate any rust. Dusts on the
steel and carbon plates were removed using acetone.
End anchorage details, (a) front view showing the positions
of end anchorage and (b) cross-section of the end of the beam showing
the end anchorage
|| Typical strengthened beams showing the position of
LVDT and electrical resistance strain gauge (front elevation)
After the surface treatment, a putty was applied to fill up the cavities
or holes on the concrete surface. The well mixed sikadur (an adhesive
produced by Sika) was then troweled on to the surface of the concrete
specimens to form a thin layer. It was also applied onto the steel plate
and CFRP laminate using a special dome-shape spatula. The thickness of
the applied sikadur ranged from 2-3 mm in line with the varying of roughness
of the beams surface. The strengthening plates were then placed on to
the soffit face of the beam (Fig. 3). Using a rubber
roller, the plates were then gently pressed into the adhesive until the
material was forced out on both sides of the laminates. The surplus adhesive
was then removed.
End anchoring: Sikadur was also applied to the inner face of the
U-shaped anchoring plates. An excessive amount of adhesive was applied
to the beam surface as well as inner face of the anchoring plate to avoid
any risk of gaps in the adhesive. The anchor-plates were then positioned
on the beam and clamped for 3 days. The end anchorage detailing is shown
in Fig. 4.
Typical strengthened beams showing the position of the
strengthening laminate and electrical resistance strain gauge (bottom
Instrumentation and test procedure: Figure 5
and 6 show the location of the instruments used to record
data during the test. Thirty millimeter electrical resistance strain gauges
were used to measure the tensile strains of steel bar, steel plate and
CFRP laminate. These strain gauges were also placed at the top of the
beam to measure the concrete compressive strains. All strain gauges were
positioned at the mid-span of the beam.
Demec gauges were attached along the depth of beam at mid span region
to measure the concrete strains. Three Linear Variable Displacement Transducers
(LVDTs) were used to measure the deflection of the beam at mid-span and
under the load points (Fig. 5).
The beams were tested in four-point bending as shown in Fig.
2. Initially the load was applied step-by-step using the load control
command of the Universal Testing Machine (UTM). Before failure of the
test specimens, the load was applied using the position control command
of UTM to observe the actual failure mode of the specimens.
Experimental results on the effect of end anchorage on steel plate and
CFRP laminate strengthened beams are presented and discussed subsequently
in terms of the observed mode of failure, failure load, deflection, cracking
pattern and ductile characteristic.
Mode of failure: Figure 7a-e show the failure
modes for various test beams. The control beam without strengthening plates
or laminates showed the conventional flexural failure as it was designed
to fail in flexure. The steel plate and CFRP laminate strengthened beams
without end anchorage (B2 and B4) were found to fail by debonding of the
plate or laminate even though the beams were strengthened for full-span
length between the supports. This debonding failure was initiated due
to shear cracks from the end of the plate. Due to the discontinuity of
the joint between the r.c. beams and the plate, excessive shear stresses
normally develop near the end of the plate. When this shear stress exceeds
the shear resisting capacity of the concrete shear crack would occur.
Because of this shear crack, the plate debonded either at the level of
the reinforcement or at the level of the bonding interface. Both of these
debonding failure modes occurred suddenly and showed brittle characteristics.
However, U-shaped end anchored steel plate and CFRP laminate strengthened
beams (B3 and B5) failed in flexure and shear, respectively. U-shaped
end anchor with designed anchorage length prevented premature end peeling.
The mechanism could be explained in such a way; because of the U-shape
of anchor both sides of the plates were attached firmly on the beams and
resulted in an increase in the shear strength in that portions of the
beams. As the shear strength increased, the number of shear cracks at
the ends of the plate was reduced. Furthermore, due to the U-shaped of
end anchorages, the strengthened plates were firmly clamped to the beam
till time of failure. Due to this mechanism, debonding caused by normal
stress was also minimized. Thus plate debonding did not occur and the
failure modes showed ductile characteristics.
Failure mode of the r.c. beams tested. (a) Beam B1:
control beam, (b) Beam B2: Steel plate strengthened beam without end
anchorage, (c) Beam B3: Steel plate strengthened beam with end anchorage,
(d) Beam B4: FRP laminate strengthened beam without end anchorage
and (e) Beam B5: FRP laminate strengthened beam with end anchorage
Failure load: All test results are shown in Table
3. The results show that, all strengthened beams i.e., B2, B3, B4
and B5 showed 29, 55, 54 and 89% increase in failure load, respectively
compared to the control beam (B1).
The test results also showed that the end anchored strengthened beams
carried more loads compared to the unanchored strengthened beams. The
failure loads of the U-shaped end anchored steel plate and CFRP laminate
strengthened beams (B3 and B5) were 20 and 23% higher than the failure
load of the beams without end anchorage (B2 and B4).
|| Test result
|| Measured and theoretical failure loads
This was because
of the presence of U-shaped end anchorage that prevents premature plate
debonding. Table 4 shows the comparison between the
measured failure load and the theoretical failure load. The measured failure
load of the U-shaped end anchored steel plate strengthened beam (B3) was
found to be almost similar to that of the theoretical value. The strengthened
beam without end anchorage (B2) showed a smaller value compared to the
theoretical value. The explanation for these phenomena could be attributed
to the prevention of premature debonding failure of the end anchored strengthened
beams which allowed the beams to achieve their full strengths before failure.
For the strengthened beam without end anchorage, premature plate debonding
occurred before the beams could achieve their full strength.
For CFRP laminate strengthened beams (B4, B5), the experimental failure
loads were higher than the theoretical failure loads both for the end-anchored
and the non end-anchored strengthened beams. This could be explained in
terms of the properties of the CFRP laminates which possess high strength
and flexible characteristics. Due to the nature of the CFRP laminate which
is less stiffer than the steel plate, the CFRP laminate strengthened beams
without end anchorage showed a delayed laminate debonding properties which
allowed the beams to carry more loads before failure. However, due to
the high strength nature of the CFRP laminate the strengthened beam with
end anchorage (B5) also showed higher failure loads compared to theoretical
Deflection: Figure 8 shows the load versus mid-span
deflection plots for all the beams. At all load levels, all the strengthened
beams showed smaller deflections compared to the control beam due to their
||Load vs deflection plots for all beams
At failure the end anchored strengthened beams showed
higher deflections compared to un-anchored strengthened beams because
of the prevention of plate debonding.
Cracking patterns: The first crack loads, obtained experimentally,
are shown in Table 4. The strengthened beams in general
showed the higher cracking loads compared to un-strengthened control beam.
Since first crack load depends on the modulus of rupture of the concrete
and the stiffness of strengthening materials, the first crack loads of
both the steel plate strengthened beams (end anchored and un-anchored)
were found to be similar. The same was also noted on the CFRP laminate
Figure 9 shows the load versus crack width of all the
beams. The strengthened beams showed smaller crack widths compared to
the control beam. The crack widths of all steel plate strengthened beams
were similar. This was also the case for the CFRP laminate strengthened
Ductile characteristics: Figure 10 shows the
load versus steel bar strains of all the beams. Due to the higher stiffness
of the strengthened beams, at all load levels, the bar strains of all
the strengthened beams were found to be less than the control beam. Figure
10 also shows that the bar strain of the steel plate strengthened
beams (B2 and B3) was identical due to the similar material properties
of both of the beams. This was also true for CFRP laminate strengthened
||Load vs crack width plots for all beams
||Load vs bar strain plots for all beams
Figure 10 also shows that the bars of end anchored
strengthened beams were yielded before the beams failed and the approximate
yield load is 120 kN. It is the sign of ductile failure. However, bar
yield loads of the un-anchored strengthened beams (B2 and B4) were not
found due to their premature debonding failure.
From the study the following conclusions can be drawn:
||The steel plate and CFRP laminate flexurally strengthened beams
without end anchorage exhibited premature shear-type plate debonding
failure in a brittle manner. The control beam and steel plate strengthened
beam with U-shaped end anchorages showed a conventional ductile flexural
failure mode. The CFRP laminate flexurally strengthened beam with
end anchorages exhibited a conventional ductile shear failure mode
||The failure loads of the non end-anchored and the end anchored steel
plate strengthened beams were found to be 30% and 55% respectively
higher than the control beam. End anchored steel plate strengthened
beam recorded a 20% higher failure load compared to non end-anchored
strengthened beam. The failure loads of the non end-anchored and end-
anchored CFRP laminate strengthened beams were found 54 and 89%, respectively
higher than the control beam. The failure load of end anchored CFRP
laminate strengthened beam was 23% higher to that of the non end-anchored
||All strengthened beams showed less deflection compared to the control
beam due to higher stiffness of strengthened beams. End anchorage
had a significant effect on the deflection at failure load. The bars
of end anchored strengthened beams were yielded before the beams
failed. Thus, the beams with end anchors failed with ductile manner
||All strengthened beams possessed higher cracking loads and better
cracking patterns. Since, cracking loads and widths depend on the
modulus of rupture of the concrete and stiffness of the strengthening
material, the cracking loads and widths of the strengthened beams
with end anchorages were found to be similar to those strengthened
beams without end anchorages
||The designed anchorage length of end anchors is sufficient to prevent
premature end peeling
The authors would like to thank the Majlis Penyelidikan Kebangsaan Sains
Negera under the E-science Fund 13-02-03-3022 for providing the fund to
carry out the work reported in this study. Thanks are also due to Sika
Kimia and their staff for providing the technical and materials supports
for this research. The authors would also like to express their gratitude
to whomsoever had contributed to this work either directly or indirectly.