Академический Документы
Профессиональный Документы
Культура Документы
Downloaded from ascelibrary.org by Griffith University on 07/30/16. Copyright ASCE. For personal use only; all rights reserved.
INTRODUCTION
The need for strengthening (or stiffening of) reinforced concrete (RC) and prestressed concrete (PC) structures is becoming more apparent, particularly when there is an increase in
load requirements, a change in use, a degradation problem, or
some design/construction defects. Potential solutions range
from replacement of the structure to strengthening with a variety of techniques. Plate bonding is one of them and advanced
composites, such as fiber-reinforced plastic (FRP), are just one
type of material that can be used in this respect. Plates in the
fonn of flexible carbon FRP (CFRP) sheets, to be externally
bonded to the concrete surfaces, were considered in this paper.
They are made of unidirectional dry or preimpregnated (prepreg) fibers and are installed with a technique known as manuallayup.
The use of FRP materials for structural repair presents several advantages and had been recently investigated all over
the world (Saadatmanesh and Ehsani 1991a, b; Chajes et ai.
1994; Meier and Winistorfer 1995; Varastehpour and Hemelin 1995; Baaza et ai. 1996; Nakamura et ai. 1996; Crasto
et ai. 1996; CEB-FIP 1993; Arduini et aI., in press, 1997).
The writers have so far concentrated their investigations on
the strengthening of virgin RC beams. It has also been shown
that, as a result of the FRP application, the mode of failure
of a flexural member may change from ductile to brittle. For
example, shear failure in concrete may substantially reduce
the nominal flexural capacity expected with computations using standard design equations. Changing the thickness of the
FRP sheet, changing its bonded length, or adding shear reinforcement, significantly modifies the crack distribution
along the beam and changes the failure mechanism (Arduini
et aI., in press, 1997). Possible failure mechanisms are: FRP
rupture, concrete crushing, shear failure, peeling at the adhesive-concrete interface, and so on (Arduini et aI., in press,
1997).
In this paper, experimental and analytical results are pre-
~MSert..
r;::::::;:=::;:::::::;::::=;::::::::;:=::;::=~;::tfso
220
4 16
So
----IllOoi
-----1
FIG. 1.
FRP meet
SOrt60 +lSO
+-bp -f
= 1 In.)
JOURNAL OF COMPOSITES FOR CONSTRUCTION 1 MAY 1997/63
Downloaded from ascelibrary.org by Griffith University on 07/30/16. Copyright ASCE. For personal use only; all rights reserved.
TABLE 1
Beam
code
(1 )
Surface
preparation
(2)
Beam
condition
(3)
SI
SM2
SM3
SM4
SM5
SM6
ST2
ST3
ST4
Ml
MM2
MM3
MM4
MM5
MT2
MT3
MT4
MT5
none
sanding
sanding
sand blasting
sand blasting
sanding
sanding
sand blasting
sand blasting
none
sanding
sand blasting
sanding
sand blasting
sanding
sand blasting
sand blasting
sand blasting
virgin
virgin
precracked
virgin
virgin
precracked
precracked
virgin
virgin
virgin
precracked
precracked
precracked
virgin
precracked
precracked
precracked
precracked
bp
(mm)
(5)
FRP
type
(6)
500
500
500
500
450
500
500
500
300
300
300
140
300
300
300
300
M
M
M
M
M
T1
T1
T5
950
950
850
850
950
950
950
950
150
150
150
150
150
150
150
150
M
M
M
M
Tl
Tl
Tl
T1
CFRP strengthening
(7)
none
one ply at the bottom
one ply at the bottom
one ply at the bottom
one ply at the bottom
two plies at the bottom
one ply at the bottom
one ply at the bottom
one ply at the bottom
none
one ply at the bottom
one ply at the bottom
one ply at the bottom
two plies at the bottom (0) + one ply on the sides (0 and 90)
one ply at the bottom
one ply at the bottom
vertical load, one ply at the bottom
vertical load + external prestress, one ply at the bottom
RESEARCH SIGNIFICANCE
The study shows that it is possible to strengthen precracked
RC flexural members and to obtain similar results to the case
of strengthened virgin specimens. It is also possible to predict
the load-deflection behavior of precracked members using an
analytical model. The study shows that the failure mode of the
flexural member may be altered by strengthening and become
brittle. The parameters that affect the failure mechanisms are
recognized and discussed.
EXPERIMENTAL PROGRAM
RC Specimen Characteristics
The total number of test specimens was 18: nine were rectangular RC short-length beams (S series) with details and
dimensions as shown in Fig. 1(a), and nine were rectangular
RC medium-length beams (M series) with details and dimensions as shown in Fig. l(b). The S series consisted of specimens intended to simulate the geometry of a shallow beam
with a height-to-width ratio (h/b) equal to 0.5, whereas the M
series was intended to simulate a deeper section geometry (hi
b = 2.0). The test span and the position of the loading knives
for the two series were such to maintain a similar shear spanto-reinforcement depth (aid) ratio (i.e., 3.8 for s series and 3.5
for m Series). All specimens were tested under four-point loading.
All beams were constructed at a precast plant using conventional fabrication, curing, and transportation techniques.
The longitudinal reinforcement consisted of four 12 mm (0.47
in.) and four 16 mm (0.63 in.) deformed steel bars for series
Sand M, respectively. All specimens used 6 mm (0.24 in.)
deformed steel stirrups, equally spaced at 100 mm (5.9 in.).
CFRP Strengthening
Table 1 shows FRP reinforcement characteristics and surface preparation for all specimens. The code used to identify
each specimen (column 1) consists of two letters and a number. The first letter (S or M) indicates the beam series and the
second letter (M or T) indicates the FRP material system (to
be discussed later).
As indicated in Table 1, column 2, the concrete surface for
the specimens to be strengthened with CFRP was prepared by
either sanding or sand blasting. The first technique was a light
surface cleaning performed by a technician with a power
TABLE 2
Material
(1 )
E
(GPa)
(2)
v
(3)
Concrete
Steel
Adhesive-M
CFRP-M sheet
Adhesive-T
CFRP-Tl sheet
CFRP-T5 sheet
27
200
2
235
2
235
380
0.2
0.3
0.38
not applicable
0.35
not applicable
not applicable
fy
(MPa)
(5)
f,
(MPa)
(6)
(%)
(7)
not applicable
550
not applicable
not applicable
not applicable
not applicable
not applicable
2.7
640
37
3,510
49
3,480
2,940
0.05
20
1.5
1.5
4.8
1.5
0.8
(MPa)
(4)
not
not
not
not
not
not
36
applicable
applicable
applicable
applicable
applicable
applicable
Downloaded from ascelibrary.org by Griffith University on 07/30/16. Copyright ASCE. For personal use only; all rights reserved.
practical cases where strengthening is performed without removing the service load and the resulting deflection is or is
not compensated for.
Beams S 1 and M 1 were tested to failure without bonded
FRP sheets (control specimens). Unidirectional CFRP sheets
were used to strengthen the remaining beams. The sheets were
applied following the specifications of the material system
manufacturers. This included compliance with resin proportioning, mixing, application, and curing. The typical sequence
of operations for manual layup was: application of a surface
primer, application of the first layer of impregnating resin, application of the sheet (ply), removal of the paper backing, and
application of the second layer of impregnating resin. The last
two operations were repeated in the base of multiple plies.
Adhesion of the FRP sheets was conducted by working under
the specimen as it would be done in the field for the repair of
the soffit of a flexural element. No particular problems were
encountered during manual layup installation. After seven
days of resin curing, the beams were tested to failure.
Table 1, columns 4-6, presents the length, width, and type
of FRP for each beam tested. For all specimens, but two, of
series S, bonded length (lp) and width (b p ) were kept constant
at 500 and 30D mm (19.7 and 11.8 in.). For specimen SM5,
the sheet width was reduced to 140 mm (5.5 in.) and, for
specimen SM6, the sheet length was reduced to 450 mm (17.7
in.). For all specimens, but two, of series M, bonded length
(lp) and width (bp) were kept constant at 950 and 150 mm
(37.4 and 5.9 in.). For specimens MM4 and MM5, the sheet
length was reduced to 850 mm (33.5 in.). Two FRP material
systems were used and are identified as M and T. The first
material system (M) consists of a prepreg carbon fiber sheet,
whereas the second consists of a dry carbon fiber sheet. Furdebonding at concrete-adhesive interface
F/2~n".
1'?'l:--_:=\~__.C..__...............~Od".
:zs:
Mechanical Properties
Table 2 shows the mechanical properties of the constituent
materials (i.e., concrete, steel reinforcing bars, adhesives, and
CFRP sheets) as obtained experimentally or assumed. With
reference to Table 2, the following symbols and experimental
procedures were used:
E = Young's modulus derived from direct load and strain
(a)
1/2
F/2
zs:_"'!L=====c==....__=-
1...
2S
(b)
conctete
j !
:
cut angle
.................................
: :
FRP@Odos.
\..l
....
.
~~~~~y.~
F
Strengthening Schemes and Failure Mechanisms
concrete
a~
~ \
conciete
FIG. 3.
L-_co_n~cllrre_te_....I
F
TABLE 3.
Adhesive
(1 )
Cut angle a
(degrees)
Type M
Type M
Type M
TypeT
Type T
Type T
Compression Test
lJ'
(MPa)
(4)
Failure
mode
(5)
(MPa)
(6)
(MPa)
(7)
Failure
mode
(8)
1.2
2.8
not available
1.5
2.3
not available
concrete
concrete
not available
concrete
concrete
not available
not available
not available
-24.3
not available
not available
-32.9
not available
not available
24.3
not available
not available
32.9
not available
not available
adhesive-concrete interface
not available
not available
adhesive-concrete interface
lJ'
(2)
(MPa)
(3)
30
60
45
30
60
45
2.1
1.6
not available
2.7
1.3
not available
Downloaded from ascelibrary.org by Griffith University on 07/30/16. Copyright ASCE. For personal use only; all rights reserved.
Adh. siveT
..
.1I..f
..
...........
15
10
-5
r.::iIi::N1-T---I--=:;;~-I---;::;;::::;:::::;;:;::;::;
ISorieaSM I
140 t-----+----~q..---::::::lI"""~~-=====~
Tension
120
+-----F-~.,e__,.,,......::~+----_+----~
100
t---j~~---I--=::::;;,t==::::;;=-t-
80
approxh ation
-20
160
COlI jpression
r----. ~
-25
t--I~~~'""P'~=--t---V-t--r=::::::=:l
-81
6O+---I-'N~~-+-----+----_+----l-SM2
..... 5M3
...
40 f-:,,~r,L----+-----+----_+----l
\~
o
201l':1-
-+-
--SM4
-SMS
+----_+--.J.".--,.;8;;:;M.6io11
lIII deflection
[mm]
5
10
120 +-----fg..,..~~--+------+-----_l
100
20
15
t--7rpr-----r::::::::;::;;~t::::::~-t_
8Ol---~r;-_::::;;;;o
_ _=:-+---_I:.~+----__I
60 +--""7!iF-~:L-_I_----+------+--_r_
40
f-llfI-"F----+-----+-----+----i
20 :u~
+--
-I-
=m
-I-_..!=.....
=8T=4:::!...
ddlection [mm]
10
FIG. 7.
rlesST
20
15
2SO
-M1
-MM2
..... MM3
-MM4
-MM5
200
ISO
100
50
25
____;;87"1..-1
30
ISeries MT I
Load [kN]
250
-MI
-MT2
-MT3
-MT4
-MTS
200
ISO
~
~
100
~~
-,
y/
~/
'/
VI
50
....
/
IS
10
20
2S
30
Downloaded from ascelibrary.org by Griffith University on 07/30/16. Copyright ASCE. For personal use only; all rights reserved.
ISeries 8M I
Load [kN]
140
.l jti'
120
./
",f:: -::.~//' ~
100
~.4~"'i
.:~.
80
/11
--SM2alll2
'.D"SM2ala12
....... 8M3 alll2
... SM3 ala12
--SM6aIV2
...8M6ala12
I
vt / / /
.J / I /
60
40
ir/ /
20
FRP strain
0.002
0.000
FIG. 10.
250
0.004
0.006
0.008
0.010
0.012
!!!!....1~+---+----t.:=~:J--t----::::~.......,9
200 +---__---I---+--;;;Jq---+-----j
-MM2atll2
.. .. MM2ala12
....... MM3 alll2
-MM4atll2
.....
MM4 ata12
5O+---,,~#---+->tC----1~--+----+---1
-MMSatll2
100
+---,.4:..o!t~r--~--+---+--j
O~H--+--+-+_+-+_I_
o
FIG. 11.
0.002
FRP
in
__+_+-I__+-+--+-_I___+__+_+_..j..2~~!!.1
0.004
0.006
0.008
0.010
SM5 showed a debonding failure at the adhesive-concrete interface. In these specimens, the debonding started at one of
the flexural cracks in the constant moment region and propagated towards the sheet end until total delamination occurred.
Specimen SM6 failed by concrete shear at the end of the sheet.
For all strengthened specimens, the maximum strain recorded
in the FRP sheet was at best one third of its ultimate value
(see Fig. 10), indicating an inefficient use of material.
By comparing specimen SM2 with SM4, it can be concluded that sandblasting is a slightly more effective type of
surface preparation by sanding as it produces a higher load
and deflection, even though it does not change the failure
mechanism.
The effect of preloading can be observed by comparing
beam SM2 with SM3. Obviously, SM3 is less rigid in the
initial loading phase, and also its ultimate capacity is lower.
Reducing the area of FRP does not result in a proportional
reduction of strength as observed by comparing specimens
SM4 (bp = 300 mm) and SM5 (bp = 140 mm). This is because
the failure is in both cases is controlled by FRP debonding.
However, it does have an effect on structural stiffness after
yielding of the steel as one would expect.
Specimen SM6, with two FRP plies, showed a considerably
higher stiffness in the cracked region of the load-deflection
curve compared to SM3. It failed at a load higher than that of
SM3 but the mechanism was that of concrete shear failure at
the end of the sheet. In Fig. 10, the load-strain diagrams for
specimens SM3 and SM6, clearly indicate that, at any given
load, the strain in the former sample is considerably higher
than in the latter.
0.012
shear span (aI2). The location for each gage is given in the
figure legend.
For ease of interpretation, the unloading-reloading cycles
are not shown in the load-deflection or load-strain curves (with
exception for the control specimens). Fig. 3 reported typical
sketches of crack evolution and failure mechanism obtained in
the experimental program.
GroupSM
Specimens in this group (Figs. 6-10) were strengthened
with prepreg CFRP. All specimens tested as virgin (i.e., SI,
SM2, SM4, SM5) showed at first a linear-elastic behavior followed by a first crack within the constant moment region of
the beam. Thereafter, a large nonlinear phase was recorded
with the development of numerous flexural cracks. In this
phase, the strain in the FRP sheet increased considerably (see
Fig. 10) as well as the beam deflection. Control beam SI
reached failure by crushing of the concrete long after yielding
of the steel reinforcement. Specimens SM2, SM3, SM4, and
Group ST
Specimens of this group (Fig. 7) were strengthened with
dry-fiber CFRP sheets. The relative performance of specimens
ST2 and ST3 is shown in Fig. 7. The former was precracked
and had the surface prepared by sanding, the latter was virgin
and sandblasted. ST2 is less stiff than ST3 in the initial load
range because of the precracking, and fails by FRP debonding
at a lower load because of the surface preparation of lower
effectiveness.
Everything else being the same (i.e., virgin, sandblasted surface), the effect of high modulus fibers is seen in the comparison of ST4 with ST3. The overall stiffness of the former is
obviously higher; however, failure load and mode are the
same.
It is concluded that the difference between the two material
systems M and T is practically insignificant when comparing
the load-deflection curves of specimen pairs SM3 and STI,
and SM4 and ST3 (see Figs. 6 and 7).
Group MM
Specimens in this group (Figs. 8 and II) were strengthened
with prepreg CFRP. Control beam M I reached the failure by
crushing of the concrete long after yielding of the steel. Beams
MM2 and MM3 showed a debonding failure at the adhesiveconcrete interface. In these beams, the debonding started at
one of the flexural cracks in the constant moment region and
propagated towards the sheet end until total delamination occurred. The effect of different surface preparation seems to be
negligible.
Beam MM4, with three plies, showed stiffness in the
cracked region considerably higher than that of the corresponding specimen MM2 strengthened with one FRP ply.
However, it failed at a load only slightly higher due to concrete
shear failure at the end of the sheet.
The brittle failure mechanism due to sheet debonding was
corrected in beam MM5, where a ply of unidirectional fibers
JOURNAL OF COMPOSITES FOR CONSTRUCTION / MAY 1997/67
Downloaded from ascelibrary.org by Griffith University on 07/30/16. Copyright ASCE. For personal use only; all rights reserved.
(b)
GroupMT
Specimens of this group (Fig. 9) were strengthened with
dry-fiber CFRP sheets. A comparison of the five load deflection curves shown in Fig. 9 offers interesting observations with
regards to the performance of members strengthened under
different loading conditions. Considering specimens MT3,
MT4, and MT5 (i.e., sandblasting and precracked), one observes that:
Curve MT3 starts from the origin as the specimen is precracked, unloaded, strengthened, and then cyclically
loaded to failure.
Curve MT4 starts at a deflection of nearly 2 mm (0.08
in.). This is the value of the deflection at which the beam
is held under service load (44 kN = 9,879 Ib), strengthened, and then cyclically loaded to failure.
Curve MT5 starts at a deflection of nearly zero under a
service load (44 kN = 9,879 Ib). The deflection is fully
recovered as a result of the external postensioning (i.e.,
340 kN = 76,340 Ib applied at an eccentricity of 80 mm
= 3.1 in.).
Because failure in all cases is controlled by debonding of the
FRP, there is not a substantial difference in ultimate capacity
of the three specimens.
By comparing specimen MT2 with MM2 (Fig. 8), it is
shown that the change in material system is insignificant if the
specimens are strengthened under the same conditions (i.e,
sanding and precracking).
MODELING
The analytical model used to interpret and simulate the experimental data was presented in an earlier publication (Arduini et aI., in press, 1997). The most relevant features of the
model are: nonlinear constitutive relationship for concrete in
compression and tension, confinement effect due to closed stirrups, and concrete-adhesive interface properties. For this paper, the model has been modified to include the effects of
precracking and unloading-reloading.
The constitutive laws for the constituent materials considered by the model are as follows (Arduini et al., in press,
1997).
Concrete in compression is nonlinear according to the
CEB-FIP "Model Code 90" (1993). Fig. 12(a) shows the
8.
............::::....f_........
.-r;
Downloaded from ascelibrary.org by Griffith University on 07/30/16. Copyright ASCE. For personal use only; all rights reserved.
..
Load [leN]
.)' ~
/:i>
100
60
20
ISeriesSM I
::,~
~.,.
1///
---:-: ~/
1/
~
I/"'.
80
40
.....
140
120
.........
160
-Sl
--SM4
-+-SM6
...... SIA
..SM4A
"+"SM6-A
Mid-, an deflection [mm)
1(...7
,~r
10
[kN]
-1----+----+=-""--__"""'1-----+------1
-MI
-MM2
-~
...MIA
" oMM2A
ISO +----+~,,-.~_:b_O;:';':'-"'!='L---+_--_l
.....MMS-A
loo-!--SO-!--AtI"#~-"L--_1_---'~+----I___--_+-----l
10
IS
20
25
30
Beam
code Fu./Fm.x
(1 )
(2)
Sl
SM3
SM4
SM6
ST2
ST4
Ml
MM3
MM5
0.90
1.03
1.12
1.13
1.16
1.16
0.93
1.14
0.97
8u. /8m.x
(3)
N/A
0.86
0.99
1.21
0.96
1.12
N/A
0.90
0.71
Type of failure
(4)
concrete crushing
debonding adhesive-concrete interface
debonding adhesive-concrete interface
shear concrete at the end of the sheets
debonding adhesive-concrete interface
debonding adhesive-concrete interface
concrete crushing
debonding adhesive-concrete interface
shear failure at the support of the con
crete beams
.............
ISO
-MM211112
-MM2lte12
100 -I---H""".......I'"--+----+----H --MMSII1I2
" 0 " MM2-A 11112
..... MM2-A II eI2
so.J.l..&.:jl!.-+----1---+-----+-1 ..... MMSA 11112
FRP strsin
ol--I--l-+-+-+---+-+-I-+_+_-+--+--I--l-+-+_+_-+-+-II---+-+--+-I
0.002
0.004
0.006
0.008
om
0.012
20
IS
200 -!----+----1---+-...,,:J#=+----1~-__I
CONCLUSIONS
o
o
2SO f!<!!!!!~!lf---+--{.:Seri~e~I~MM~I_-+--__::k~~~-
Downloaded from ascelibrary.org by Griffith University on 07/30/16. Copyright ASCE. For personal use only; all rights reserved.
APPEND~.
REFERENCES
Crasto, A. S., Kim, R. Y., Fowler, C., and Mistretta, J. P. (1996). "Rehabilitation of concrete bridge beams with externally-bonded composite
plates. Part I." Proc., 1st Int. Con! on Compos. in Infrastruet. ICCI
96. 857-869.
Meier, U., and Winistorfer, A. (1995). "Retrofitting of structures through
external bonding of CFRP sheets." Proc., 2nd Int. RILEM Symp.
FRPRCS-2, Reunion Internationale des Laboratoires d'Essais et de Recherches sur les Materiaux et les Constructions, E & FN Spon, London,
England.
Nakamura, M., Sakai, H., Yagi, K., and Tanaka, T. (1996). "Experimental
studies on the flexural reinforcing effect of carbon fiber sheet bonded
to reinforced concrete beam," Proc., 1st Int. Con! on Compos. in 1nfrastruet. ICCI 96, 760-773.
Saadatmanesh, H., and Ehsani, M. (1991a). "RC beams strengthening
with GFRP plates: experimental study." J. Struet. Engrg., ASCE,
117(11), 3417 -3433.
Saadatmanesh, H., and Ehsani, M. (l991b). "RC beams strengthened
with GFRP plates: analytical and parametric studies." J. Struet. Engrg.,
ASCE, 117(11), 3432-3455.
Varastehpour, H., and Hamelin, P. (1995). "Structural behavior of reinforced concrete beams strengthened by epoxy bonded FRP plates."
Proe., 2nd Int. RILEM Symp. FRPRCS-2, Reunion Internationale des
Laboratoires d'Essais et de Recherches sur les Materiaux et les Constructions.